Transfer Pricing and Intangibles: Oddleif Torvik

45
DOCTORAL
SERIES Oddleif Torvik
Transfer Pricing
About the Author Transfer Pricing and Intangibles About the Series
Oddleif Torvik
US and OECD Arm’s Length Distribution of Operating Profits from IP Value Chains
True to its mission of disseminating
and Intangibles
The transfer pricing of intangibles (patents, trademarks, etc.) is an important issue in knowledge of international taxation
international tax law, because it determines how superprofits generated by multinationals and promoting the study of taxation
through the exploitation of valuable intellectual property (IP) in their worldwide value in general, IBFD has taken the initiative
chains are allocated among the jurisdictions in which they do business. For decades,
Distribution of Operating Profits from IP Value Chains

Transfer Pricing and Intangibles: US and OECD Arm’s Length
to make available to a wider audience
multinationals have used IP transfer pricing to shift taxable profits out of high-tax jurisdictions,
a series of books based on doctoral
causing serious base erosion. Both the United States and the OECD seek to combat these
research.
practices through mandatory transfer pricing rules aimed at ensuring that IP superprofits
are taxed where the intangible value was created. The profit allocation process prescribed US and OECD Arm’s Length The IBFD Doctoral Series accepts
Distribution of Operating Profits

by these rules is analysed in this text.
only contributions that enhance the
international academic tax debate and
The first part of the process determines the amount of superprofits allocable to a unique and
Oddleif Torvik is a tax lawyer with

valuable IP (royalty amount). The US and OECD transfer pricing methods that govern this
determination are analysed, applying a distinction between unique and non-unique value
from IP Value Chains meet the highest academic standards.
In order to ensure top quality, every
chain contributions, and it is observed that the methodology has evolved significantly over thesis published in the series underwent
Advokatfirmaet Schjødt AS, one of peer reviewing in line with the strictest
the years, from primarily relying on imprecise third-party benchmarking to more substance-
Norway’s largest corporate law firms, selection standards.
based approaches that seek to ensure results that adhere to the realistic alternatives of the
where he advises primarily on transfer
controlled parties.
pricing and international tax issues. He
also holds a position as Associate Professor Books published in the IBFD Doctoral
The second part of the profit allocation process determines to which group entity, and thus
at the Department of Accounting, Auditing indirectly also to which jurisdiction, the amount of IP superprofits will be allocated. The US Series should:
and Law at the Norwegian School of and OECD intangible ownership provisions that govern this determination are analysed,
• achieve particularly original
Economics (NHH) in Bergen and is an applying an original analytical distinction between manufacturing and marketing IP. The
research results
affiliated member of the Norwegian analysis shows that, while both the US and OECD rules go a long way towards aligning
• include a significantly innovative
Centre for Taxation (NoCeT) at NHH. the allocation of superprofits from R&D-based manufacturing IP with value creation, the
component
He obtained his PhD at the Faculty of allocation of superprofits from marketing IP still largely hinges on formal legal ownership
• be based on a thorough knowledge
Law at the University of Bergen. His and thus opens the opportunity for tax planning from multinationals and should be ripe
of the existing literature on the
primary research interest lies in the field for future reform.
topic(s)
of transfer pricing.
This book is suited for those that have an interest in transfer pricing analysis, e.g. students, • have a particularly strong, long-
lawyers, accountants and economists. The historical background of the current transfer lasting impact on European and/or
pricing rules is explained, allowing for an “all-in-one” solution for catching up with the US international tax law
and OECD transfer pricing development over the last decades.
Your Portal to Cross-Border Tax Expertise

IBFD DOCTORAL SERIES 45
18-10-2018 Doctoral 45 - 21,9 mm rug indd.indd 1 02-11-18 14:53

Transfer Pricing and Intangibles
IBFD Doctoral Series
Editor-in-Chief: Pasquale Pistone

Managing Editor: Craig West
True to its mission of disseminating knowledge of international taxation and

promoting the study of taxation in general, IBFD has taken the initiative
to make available to a wider audience a series of books based on doctoral
research.
The IBFD Doctoral Series accepts only contributions that enhance the in
ternational academic tax debate and meet the highest academic standards.
In order to ensure top quality, every thesis published in the series underwent
peer reviewing in line with the strictest selection standards.
Books published in the IBFD Doctoral Series should:

– achieve particularly original research results;
– include a significantly innovative component;
– be based on a thorough knowledge of the existing literature on the
topic(s);
– have a particularly strong, long-lasting impact on European and/or in
ternational tax law.
The IBFD Doctoral Series is indexed in Thomson Reuters’ Book Citation

Index.
Transfer Pricing and Intangibles
US and OECD arm’s length distribution of

operating profits from IP value chains
Oddleif Torvik
Thesis
submitted to the University of Bergen
in fulfilment of the requirements for the degree of
Philosophiae Doctor (PhD)
Degree awarded on 9 December 2016
Volume 45
IBFD Doctoral Series
IBFD
Visitors’ address:
Rietlandpark 301
1019 DW Amsterdam
The Netherlands
Postal address:
P.O. Box 20237
1000 HE Amsterdam
The Netherlands
Telephone: 31-20-554 0100

Fax: 31-20-622 8658
www.ibfd.org
© 2018 IBFD
All rights reserved. No part of this publication may be reproduced, stored

in a retrieval system or transmitted in any form or by any means, electronic,
mechanical, photocopying, recording or otherwise, without the written prior
permission of the publisher. Applications for permission to reproduce all or
part of this publication should be directed to: permissions@ibfd.org.
Disclaimer
This publication has been carefully compiled by IBFD and/or its author, but
no representation is made or warranty given (either express or implied) as
to the completeness or accuracy of the information it contains. IBFD and/or
the author are not liable for the information in this publication or any deci-
sion or consequence based on the use of it. IBFD and/or the author will not
be liable for any direct or consequential damages arising from the use of the
information contained in this publication. However, IBFD will be liable for
damages that are the result of an intentional act (opzet) or gross negligence
(grove schuld) on IBFD’s part. In no event shall IBFD’s total liability exceed
the price of the ordered product. The information contained in this publica-
tion is not intended to be an advice on any particular matter. No subscriber or
other reader should act on the basis of any matter contained in this publication
without considering appropriate professional advice.
The IBFD and/or the author cannot be held responsible for external content,
broken links or risks within the external websites that are referenced as
hyperlinks within this publication.
Where photocopying of parts of this publication is permitted under article 16B of the 1912
Copyright Act jo. the Decree of 20 June 1974, Stb. 351, as amended by the Decree of
23 August 1985, Stb. 471, and article 17 of the 1912 Copyright Act, legally due fees must
be paid to Stichting Reprorecht (P.O. Box 882, 1180 AW Amstelveen). Where the use of
parts of this publication for the purpose of anthologies, readers and other compilations
(article 16 of the 1912 Copyright Act) is concerned, one should address the publisher.
ISBN 978-90-8722-495-0 (print)

ISBN 978-90-8722-497-4 (eBook, ePub); 978-90-8722-496-7 (eBook, PDF)
ISSN 1570-7164 (print)
NUR 826

Table of Contents
Preface xxvii
Abbreviations xxxi
Part 1
Chapter 1: Research Questions, Methodology and Sources

of Law 3
1.1. Introductory comments 3
1.2. Key terminology and contextualization 6
1.3. Research questions and structure 12
1.4. Methodology 16
1.5. The relevant OECD sources of law 22

1.5.1. Introduction 22
1.5.2. Article 9 of the OECD MTC 23
1.5.3. The OECD Commentaries on Article 9 and
the OECD Transfer Pricing Guidelines 24
1.5.4. Article 7 of the OECD MTC 28
the 2010 OECD Report 29
1.5.6. Case law in connection with articles 9 and 7 30
1.6. The relevant US sources of law 30

1.6.2. IRC section 482 31
1.6.3. The IRC section 482 US Treasury Regulations 33
1.6.4. Case law 35
1.6.5. The OECD TPG 36
1.7. A few words on the 2017 US tax reform 37
1.8. The relationship between the book and other

transfer pricing literature 41
1.9. Reference register and source abbreviations 45
v
Table of Contents
Chapter 2: Business and Tax Motivations for Intangible

Value Chain Structures 47
2.1. Introduction 47
2.2. Horizontal and vertical FDI 48
2.3. To stay home (outsource) or to go out (FDI)? 50
2.4. The centralized principal model for profit allocation 52
2.5. IP regimes and the 2015 OECD nexus approach 54
Chapter 3: Controlled Intangibles Transfers 63
3.2. The US intangibles definition 64

3.2.2. The pre-2018 version of the US IP definition 65
3.2.2.1. Introductory comments 65
3.2.2.2. The relationship between the 936 definition and
profit allocation 68
3.2.2.3. Are goodwill, going concern value and workforce
in place encompassed by the pre-2018 version of
the 936 definition? 71
3.2.2.4. Is goodwill distinguishable from synergy value
attributable to a group of identifiable 936-definition
intangibles valued in the aggregate? 79
3.2.2.5. Concluding comments on the pre-2018 version of
the US IP definition 82
3.2.3. The 2018 version of the US IP definition (the 2017
tax reform amendment) 84
3.3. The OECD intangibles concept 85
3.4. Useful distinctions on the intangibles concept 88

3.4.2. Manufacturing and marketing intangibles 88
3.4.3. Unique and non-unique value chain contributions 90
vi
Table of Contents
3.5. Controlled intangibles transfers subject to

transfer pricing under US law 95
3.5.1. The taxation of US inbound and outbound
intangibles transfers 95
3.5.2. The context in which IRC section 367 applies:
Non-recognition transactions 97
3.5.3. The historical background of IRC section 367 98
3.5.4. Current gain recognition under IRC section 367(a) 99
3.5.5. Deemed royalty inclusions under IRC section 367(d) 104
3.5.5.1. Historical background 104
3.5.5.2. The material content of IRC section 367(d): Sale of
contingent payments 105
3.5.6. Income recognition under section 367(a) or (d) for
intangible transfers? 108
3.5.7. The further relationship between profit allocation
under sections 482 and 367 109
3.5.8. The relationship between profit allocation under
the section 482 cost-sharing regulations and
section 367(d) 112

transfer pricing under the OECD TPG 116
3.7. Concluding comments 117
Chapter 4: Introduction to Part 2 119
Part 2
Chapter 5: The Historical Development of

Profit-Based Transfer Pricing Methodology 125
5.2. Development of the US PSM and the contract

manufacturer theory through case law 126
5.2.2. The 1968 regulations and their background 127
5.2.3. Three inbound cases: Nestlé, French and Ciba 129
5.2.3.1. Introduction 129
vii
Table of Contents
5.2.3.2. Nestlé (1963) 130

5.2.3.3. French (1963) 131
5.2.3.4. Ciba (1985) 135
5.2.4. Three outbound cases: Eli Lilly, Searle and Merck 136
5.2.4.2. The historical tax treatment of investments in
US possessions 137
5.2.4.3. Eli Lilly (1985) 137
5.2.4.4. Searle (1987) 142
5.2.4.5. Merck (1991) 143
5.2.5. Four roundtrip cases: Bausch, Sundstrand, Perkin
and Seagate 145
5.2.5.2. Bausch (1989) 146
5.2.5.3. Sundstrand (1991) 148
5.2.5.4. Perkin-Elmer (1993) 149
5.2.5.5. Seagate (1994) 150
5.2.6. Two cases on controlled services and
sales contracts: DuPont (1979) and Hospital
Corporation of America (1983) 152
5.3. US legislative and regulatory implementation of

“profit-based” methods 153
5.3.2. The 1986 tax reform 153
5.3.3. The 1988 White Paper 155
5.3.4. The 1994 US regulations 159
5.4. OECD implementation of “profit-based” transfer

pricing methodology 161
Chapter 6: Metaconcepts Underlying the US and OECD

Profit Allocation Rules 165
6.2. The relationship between operating profits and

the transfer pricing methods 167
6.2.2. The concept of operating profits 167
6.2.3. Delineating the components of operating profits 169
viii
Table of Contents
6.2.3.1. Sales 169

6.2.3.2. Costs of goods sold 170
6.2.3.3. Gross profit 170
6.2.3.4. Operating expenses 170
6.2.3.5. Net profit 171
6.2.4. Information on gross profits may be unavailable 172
6.2.5. Information on transaction-level profits may be
unavailable173
6.3. The relationship between gross and net profit methods 174
6.3.2. Common methodological traits among the gross
and net profit methods 175
6.3.3. Relevant parameters under the gross and net profit
methods and their impact on reliability 178
6.3.4. Are operating expenses relevant under
the transactional pricing methods (CUT, resale
and cost-plus)? 181
6.3.5. Are comparability adjustments under the gross
profit methods more reliable than under the net
profit methods? 184
6.4. Which transfer pricing method should govern

the profit allocation among value chain inputs? 187
6.5. The arm’s length range 190

6.5.2. The level of comparability required to include
an uncontrolled transaction in the arm’s length range 191
6.5.3. On which point within the arm’s length range may
a reassessment be based? 194
6.6. Comparability 194

6.6.1. Introductory comments 194
6.6.2. The standard of comparability 197
6.6.3. Does the degree to which comparability is required
vary among the pricing methods? 198
6.6.4. The relationship between comparability and
the rules for determining ownership of
intra-group-developed intangibles 199
ix
Table of Contents
6.6.5. Comparability factors 200

6.6.5.2. Contractual terms 201
6.6.5.2.1. Introduction 201
6.6.5.2.2. Comparability of contractual terms 201
6.6.5.2.3. Economic substance and non-recognition 202
6.6.5.3. Functions 207
6.6.5.4. Economic conditions 208
6.6.5.4.1. Introduction 208
6.6.5.4.2. Use of comparables from other markets 209
6.6.5.4.3. Location savings 209
6.6.5.4.4. Temporary pricing strategies 210
6.6.5.5. Risks 211
6.6.5.5.1. Introductory comments on risk 211
6.6.5.5.2. Contractual risk allocation among group entities 213
6.6.5.5.3. Risks affect pricing, not the other way around 217
6.7. The aggregation of controlled transactions 217

6.7.2. The US regulations 218
6.7.3. The OECD TPG 220
6.7.4. GlaxoSmithKline (Canada) 221
6.7.4.2. The factual pattern 222
6.7.4.3. The 2008 Tax Court ruling 225
6.7.4.4. The 2012 Supreme Court ruling 230
6.7.4.5. Observations on the Supreme Court ruling 232
Chapter 7: Direct Transaction-Based Allocation of

Residual Profits to Unique and Valuable IP:
The CUT Method 237
7.2. The US CUT method 238

7.2.2. The purported CUT pertains to a transfer
of the same intangible as transferred in
the controlled transaction 238
x
Table of Contents
7.2.3. The purported CUT pertains to a transfer of

a different intangible than that transferred in
the controlled transaction 239
7.2.3.2. Direct assessment of profit potential 241
7.2.3.3. Indirect assessment of profit potential 241
7.2.3.4. Assessment of profit potential in other cases 242
7.2.4. There are no CUTs available 246
7.3. The OECD CUT method 247

7.3.2. Comparability requirements for unique IP
under the CUT method 248
7.3.3. Comparability adjustments for unique IP
under the CUT method 249
7.3.4. Commercial databases 250
7.3.5. Concluding comments 251
Chapter 8: Indirect Profit-Based Allocation of Residual

Profits for Unique and Valuable IP:
The CPM (US) and TNMM (OECD) 253
8.2. A lead-in to the methodology 254
8.3. The scope of application of the methodology 257
8.4. How operating profits may be allocated to

the tested party under the methodology 259
8.4.2. Selecting an appropriate profit level indicator 259
8.4.3. Extracting the profit level indicator data from
comparable independent enterprises 262
8.4.4. Applying the extracted profit level indicator data
to the tested party 264
8.5. Comparability under the CPM 265
8.6. Comparability under the TNMM 267

xi
Table of Contents
8.6.2. The concept of blended profits illustrated

by an example 268
8.6.3. The 1995 consensus text on the TNMM with
respect to aggregation of transactions 273
8.6.4. The 2006 comparability report 277
8.6.5. The 2008 discussion draft 282
8.6.6. The final 2010 OECD TPG on the use of
aggregated third-party profits as comparables 284
8.6.6.2. The first norm: Aggregated third-party profits
may be used as comparables as long as they are
the result of “similar” third-party transactions 284
8.6.6.3. The second norm: Aggregated third-party profits
may be used as comparables as long as they are
not the result of “materially different” third-party
transactions285
8.6.6.4. The third norm: Aggregated third-party profits
may be used as comparables if the total functions
performed by the third party are closely aligned
with the functions performed by the tested party
with respect to the controlled transaction 287
8.6.6.5. Harmonizing the three norms through interpretation 289
8.6.6.6. Conclusion 292
8.6.7. Applying the 2010 OECD TPG rule to the golf ball
example294
8.7. Has the TNMM converged towards the 1988

White Paper BALRM? 295
8.8. Is reduced transactional comparability under

the TNMM a significant problem? 297
8.9. Closing comments on comparability under

the one-sided, profit-based methodology 299
Chapter 9: Direct Profit-Based Allocation of Residual

Profits to Unique and Valuable IP:
The Profit Split Method 301
9.2. The scope of application of the methodology 302
xii
Table of Contents

9.2.2. The US PSM 302
9.2.3. The OECD PSM 304
9.3. How operating profits may be split under

the methodology 311
9.3.2. Profit split allocation patterns allowed under
the US regulations 311
the OECD TPG 317
9.4. The PSM in valuation scenarios 320
9.5. The OECD PSM is limited to information known

or reasonably foreseeable at the outset 323
Chapter 10: Location Savings, Local Market Characteristics

and Synergies 327
10.2. A lead-in to the topic: The incremental operating

profits at stake 328
10.3. Which jurisdiction should be entitled to tax

incremental operating profits: The basic arguments 330
10.4. What are location savings? 331
10.5. The allocation of cost savings 333

10.5.2. Local comparables are available 333
10.5.3. Local comparables are unavailable 339
10.6. What are other LSAs? 341
10.7. The allocation of location rents 343

10.7.2. Local comparables are available 343
10.7.3. Local comparables are unavailable 343
10.8. Synergies 344
xiii
Table of Contents
Chapter 11: Transfer Pricing of Intangibles in

the Post-BEPS Era under the OECD TPG 347
11.2. A transfer pricing paradigm under pressure 347
11.3. Shall something more now be allocated to source

jurisdictions?350
11.4. The relative roles of the CUT method, TNMM

and PSM for the transfer pricing of intangibles 355
Chapter 12: Allocation of Residual Profits for Unique

and Valuable IP Based on Unspecified
Pricing Methods 359
12.1. Introductory comments 359
12.2. Unspecified methods under the US regulations 361
12.3. Unspecified methods under the OECD TPG 363
Chapter 13: Allocation of Residual Profits to Unique and

Valuable IP through Valuation 367
13.2. The valuation techniques accepted under

the OECD TPG 369
13.3. The valuation parameters in DCF-based valuation 371

13.3.2. The estimation of future operating profits 371
13.3.3. The estimation of useful life, growth rates and
terminal value 372
13.3.4. The estimation of discount rates 375
13.3.5. What are the consequences of using unreliable
valuation parameters? 376
xiv
Table of Contents
13.4. Allocation of the valuation amount among

the controlled value chain contributions 376
13.5. The options realistically available as a restriction

on the possible allocation outcomes 379
13.6. Is there a legal basis for applying a discount

to the transfer price? 382
Chapter 14: Allocation of Residual Profits to Unique and

Valuable IP through Cost-Sharing Structures 383
14.2. Buy-in pricing under the US regulations 385

14.2.2. The development of the US cost-sharing regulations 385
14.2.3. The 2007 Coordinated Issue Paper 389
14.2.4. Case law: Veritas (2009) 394
14.2.5. Case law: Amazon.com (2017) 401
14.2.6. Concluding comments on Veritas and Amazon.com 409
14.2.7. Key concepts under the current cost-sharing
regulations412
14.2.7.2. Cost-sharing transactions 412
14.2.7.3. RAB share 412
14.2.7.4. Non-overlapping (ownership) interests in
intangibles d eveloped under the CSA 413
14.2.7.5. Platform contributions 414
14.2.8. The buy-in pricing methods 416
14.2.8.2. The CUT method 417
14.2.8.3. The income method 418
14.2.8.4. The acquisition price method 425
14.2.8.5. The market capitalization method 426
14.2.8.6. The RPSM 427
14.2.8.7. Unspecified methods 430
14.3. Buy-in pricing under the OECD TPG 430
14.4. The relationship between US and OECD

buy-in pricing 435
xv
Table of Contents
Chapter 15: Taxpayer-Initiated Compensating Adjustments

to Indirect IP Pricing 437
15.2. A lead-in to compensating adjustments 438
15.3. Taxpayer-initiated adjustments under US law 443
15.4. Year-end adjustments under the OECD TPG 447
15.5. Year-end adjustments in the European Union 451
15.6. Case law: Vingcard (Norway, 2012) 452

15.6.2. The factual pattern of the case 453
15.6.3. The comparables supporting the taxpayer profit
allocation454
15.6.4. The contractual risk allocation 457
15.7. Case law: ITCO (Italy, 2010) 460
15.8. Case law: H1 A/S (Denmark, 2010) 460
15.9. Case law: H1.1.1 A/S (Denmark, 2012) 461
15.10. The US GlaxoSmithKline settlement (2006) 462
Chapter 16: Periodic Adjustments of Controlled IP

Transfer Pricing 465
16.2. The development of the US periodic adjustment

concept465
16.2.2. The 1985 House Report and the 1988 White Paper 466
16.2.3. The relationship between the transfer pricing
methods and the periodic adjustment provision
in the White Paper 470
xvi
Table of Contents
16.2.4. Exceptions from the White Paper’s periodic

adjustment provision 471
16.3. The US periodic adjustment provision 472

16.3.2. The main rule: The profit allocation must be
commensurate with income 472
16.3.3. Exceptions to the periodic adjustment rule 479
16.3.3.2. First exception: Initial fixed pricing based
on CUTs i nvolving the same intangible
(“genuine” CUT exception) 480
16.3.3.3. Second exception: Initial fixed pricing based
on CUTs i nvolving a comparable intangible
(inexact CUT exception) 481
16.3.3.4. Third exception: Initial fixed pricing based on
methods other than the CUT method 487
16.3.3.5. Fourth exception: Extraordinary events 488
16.3.4. Five-year cut-off rule 490
16.3.5. Concluding remarks on the exceptions 490
16.4. Periodic adjustments to lump-sum IP transfers

under US law 491
16.5. The OECD periodic adjustment provision 493
16.6. Periodic adjustments of buy-in pricing under

the US regulations 503
16.6.1. Introductory remarks 503
16.6.2. A periodic adjustment is triggered if the AERR
is greater than the PRRR 504
16.6.3. Making a periodic adjustment: Applying
the adjusted RPSM 507
16.6.4. The procedure for making a periodic adjustment
to a buy-in payment illustrated 508
16.6.5. Exceptions to the periodic adjustment provision 514
16.6.5.2. Exception: The initial buy-in pricing is based on
a CUT involving the same platform contribution 514
xvii
Table of Contents
16.6.5.3. Exception: Extraordinary events 514

16.6.5.4. Exception: Reduced AERR does not cause
a periodic trigger 515
16.6.5.5. Exception: Increased AERR does not cause
a periodic trigger 516
16.6.5.6. Exception: Cut-off rule 517

the OECD TPG 517
Chapter 17: The Allocation of Residual Profits to

a Permanent Establishment 519
17.2. Historical background 520
17.3. The relationship between articles 7 and 9 522
17.4. The article 7 profit allocation system 523

17.4.2. Assignment of assets, capital and risk to the PE 524
17.4.3. Assignment of transactions and dealings to the PE 528
17.5. Transfer pricing of the IP transactions and

the dealings of a PE 529
17.6. Allocation of operating profits to a dependent

agent PE 532
17.6.2. A lead-in to the discussion 533
17.6.3. The OECD approach for allocating operating
profits to a dependent agent PE 535
17.6.4. Is the dependent agent PE relevant? 537
17.7. The UN approach for allocating profits to a PE 543

17.7.2. The allocation norm under article 7 of the 2011
UN MTC 544
17.7.3. The relationship between articles 7 and 9 of
the UN MTC 547
xviii
Table of Contents
Chapter 18: Introduction to Part 3 549
Part 3
Chapter 19: The Evolution of the US and OECD Approaches

to Intangible Ownership 555
19.2. Determination of IP ownership under

the 1968 US regulations 556
19.2.2. The DA rule 557
19.2.3. The DA rule was geared towards, but not limited
to, manufacturing IP 559
19.2.4. Was the DA rule limited to the development of
entirely new IP? 560
19.2.5. Case law on the DA rule 561
19.2.5.2. GlaxoSmithKline Holdings v. CIR (2006) 561
19.2.5.2.1. Introductory comments 561
19.2.5.2.2. A lead-in to the case 562
19.2.5.2.3. The principal IRS argument that the US subsidiary
was the developer 563
19.2.5.2.4. The secondary IRS argument that the US
subsidiary was the assister 564
19.2.5.2.5. How would the case have been assessed under
the DA rule? 566
19.2.5.3. DHL Corporation v. CIR (1998) 567
19.3. Determination of IP ownership under the 1992

proposed and 1993 temporary US regulations 571

final US regulations 572
19.4.2. Reason 1 for replacing the DA rule: Criticism
against its treatment of the legal owner 573
19.4.3. Reason 2 for replacing the DA rule:
OECD conformity 575
xix
Table of Contents
19.4.4. The point of departure under the 1994 IP

ownership rules: The legal owner (of IP subject
to legal protection) and the developer (of IP not
subject to legal protection) are entitled to
residual profits 577
19.4.5. The first exception to the legal ownership rule:
Economic substance 580
19.4.6. The second exception to the legal ownership rule:
The multiple owners rule (1994 cheese examples) 581
19.4.7. Concluding remarks on the 1994 US approach to
intangible ownership 584
19.5. Determination of IP ownership under the historical

OECD TPG 585
Chapter 20: A Lead-In to the Determination of IP

Ownership under the US Regulations and
OECD TPG: A Story about Legal Ownership,
Control and Economic Substance 589
20.2. The US regulations 590

20.2.2. Legal ownership 590
20.2.3. The treatment of licensees 591
20.3. The OECD TPG 596
20.4. The determination of ownership of intangibles

not subject to legal protection 597
20.5. Ownership and economic substance 597
Chapter 21: The Distribution among Group Entities

of Residual Profits Generated through
Exploitation of Internally Developed
Manufacturing IP under the US Regulations 607
21.2. The value drivers in manufacturing IP development 608
xx
Table of Contents
21.3. The economic substance exception applied to

manufacturing IP 610
21.3.2. The economic substance exception should not
replace the transfer pricing methods 610
21.3.3. Imputation of contingent payment terms: A lead-in 612
21.3.4. Imputation of contingent payment terms base
example: Successful contract R&D arrangement
with contingent profit split payment structure 614
21.3.5. Imputation of contingent payment terms example:
Unsuccessful contract R&D arrangement with
contingent profit split payment structure 615
Successful contract R&D arrangement with
cost-plus-based contingent payment structure 616
21.3.7. Concluding remarks on the application of
the economic substance exception to impute
contingent payment terms 617
21.4. The US stance on contract R&D arrangements in

light of the 2015 provisions on the arm’s length
standard and best-method rule 619

Exploitation of Internally Developed
Manufacturing IP under the OECD TPG 625
22.2. A lead-in to the profit allocation problem for

internally developed manufacturing IP 626
22.3. Profit allocation for IP development contributions:

Functions628
22.3.2. The “important functions” doctrine 629
22.3.3. Outsourcing: Contract R&D arrangements 630
xxi
Table of Contents
22.3.3.2. The relationship between the 2015 OECD

important-functions doctrine and the historical
2009 business restructuring and intra-group
services guidance 631
22.3.3.3. The important-functions doctrine and contract
R&D arrangements 634
22.3.3.4. Performance of important R&D functions through
geographically dispersed employees 639
22.4. Profit allocation to IP development contributions:

Funding640
22.4.2. A lead-in to the issue 640
22.4.3. Control over financial risk 643
22.4.4. The point of departure for determining the risk-
adjusted rate of return 646
22.4.5. Uncontrolled transaction analogy for determining
the risk-adjusted rate of return: Venture capital
financing647
22.4.6. Do the OECD TPG examples contribute to
the determination of the risk-adjusted rate of return
allocable to the funding entity? 652
22.4.7. Do the US CSA regulations contribute to
22.4.8. Does the cost of capital contribute to
22.4.9. Do the financing alternatives realistically available
contribute to the determination of the risk-adjusted
rate of return allocable to the funding entity? 659
22.4.10. Does the financial risk assumed contribute to
22.4.11. Concluding comments on IP development funding 667

Pre-existing unique IP 672
22.5.2. The point of departure for pricing the contribution
of the pre-existing IP 673
xxii
Table of Contents
22.5.3. The pre-existing IP is contributed by the same

group entity that carries out the ongoing R&D
functions675
22.5.4. The pre-existing IP is contributed by a group entity
different from that which carries out the ongoing
R&D functions 677
22.5.5. Concluding remarks on the contribution of
pre-existing IP to intangible development processes 679

Exploitation of Internally Developed Marketing
IP under the US Regulations 681

internally developed marketing IP 682
23.3. When are the US regulations unwilling to give

pricing effect to foreign legal ownership of
marketing IP due to a lack of economic substance? 684
23.3.2. The wristwatch example: US distributor incurs
incremental marketing expenditures to build local
value of foreign-owned trademark 684
23.3.3. The athletic gear base example: US subsidiary
incurs incremental marketing expenditures to build
local value of foreign-owned trademark 685
23.3.4. The athletic gear example: The 2006 extension 686
23.3.5. Concurrent remuneration during the IP
development phase as a safe harbour from
the economic substance exception 687
23.3.6. A premise for triggering the economic substance
exception: “Incremental” marketing expenditures 688
23.3.7. The economic substance exception should be
relevant only for incoherent pricing structures 691
23.3.8. The economic substance exception is balanced
relative to its 1994 predecessor 694
23.4. Remuneration of a US distribution entity when

the economic substance exception is not triggered 696
xxiii
Table of Contents

23.4.2. Arm’s length marketing expenditures are incurred 697
23.4.3. Above-arm’s length marketing expenditures
are incurred 699
23.4.3.2. Scenario 1: The US subsidiary is deemed owner
of a licence 699
23.4.3.3. Scenario 2: The US subsidiary is compensated
under a separate service agreement 702
23.4.3.4. Scenario 3: The foreign legal owner is
compensated under a separate service agreement 703
23.4.4. What is the appropriate transfer pricing
methodology to remunerate an assister? 705
23.5. Concluding remarks 707
Chapter 24: Distribution among Group Entities of Residual

Profits Generated through Exploitation of
Internally Developed Marketing IP under
the OECD TPG 711
24.2. The 2017 OECD TPG require differentiation of

market-based super profits 712
24.3. Scenario 1: The local group distribution entity

is reimbursed on a cost-plus basis 714

bears an arm’s length level of marketing costs 719
24.4.1. The subsidiary must earn a normal market return
throughout the IP development phase 719
24.4.2. The first twist: The duration of the distribution
agreement is shortened; can residual profits then
be allocated to the subsidiary? 724
24.4.3. The second twist: A royalty payment is introduced;
can residual profits then be allocated to
the subsidiary? 726

bears an above-arm’s length level of marketing costs 729
xxiv
Table of Contents

24.5.2. The OECD TPG threshold for additional profit
allocation to the source-state distribution subsidiary 729
24.5.3. The material content of the profit allocation 734
24.5.4. The profit allocation reservation 741
24.6. Concluding comments 742
Chapter 25: The Allocation of Residual Profits from Unique

and Valuable IP to Permanent Establishments 745
25.2. Assigning economic ownership to internally

developed manufacturing IP 745

developed marketing IP 747
25.4. Assigning economic ownership of acquired

manufacturing IP 749

marketing IP 750
25.6. The cliff effect under article 7 of the OECD MTC

and the important-functions doctrine 751
25.6.2. Philip Morris (Italy, 2001) 753
25.6.3. Rolls Royce (India, 2007) 755
25.6.4. Zimmer (France, 2010) 756
25.6.5. Dell (Norway, 2011) 758
25.6.6. Boston Scientific (Italy, 2012) 761
Chapter 26: Concluding Remarks 767
26.2. The methodology for the transfer pricing of IP 767
26.3. Remuneration of IP development funding 768
xxv
Table of Contents
26.4. Allocation of profits for foreign-owned marketing IP 771
26.5. The identification of local marketing IP 773
26.6. Is the OECD arm’s length standard heading

towards formulary apportionment? 773
References 777
xxvi
Preface
The research project that resulted in this book commenced in September

2011, when the author began his studies for the PhD programme at the
Faculty of Law of the University of Bergen (UiB). The project was made
possible by financing provided through a PhD research scholar position at
the Department of Accounting, Auditing and Law (IRRR) at the Norwe-
gian School of Economics (NHH). The project was initially intended to
be finalized in September 2015, when my position at NHH ended. As life
would have it, however, this was around the time that the US tax authori-
ties issued relevant new regulations and just before the OECD finalized
its BEPS deliverables on transfer pricing of intangibles and related topics
in October 2015. The author found withholding the thesis until he could
fully incorporate these developments to be the only rational solution. The
thesis was submitted for evaluation at the Faculty of Law at UiB on 3 May
2016 and was defended on 9 December 2016. This book is an updated,
and somewhat edited, version of that thesis. The text takes into account
material published before 31 March 2018 (the date on which the text was
submitted for publication).
The author’s research has benefited from the input of many. When he began
working as a tax lawyer in Oslo after his studies in 2006, he found it inspir-
ing to observe how those he worked under mastered the tax law discipline,
including, in particular, tax lawyer and Professor Arvid Aage Skaar and
tax lawyer Terje Hoffmann, both with Wiersholm then, and tax lawyers
Christian Bruusgaard, Sverre Koch and Henning Naas, all with Thommes-
sen then. In this sense, they have contributed to the author’s research, as
there perhaps would be none without their inspiration. The author thanks
the resources at IRRR, where he had his daily workplace throughout the
project. He also thanks Professor Katarina Kaarbøe and Professor Gut-
torm Schjelderup at the Norwegian Centre for Taxation (NoCeT), for be-
lieving in the project and offering strong initial support; Professor Trond
Bjørnenak, Professor Kjell Henry Knivsflå and Professor Frøystein Gjesdal
for management accounting, valuation and transfer pricing discussions, re-
spectively; Assistant Professor Dirk Schindler for insights on the econom-
ics of profit shifting; and PhD candidate Kjell Ove Røsok for financial
accounting insights.
The author also benefited from discussions with tax economists connected
to, and from several interesting seminars arranged by, NoCeT. He thanks
the European Association of Tax Law Professors (EATLP) and the Inter-
national Fiscal Association (IFA) for allowing him to participate in the
xxvii
Preface
2013 EATLP Poster Program for Doctoral Students in Lisbon and the 2014
IFA Mumbai Congress Poster programme, respectively, and would also
like to thank everyone who engaged in discussions with him there. The
author made several trips to the OECD in Paris throughout the project,
participating in discussion draft consultations. He thanks those who shared
their knowledge with him there, in particular Arthur Kristoffersen, Trude
Sønvisen and Stig Sollund, with the Norwegian Ministry of Finance, and
Matthew Wall, with MDW Consulting in Canada. The author also partici-
pated in several PhD seminars. At a 2012 seminar at BI Norwegian Busi-
ness School in Oslo, Professor Ole Gjems-Onstad offered helpful com-
ments at an early stage. At a 2013 seminar arranged by the Nordic Tax
Research Council, the author learned from a lecture held by tax lawyer
PhD Andreas Bullen. At a 2013 David Doublet seminar in Solstrand, Pro-
fessor Ragna Aarli at UiB provided the author with helpful guidance. At a
2014 seminar at the Uppsala Center for Tax Law, the author benefitted tre-
mendously from the vast international tax law insights of Professor Hugh J.
Ault from Boston College Law School and from discussions with Professor
Bertil Wiman and Associate Professor Jérôme Monsenego, both with Upp-
sala University. The author participated in several seminars arranged by
the International Bureau of Fiscal Documentation (IBFD) in Amsterdam.
He thanks those who engaged in discussions with him there, in particular
Antonio Russo, with Baker & McKenzie, and Patrick Ellingsworth, trustee
of IBFD. The author thanks NHH, NoCeT and the Meltzer Research Fund
for financial contributions for making these research travels possible.
Others that kindly offered him comments, practical help or other support
during the project include Professor Richard T. Ainsworth with the Bos-
ton University School of Law; Lee Sheppard with Tax Analysts, whom
he met at a seminar at NHH in the fall of 2012 and provided interesting
comments on US transfer pricing law; tax lawyer Michael Lebovitz, with
White & Case, for helpful comments on the US cost-sharing regulations
at a 2012 seminar he attended in Amsterdam; tax lawyer Leif Drillestad,
then with PwC, for interesting practical insights on transfer pricing; tax
lawyer PhD Hugo P. Matre, with Schjødt, for initial talks in 2011; tax
lawyer Christian Svensen, with Simonsen Vogt Wiig, for transfer pricing
discussions; tax lawyer Kristine Ilstad, with DnB; tax lawyer Bjørn Chris-
tian Lilletvedt Tovsen, with Thommessen; Associate Professor Emeritus
Arthur J. Brudvik, at NHH; and Susanne Tollefsen Log, with Skatt Vest.
The author is grateful for the help he received from University Librarian
Jørn Wangensten Ruud at UiB and University Librarian Fredrik Andersen
Kavli at NHH. He also thanks the IBFD library staff for their help. He also
thanks Administration Manager at IRRR, Maren Dale Raknes, who has
xxviii
Preface
always been helpful, and Senior Consultant Mari Myren, then at UiB, for
her kind assistance. He also thanks his fellow tax law PhD candidates for
friendly discussions throughout the project. These are Tormod Torvanger,
Henrik Skaar, Ingebjørg Vamråk, PhD Eivind Furuseth, Sarah Lindeberg
and Blazej Kuzniacki. He also thanks Katriina Pankakoski, now with the
Finnish Tax Administration, as well as his PhD candidate friends at NHH,
in particular Øivind Schøyen and Martin Evanger. He thanks Julie Wille
for proofreading the manuscript with impressive precision and haste.
He thanks tax lawyer and Professor Jens Wittendorff, with EY and Aarhus
University, for thorough, insightful and helpful comments and discussions
in connection with the midway evaluation of this project.
The person the author is most indebted to is his supervisor, Professor Fred-
erik Zimmer, with the Faculty of Law at the University of Oslo. He con-
tributed significantly during every stage of the project, offering immense
help, always being kind, patient and elegant. It was a privilege to benefit
from his teachings.
He further thanks his mother, Ingvild, for her kind support throughout the
entire project.
He dedicates this work to his beloved daughter, Kamille:
Denne boken er til deg, Kamille. Bestevenna for alltid.
Oddleif Torvik
Førde i Sogn og Fjordane, 31 March 2018
The author will be thankful for comments or questions on the book and can
be contacted by email at oddleif.torvik@nhh.no.
xxix
Abbreviations
10th Cir. Tenth Circuit (US Court of Appeals)

11th Cir. Eleventh Circuit (US Court of Appeals)
1st Cir. First Circuit (US Court of Appeals)
2nd Cir. Second Circuit (US Court of Appeals)
3rd Cir. Third Circuit (US Court of Appeals)
4th Cir. Fourth Circuit (US Court of Appeals)
5th Cir. Fifth Circuit (US Court of Appeals)
6th Cir. Sixth Circuit (US Court of Appeals)
7th Cir. Seventh Circuit (US Court of Appeals)
8th Cir. Eight Circuit (US Court of Appeals)
9th Cir. Ninth Circuit (US Court of Appeals)
2010 OECD Report 2010 OECD Report on the Attribution of Profits to
Permanent Establishments
AC EU Arbitration Convention
ADR Applicable discount rate (United States, cost shar-
ing)
AERR Actually experienced return ratio (United States)
A.F.T.R.2d American Federal Tax Reports, 2nd Series
AOA Authorized OECD approach
AOD Action on decision (United States, Internal R
evenue
Service)
APA Advance pricing agreement
ATB Active trade or business
BALRM Basic arm’s length return method (United States)
BC Basis for conclusions (International Financial Re-
porting Standards)
BEPS Base erosion and profit shifting
BIAC Business and Industry Advisory Committee to the
OECD
BRIC countries Brazil, Russia, India and China
BTA US Board of Tax Appeals
CAPM Capital asset pricing model
xxxi
Abbreviations
CATM Comparable adjustable transaction method (United

States)
CB Cumulative bulletin (United States)
CBCR Country-by-country reporting
CBDT Central Board of Direct Taxes (India)
CBP US Customs and Border Protection
CC Cost centre
CCA Cost contribution arrangement (OECD)
CCA US Circuit Court of Appeals
CCA2 US Circuit Court of Appeals, Second Circuit
CCC Community Customs Code
CCCTB Common consolidated corporate tax base (Euro-
pean Union)
CCED.Mo. US Circuit Court, Missouri, Eastern Division
CCPA US Court of Customs and Patent Appeals
CD.Cal. US District Court, Central District of California
CFA Committee on Fiscal Affairs (OECD)
CFC Controlled foreign corporation
CFO Chief Financial Officer
CFR Code of Federal Regulations (United States)
CIP 2007 IRS Coordinated Issue Paper
CIR Commissioner of Internal Revenue (United States,
Internal Revenue Service)
CIT Commissioner of Income Tax (India)
Cl. Ct. US Claims Court Reporter
CMO Contract manufacturing organization
COGS Cost of goods sold
COI Constructive operating income (United States)
CPI Comparable profit interval (United States)
CPM Comparable profits method (United States)
CRO Contract research organization
CSA Cost sharing agreement (United States)
CUP Comparable uncontrolled price
CUT Comparable uncontrolled transaction
xxxii
Abbreviations
CWI Commensurate with income

DA rule Developer-assister rule
DCF Discounted cash flow
DDIT Deputy Director of Income Tax (Indian tax au-
thorities)
D.Kan. US District Court, Kansas
D.Minn. US District Court, Minnesota
D.N.J. US District Court, New Jersey
EBIT Earnings before interest and taxes
EBITDA Earnings before interest, tax, depreciation and
amortization
ECJ Court of Justice of the European Union
ECR European Court Reports
EDNC US District Court, North Carolina, Eastern Division
EFTA European Free Trade Area
ESO Employee stock option
F.2d Federal Reporter 2nd (United States)
F.3d Federal Reporter 3rd (United States)
FASB Financial Accounting Standards Board (United
States)
FCA Federal Court of Appeal (Canada)
FCMU Full cost mark-up
FDA US Food and Drug Administration
FDI Foreign direct investment
Fed. Reg. Federal Register (United States)
Fed. TD Federal Court, Trial Division (Canada)h
FIFO First in, first out
Fin. Treas. Reg. Final Treasury Regulations (United States)
FRD Federal rules decisions (United States)
FSA Field service advisory
GAAP Generally Accepted Accounting Principles
GATT General Agreement on Tariffs and Trade
GLAM Generic Legal Advice Memorandum (United States,
Internal Revenue Service)
GP General partner
xxxiii
Abbreviations
GSMM Gross services margin method (United States)

GVC Global value chain
HMRC Her Majesty’s Revenue and Customs (United King-
dom)
HR Conf. Rep. US House Conference Report
HR Rep. US House Report
HTVI Hard-to-value intangible
IAS International Accounting Standards
IASB International Accounting Standards Board
IBFD International Bureau of Fiscal Documentation
ICC International Chamber of Commerce
IDA Intangible development activity
IFA International Fiscal Association
IFRS International Financial Reporting Standards
IMC Incremental marketing costs
IP Intangible property
IPO Initial public offering
IPR Intellectual property rights
IRB Internal Revenue Bulletin (United States)
IRC Internal Revenue Code (United States)
IRR Internal rate of return
IRS Internal Revenue Service (United States)
IRS RR IRS Revenue ruling (United States)
IRS TD IRS Treasury Decision (United States)
IT Information technology
ITAT Income Tax Appellate Tribunal (India)
ITPL International transfer pricing legislation (Japan)
KERT Key entrepreneurial risk-taking
LIFO Last in, first out
LMSB Large and Mid-Size Business Division (United
States, Internal Revenue Service)
LOC Local operating company
LoN League of Nations
LP Limited partner
xxxiv
Abbreviations
LRD Low-risk distributor

LSA Location-specific advantage
M&A Mergers and acquisitions
MAP Mutual agreement procedure
MNE Multinational enterprise
MTM Matching transaction method (United States)
N.D. Cal. US District Court, Northern California
N-GAAP Norwegian Generally Accepted Accounting Prin-
ciples
NPV Net present value
OECD Organisation for Economic Co-operation and De-
velopment
OECD MTC OECD Model Tax Convention on Income and on
Capital
OECD TPG OECD Transfer Pricing Guidelines for Multina-
tional Enterprises and Tax Administrations
OEEC Organisation for European Economic Co-opera-
tion
OEM Original equipment manufacturer
ORA Options realistically available
P&L statement Profit-and-loss statement (financial accounting)
PC Profit centre
PCT Platform contribution transaction (United States,
cost sharing)
PE Permanent establishment
PFA Post-formation acquisition (United States, cost
sharing)
PLI Profit level indicator
PPA Purchase price allocation (financial accounting)
Prop. Treas. Regs. Proposed Treasury Regulations (United States)
PRRR Periodic return ratio range (United States, cost
sharing)
PSM Profit split method
Pub.L. US Public Law
PV Present value
xxxv
Abbreviations
PVI Present value of investments (United States, cost

sharing)
PVTP Present value of total profits (United States, cost
sharing)
R&D Research and development
RAB share Reasonable anticipated benefits share (United
States, cost sharing)
RC Revenue centre
RDD OECD revised discussion draft on transfer pricing
aspects of intangibles
Rev. Proc. Revenue procedure (United States, Internal Rev-
enue Service)
Rev. Rul. IRS revenue ruling
ROCE Return on capital employed
RPSM Residual profit split model
RT Reference transaction (United States, cost sharing)
SCBM Simplified cost-based method
SCC Supreme Court of Canada
S.Ct. Supreme Court of the United States
S.D.Fla. US District Court, Southern District of Florida
S.D.Tex. US District Court, Southern District of Texas, Hous
ton Division
SE Societas Europaea
SEC Securities and Exchange Commission (United
States)
SEP Standard essential patent
SIC Code Standard Industrial Classification Code
SpA Società per Azioni (Italian corporate form used by
public corporations)
SPF Significant people functions
TAM IRS National Office Technical Advice Memoran-
dum
Tax Ct. US Tax Court
TCC Tax Court of Canada
TC Memo Tax Court Memorandum Decision (United States)
xxxvi
Abbreviations
TD Treasury Decisions (United States)

Temp. Treas. Regs. Temporary Treasury Regulations (United States)
TMTPR Tax Management Transfer Pricing Report)
TNMM Transactional net margin method (OECD)
TPO Transfer pricing operations, IRS Large Business
and International Division
Treas. Regs. Treasury Regulations (United States)
True-up Taxpayer-initiated compensating adjustment
UN United Nations
UNIDROIT International Institute for the Unification of Private
Law
UNM United Nations Practical Manual on Transfer Pri
cing for Developing Countries
UN PMTP United Nations Practical Manual on Transfer Pri
cing for Developing Countries
USCA US Code Annotated
VAT Value added tax
VC Venture capital
VCLT 1969 Vienna Convention on the Law of Treaties
WACC Weighted average cost of capital
WD Wash. US District Court, Western District of Washington
WP6 Working Party No. 6 on the Taxation of Multina-
tional Enterprises (OECD)
WTO World Trade Organization
xxxvii
Part 1
Chapter 1
Research Questions, Methodology and Sources of Law
1.1. Introductory comments
Multinational enterprises are profitable. They make and sell products and
services in multiple geographical markets. Behind the profits realized from
selling a product in one particular jurisdiction may lie contributions from
group companies resident in other countries or permanent establishments
(PEs) of such companies in source jurisdictions. These different taxable
entities within the multinational enterprise are all part of the same eco-
nomic totality and do not have conflicting economic interests. Their contri-
butions are priced, thereby effectively extracting profits from the jurisdic-
tion where the product is sold. These controlled prices may deviate from
those that would have been yielded by the normal supply-and-demand
market mechanism that ensures balanced pricing among third parties with
conflicting interests.1 Most jurisdictions have, for this reason, enacted man-
datory profit allocation rules that govern how a multinational must distrib-
ute its profits among its entities and have entered into tax treaties, which,
also via profit allocation rules, ensure that there is no double taxation on
such profits.2
This book is a study of how the profits from multinationals’ sales of prod-
ucts and services based on unique intangibles (valuable patents, trade-
marks, etc.) are allocated among jurisdictions under two of the most signif-
icant and influential transfer pricing systems in the world:3 (i) the transfer
pricing regime under US law under section 482 of the Internal Revenue
Code (IRC); and (ii) the transfer pricing regime under articles 7 and 9 of
tax treaties based on the OECD Model Tax Convention on Income and on
Capital (OECD MTC).
There are important interactions between the two regimes. First, both
are based on the same meta-norm, i.e. the arm’s length standard, aimed
1. Or, as stated in Schön (2010a), at p. 236, “[B]etween independent taxpayers,

transaction prices therefore truly allocate income.”
2. Schön (2010a), at p. 232, applies the terminology “transfer pricing control”,
which conveys the purpose of such mandatory profit allocation rules well.
3. See Schoueri (2015) for principal reflections on the arm’s length principle. See
Schön et al. (2011), at pp. 47-67, for an insightful overview of the concept of transfer
pricing across different legal contexts.
3
Chapter 1 - Research Questions, Methodology and Sources of Law
at achieving parity in the taxation of related and unrelated enterprises.4

Second, the overarching legal structure of the systems is similar. The ba-
sic principle for profit allocation is expressed in a few sentences in IRC
section 482 and articles 7 and 9 of the OECD MTC and elaborated in
the comprehensive section 482 of the US Treasury Regulations and the
OECD Transfer Pricing Guidelines for Multinational Enterprises and Tax
Administrations (OECD TPG). Third, even though the main principles for
profit allocation under the two systems have remained intact since the first
part of the 20th century, the more specific methodology that is decisive for
the actual allocation of profits among jurisdictions is constantly evolving,
with revisions typically aimed at avoiding base erosion and profit shifting
(BEPS). Fourth, the systems are largely self-contained. Their allocation of
intangible profits generally do not depend on underlying private law classi-
fications. Fifth, there have historically been significant “spill-over” effects
of the US regime on the OECD TPG. For instance, central TPG concepts,
such as the transactional net margin method (TNMM) and the periodic
adjustment authority, are more or less direct imports from US law.
The division of taxing rights among jurisdictions under both systems is

effectively carried out through these profit allocation rules. The United
States will tax the amount of profits from controlled transactions calcu-
lated pursuant to section 482 of the US Treasury Regulations. Indirectly,
these rules also determine the amount of profits that the residence jurisdic-
tion of the other group entity involved in the controlled transaction may
tax without resulting in double taxation, disregarding tax treaties. Simi-
larly, under article 9(1) of the OECD MTC, the amount of business profits
from controlled transactions calculated pursuant to the OECD TPG may
be taxed by the residency jurisdiction of the relevant group entity, while the
residence jurisdiction of the other group entity involved in the controlled
transaction shall, in principle, exclude an identical amount from taxation
under article 9(2), thereby avoiding economic double taxation.
4. See Brauner (2016), at p. 108, where a background is provided through these

fitting words: “There is no inherent justification for treating related and unrelated trans-
actions alike beyond simplistic symmetry. One could have perhaps made an efficiency-
based justification for such symmetry in some circumstances; however, such a case
has not been made, and the arm’s length transfer pricing rules hardly follow efficiency
goals. Arm’s length, therefore, is not a principle; it is a standard. It serves as a basis
for the specific rules that implement it and is justified by other principles. In the case
of arm’s length, it is justified as a method for allocation of profits. Even we know that,
historically, it was one of several standards that could be used for achieving the goals
that underlie our tax systems. It is perhaps the most desirable standard, yet it is not a
principle”.
4
Introductory comments
The same basic system, even though operationalized through a different

legal mechanism,5 is put in place to govern the distribution of taxing rights
to business profits among residence and source jurisdictions in the context
of PEs under article 7(2) of the OECD MTC, according to which the profits
allocated to the source jurisdiction pursuant to the OECD TPG and the
2010 OECD Report on the Attribution of Profits to Permanent Establish-
ments (2010 OECD Report) shall be excluded from taxation in the resi-
dence jurisdiction through the provision of double taxation relief, thereby
avoiding juridical double taxation.
This book is an analysis of the following two research questions:
(1) The primary question is how the taxing rights to operating profits from
intangible value chains shall be allocated among jurisdictions under
IRC section 482 in US law and articles 7 and 9 of tax treaties based on
the OECD MTC.
(2) The secondary question, which is dependent on the results from the
analysis of the primary research question, is to provide a critical as-
sessment of whether the current US and OECD profit allocation solu-
tions are useful or if they ideally should be altered, and if so, to pro-
pose the relevant amendments.
The author will further develop the research questions and outline the struc-
ture of the book in section 1.3., after introducing key terminology and pro-
viding necessary contextualization for the research questions in section 1.2.
This book will not address possible alternatives to arm’s length transfer
pricing, e.g. so-called “formulary apportionment” (distribution of world-
wide operating profits based on predetermined allocation keys).6 Arm’s
length transfer pricing is the international consensus for profit allocation. It
does not seem realistic that this will change in the near future. Analytical
efforts therefore seem better spent contributing to legal clarification of the
current regime rather than discussing more loosely based notions of possi-
5. For a multinational’s tax planning purposes, an art. 9 allocation may yield a more
favourable profit allocation, in the sense that double taxation relief is not contingent
on the extent to which the profits are actually taxed in the other residence jurisdiction,
much akin to the result of the exemption method under art. 7 (see art. 23A). This stands
in contrast to an art. 7 allocation, where relief nowadays tends to be provided through
the credit method (see art. 23B) and thus is contingent on the extent of taxation in the
source state (of course, the exception method is still applied in some treaties).
6. For a recent overview of the features of formulary apportionment, see Andrus et
al. (2017), at p. 96; and Pankiv (2017), at pp. 38-42. For further discussions on formu-
5
ble alternative allocation regimes. An analysis of the formulary apportion-

ment alternative would also expand the scope of this book beyond what
could be addressed within the time constraints of the research project.7 The
book will nevertheless provide some limited comments on certain aspects
of the relationship between arm’s length transfer pricing and formulary ap-
portionment, as this is deemed to contribute to the analysis of the research
questions.8
1.2. Key terminology and contextualization
Operating profits are business profits before interest expenses and taxes, i.e.
sales revenues minus the cost of goods sold and other operating expenses.9
The author’s analysis is limited to the allocation of operating profits gener-
ated through the sale of products or services based on unique intangibles,
e.g. a pharmaceutical preparation manufactured on the basis of a patent and
sold under a trademark. In the context of transfer pricing, operating profits
are determined and benchmarked at the level of the value chain for a par-
lary apportionment (versus the arm’s length principle), see, e.g. Langbein (1986); Turro
(1994); Lebowitz (1999); Kauder (1993); Hellerstein (1993); Sadiq (2001); Hamaekers
(2001), at p. 38; Ackerman et al. (2002); McLure (2002); Vincent (2005), at p. 414 (on
global profit splits); Hellerstein (2005a); Hellerstein (2005b); Hardy (2006); Bensha-
lom (2007); Roin (2008); Benshalom (2009); Mayer (2009); Angus et al. (2010); Durst
(2010); Morse (2010); Durst (2012a); Kroppen et al. (2011); Fleming et al. (2014); Avi-
Yonah (2015); White (2016), at p. 216; Lebowitz (2008); Luckhaupt et al. (2011), at
pp. 100 and 107; Gresik (2011); Wilkie (2011), at p. 152; as well as the more sceptical
view expressed in Burke (2011). On global tax reform, see Brauner (2003). See also re-
cent reflections on the usefulness of the arm’s length principle in Biegalski (2010); and,
in particular, Schoueri (2015). For a theoretical proposal to address intangible property
(IP) profit shifting through cost sharing agreements by way of formulary apportion-
ment pricing, see Benshalom (2007), at pp. 648 and 679. See Brauner (2008), at p. 160,
on the use of a formulary apportionment approach to IP valuation.
7. For an interesting economic analysis of the relationship between the separate
entity approach and formulary apportionment, see Altshuler et al. (2010), a study that
also highlights some of the problems associated with formulary apportionment.
8. See the discussion in secs. 11.2. and 26.6.
9. This description will suffice for now. A more in-depth understanding of the con-
cept is primarily necessary for the purpose of analysing the one-sided transfer pricing
methods (the gross [resale price and cost plus] and net [comparable profits method
(CPM)/transactional net margin method (TNMM)]) and for understanding the his-
torical context in which the TNMM was introduced into the OECD Transfer Pricing
Guidelines for Multinational Enterprises and Tax Administrations (OECD TPG) in
1995, in particular the OECD arguments against the method. For a further analysis of
the concept of operating profits, see section 6.2. It should be noted that operating profits
do not reflect the costs of debt financing. The issue of intra-group debt financing will
not be discussed in this book, as it falls outside the scope of the research questions.
An important nuance here is that the profit allocation rules under art. 7 of the OECD
6
Key terminology and contextualization
ticular product or service (transactional level), not at the total level for all
products and services sold by the relevant group entity (aggregated level).
This is fundamentally due to the fact that intangibles are normally used in
connection with the creation and sale of specific products or services and
contribute to their profits (e.g. the patent for a blockbuster drug or the code
for a best-selling software package).10 A value chain is a set of activities that
an enterprise performs in order to deliver a valuable product or service to the
market.11 As the author’s focus is on profits from products based on intangi-
bles, he will refer to the relevant value chain as an “intangible value chain”.
In order to deliver a product to the marketplace, a multinational will per-

form functions (research and development (R&D), manufacturing, sales,
etc.) and apply tangible and intangible assets (plant and property, patents,
trademarks, etc.). In doing so, it will incur expenses (for R&D, manufac-
turing, distribution and marketing, etc.), and thereby also financial risks.
All functions performed, assets used and risks assumed along the value
chain, from early-phase R&D to the sale of the final product to the end
consumer, contribute to the value of the product (and thereby the operating
profits derived from its sale) and are, in this sense, value chain contribu-
tions (or inputs).
Value chain contributions are conducive to operating profits to varying

degrees. Both the US and OECD rules rely on the fundamental distinction
Model Tax Convention on Income and on capital (OECD MTC) allow the allocation
of external interest expenses to the permanent establishment (PE) for the purpose of
determining its operating profits. This is not a pricing issue, as the interest expenses are
at arm’s length, but a matter of allowing for external financing of a PE for profit calcula-
tion purposes (see the comments in section 17.4.2).
10. For example, in Eli Lilly v. Commissioner of Internal Revenue (84 T.C. No. 65
[U.S. Tax Ct., 1985], affirmed in part, reversed in part by 856 F.2d 855 [7th Cir., 1988],
the question was how to allocate the operating profits connected to a patent and a trade-
mark employed in the value chain for the drug Darvon and Darvon-N. In both the
US Glaxosmithkline settlement (see the analysis in sec. 19.2.5.2.) and the Canadian
Supreme Court ruling in GlaxoSmithKline Inc. v. R. (2012 SCC 52 [2012], which af-
firmed 2010 CAF 201, F.C.A., [2010], which reversed 2008 TCC 324 [T.C.C., 2008];
see the analysis of 2012 SCC 52 in sec. 6.7.4.), the question was how to allocate operat-
ing profits from sales of the Zantac drug to the connected patents and trademarks. In
Veritas Software Corporation & Subsidiaries v. CIR (133 T.C. No. 14 [U.S.Tax Ct.
2009], US Internal Revenue Service (IRS) nonacquiescence in AOD-2010-05; see the
analysis of the ruling in sec. 14.2.4.), the question was how to allocate profits to the
intangibles connected to a software package contributed to a cost sharing agreement
(CSA).
11. The general concept was introduced by Porter (1985), but there has been a con-
scious focus on the structure of value chains in transfer pricing jurisprudence for far
longer.
7
between unique (or non-routine) and non-unique (or routine) value chain
contributions.12 This distinction is the heart of modern transfer pricing
and will be a red thread throughout the different profit allocation contexts
discussed in this book. The point of the distinction is that routine value
chain inputs only contribute operating profits equal to normal market re-
turns, while non-routine inputs (in practice, unique intangibles) may con-
tribute above-normal returns, which are so-called “super profits”.13 Thus,
there is a significant profit allocation “cliff effect” associated with the
distinction.
Routine contributions typically include contract manufacturing, distribu-

tion, marketing and sales functions. Due to their relatively generic nature, a
range of enterprises will compete to offer these inputs, driving prices down
to a level where there are no super profits, only normal market returns. In
contrast, unique intangibles represent market imperfections. Their pres-
ence in a value chain may cause the supply-and-demand market mecha-
nism to fail in setting a price that provides mutually beneficial outcomes
for the contracting parties. An enterprise that owns unique intangibles may
then reap super profits by exploiting them, as competitors will not have
access to equivalent input factors. This exclusive market entry barrier po-
sition may be shielded by legal protection (patents, trademarks, etc.) or
business secrets (e.g. the Coca-Cola recipe), effectively securing the enter-
prise in a monopoly position to sell certain products or services. This can
normally be sustained only for so long due to time-limited legal protection
or the emergence of new and superior products or services that render the
unique intangibles economically obsolete.
Super profits are known as “residual profits” in the transfer pricing juris-
prudence of US law and the OECD MTC.14 These are the operating profits
that are allocated to a group entity that is deemed to own a unique in-
tangible after all other group entities that have contributed to the relevant
value chain have been compensated with a separately determined normal
market return for their routine contributions. In general, residual profits
12. See the US and OECD definitions in para. 6.17 OECD TPG; and US: Treasury
Regulations (US Treas. Regs.) § 1.482-6(c)(3)(i)(B), and the analysis of the US and
OECD concepts of unique and non-unique value chain contributions in section 3.4.3.
13. Pankiv (2017), at p. 198, touches on this. See also Roberge (2013), at p. 220.
14. Super profits go by different names, depending on the discipline in which the
concept is referred to. Economists normally refer to it as an “economic rent”, meaning
a profit in excess of the market return to the factors of production (labour and capital).
Under perfect competition, this rent will be zero. Financial economists and accountants
normally refer to super profits as the rate of return in excess of the capital requirement
(risk-adjusted cost of capital), yielding a positive net present value for an investment.
8
represent a significantly greater amount of operating profits than those al-

located as normal market returns.15 Further, entitlement to residual profits
is an ongoing interest in the operating profits generated by the intangible.16
Thus, a group entity entitled to residual profits will receive such allocation
throughout the life of the intangible.
Both the US and OECD rules have traditionally assumed that the oper-
ating profits remaining after all routine functions, assets and risks have
been remunerated are due solely to the unique intangibles exploited in the
value chain. The implication of this approach is that all remaining prof-
its are classified as residual profits, and the right to tax this profit is allo-
cated to the jurisdiction where the group entity that is assigned ownership
(for transfer pricing purposes) of the unique intangibles is resident (or the
source state in the case of a PE).17 Such an assumption is normally un-
realistic. Parts of the remaining profits in an intangible value chain may
be incremental profits due to location savings, local market characteristics
and synergies. These profits are, in principle, distinguishable from those
generated by unique intangibles.18 The new OECD rules seek to amend the
historical flaw that the transfer pricing rules have not sufficiently distin-
guished operating profits in this manner.19
The larger the normal market return and incremental operating profits, the
smaller the residual profits will be. The question of how the taxing rights
to residual profits generated by unique intangibles are allocated among ju-
risdictions under US law and the OECD MTC will therefore not be pos-
sible to analyse without also addressing how normal market returns from
the same intangible value chain are allocated among routine value chain
contributions and how incremental operating profits due to cost savings, lo-
cal market characteristics and synergies are allocated among the involved
jurisdictions. This is because the residual profits, due to unique intangibles,
15. This line may be blurred in some scenarios. For instance, in the context of intan-
gible development under the OECD TPG, the profits allocable to research and devel-
opment (R&D) financing may become significant, resembling residual profits (as the
author will revert to in sec. 22.4.).
16. This stands in contrast to a separate normal market return to routine value chain
contributions. If no such contributions are rendered in a given income period, no com-
pensation will be allocated.
17. See also Francescucci (2004a), at p. 72.
18. See, however, Kane (2014) for an interesting discussion of whether synergy value
should be seen as an intangible.
19. See the analysis in ch. 10. See also Francescucci (2004a), at p. 72, for a discus-
sion of the allocation of incremental profits (in the historical context of the 1995 OECD
TPG).
9
are the operating profits that remain after these two groups of profits have
been allocated.
It has been claimed that the arm’s length principle is “flawed”, as it sup-
posedly is unable to account for and allocate parts of the profits that big
multinationals generally make, i.e. residual profits from unique intangible
property (IP) and incremental profits from economies of scale and integra-
tion.20 The rationale is that multinationals are able to create such profits
while unrelated parties are not. Thus, if the profits of a multinational are
allocated among its group entities, and thus among jurisdictions, based on
comparison (benchmarking) with the pricing applied between unrelated
parties, the intra-group pricing will always “miss out” on the residual and
incremental profits, as such profits do not exist among third parties. The
author is sceptical as to whether the bulk of this criticism is indeed justi-
fied, taking into account the transfer pricing methodologies currently at
offer under the US and OECD arm’s length regimes for allocating taxing
rights to business profits.21
All benefits derived by multinationals due to their assets and organization

(unique IP, integrated value chains, synergies, cost savings, etc.) material-
ize in profits through the sale of products and services to third parties in
market jurisdictions where the multinational does business. The US and
OECD transfer pricing methods will allocate all of these profits to the
group entities that have contributed to the value chains through which the
profits were created. The critics of the arm’s length principle claim – and
rightfully so – that this allocation is difficult (if not impossible) to carry out
if it is to be based on third-party comparables for the unique value chain
contributions (unique IP), as such comparables simply do not exist. On this
point, however, it is important to remember (as the critics not always do)
that multinationals do not only use unique value chain inputs, but they also
use a lot of generic (or routine) inputs, for which there indeed are third-
party comparables available. Thus, the key is to recognize that allocation
of profits from controlled transactions can then be based on benchmarking
such routine value chain inputs, resulting in a normal market return profit
allocation to the tested party and treating the remaining profit as a residual
that either shall be allocated fully to the controlled party that contributes
the unique value chain input (under the comparable profits method (CPM)/
20. For an overview of the debate, see, in particular, Schön (2010a), at pp. 233-234;
and Schoueri (2015), at p. 698. For further discussions, see, e.g. Durst (2010); Kobetsky
(2008); Lebowitz (2008); and Francescucci (2004a); as well as much of the formulary
apportionment discussions referred to in the works mentioned in supra n. 6.
21. The transfer pricing methods are analysed in part 2 of the book.
10
TNMM) or be split among the controlled parties if they both contribute

such unique inputs (profit split method).22 Both the normal return and re-
sidual profits will be effectively allocated among jurisdictions through the
application of these transfer pricing methods that operationalize the arm’s
length principle. None of the multinational’s profits will then be “missed”.
This fact seems to be recognized by at least some now.23
My impression is that the critics that claim that the arm’s length principle
is “flawed” may have based their reasoning on an inaccurate understand-
ing (likely influenced by the historical dominance of the comparable un-
controlled transaction (CUT) method) of how the current transfer pricing
methods actually work in practice. For instance, critics often focus solely
on the CUT method without recognizing that other transfer pricing (the
“profit-based”) methods in fact dominate the transfer pricing practices of
both tax authorities and taxpayers worldwide nowadays. In order to facili-
tate a more nuanced debate, critics should, in the author’s view, take into
account that the arm’s length principle does not equal the CUT-method,
but encompasses also a range of other – and effective – pricing method-
ologies.24 In fact, the CUT method will only rarely be applicable at all
to allocate profits from the typical IP-dominated value chains of multina-
tionals.25 The key methods in practice are the CPM/TNMM and the profit
split method,26 but the workings of these are seldom highlighted by critics.
Further, in light of the fact that the 2017 OECD TPG contain elaborate
provisions for allocating residual profits from unique IP27 and also address
how incremental profits from cost savings, local market characteristics and
synergies shall be distributed among jurisdictions,28 there should, in the
author’s view, be little doubt that the arm’s length principle – as it today is
operationalized through the methodology set out in the OECD TPG – actu-
ally does allocate such profits among jurisdictions and thus, in this sense at
least, should not be regarded as “flawed”.
22. This approach is the core of the profit-based methodology paradigm introduced
in the 1988 US White Paper; see the analysis in sec. 5.3.3. (with further references).
23. See Peng (2016), at p. 383 (see also p. 380) with respect to TNMM allocation,
and p. 385 for profit-split-method allocation. See also Schoueri (2015), at p. 699.
24. The current US and OECD transfer pricing methodologies, as applied to IP value
chains, are analysed in part 2 of the book.
25. See the analysis of the comparable uncontrolled transaction (CUT) method in ch.
7.
26. See the analyses in ch. 8 and ch. 9, respectively.
27. See the analysis of the 2017 OECD provisions for allocating residual profits from
intra-group developed manufacturing and marketing IP in ch. 22 and ch. 24, respec-
tively.
28. See the analysis of the OECD guidance in ch. 10.
11
There is no doubt that the arm’s length principle can be criticized for a
whole range of issues (e.g. ambiguous and often imprecise allocation rules,
significant compliance costs due to documentation requirements, etc.), but
the author does find it very difficult to see that the arm’s length principle is
unable, as the critics claim, to allocate all of a multinational’s profits due
to the absence of third-party comparables that reflect residual profits from
unique IP and incremental profits from local market characteristics and
synergies.
1.3. Research questions and structure
The primary research question is how the taxing rights over operating
profits from intangible value chains shall be allocated among jurisdictions
under IRC section 482 in US law and Articles 7 and 9 of tax treaties based
on the OECD MTC.
Applied in the context of intangible value chains, the US and OECD profit
allocation rules will generally be relevant in the exploitation phase of an
intangible’s life. There will normally be no need to allocate profits before
an intangible has been successfully developed and commercialized, as it
will generate profits first when it is exploited through the sale of products
and services.29 Prior to this phase, there will be no profits to allocate.
The profit allocation assessment begins by splitting the total operating

profits from the intangible value chain among the value chain contribu-
tions; in other words, among:
(1) routine value chain contributions (manufacturing, distribution, etc.),
which are assigned normal market return profits;30
(2) location savings, local market characteristics and synergies, which are
assigned incremental profits; and
(3) non-routine value chain contributions (i.e. unique intangibles), which
are assigned the residual profits.
29. It may, however, be that an in-development intangible is transferred among

group entities. An arm’s length charge for the transfer, in the absence of a CUT, will
likely need to rely on a valuation in which one of the key parameters will be an estimate
of the profits that can be generated through future exploitation of the intangible. See the
analysis of the OECD guidance on IP valuation in ch. 13.
30. For the purpose of this overview, the author deems a separately determined risk-
adjusted rate of return to intangible-development-funding contributions (as discussed
in sec. 22.4.) to be included in this category of profits.
12
Research questions and structure
This split is governed by the US and OECD transfer pricing methodologies,31

the analysis of which may be broken down into a range of different sub-
questions, depending on the context and specific methodology used. The
purpose of the methodologies is twofold.
First, they aim to split the total operating profits among the above three
categories of value chain contributions. This is a question both of causality
and value, i.e. the value chain inputs that have contributed to the total prof-
its must be identified and the degree of profit contribution from each input
must be determined (the amount of profits allocable to the input).
Second, the methodologies aim to assign the categorized profits to the

group entities (or headquarters or PE) that have contributed the relevant
value chain inputs. The assignment will normally be straightforward for
routine value chain contributions. For example, it will be causally clear
which group entity has performed contract manufacturing or marketing
functions. It may, however, be more complex to assign incremental operat-
ing profits due to cost savings, local market characteristics and synergies to
a specific group entity, as the determination does not depend on causality,
but rather on how third parties would have allocated the profits. For unique
intangibles, the transfer pricing methodologies only determine the amount
of residual profits to be allocated to a specific unique intangible. They do
not provide a link between the determined residual profit amount and the
group entity to which the amount is to be allocated. That task is left for the
US and OECD intangible ownership provisions to deal with. These latter
provisions connect the residual profits to specific group entities within the
multinational.
The basic principle underlying both the US and OECD intangible own-
ership provisions is that the residual profits generated in the exploitation
phase shall be allocated among the group entities that participated in the
creation of the intangible. This profit allocation shall be carried out in pro-
portion to the relative values of the involved group entities’ routine and
non-routine contributions in the development phase of the intangible’s life.
Through this profit assignment to a specific group entity (or headquarters
or PE), the residence (or source) jurisdiction of the relevant entity is al-
located the right to tax the residual profits. In this way, taxing jurisdiction
31. For the purpose of this overview, the author uses the term “transfer pricing meth-
odologies” broadly to not only encompass the pricing methods, but also the OECD
profit allocation guidance on incremental profits from location savings, market charac-
teristics and synergies.
13
over operating profits from intangible value chains is divided among the
jurisdictions through which the multinational routes its value chains.32
The primary research question must therefore be answered through an

analysis of the US and OECD transfer pricing and intangible ownership
provisions relevant to intangible value chains, as these provisions in con-
cert determine the profit allocation. The author seeks to illustrate this profit
allocation process in figure 1.1.
Figure 1.1
The allocation of the total The allocation of residual profits

operating profits among the among group entities
The total operating profits to be value chain contributions is (that provided intangible
allocated governed by the TP methods development contributions) is
(and the provisions for location governed by the intangibles
savings, location specific ownership rules
advantages and synergies)
Normal market return to

non-unique value chain
contributions
Incremental operating
Operating profits from profits to location savings,
intangible value chain location specific
advantages and synergies
Residual profits to unique

value chain contributions Group entities
(unique IP)
Value chain contributions IP development contributions
Thus, the transfer pricing provisions determine the amount of profits that
shall be assigned to an intangible, and the ownership provisions determine
which group entity, and therefore which jurisdiction, the amount shall be
assigned to.
While the transfer pricing rules are mainly relevant in the exploitation
phase of an intangible’s life, they are, however, also of relevance in the de-
velopment phase. Group entities that contribute routine development inputs
(e.g. laboratory equipment and research facilities) to the creation of an in-
tangible but are not assigned entitlement to subsequent residual profits un-
32. See Schön (2010a), at p. 230, on the two-sided function of the arm’s length princi-
ple with respect to the allocation of taxing rights (income allocation first to persons and
then jurisdictions). See also Schön (2010b) on the topic of the allocation of taxing rights.
14
Research questions and structure
der the intangible ownership provisions are assigned a concurrent normal

market return in the development phase for their contributions through the
transfer pricing methods. Comparatively, the intangible ownership provi-
sions (for the purpose of allocating residual profits in the exploitation phase
of an intangible’s life) look towards which functions, assets and risks were
contributed to the creation of the IP by the involved group entities in the
development phase and allocate profit in a way much akin to the profit split
methodology. Thus, there is an interplay between the transfer pricing and
intangible ownership provisions in both the intangible development and
exploitation phases.
Nevertheless, as the focus of the US and OECD transfer pricing and intan-
gible ownership provisions is on the remuneration of value chain contribu-
tions in the exploitation phase and intangible development contributions in
the development phase, it is necessary to analyse the provisions separately,
which the author does in parts 3 and 4 of this book, respectively.33 This ba-
sic structure of the book mirrors that of a practical transfer pricing analy-
sis. The structure, however, departs from the chronology of section 482 of
the US Treasury Regulations and the OECD TPG, where the intangible
ownership issue is addressed before the transfer pricing issues. The author
finds that the structure of this book is more appropriate for analytical pur-
poses. It reflects the fact that the material content of the ownership rules
has converged significantly with the transfer pricing methodologies34 and is
best seen as a specific application of these. Thus, the structure of this book
offers the benefit of seeing these applications in light of more general prin-
ciples. It is also the author’s view that this makes the book easier to read,
as it otherwise would have been necessary to refer to the transfer pricing
analysis when analysing the intangible ownership provisions.35
33. The author refers to the introductions to parts 3 and 4 of this book for detailed
outlines of the analysis in each respective part.
34. In particular, the profit split method, which is analysed in ch. 9.
35. As mentioned, group entities that contribute to the development of an intangible
and are not compensated with residual profits shall be allocated a concurrent normal
market return compensation for their efforts. In other words, such compensation will
not be drawn from the operating profits generated through the exploitation of the intan-
gible once fully developed. Thus, the compensation of such entities will, in principle, be
triggered before the profit allocation issues discussed above in this section. This does
not, however, apply for the remuneration of intangible development financing under the
OECD TPG, which is linked to the profits generated through the exploitation of the
developed intangible. This entails that, in practice, it will only be the remuneration of
group entities that have rendered routine development contributions that shall be allo-
cated compensation concurrently throughout the R&D phase. While it could be argued
that it would be beneficial to discuss the remuneration of these entities before the main
15
The secondary research question, which is dependent on the results of the

analysis of the primary research question, is to provide a critical assess-
ment of whether the current US and OECD profit allocation solutions are
useful or if they ideally should be altered, and if so, to propose relevant
amendments. This is addressed throughout the book concurrently and in
connection with the analysis of each sub-question under the primary re-
search question.
The author will introduce fundamental concepts in part 2 of the book. The
topics discussed there are closely interwoven with the subsequent analysis
of the profit allocation rules and form the platform for, and should be seen
as an integrated part of, the analysis of the research questions. The au-
thor will outline the business and tax reasons for intangible value chains,
with a focus on the concept of foreign direct investments and how they
relate to super profits.36 He will also introduce the centralized principal
model, which is commonly applied by multinationals for profit allocation
purposes. A discussion of the 2015 OECD nexus approach for preferential
taxation of super profits under IP regimes is also provided. Further, the
author will discuss the types of controlled intangible transactions that are
encompassed by the US and OECD profit allocation rules, as well as the
US and OECD intangibles definitions.37
1.4. Methodology
This book is a legal analysis carried out under the academic traditions of
the discipline of law. The main object of legal research is text. The main
research activity is interpretation. Hermeneutics is, broadly stated, the phi-
losophy and methodology of text interpretation.38 Thus, legal research can
be seen as a hermeneutical discipline.39 Its closest academic parallels are
likely theology and the study of literature. Legal research may also be seen
as a normative discipline.40 The researcher will not always be able to find
a legal norm that exists independently of his own interpretative contribu-
allocation issues discussed above in this section, the author finds it to be a small sacri-
fice to delay the discussion of this issue in order to attain, in his view, an undoubtedly
better overall structure of the book.
36. See ch. 2.
37. See ch. 3.
38. For a somewhat diverging definition, cf. Bernt & Doublet (1998), at p. 181.
39. For a fascinating hermeneutical perspective on legal research, see Bernt & Dou-
blet (1998), at p. 178. See also Hoecke (2011), ch. 1, p. 4.
40. See Hoecke (2011), p. 10.
16
Methodology
tions. When there is more than one interpretation alternative available, the
researcher may have to take a normative position based on his own dis-
cretionary balancing of conflicting interests. By doing so, the researcher
will unavoidably contribute to the formation of the law. The notion that
a legal researcher passively applies a set of tools (legal principles for in-
terpretation) to given material (sources of law) and a solution (legal rule)
just reveals itself, independently of the choices he makes throughout the
interpretation process, is groundless.
Legal research is not an empirical discipline.41 While sources of law in

themselves (statutory texts, preparatory works, court rulings, etc.) clearly
are empirically verifiable, their legal meaning is not. Legal research is, in
the author’s view, a hermeneutical-normative discipline, the goal of which
is to arrive at soundly founded interpretations of the relevant legal text.
Empirical work may be useful in legal analysis as a supplementary element
(e.g. to substantiate factual assertions, provide contextualization for legal
discussions or indicate whether the law works as intended), but will gener-
ally have no place in the legal interpretation process as such. Legal analysis
is, and should remain, an art of authoritative text interpretation, providing
continuity to a tradition with historical roots dating back over 2 millennia
to the Roman legal doctrine.42
An implication of this perspective is that legal research is not an academic

discipline that yields objective answers. There may be no right or wrong
answer as to how a legal text should be interpreted – or, put differently, it
will normally not be feasible to verify in an objective manner whether the
output of the legal analysis, i.e. the interpretation result, is “correct”.43
This stands in contrast to the natural sciences (biology, chemistry, phys-

ics, etc.) and formal sciences (mathematics, statistics, logic, etc.), where a
research problem may be broken down into a research hypothesis capable
of falsification, i.e. of being proven false. Philosopher of science Karl Pop-
per regarded all theories without potential for falsification as non-scientif-
ic.44 The logical positivists supplemented the idea of falsification with the
verifiability requirement, entailing that only theories that were verifiable
41. On the relationship between legal science and empirics, see, e.g. Burns et al.
(2009), at p. 153; and Sandgren (1995), at p. 726.
42. On the Roman roots, see, e.g. Hoecke (2011), at p. 1; and Samuel (2003), at p. 25.
43. It will, however, likely be feasible to verify whether the interpretation itself is
valid. This will be a question of whether the interpretation has respected the boundaries
set by the governing hermeneutic interpretation principles.
44. Popper (1962).
17
through empirical means, or logically necessary, should be considered sci-

entific.45 Contrary to the view of Alf Ross that the principle of verification
must apply also to legal research, in the sense that assertions in legal doc-
trine should be verified against the actions of the courts,46 the author finds
that the notion of verification has no place in legal research.47
As there is no manner in which the interpretation of a legal text may be

proven false or verified empirically or logically, there seems to be little
sense in formulating research hypotheses in legal research. A more useful
approach is to formulate relevant interpretation questions, which may serve
as research questions.48 The author has acted in accordance with this logic
and formulated the two research questions presented in section 1.1. In doing
so, he recognized that there are no clear norms governing how such ques-
tions should be structured. The guiding parameters that he used are that
the research questions should make sense and be highly relevant within the
legal field to which the research pertains (i.e. international transfer pricing).
The research questions are answered through legal analysis. With this, the
author refers to:
− an interpretation of the relevant sources of law aimed at delineating
their meaning for the purpose of determining a legally binding rule
through a harmonization of the source contributions, i.e. to find the law
“as it is” (de lege lata approach); and
− an assessment of whether the derived legally binding rule is useful as
it is or if it should ideally be altered, and if so, proposals for the rel-
evant amendments (de lege ferenda approach), i.e. to find the law “as it
should be”.
45. See https://en.wikipedia.org/wiki/Logical_positivism#Verification (6 Oct. 2015).

46. Ross (1958), at p. 40. See also Samuel (2008), at p. 319.
47. It could be argued, however, that if the positions taken in legal research are not
given any weight by the courts in their applications of the law, the positions would be of
little value. Contrary, if they are taken into account by the courts, they thereby, in some
sense at least, are “verified”. This argument, however, does not pertain to verifiability
as the term is used in the natural and formal sciences.
48. There are nevertheless examples of research hypotheses being used in legal re-
search of tax law issues, as done in Skaar (1991). If research hypotheses are used in
legal research, it should, in the author’s view, be made clear that they should not be
read as assertions that are capable of objective scientific falsification, but rather as a
way to formulate the relevant interpretation questions. A hypothesis in this field of re-
search does not seem to be anything other than an interpretation question “in disguise”,
nor does it add anything (relative to an interpretation question), and it may provide a
misleading impression (that the hypothesis is capable of scientific falsification). The
author is therefore of the view that it could be for the better to just discard the notion of
research hypotheses in the field of legal research altogether.
18
Methodology
The primary research question is answered through a de lege lata analysis of

the current US and OECD profit allocation provisions, encompassing both
the transfer pricing methodologies and the intangible ownership provisions.
Such an analysis is, in the author’s view, relevant, as there is considerable
ambiguity associated with these provisions, triggering a need for legal clari-
fication. Also, many of the relevant provisions were introduced recently and
have therefore not previously been thoroughly analysed for academic pur-
poses.49
The author’s interpretations of the relevant texts are restrained by the fun-
damental legal interpretation principles that govern which sources of law
are relevant, how interpretation arguments may be derived from them and
how the arguments should be harmonized when they provide diverging di-
rections for the interpretation. Application of these principles validates the
author’s interpretations as legal research, distinguishable from text analy-
ses carried out in other hermeneutical disciplines.50
The US rules are domestic tax law, while the OECD rules are international
treaty law. The principles for interpretation of the two sets of rules have dif-
ferent legal foundations. US tax law shall be interpreted in accordance with
the prevailing interpretation principles developed within this legal system
for the particular field of tax law. The IRC is the primary source of law, as
interpreted by the US Treasury Regulations and court decisions. The Su-
preme Court is the highest court in the United States,51 above the US courts
of appeals52 and the trial courts.53 Decisions from the US Tax Court are
generally deemed to have more weight than decisions from a trial court due
to its technical expertise on tax law issues.54 The OECD MTC provisions
shall be interpreted in accordance with the principles developed under in-
ternational law, as codified in articles 31-33 of the 1969 Vienna Convention
on the Law of Treaties (VCLT).55 Primary weight is placed on the wording
49. For a brief and recent overview of the 2015 OECD profit allocation rules, see
Wittendorff (2016). See also Petruzzi (2016). For critical comments, see Avi-Yonah et
al. (2017), at sec. 3.12 (The limits of Actions 8-10); and Musselli et al. (2017).
50. See also Bernt & Doublet (1998), at p. 205.
51. The US Supreme Court has its legal basis in art. III of the US Constitution. The
federal court system is made up of 94 district-level trial courts and 13 courts of appeals
that sit below the US Supreme Court.
52. These are 13 appellate courts that sit below the US Supreme Court. The 94 federal
judicial districts are divided into 12 regional circuits, each with their own court of appeals.
53. The 94 districts (or trial courts) are called US District Courts.
54. The US Tax Court is established under art. I of the US Constitution.
55. See, in particular, Bullen (2010), at pp. 27-32 and pp. 42-53 on the application of
the 1969 Vienna Convention on the Law of Treaties (VCLT) interpretation principles
for the interpretation of the OECD MTC and the OECD TPG, respectively.
19
of the relevant treaty provisions in light of their context and the purpose of
the treaty.
These formal points of departure, however, offer little in terms of useful

guidance – or meaningful restrictions, for that matter – with regard to
the interpretation process. The core principle for profit allocation in both
regimes, i.e. the arm’s length principle, is (as touched upon in section
1.1.) expressed in a few sentences in the US IRC, section 482 and the
OECD MTC, articles 7 and 9. The principle is developed and operation-
alized through comprehensive, detailed and often ambiguous provisions
in the US Treasury Regulations and the OECD TPG in implementing
the arm’s length metanorm for a wide range of different transfer pricing
contexts. As the author will revert to in sections 1.5.6. and 1.6.4., case
law may offer useful analogies but is rarely decisive for the interpreta-
tion process.
The lion’s share of the interpretation problems raised in this book are re-
solved by way of interpreting the US Treasury Regulations or the OECD
TPG, not the wording of IRC section 482 or articles 7 and 9 of the OECD
MTC as such. The former secondary texts have their own distinct ter-
minology, structure and inherent logic, making up complex micro legal
systems. While the wording of their provisions often may yield little in
the way of specific direction, a range of contextual, transfer pricing and
economic arguments generally aid in the interpretation. Such arguments
include:
– internal consistency (within the US Treasury Regulations and the
OECD TPG);
– direction provided through examples (included in the US Treasury
Regulations and the OECD TPG);
– the core purpose behind the US and OECD profit allocation rules for IP
that intangible profits should be allocated according to value creation;
– the degree of correlation between the different interpretation alter-
natives and the fundamental transfer pricing principles deducted
from the arm’s length metanorm, e.g. parity in the taxation of related
and unrelated enterprises, the realistic alternatives available, etc.;
and
– fundamental economic reasoning, e.g. that routine and non-routine
value chain contributions attract normal market returns and residual
profits, respectively.
The US and OECD profit allocation rules are constantly evolving. It is im-
portant to have a clear understanding of the material content of historical
20
Methodology
rules, as new rules are normally adopted to alter the profit allocation con-
sequences of their predecessors. It will therefore often be relatively easy to
identify the purpose of new rules, which may then aid in their interpreta-
tion. An understanding of the historical context of the current rules also
facilitates a better understanding of development trends in transfer pricing
jurisprudence. The historical rules are therefore analysed where the author
finds them to be relevant.
The secondary research question will be analysed using a de lege ferenda

approach for the purpose of determining whether the law is as it should
be. The relevance of such an analysis is accentuated by the likelihood that
multinationals will continue to adapt to the US and OECD profit allocation
rules, resulting in a potential for continued BEPS and a need to revise the
current rules. Analytically founded suggestions for more useful profit al-
location rules should therefore, in the author’s view, be relevant. Also, the
OECD has not yet fully completed its BEPS Project revision of the transfer
pricing guidance. The OECD work on significant topics, such as the ap-
plication of the profit split method in the context of global value chains and
the allocation of profits to R&D funding contributions, is ongoing.56 The
viewpoints expressed in this book could be relevant for the ongoing OECD
work.
This book is not an interdisciplinary work. The author’s analysis of the

research questions is based solely on a hermeneutical-normative legal
methodology. No theoretical or empirical methodology from other aca-
demic disciplines is applied. This is not to say that an interdisciplinary
approach would not possibly be fruitful; perhaps it would have been. Such
an analysis, however, would raise principal methodological and practi-
cal issues and should likely be carried out in cooperation with research-
ers from other disciplines (e.g. economics, financial accounting and
finance).57
Important transfer pricing research has been done within the field of eco-
nomics in particular. The author will present an overview of some of the
basic findings from this research in chapter 2. This will provide useful
perspectives and facilitate the subsequent analysis through a framing of
56. See the analyses in ch. 9 and sec. 22.4. for the profit split method and R&D fund-
ing remuneration, respectively.
57. It would have been necessary to have separate parts on legal analysis and empiri-
cal work. One approach would, for instance, be that empirical data was presented after
the legal interpretation of a provision to illustrate how multinationals actually have
adapted to the rule. Whether such use of empirical data would add much value to the
21
the relevant legal issues in an economical and empirical context. It will

also ground factual assertions that the author employs in his analysis
(e.g. on the realistic alternatives to a controlled transaction) of empirical
facts.
On a related topic, the author would like to add that some of his discussions
pertain to rules that govern somewhat technical financial concepts, such
as transfer pricing valuation. It is necessary to have a basic understanding
of these concepts in order to gauge how the law governing them should
be read. The author will provide brief outlines of the underlying concepts
where necessary.
1.5. The relevant OECD sources of law
1.5.1. Introduction
The analysis of both research questions, with regard to the OECD MTC,
depends on an interpretation of articles 7 and 9. Together, they govern the
international allocation of business profits under OECD MTC-based trea-
ties, eliminating double taxation. Their purpose, however, goes far beyond
the mere avoidance of double taxation, as they are set in place to ensure that
business profits are allocated pursuant to the specific arm’s length pattern,58
book is another question. This does not seem clear, given the considerable empirical
work already done by the OECD on the extent of BEPS; see OECD, Measuring and
Monitoring BEPS – Action 11: 2015 Final Report (OECD 2015), International Organi-
zations’ Documentation IBFD. See also, somewhat in the opposite direction, Musselli
et al. (2012); Musselli et al. (2013); and Hines (2014).
58. There is a clear distinction between profit allocation aimed solely at avoid-
ing double taxation and that which is aimed at avoiding double taxation by way of
an arm’s length profit allocation. The former is adhered to as long as the profit is
not taxed in both jurisdictions, regardless of the underlying contributions of func-
tions, assets and risks, for example, if 100% of the profits are allocated to one of the
jurisdictions. The latter type of allocation is achieved if total profits are split among
the involved jurisdictions pursuant to the OECD profit allocation rules. For example,
this could be the case if the contract manufacturing performed by a group entity in
jurisdiction 1 is responsible for 10% of the total operating profits while the patent
and trademark owned by a group entity in jurisdiction 2 is responsible for 90% of the
operating profits from the value chain and the allocation of income reflects this value
contribution. Through the allocation of 10% to jurisdiction 1 and 90% to jurisdiction
2, double taxation will be avoided and the arm’s length standard adhered to. On a sim-
ilar note, see Schön (2010a), at p. 232, where he points out that “this does not explain
why it is exactly the comparison between a dependent and an independent enterprise
which provides the standard benchmark for jurisdictional allocation. The same holds
true when the arm’s length standard is regarded primarily as a means to administrate
22
The relevant OECD sources of law
i.e. as unrelated enterprises would have done under similar conditions. The
author will tie some comments to the sources of law relevant to the inter-
pretation of these provisions in sections 1.5.2.-1.5.6., beginning with arti-
cle 9, as the OECD TPG were designed to elaborate on this provision.
1.5.2. Article 9 of the OECD MTC
The allocation of operating profits among group entities resident in differ-

ent jurisdictions is governed by article 9 of the OECD MTC. Article 9(1)
expresses the arm’s length standard for profit allocation as follows:
[C]onditions are made or imposed between the two enterprises in their com-
mercial or financial relations which differ from those which would be made
between independent enterprises, then any profits which would, but for those
conditions, have accrued to one of the enterprises, but, by reason of those con-
ditions, have not so accrued, may be included in the profits of that enterprise
and taxed accordingly.59
The question is how this standard should be understood in the context of

controlled intangibles transfers.60 The language of article 9 offers little
guidance. It should be seen as merely expressing the principle that profits
should be allocated in a manner that reflects the allocation that would have
been carried out among third parties with conflicting interests. The deline-
ation of the further content of article 9 must therefore rely on other sources
of law.
double taxation: while this requires the consent to apply one single standard by both
countries, it is not self-evident that this measuring rod should be exactly the arm’s
length standard”. On the status of the arm’s length principle in treaty law, see Lepard
(1999/2000).
59. The system of art. 9 is that the profits allocable to a residence jurisdiction un-
der art. 9(1) shall be excluded from taxation in the other residence jurisdiction under
art. 9(2). This avoids economical double taxation and ensures that the profits are al-
located among the residence jurisdictions pursuant to the arm’s length standard. With
respect to the area of application of art. 9, see Dwarkasing (2011) and Rotondaro (2000)
for an analysis of the concept of associated enterprises under art. 9 of the OECD MTC.
60. See Schön (2010a), at p. 231, for reflections on the origins of the OECD arm’s
length standard as an allocation norm for attributing income to PEs, a norm which is now
applied to allocate income among different group entities and has had immense “reverse”
influence on the OECD PE allocation regime (for the TPG analogy approach taken by the
Authorized OECD Approach doctrine, see the discussion in sec. 17.4.). On the interpreta-
tion of art. 9, see, e.g. Vogel et al. (1993); Wittendorff (2010a), at pp. 112-145; and Bullen
(2011), at pp. 27-30, with further references. For interesting analysis of the arm’s length
principle and EU law, see, in particular, Schön (2011a); and Almendral (2013).
23

the OECD Transfer Pricing Guidelines
The main sources of law for interpreting article 9 are the OECD com-
mentaries.61 They do not contain any transfer pricing guidance in and of
themselves, but effectively incorporate the TPG by way of reference.62 The
TPG represent the OECD’s interpretation of the arm’s length principle,
thereby expressing the consensus approach of the OECD member coun-
tries on the allocation of operating profits. This document enjoys an im-
mense and unique position as the dominating source of law in international
transfer pricing jurisprudence.63 Both tax authorities and multinationals
follow the development of the TPG closely, often adapting their practices to
new OECD positions even before these are finalized in a consensus text.64
The text is influential for the design and interpretation of domestic transfer
pricing provisions worldwide, as well as for other model treaties, most no-
tably the UN MTC.65
The TPG describe a transfer pricing system, with well-developed and de-
tailed concepts. There is, in the author’s view, no room for illusions here:
it will simply not be possible to carry out a complex profit allocation as-
sessment in a uniform manner based on the arm’s length metanorm ex-
pressed in article 9(1) alone.66 This comprehensive document spans over
61. See Bullen (2011), at pp. 31-33 for a discussion of the status of the OECD Com-
mentaries as a source of law.
62. Sec. 1 of the OECD Commentary on Article 9. For discussions pertaining to the
status of the OECD TPG as a source of law for the interpretation of art. 9, as well as on
the interpretation of the OECD TPG themselves, see Bullen (2011), at pp. 33-56. For a
critical view on the (2010) OECD TPG, see Li (2011). For interesting de lege ferenda
reflections on the OECD TPG, see Schön et al. (2011), in particular pp. 71-89 and 123-
136. For an informed historical discussion of why the OECD arm’s length regime has
persisted, see Durst (2011), at p. 128.
63. See Rocha (2017), at p. 193, where it is stated that the 2017 IFA branch reports
demonstrated that the OECD TPG comprise the most important transfer pricing stand-
ard for the IFA branches. For a thorough analysis of the influence of the OECD TPG on
international tax law, see Calderón (2007).
64. The transfer pricing influence of the OECD should be seen in the broader context
of the significant role that the organization plays in developing international tax rules.
On this issue, see, in particular, Ault (2008). See also Ernick (2013).
65. See, e.g. Vega (2012).
66. For instance, it is self-evident that transfer pricing concepts, such as the TNMM,
profit split method (PSM), intangible ownership and periodic adjustment, must be de-
veloped with detailed guidance in order to be applied in a uniform manner that avoids
double taxation and reflects a third-party profit allocation pattern. It is therefore ironic
that the OECD, in the historical Commentary on Article 9 of the draft 1963 OECD
MTC, stated that “the Article seems to call for very little comment”.
24
a multitude of topics, ranging from general guidance on the arm’s length

principle,67 the transfer pricing methods,68 comparability analysis,69 ad-
ministrative procedures and documentation,70 as well as guidance tailored
to specific types of controlled transactions, including intangibles,71 intra-
group services,72 cost sharing agreements73 and business restructurings.74
The author will make two general observations of this document here and
address more specific issues as they arise in later analysis.
First, the arm’s length standard in article 9(1) will generally be met if the
TPG are adhered to.75 It is difficult to imagine a realistic scenario in which
there is direct conflict between the language of article 9 and the material
content of the TPG. Nevertheless, it is not inconceivable that some of the
solutions drawn up in the TPG would likely be difficult for third parties to
accept.76 Even in these instances, however, it seems unrealistic to imagine
that a convincing legal argument could be made that the positions taken in
the TPG should be interpreted as conflicting with article 9 and that the lat-
ter should prevail.77 The reason for this is that article 9(1) only expresses a
principle, while the more specific content of that principle is developed in
the TPG and is subject to more or less constant changes.
67. OECD TPG, ch. 1.

70. OECD TPG, chs. 4-5.
75. See, however, Pankiv (2017), at p. 148, where it is asserted that “the transfer
pricing analysis should not only be limited to the pricing of particular transactions and
the methods provided by the OECD Guidelines, but should also take into account the
financial or commercial conditions under which the transactions occur”. This assertion
must be based on a shallow view of the transfer pricing methods, as the methods clearly
do take into account relevant financial and commercial conditions. Thus, the “conflict”
that the author seems to indicate between art. 9 and the OECD TPG is, in the author’s
view, one that does not exist.
76. As the author will revert to at several points in this book, some of the positions
taken in the OECD TPG do seem difficult to reconcile with third-party behaviour.
For instance, the now historical 2009 guidance on contract R&D agreements in effect
allocated residual profits in a manner that was not aligned with intangible value crea-
tion; see 2010 OECD TPG, at paras. 7.41 and 9.22-9.28; and the discussions in secs.
22.3.3.2., 25.2. and 22.4.2. of this book.
77. See, however, Musselli et al. (2017), at p. 341, where it seems to be argued that
the new OECD approach to IP ownership in the context of intra-group contract R&D
agreements (see the analysis in sec. 22.3.3.) could be contrary to the wording of art. 9
of the OECD MTC. The author does not find the argument convincing.
25
Second, the TPG is a document that raises problematic interpretation is-

sues. This is partly due to incoherence. Some of its positions, typically on
controversial topics, are difficult to reconcile with each other. Much of the
reason for this is to be found in the fact that the TPG is a “patchwork”, con-
sisting of texts written at different times. The core foundation of the current
TPG was largely put in place in connection with the revision of the 1979
OECD Report on Transfer Pricing and Multinational Enterprises (1979
OECD Report), which resulted in the 1995 TPG. Among the “big issues”
at that time was the adoption of the profit-based transfer pricing methodol-
ogy (the TNMM, profit split method and the periodic adjustment authority)
as a reaction to section 482 of the 1994 US Treasury Regulations. In 2009,
the (now previous generation of) business restructuring guidance was fi-
nalized, with important positions taken with regard to, for instance, the
allocation of profits from contract R&D agreements. The 2015 implemen-
tation of the BEPS work introduced a range of new positions with regard
to the allocation of residual profits.78 The composition of Working Party
No. 6 of the OECD’s Committee on Fiscal Affairs was different in each of
these revisions, as were the political and economic contexts in which the
texts were written.
For instance, the OECD took relatively conservative positions in the 1995
revision, offering taxpayers significant leeway in pricing their controlled
transactions. The pre-financial crisis text on business restructurings, writ-
ten a good 10 years after the 1995 text, addressed a specific type of con-
trolled transaction, and its scope was therefore narrower than that of the
general 1995 revision. While geared towards curtailing profit shifting, it
still contained a range of MNE-friendly positions relevant to the transfer
pricing of intangibles (the text came perhaps a little late, as many multi-
nationals by that time had already restructured their operations to accom-
modate a tax-efficient structure). The 2015 BEPS text was written in the
post-financial crisis climate, with public financing in most OECD juris-
dictions declining and on the basis of a broad political consensus that the
transfer pricing rules were ripe for a major revision aimed at preventing
widespread base erosion. As the author will revert to at relevant points in
the book, the positions taken in these three major revisions are not always
easily reconcilable.
78. For informed overviews of the BEPS Project, see, in particular, Christians et al.
(2017), at pp. 36 and 46 with respect to transfer pricing; as well as Brauner (2014a);
Andrus et al. (2017), at p. 89; Ernick (2011); Wilkie (2014b); and Brauner (2014b). On
changes to the OECD profit allocation rules for residual profits from manufacturing
(R&D-based) intangibles in particular, see Musselli et al. (2017).
26
Adding to this that the key controversial issues in each of the three re-
visions were problematic to achieve clear consensus on,79 resulting in
relatively ambiguous language, one is faced with a document that is not
straightforward for interpretation. In some scenarios, ambiguous text may
conceal a lack of real consensus on a contentious issue. In these cases, it
may be that the examples included in the TPG, the relationship between the
relevant language and other parts of the TPG, “preparatory works” (i.e. dis-
cussion papers issued prior to the enactment of the final text), fundamental
transfer pricing principles, basic economic reasoning, etc. may contribute
to the interpretation.80
The OECD BEPS Project resulted in the finalization of the following guid-
ance (relevant to the analysis in this book) in October 2015: 81
− revisions to section D of chapter I of the OECD TPG on, inter alia,
controlled risk allocations (resulting in new paragraphs 1.33-1.173);
− revisions to chapter VI of the OECD TPG on intangibles (resulting in
new paragraphs 6.1-6.212), including a new annex to the chapter illus-
trating the intangibles guidance (resulting in new paragraphs 1-111);
− revisions to chapter VII of the OECD TPG on intra-group services
(resulting in new paragraphs 7.1-7.65); and
− revisions to chapter VIII of the OECD TPG on cost contribution ar-
rangements (paragraphs 8.1-8.53), including a new annex to the chap-
ter illustrating the cost-sharing guidance (resulting in new paragraphs
1-22).
These 2015 texts represented the consensus view of the OECD member
countries at the time at which they were issued82 and were formally ap-
proved by the OECD Council for incorporation into the OECD TPG in
May 2016.83 In June 2016, the OECD issued a revised version of the busi-
ness restructuring guidance in chapter 9 of the OECD TPG under the label
79. Relevant examples include the implementation of the TNMM and the periodic
adjustment authority in the 1995 revision, the treatment of contract R&D agreements
in the 2009 revision and the allocation of intangible profits pursuant to the concept of
important functions in the 2015 revision.
80. See Bullen (2011), at pp. 50-53, for a discussion of the significance of other
OECD publications addressing the interpretation and application of the arm’s length
principle as sources of law.
81. The texts are printed in the OECD, Aligning Transfer Pricing Outcomes with
Value Creation – Actions 8-10: 2015 Final Report (OECD 2015), International Organi-
zations’ Documentation IBFD.
82. Id., at p. 10.
83. See http://www.oecd.org/ctp/transfer-pricing/oecd-council-approves-incorpora
tion-of-beps-amendments-into-the-transfer-pricing-guidelines-for-multinational-en
terprises-and-tax-administrations.htm (accessed 15 June 2016).
27
“conforming amendments”, which removed the sections of the 2009 text

on business restructuring that were deemed incompatible with the profit
allocation positions taken in the above 2015 texts.84 The revised chapter 9
was approved by the OECD Council in April 2017. In June 2017, the OECD
released the 2017 edition of the OECD TPG, which incorporated these
revisions to the 2010 OECD TPG.85
References throughout the book to the OECD TPG are references to the
current 2017 version. The author will specifically state when he refers to
the historical 1995 or 2010 versions of the OECD TPG.
Further, the OECD has now issued discussion drafts containing new guid-
ance for the profit split pricing method, as well as for the attribution of
profits to PEs. While these drafts have not yet resulted in consensus texts
that have been approved by the OECD Council, they are, at the time of
writing, well developed, and there will likely be no major changes in the
final consensus texts. The author therefore assumes that his discussions on
the drafts will be relevant also after the consensus texts have been issued
and approved by the OECD Council.
1.5.4. Article 7 of the OECD MTC
The allocation of operating profits among the headquarters of an enterprise

in its residence jurisdiction and a PE in a source jurisdiction is governed
by article 7 of the OECD MTC. Article 7(2) expresses the arm’s length
standard in the PE context and states that the profits allocable to a PE are:
[profits that] it might be expected to make, in particular in its dealings with
other parts of the enterprise, if it were a separate and independent enterprise
engaged in the same or similar activities under the same or similar conditions,
taking into account the functions performed, assets used and risks assumed
by the enterprise through the permanent establishment and through the other
parts of the enterprise.86
84. See http://www.oecd.org/ctp/transfer-pricing/public-review-sought-of-beps-con

forming-changes-to-chapter-ix-of-the-oecd-transfer-pricing-guidelines.htm (accessed
4 July 2016).
85. See http://www.oecd.org/ctp/transfer-pricing/oecd-releases-latest-updates-to-the-
transfer-pricing-guidelines-for-multinational-enterprises-and-tax-administrations.htm
(accessed 10 July 2017).
86. The system of art. 7 is that the profits that are allocable to the source jurisdic-
tion under art. 7(2) shall be excluded from taxation in the residence jurisdiction under
art. 7(3) through double taxation relief, avoiding juridical double taxation. Relief is
provided in the form of exclusion of income under the exception method in the OECD
28
The question is how the arm’s length standard in article 7(2) shall be un-
derstood in the context of controlled intangibles transfers. The language
of the provision is from 2010 and reflects the essential building blocks of
modern transfer pricing jurisprudence through its reference to functions,
assets and risks, in contrast to the old-fashioned and obtuse language of ar-
ticle 9. Nevertheless, there is realistically little concrete guidance that can
be deduced from the language of article 7(2). The delineation of its further
content must therefore rely on other sources of law.

the 2010 OECD Report
The main sources of law for interpreting article 7(2) are the OECD Com-
mentaries.87 These largely contain a summary of the 2010 OECD Report.
The Report is a significant document and was a long time coming.88 There
are some points that should be noted with respect to the 2010 OECD Re-
port as a source of law for interpreting article 7(2).
First, its basic approach is to apply the OECD TPG analogically to allocate
operating profits to a PE. This means that the new 2015 text on intangibles
in the OECD TPG is directly relevant for the interpretation of article 7(2).
Second, most of the 2010 OECD Report pertains to the allocation of profits
in three specific sectors: (i) banking; (ii) global trading; and (iii) insurance.
The incurring of financial risk is the hallmark of these sectors, driving
profits. This has also influenced the content of the allocation rules that ap-
ply to other sectors. These latter rules are the ones that are relevant for the
purposes of this book, as they govern the allocation of residual profits from
unique intangibles in the PE context, together with the OECD TPG. The
2010 OECD Report was written in the same pre-financial crisis environ-
ment as the 2009 (previous generation) business restructuring guidance of
the OECD TPG. The new 2017 OECD TPG on intangibles tones down the
importance of financial risk for the purpose of allocating residual profits,
MTC, art. 23A or under the credit method in the OECD MTC, art. 23B. Thus, art. 7
determines the maximum amount of profit allocable to the source jurisdiction, as well
as the maximum amount of profits for which the residence jurisdiction must provide
double taxation relief.
87. For a general discussion of the relevance of the OECD commentaries, see, e.g.
Wittendorff (2010a), at pp. 122-131, with further references.
88. It essentially reiterates the positions taken in the 2008 OECD Report on the At-
tribution of Profits to Permanent Establishments.
29
triggering tension regarding the 2010 OECD Report. The author will revert
to these interpretation issues in his later analysis of the allocation of profits
under article 7.
1.5.6. Case law in connection with articles 9 and 7
There is some relevant case law from various countries with regard to arti-
cles 9 and 7 that will be touched upon in this analysis.89 Examples include
cases on the allocation of profits to asserted commissionaire PEs, as well
as on the application of the TNMM in the context of taxpayer-initiated
compensating adjustments. These cases were decided by domestic courts
for the purpose of determining domestic tax obligations but contain view-
points on articles 7 or 9 that arguably may be relevant to the interpreta-
tion of these articles.90 These cases are, however, pronouncedly fact-driven
and therefore normally readily distinguishable from similar matters. They
may therefore mostly be used as illustrations of practical profit allocation
scenarios. Another point is that this case law generally trails behind the
rapidly evolving profit allocation rules. Some of the problems at issue are
no longer directly relevant in light of the 2015 revision of the OECD TPG.91
1.6. The relevant US sources of law
1.6.1. Introduction
The analysis of both research questions, with respect to US law, depends on

an interpretation of IRC section 482. The author will tie some comments to
the sources of law relevant to this in sections 1.6.2.-1.6.5.
89. For a general discussion on the relevance of case law pertaining to art. 9, see
Wittendorff (2010a), at pp. 143-145, with further references. See also Bullen (2011), at
pp. 64-66, for a discussion on the role of domestic law for the interpretation of art. 9.
For a comprehensive comparative discussion of transfer pricing case law, see Roin
(2011).
90. See Bullen (2011), at p. 12 on the principle of common interpretation.
91. It has been argued that it is a problem that the material content of the arm’s
length standard has not been subject to more delineation through case law; see Schoueri
(2015), at p. 699; and Baistrocchi (2006), at p. 949. The author does not share this con-
cern, as it is difficult for case law to contribute significantly to the clarification of the
content of the arm’s length standard due to the fact-sensitive nature of transfer pricing
cases. Clarification should, first and foremost, be done through the OECD TPG, legal
literature and transfer pricing practice.
30
The relevant US sources of law
1.6.2. IRC section 482
IRC section 482 governs the allocation of operating profits among related
parties under US law.92 It reads as follows:
In any case of two or more organizations, trades, or businesses (whether or
not incorporated, whether or not organized in the United States, and whether
or not affiliated) owned or controlled directly or indirectly by the same inter-
ests, the Secretary may distribute, apportion, or allocate gross income, deduc-
tions, credits, or allowances between or among such organizations, trades,
or businesses, if he determines that such distribution, apportionment, or al-
location is necessary in order to prevent evasion of taxes or clearly to reflect
the income of any of such organizations, trades, or businesses. In the case of
any transfer (or license) of intangible property (within the meaning of section
936(h)(3)(B)), the income with respect to such transfer or license shall be
commensurate with the income attributable to the intangible. For purposes of
this section, the Secretary shall require the valuation of transfers of intangi-
ble property (including intangible property transferred with other property or
services) on an aggregate basis or the valuation of such a transfer on the basis
of the realistic alternatives to such a transfer, if the Secretary determines that
such basis is the most reliable means of valuation of such transfers.93
Unaltered since its introduction in 1928, the first sentence provides the
IRS with the authority to reallocate income among related parties if nec-
essary to prevent “evasion of taxes or clearly to reflect the income”.94 The
language does not indicate the profit allocation standard to be applied to
reallocate income, but the regulations in section 482 of the IRC clarify that
“the standard to be applied in every case is that of a taxpayer dealing at
arm’s length with an uncontrolled taxpayer”.95
This arm’s length standard is, as the point of departure, met if the con-
trolled pricing adheres to the profit allocation rules in the IRC section 482
regulations. This will, for instance, be the case if its results are consistent
92. For an introduction to the Internal Revenue Code (IRC), sec. 482, see, e.g. Wit-
tendorff (2010a), at pp. 24-25 and 55-56; and Levey et al. (2010), at pp. 6-14.
93. On the control criterion of IRC sec. 482, see Gazur (1994).
94. Sec. 45 of US: Revenue Act of 1928 (45 Stat. 806). On the development of IRC
sec. 482 from being an anti-abuse rule to becoming the basis for the current and com-
prehensive US transfer pricing regime, see Brauner (2017), at sec. 2.3.
95. Treas. Regs. § 1.482-1(b)(1). The US arm’s length standard concept has its his-
torical roots in US: War Revenue Act of 1917, ch. 63, 40 Stat. 300 (1917), pursuant to
the regulations of which (Regulation 41, arts. 77-78 (T.D. 2694, 20 Treas. Dec. Int.
Rev. 294, 321 [1918])) the Commissioner was granted the authority to require related
companies to file consolidated returns when “necessary to more equitably determine
the invested capital or taxable income”. Some years later, in 1921, new legislation was
31
with those that would have been realized if uncontrolled taxpayers had
engaged in the same or a comparable transaction under the same or com-
parable circumstances.96 While the regulations are clarifying in linking the
reallocation authority in IRC section 482 to the arm’s length standard, their
reference to uncontrolled transactions provides a somewhat misleading im-
pression of the material content of their current profit allocation rules for
profits from intangibles. As the author will revert to in detail later in the
book (in particular, in chapters 8, 9 and 14), the reality is that the allocation
of intangible profits under IRC section 482, to a considerable degree, does
not rely directly on uncontrolled transactions, but rather on indirect meth-
ods (the one-sided transfer pricing methods, in particular the comparable
profits method), subjective assessments (the residual profit split method)
or the application of valuation techniques premised on the best realistic
alternatives to the controlled transaction (e.g. the income method under the
IRC section 482 cost-sharing regulations).
Further, the second sentence contains the so-called “commensurate with

income standard” for intangibles, which was added to the IRC in 1986.97 It
has two sides: (i) that the profits from controlled multi-period intangibles
transactions must be allocated pursuant to the relative values of the con-
enacted by Congress, allowing the government to require consolidated accounting from

groups “for the purpose of making an accurate distribution or apportionment of gains,
profits, income deductions, or capital between or among such related trades or busi-
ness” (US: Revenue Act of 1921, ch. 136, s. 240(d), 42 Stat. 260 (1921), re-enacted in
US: Revenue Act of 1924, ch. 234, s. 240(d), 43 Stat. 288 (1924); and US: Revenue Act
of 1926, ch. 27, s. 240 (f), 44 Stat. 46 (1926)). In 1928, this provision was replaced by
sec. 45 on the allocation of income and deductions (based on sec. 240(f) of the 1926
Revenue Act) in the Revenue Act of 1928, ch. 852, s. 45, 45 Stat. 806 (1928). The
wording of this provision equals the first sentence of the current IRC sec. 482. The
1928 provision was, in 1934, supplemented by regulations applying the arm’s length
standard (Reg. 86, art. 45 (1935)). The whole regulation is quoted in Essex Broadcast-
ers, Inc. v. CIR, 2 T.C. 523 [Tax Ct., 1943]). These prevailed for almost 35 years, until
the 1968 regulations were issued. The tax code provisions were re-codified in 1939 as
the Internal Revenue Code of 1939. 15 years later, the Internal Revenue Code of 1954
was enacted in the form of a separate code by the Act 16 August 1954, ch. 736, 68A
Stat. 1, changing the lettering and numbering of subtitles, sections, etc. of the 1939
Revenue Code. Several provisions of the Revenue Code were revised in a tax reform in
1976 (4 October 1976, Pub.L. 94-455, Title XIX, § 1906(b)(13)(A), 90 Stat. 1834). The
Tax Reform Act of 1986 changed the name of the 1954 Revenue Code to the Internal
Revenue Code of 1986. See, e.g. Avi-Yonah (1995); and Avi-Yonah (2012) for historical
walk-throughs of the development of the US IRC. The last sentence of the provision
was added in 2017, Pub. L. 115–97, title I, §  14221(b)(2), Dec. 22, 2017, 131 Stat. 2219.
96. Treas. Regs. § 1.482-1(b)(1).
97. US: Tax Reform Act of 1986, Pub. L. No. 99-514, Sec. 1231(e)(1), 100 Stat. 2085,
2562-63 (1986).
32
trolled value chain contributions in each income year; 98 and (ii) that the
controlled ex ante pricing may be adjusted in light of ex post profits, within
certain limits.99 The author will revert to the applications of the commen-
surate with income standard in detail later in the book (in particular, in
chapters 8, 9 and 16).100
The third sentence of section 482 of the IRC was added in the 2017 US
tax reform and codifies the aggregated valuation approach based on the
realistic alternatives of the controlled parties, which the IRS has gradually
developed in practice over the last decades.101
1.6.3. The IRC section 482 US Treasury Regulations
The detailed provisions of the US Treasury Regulations represent the IRS

interpretations of IRC section 482.102 The US Treasury Regulations are, in
practice, the most significant sources of law for the interpretation of sec-
tion 482 of the IRC. Their interpretation will play a key role in the author’s
analysis.
The US Treasury Regulations are generally treated as authoritative. The

Supreme Court of the United States has established a high threshold for
disregarding the Treasury Regulations. As long as Congress, through the
wording of the relevant IRC provision, cannot be deemed to have spoken
directly in respect of the precise tax question at issue (which likely will
be extremely rare in transfer pricing cases due to the brief and ambiguous
wording of IRC section 482) and the regulations thus fill a gap left by the
statues, a court may not substitute its own construction for the reasonable
interpretation of an agency.103
98. This result normally follows directly from a faithful application of the relevant
pricing method (in practice, either the CPM or PSM).
99. This is, in some contexts, a necessary consequence of the first point.
100. It should, however, already at this stage be recognized that the standard is ef-
fectively incorporated into the CPM and PSM, as these methods allocate actual profits.
101. See the discussions in secs. 1.7. and 3.2. of this book, with further references.
102. For a thorough analysis of the historical development of the US regulations, see,
in particular, Culbertson et al. (2003).
103. See Mayo Found. v. US, 562 US (2011), in which the Supreme Court applied the
so-called “Chevron doctrine”; see Chevron USA Inc. v. Natural Resources Defence
Council Inc., 467 US 837. See White (2016), at pp. 214-215, for comments. See also
Altera Corp. & Subsidiaries v. CIR (145 TC No. 3, currently on appeal to the Ninth
Circuit), where the Tax Court held that the Treasury had not engaged in reasoned de-
cision-making with the result that the sec. 482 (2003 cost sharing) regulations at issue
were found to be “arbitrary and capricious and therefore invalid”. See Kantowitz (2018)
33
The core framework of the current US Treasury Regulations was estab-

lished in 1994, when new regulations were issued to replace the 1968 pre-
decessor regulations.104 A significant driving force behind the 1994 regula-
tions was to better address the allocation of residual profits from unique
intangibles.105 Key concepts of the 1994 regulations included the operation-
alization of the commensurate-with-income standard, as manifested in the
new transfer pricing methods (i.e. the CPM and profit split method) and the
periodic adjustment authority. The regulations have been amended several
times since then, most notably through the potent 2009 cost sharing regula-
tions and the 2015 regulations clarifying the application of the arm’s length
standard and the best-method rule.
The basic structure of the Treasury Regulations is as follows:

(1) the first part deals with general transfer pricing issues (the arm’s length
standard, the best-method rule, comparability, etc.);106
(2) the second part discusses the allocation of profits in specific situations
(e.g. loans);107
(3) the third part discusses the allocation of profits for tangibles;108
(4) the fourth part contains a general discussion of the allocation of profits
for intangibles;109
for comments on Altera. Without going into depth on the Altera ruling, the issue of
whether the applicable cost sharing regulations (which required stock-based compen-
sation to be treated as a cost and shared among the CSA participants) were compliant
with the arm’s length standard in IRC sec. 482 was, in effect, treated as a question of
whether it could be documented (by the IRS) that third parties – in so-called “com-
parable” CSAs – also shared such equity-based compensation costs. The author finds
this entire reasoning puzzling, as it must be regarded as rather obvious that the typical
IP-tax-planning CSAs, which have been so popular among US-based multinationals,
in fact have no real third-party comparables (see fn. 1986, as well as fn. 1294 of the
Xilinx ruling on the Xilinx ruling). It was thus not possible for the IRS to offer any such
third-party evidence. However – as always in transfer pricing – the lack of third-party
comparables cannot entail that the profits from controlled transactions should not be
allocated among group entities in an arm’s length manner (based on, e.g. the realistic
alternatives of the parties).
104. The US section 482 regulations have developed as follows: the 1968 regulations
were published in the Federal Register (33 FR 5848) on 16 April 1968. Subsequent
revisions and updates of the transfer pricing regulations were published in the Federal
Register on 8 July 1994 (59 FR 34971), 20 December 1995 (60 FR 65553), 13 May 1996
(61 FR 21955), 26 August 2003 (68 FR 51171), 4 August 2009 (74 FR 38830), 22 De-
cember 2011 (76 FR 80082) and 27 August 2013 (78 FR 52854).
105. See US: Notice 88-123 (White Paper).
106. Treas. Regs. § 1.482-1.
107. Treas. Regs. § 1.482-2.
108. Treas. Regs. § 1.482-3.
109. Treas. Regs. § 1.482-4.
34
(5) the fifth part discusses the CPM;110

(6) the sixth part discusses the profit split method;111
(7) the seventh part discusses the allocation of profits for cost sharing
agreements;112
(8) the eight part contains examples of the best-method rule;113 and
(9) the ninth (and last) part discusses the allocation of profits for intra-
group services.114
Apart from the second and third parts, all of these provisions are relevant
for the analysis of the research questions.
When the author refers to the US Treasury Regulations, he refers to the

current regulations unless otherwise specifically stated. As he will refer to
a range of historical US regulations throughout the book (including pro-
posed, temporary and final regulations from all relevant revision rounds),
he has included detailed references to these texts in the References.115
1.6.4. Case law
There is relatively comprehensive case law from US courts on the alloca-

tion of intangible profits under section 482 of the Treasury Regulations.
Most is decided by the US Tax Court. Some cases are appealed, but the
lion’s share is not, due to the significant costs and time required to litigate
comprehensive transfer pricing cases and the inherent unpredictability of
outcomes caused by their fact-sensitive nature.116 The author is not aware
of any rulings by the US Supreme Court on the core profit allocation rules
analysed in this book.
This case law is generally binding on both taxpayers and the IRS with
regard to the interpretation of IRC section 482, absent of changes in the
law or regulations.117 Due to the fact-driven nature of these cases, however,
it is often possible to distinguish the facts of a pending case from those of
110. Treas. Regs. § 1.482-5.

111. Treas. Regs. § 1.482-6.
112. Treas. Regs. § 1.482-7.
113. Treas. Regs. § 1.482-8.
114. Treas. Regs. § 1.482-9.
115. See p. 777 et seq.
116. For critical reflections on the lack of ability of transfer pricing case law to pro-
vide guidance for future cases, see Baistrocchi (2004a); and Baistrocchi (2006).
117. Andrus et al. (2012), at p. 4.
35
a court case. The general impression in the literature is that transfer pric-
ing case law now plays a less important role in relation to the US Treasury
Regulations than historically has been the case, due to the comprehensive
and up-to-date nature of the Regulations.118
Further, case law generally trails behind the development of the regula-
tions. For instance, a reassessment may uncover shortcomings in the regu-
lations, leading the IRS to revise them. A subsequent court ruling on the
reassessment will then not be directly relevant for new cases, but may pro-
vide some analogical value. This was the situation in Veritas119 and the
recent Amazon.com case120 pertaining to the determination of a buy-in
amount under the predecessor of the current cost sharing regulations. The
IRS argued that a so-called “income method”, which was first codified in
the new 2009 regulations, should be applied. Even if the IRS argument
only had limited support in the then-applicable 1995 regulations, its mate-
rial content is similar to that of the current income method. Thus, even if
the matter at issue was governed by the previous cost-sharing regulations,
the rulings contributed to the interpretation of the current rules. Also, the
amounts involved in some cases may justify litigation for the IRS even if
the ruling will have no precedent value for future cases. This was the case
in Amazon.com, argued before the Tax Court in 2017, essentially retrying
Veritas.
1.6.5. The OECD TPG
US transfer pricing assessments are founded in section 482 of the IRC, its
regulations and case law. The IRS is nevertheless of the opinion that US
transfer pricing law is generally aligned with the OECD TPG.121 The IRS
competent authority accepts the OECD TPG as a basis for dispute resolu-
tion under treaty mutual agreement provisions.122 Historically, the percep-
tion has been that the material transfer pricing positions taken in the US
Treasury Regulations differ from those of the OECD TPG with respect
to intangibles, intra-group services and cost sharing arrangements.123 As
118. Id. See also Wittendorff (2010a), at p. 23.

119. Veritas Software Corporation & Subsidiaries v. CIR, 133 T.C. No. 14 (Tax Ct.,
2009), IRS nonacquiescence in AOD-2010-05, 12 Nov. 2010. The ruling is analysed in
sec. 14.2.4. of this book.
120. 148 TC No. 8; see the analysis in sec. 14.2.5. of this book.
121. See Bullen (2011), at pp. 40-41; and Wittendorff (2010a), at p. 28 on this topic.
122. See US: Rev. Proc. 2006-54.
123. Andrus et al. (2012), at p. 4. The US regulations have no direct parallel to OECD
TPG ch. 9 on business restructurings.
36
A few words on the 2017 US tax reform
the author will revert to in detail later throughout the book, there are now
pronounced similarities between the systems.
In any case, the OECD TPG will generally not be directly relevant for the
interpretation of IRC section 482. Apart from the obvious fact that the in-
terpretation of domestic law principally depend on domestic sources of law
(in this case, the regulations in IRC section 482 and case law), it should be
recognized that the design of the OECD TPG differs from that of the regu-
lations. The OECD TPG often contain rather brief and ambiguous com-
ments on specific transfer pricing issues, while the regulations generally
offer comprehensive, detailed and instructive guidance with elaborate nu-
merical examples. It should, however, be noted that the 2017 OECD TPG
go significantly in the direction of the US Treasury Regulations in this
respect. The current OECD TPG on intangibles now offer comprehensive
examples, clearly inspired by the US Treasury Regulations. Nevertheless,
due to the general difference in design, it may be difficult in practice to de-
rive meaningful interpretation contributions from the OECD TPG. There
are also notable material differences in the approaches of the two regimes
to particular transfer pricing issues, which may render the OECD TPG un-
fit to contribute to the interpretation of IRC section 482.124 In the unlikely
case that it should be possible to identify a direct conflict between the US
Treasury Regulations and the OECD TPG, the former will of course pre-
vail for the purpose of interpreting domestic US tax law.
1.7. A few words on the 2017 US tax reform
The US enacted a comprehensive federal tax reform in the end of Decem-

ber 2017 through the Tax Cuts and Jobs Act.125 As this reform came just be-
fore this manuscript was to be delivered for publication, there was no time
for a thorough analysis. A basic overview of the significance of the reform
for the discussions contained in this book is, however, provided below.
In general, the 2017 US tax reform introduced only two, but powerful,
direct amendments to the US transfer pricing rules.126
124. An example includes the US and OECD approaches to the allocation of profits
from internally developed manufacturing of IP in general, as analysed in chs. 21 and
22, respectively, and the approaches to the remuneration of R&D-funding contributions
in particular (see sec. 22.4.7.).
125. US: Pub. L. No. 115-97, 22 Dec. 2017. The provisions discussed below in this
section will have effect from and including 2018.
126. See the description in the proposal JCX-51-17, at pp. 236-237.
37
First, the intangibles definition referred to in IRC section 482 (and section
367) has been changed so that it now also encompasses residual intangibles
(goodwill, going concern value and workforce in place).127 This revision is
significant and will bring an end to years of controversy pertaining to the
interpretation of the previous version of the definition (see the discussion
in section 3.2.).
Second, a new sentence has been incorporated into section 482 of the IRC
clarifying that the IRS may, going forward, base its transfer pricing assess-
ments on aggregated valuations and on the realistic alternatives principle.
This amendment is closely linked to the IP definition revision and is also
significant. As the author will revert to in detail later in the book, the IRS
has already based significant reassessments on this aggregated valuation
approach, but has done so without a clear basis in IRC section 482 or the
US Treasury Regulations and has lost repeatedly in court.128
These two legislative amendments entail that the US IRC now has “caught
up” with the transfer pricing doctrine developed by the IRS through prac-
tice and ensure that the IRS going forward will have a crystal clear basis in
the law on which to further develop its CPM (see the discussion in chapter
8) and income method-based (see section 14.2.8.3.) aggregated valuation
approach. In particular, the legislative codification of the realistic alter-
natives pricing principle will likely lead to further refinement of the IRS
valuation approach through future regulations. Until such regulations are
issued, there will be a solid basis in the current regulations for ascertain-
ing the material content of the aggregated valuation approach based on the
realistic alternatives of the controlled parties.129 These two 2017 amend-
ments to the law should be seen as nothing less than a true victory for the
IRS and a recognition of the work that it has performed for decades with
regard to advancing transfer pricing doctrine in general, and IP valuation
doctrine in particular.
Further, in addition to these two direct amendments to US transfer pricing

law, the 2017 tax reform also introduced three sets of rules (FDII, GILTI
and BEAT, as discussed below) that, in practice, will be highly significant
127. See IRC sec. 936(h)(3)(B)(2)(vi) and (vii).

128. See the analysis of the Veritas and Amazon.com rulings in secs. 14.2.4. and
14.2.5., respectively.
129. See the discussion in sec. 6.7.2. on the 2015 expansion of the US aggregation
provisions; sec. 14.2.3. on the 2007 IRS Coordinated Issue Paper (CIP) valuation ap-
proach; sec. 14.2.8.3. on the income method; and secs. 14.2.4. and 14.2.5. of the Veritas
and Amazon.com Tax Court rulings, respectively.
38
A few words on the 2017 US tax reform
with respect to US taxation of multinationals, but are not transfer pricing

rules as such. These new rules do not determine how profits from intra-
group transactions shall be calculated and allocated among group entities
(transfer pricing), but pertain to how certain items of income shall be taxed
in the United States. These rules come in addition to the US transfer pric-
ing regime and have important interactions with the regime, as the income
calculated under these three new sets of rules can be influenced by transfer
pricing. The author will provide a brief overview of these additional three
sets of rules.
First, the 2017 US tax reform introduced what in effect can be seen as a
preferential regime for IP income derived from export activities (i.e. for
non-US IP income) through the special deduction set out in section 250
of the IRC for foreign-derived intangible income (FDII).130 This provision
entails, in a somewhat simplified manner, that excess IP returns (above a
normal market return) earned directly by a US corporation from its foreign
sales (including licensing of IP) can be deducted from its US income.131
Under the provision, 37.5% of the IP income may be deducted, resulting in
effective US taxation of the foreign-source IP income of only 13.125%.132
This US IP box provision will likely encompass IP income from both man-
ufacturing intangibles that have been developed through R&D carried out
abroad as well as from marketing intangibles and may therefore arguably
be in conflict with the OECD’s modified nexus approach for IP regimes
(discussed in section 2.5.).133
Second, the tax reform introduced an add-on to the existing US controlled

foreign corporation (CFC) rules (the subpart F rules) in IRC section 951A,
which requires current-year inclusion of global intangible low-taxed in-
130. See the description in the proposal JCX-51-17, at p. 229.

131. See IRC sec. 250(b) on foreign-derived intangible income.
132. Which is the ordinary tax rate of 21% multiplied by the remaining part of the IP
income (1-0.375). The deduction will be phased down from 37.5% to 21.875% from and
including 2026, which will entail an effective tax rate of 16.4%.
133. It should be noted that “excess IP returns” under the new provision are not the
same as residual profits from unique and valuable IP in a transfer pricing context. “Ex-
cess IP returns” may encompass also a range of other income items, as the profits,
somewhat simplified, can be said to be determined as income minus a normal market
return on tangible assets. The amount of income subject to the foreign-derived intan-
gible income (FDII) provision can be influenced through transfer pricing (e.g. through
increasing the royalty rate on IP licensing from a US parent to the controlled foreign
corporation (CFC) in question or through charging for services (from the United States)
through a transfer pricing method that includes a profit margin (as opposed to the ser-
vices cost method, or SCM), thereby reducing the profits of the CFC).
39
come (GILTI) by US shareholders.134 GILTI entails (somewhat simplified)

excess foreign IP returns above a normal market return.135 This rule en-
sures that US tax on such IP income cannot be deferred until the profits are
distributed to the US shareholder (but must be taxed on a current basis).
As opposed to the FDII rule described above, which applies to foreign IP
income earned directly by a US company, the GILTI-rule applies to foreign
IP income earned through a CFC. GILTI is subject to tax at the ordinary
US rate of 21%, but in the period of 2018-2025, a 50% GILTI deduction
is allowed, making the effective tax rate on such income 10.5%.136 This
GILTI rule results in a US CFC regime that is likely more burdensome
than the CFC rules of the key trading partners of the United States and thus
represents a comparative disadvantage for US investors.
Third, the tax reform introduced a tax on “base erosion payments” car-
ried out by large corporate taxpayers,137 i.e. the so-called “base erosion
and anti-abuse tax” (BEAT) in IRC section 59A.138 The stated purpose of
this tax is to level the playing field between US and foreign-headquartered
parent companies.139 The provision effectively denies deductions for a por-
tion of outbound payments (including amounts paid for the acquisition of
depreciable property) made to foreign related entities for which US deduc-
tions are available.140 The BEAT rule is a significant provision, both with
134. See the description in the proposal JCX-51-17, at p. 227. This add-on was nec-
essary in order for the CFC rules to catch active global intangible low-taxed income
(GILTI) (as opposed to the passive income otherwise encompassed by the Subpart F
CFC rules). See also initial reflections on ambiguous aspects of the GILTI provisions
in Stevens et al. (2018).
135. IRC sec. 951A(b)(1). This provision encompasses more income than what is
normally deemed to be IP income in a transfer pricing context (residual profits). The
amount of income subject to the GILTI provision can be influenced through transfer
pricing arrangements with other group entities.
136. IRC sec. 250(a)(1)(B).
137. The provision applies to domestic US companies that are not taxed on a flow-
through basis, that are part of a group with a minimum of USD 500 million in revenues
and that have a “base erosion percentage” (essentially, the aggregate amount of base
erosion tax benefits divided by the aggregate amount of total deductions for the tax-
payer) of a minimum of 3%.
138. See the description in the proposal JCX-51-17, at p. 242. It can also be noted
(as a side matter) that the new 2016 US Model Tax Convention denies treaty benefits
for deductible related-party payments of mobile income in cases where the receiving
beneficial owner pays little or no tax on the income due to a “special tax regime”. See
Brauner (2017), at sec. 2.2.1.2 on this issue.
139. See the report entitled Unified framework for fixing our broken tax code, re-
leased by Republican leadership, dated 27 September 2017, at p. 9.
140. Important exceptions to the base erosion and anti-abuse tax (BEAT) provision
are available for payments of: (i) amounts for services that qualify for pricing under
the services cost method of the US IRC sec. 482 regulations and that reflect the total
40
The relationship between the book and other transfer pricing literature
regard to its high degree of complexity and potential positive US revenue

effect. It is also controversial. As it denies deductions for payments made
to foreign entities where similar payments made to similarly situated do-
mestic US entities would remain fully deductible, it will be problematic in
the context of tax treaty non-discrimination provisions. The rule may also
be in conflict with the World Trade Organization trade law obligations of
the United States.
1.8. The relationship between the book and other

transfer pricing literature
The legal analysis in this book is based on interpretations of the primary
legal sources.141 Transfer pricing literature is referred to where it provides
meaningful contributions to these interpretations. The author’s experience
is that while there certainly is a significant body of international literature
on general transfer pricing issues with regard to US and OECD law,142 the
literature is, unfortunately, a bit more limited when it comes to the specific
problems analysed in this book, i.e. the profit allocation sides of the US and
OECD transfer pricing provisions for IP in particular.143
With regard to books that, in whole or in part, deal with transfer pricing of
intangibles under US and OECD law, the author’s experience is that these
are generally limited to high-level discussions of (now) mostly historical
US and OECD provisions. These books therefore provide somewhat mod-
cost of the services without any mark-up; and (ii) qualified derivative payments for
taxpayers that annually recognize ordinary gains or losses on such instruments (subject
to exceptions). Further, costs of goods sold seem to fall outside the scope of the BEAT
provision. Transfer pricing arrangements may affect the BEAT, in particular the choice
of pricing methods to remunerate services; e.g. if more than one method (including the
SCM) can be used to price a particular service, it could be strategic to choose the SCM
in order to fall under the exception.
141. I.e. IRC sec. 482 and the accompanying Treasury Regulations with respect to the
United States, and articles 9 and 7 of the OECD MTC, the OECD Commentaries and
the OECD TPG with respect to OECD law.
142. E.g. literature covering discussions of the arm’s length principle versus formu-
lary apportionment, intra-group financing, mutual agreement procedure issues, ad-
vance pricing agreement issues, documentation issues, general reflections on transfer
pricing methodology, etc.
143. In other words, the analysis of how these transfer pricing provisions dictate how
the total profits from an intangible value chain must be split into normal return profits,
incremental operating profits from local cost savings, market conditions and synergies
and residual profits from unique intangibles and allocated among the group entities that
contribute the relevant value chain inputs.
41
est contributions to a more detailed analysis of the current US and OECD

provisions. For instance, Markham’s 2005 book takes a broad approach
to the subject matter,144 as does the book written by Levey and Wrappe.145
Wittendorff’s book touches on a range of relevant issues but does not go
into depth on profit allocation problems for IP.146 The book written by Boos
is primarily an economic analysis.147 The 2007 IFA report on transfer pric-
ing of IP has been useful, but offers only high-level discussions. These pub-
lications are now partly outdated.148 There are also “practical manuals” on
transfer pricing of IP, but these are typically high-level overviews produced
by big audit firms and are not written as academic analyses.
With regard to articles, the author has, through extensive searches of,
among others, the IBFD, Tax Notes International and BNA Transfer Pric-
ing archives, identified a wide range of highly relevant publications on
transfer pricing of IP that are referred to in his discussions.
It is the author’s assertion that he has identified and referred to the most
significant contributions in international (English) legal literature that
deals with the profit allocation sides of the US and OECD transfer pricing
provisions for IP. Nevertheless, it must be said that the amount of relevant
literature that he has identified is more sparse than one could expect. He
attributes this to two factors.
First, there has been a historically unprecedented and anti-BEPS-driven

development in US and OECD transfer pricing law in recent years, which
has resulted in significant revisions of the law, altering historical profit al-
location positions.149 With regard to the United States, the final versions of
144. Markham (2005). In addition to the transfer pricing of intangibles, Markham’s

book also deals with a range of additional issues, such as documentation requirements,
APAs and formulary apportionment.
145. Levey et al. (2010).
146. Wittendorff (2010a).
147. Boos (2003).
148. IFA (2007).
149. See, for instance, Wittendorff (2016), at p. 331, where it is stated that “its final
report on BEPS actions 8-10 … involves significant changes to the OECD’s transfer
pricing guidelines and the birth of a new arm’s-length principle that replaces the tradi-
tional principle adopted by the League of Nations in 1933”. See also Lang et al. (2016),
preamble, where it is stated that “the work of the OECD on Transfer Pricing, Alloca-
tion of Risk and Recharacterization (Actions 8, 9 and 10 of the BEPS Action Plan) has
developed proposals for a substantial revision of the OECD Transfer Pricing Guidelines
with regard to a number of closely related topics which are ‘at the heart of the arm’s
length principle’”. See also, in this direction, Pankiv (2017), at pp. 209-210.
42
The relationship between the book and other transfer pricing literature
(i) the intra-group service regulations were issued in 2009;150 (ii) the cost
sharing regulations were issued in 2011;151 and (iii) the new regulations on
outbound transactions and coordination with the best-method rule were
issued in 2015 (finalized in 2016). The December 2017 tax reform has also
had an impact on US profit allocation through codification in IRC sec-
tion 482 of the aggregated valuation principle based on the best realistic
alternatives of the controlled parties (as well as amending the US IP defini-
tion for transfer pricing purposes).152 With respect to the OECD, the 2017
TPG contain revised language on key profit allocation aspects through new
chapters on intangibles, intra-group services, cost sharing arrangements
and business restructurings. Important parts of chapter I of the OECD TPG
(on allocation of risk, local market conditions, etc.) have also been signifi-
cantly revised. As the material differences between the 1995/2010 OECD
TPG and 2017 OECD TPG provisions on many points run deep with re-
spect to profit allocation for controlled IP transfers, legal literature analys-
ing the historical 1995/2010 OECD TPG is not always capable of providing
meaningful contributions to the interpretation of the current 2017 IP profit
allocation provisions.
Certain aspects of the 2017 OECD TPG rules do not even have any direct
parallels in the 1995/2010 versions of the OECD TPG, thereby rendering
legal literature that analyses the latter provisions unable to provide any
clear contributions to the author’s interpretation of the relevant OECD pro-
visions. This is the case, for instance, with respect to the new guidance
on profit allocation to R&D funding,153 IP valuation,154 local cost savings
and market conditions and synergies,155 application of the transfer pricing
methods in the context of intangibles,156 periodic adjustments (the guidance
for which is now linked to the new language on valuation and hard-to-value
intangibles) 157 and guidance directly aimed at restricting profit allocation
alternatives for cost sharing arrangements.158 It can also be mentioned that,
as a first for the OECD TPG (and clearly inspired by the US regulatory
style), the OECD has provided a range of detailed examples delineating the
150. These regulations contain key provisions for determining IP ownership under
US transfer pricing law; see the analyses in ch. 21 and ch. 23.
151. See the analysis of the current US cost sharing regulations in sec. 14.2.
152. See IRC sec. 482, at fn. 129 on aggregated valuation and the discussion in sec.
3.2. of the revised US IP definition.
153. See the discussion in sec. 22.4.
154. See the discussion in ch. 13.
155. See ch. 10.
156. See ch. 11.
157. See sec. 16.5.
158. See sec. 14.3.
43
more specific content of the 2017 profit allocation rules applied to intangi-
ble value chains. Parts of the OECD revision work is also still ongoing, e.g.
the revision of the profit split method guidance.159
This book is likely among the first academic legal texts that analyse these
new provisions in depth.
Second, this thesis takes a (somewhat at least) original analytical approach.

First, the author places his profit allocation analysis into a general frame-
work, where he establishes a link between the concepts of unique (or non-
routine) and non-unique (or routine) value chain contributions on the one
side,160 and the allocation of residual profits and normal market returns
on the other side.161 This link grounds his analysis of the profit allocation
consequences of the US and OECD transfer pricing provisions in economic
concepts and reality and is used as a red thread throughout the entire book.
Also, he places his analysis of the US and OECD IP ownership provisions
into a more specific framework that distinguishes between manufacturing
and marketing IP.162 He does so because there are important differences
between these groups of IP that significantly affect how they should be
analysed for transfer pricing purposes. This book also identifies and analy-
ses the link between profit allocation under cost sharing provisions and
the IP ownership rules.163 It is the author’s view that both frameworks have
clear analytical value, and he is not familiar with any prior legal literature
applying similar frameworks in this comprehensive manner.
This book also offers some new perspectives on the historical US and
OECD rules. First, the author provides a relatively detailed analysis of the
profit allocation sides of the historical rules in order to gauge how the pro-
visions dictated that intangible profits should be allocated among juris-
dictions, thus enabling meaningful comparisons with the profit allocation
positions taken in the current US and OECD rules. This historical analysis
deepens the understanding of the context in which the current rules were
introduced and contributes to their interpretation. It is also the author’s
impression that several of the discussions in this thesis of “old” (but still
current) US and OECD rules should be able to contribute to new insights.
An example is his in-depth analysis of the OECD transfer pricing compa-
rability requirements with respect to whether third-party “blended” profit

160. On this classification, see sec. 3.4.3.
161. See the discussions in secs. 1.2.-1.3.
162. On this classification, see sec. 3.4.2. and the analysis in ch. 20.
163. See sec. 20.5.
44
Reference register and source abbreviations
data is legally acceptable for the purpose of benchmarking controlled profit

allocations under the one-sided transfer pricing methods.164 the author is
not familiar with any similar analyses in other legal literature.
1.9. Reference register and source abbreviations
Throughout the book, the author applies abbreviations for the most fre-
quently used legal sources (OECD TPG, US regulations, etc.) in order to
avoid unnecessary repetition. This should be particularly useful with re-
spect to the many different versions of the US regulations referred to in
the book.
164. See sec. 8.6.
45
Chapter 2
Business and Tax Motivations for

Intangible Value Chain Structures
2.1. Introduction
The research questions analysed in this book would not have been relevant
had enterprises not engaged in foreign direct investment (FDI). The OECD
defines FDI as a “category of cross-border investment made by a resident
in one economy (the direct investor) with the objective of establishing a
lasting interest in an enterprise (the direct investment enterprise) that is
resident in an economy other than that of the direct investor”.165
A multinational is characterized by having activities that are geographi-

cally dispersed throughout different countries and internalized, as op-
posed to being outsourced to local third parties.166 It is the internali-
zation of activities (i.e. FDI) that triggers the need for transfer pricing
rules, as the allocation of operating profits among jurisdictions would be
governed by the supply-and-demand market force – and thus be at arm’s
length – if enterprises, instead of carrying out FDI, had chosen to out-
source foreign activities to local third parties with conflicting economic
interests.
In this chapter, the author will provide an overview of the economic find-
ings as to which factors are relevant when an enterprise decides whether to
organize its business operations in another country through FDI or third-
party outsourcing167 – in other words, why an enterprise may choose to
become a multinational. These findings are relevant to the later analysis of
the profit allocation rules, as the alternative of third-party outsourcing is
165. OECD Benchmark Definition of Foreign Direct Investment (4th ed., 2008), para.
11.
166. A seminal, economical, analytical framework for understanding this second as-
pect is the so-called “OLI-framework”, which states that an enterprise will decide to
invest abroad if it (i) has market power given by the ownership of products or produc-
tion processes (O); (ii) has a location advantage in locating its plant to a foreign, rather
than its home, country (L); and (iii) has an advantage from internalizing its foreign
activities in fully owned subsidiaries rather than carrying them out through third-party
arm’s length market transactions (I). See Dunning (1977).
167. These discussions are based on findings referred in Navaretti et al. (2004), at
pp. 23-48; and Dunning et al. (2008), at pp. 116-144.
47
Chapter 2 - Business and Tax Motivations for Intangible Value Chain
Structures
a recurring argument in transfer pricing discussions. Horizontal and verti-

cal FDI are discussed in section 2.2. The trade-offs present in the choice
between FDI and outsourcing are discussed in section 2.3.
After having discussed the business reasons for why multinationals organ-
ize themselves as they do, the author turns to a discussion of how large,
intangible property (IP)-driven multinationals typically structure their
value chains for tax purposes. The so-called “centralized principal model”
is introduced in section 2.4. The author will refer actively to this model
throughout the book in his analysis of the US and OECD profit allocation
provisions.
The chapter will end with a discussion of the new OECD approach towards
IP regimes – the so-called “modified nexus approach” – in section 2.5. The
OECD IP regime rules have an interesting relationship to the 2017 OECD
TPG on IP ownership (which are analysed in part 3 of the book), and thus
form a valuable backdrop for the later discussions.
2.2. Horizontal and vertical FDI
The main motives for FDI are access to foreign markets and the reduction
of manufacturing costs. A local presence avoids transportation costs and
trade barriers that might be incurred through imports, and it better enables
the enterprise to tailor its products to local preferences and to respond to
changes in local market conditions. It may also affect the interaction with
local competitors.168 The costs of primary inputs vary across locations, and
access to low-cost inputs is a significant reason as to why an enterprise may
choose to geographically disperse its business activities.169
168. The equilibrium prices and sales volumes of firms typically depend on the mar-
ginal costs of all firms supplying the market. An enterprise that supplies the market
through a local presence will save on trade costs relative to the costs that it would have
incurred through imports, thus reducing the marginal cost of supplying the market.
The firm with the lower marginal cost may expand, and the sales volumes and prices of
competing firms may be reduced.
169. However, this argument is not absolute. First, the factor prices must be adjusted
for the quality of the factor input. Empirical evidence shows that foreign direct invest-
ment (FDI) seldom goes to the lowest-wage economies, but instead goes to countries
where labour with basic education is available. Research and development (R&D)-in-
tensive activities will generally take place in countries where scientists are relatively
abundant. Factor costs will be relatively more important for earlier stages of the value
chain.
48
Horizontal and vertical FDI
Geographical dispersion, however, gives rise to costs in the form of forgone

economies of scale. An enterprise that carries out an integrated production
process in its home country will have to duplicate its activities if it expands
its operations to other countries. It will, logically, not choose to duplicate
its entire activities in another country if it owns “firm-level” intangibles
(patents, trademarks, know-how, goodwill, etc.) that generate firm-level in-
creasing returns to scale. Once developed, they can be spread throughout
the organization at no cost, i.e. without being duplicated. Thus, an enter-
prise will only duplicate specific business activities, such as manufacturing
and sales.
This is horizontal FDI, where the same stage of the value chain is du-
plicated in two or more countries.170 Some economies of scale will then
be relinquished, but these are tied to particular levels in the value chain
– so-called “plant-level economies of scale” – as opposed to firm-level.
FDI is likely to occur when there are high firm-scale economies combined
with relatively low plant-scale economies. The main economic trade-off
in horizontal FDI is increased sales versus foregone economies of scale.
Horizontal FDI is likely to be drawn to locations with good market access
and where sales are large enough to cover plant-level fixed costs, and are
likely to occur when there are high transportation costs associated with the
final goods.
Enterprises can alternatively split their value chains across countries with
so-called “vertical FDI”;171 e.g. production of a specific component is
placed in a foreign subsidiary. This results in a loss of economies of inte-
gration, giving rise to trade costs (freight expenses, import duties, etc.).172
The main economic trade-off in vertical FDI is lower input costs versus
increased trade costs. Vertical FDI will be drawn to locations with lower
factor costs and efficient transportation and trade links and is expected to
occur when trade and disintegration costs are low.
170. On this topic, see also Boos (2003), at pp. 40-41; Brauner (2008), at p. 92; and
Schön (2010a), at p. 228.
171. There is not always a bright line between horizontal and vertical FDI. For in-
stance, a typical horizontal FDI (e.g. assembly plant duplication) will result in loss of
economies of scale. However, if parts must be shipped to the plant, there will also be
disintegration costs. Further, if a vertical FDI does not involve 100% of the production
of the relevant component, there will be some degree of duplication.
172. Trade costs may be substantial, for instance, with freight charges ranging from
5-28%.
49
Structures
2.3. To stay home (outsource) or to go out (FDI)?
The question in this section is that of what influences an enterprise’s deci-

sion to internalize its geographically dispersed business activities through
horizontal or vertical FDI as opposed to outsourcing them to local third
parties. In other words, why do enterprises become multinationals?173
Internalization of business activities entails a cost penalty, as the enterprise

alternatively could have contracted with the lowest-cost local third-party
supplier to perform the relevant activities.174 Local third parties would like-
ly possess better information on regional conditions (labour skills, prod-
uct demand and administrative procedures), be capable of producing more
cheaply, have specialized expertise within the relevant activity and have
significant plant-level economies of scale, in the sense that expanding its
production would come at a lower cost than setting up a new plant owned
by the multinational.
If the enterprise alternatively was to outsource its business activities to

local third parties, this could trigger three types of costs. First, potential
costs from what economists refer to as incomplete contracts and special-
ized investments, i.e. the so-called “hold-up problem”.175 The logic is that
the local third-party supplier may fear that, after having made the neces-
sary investments to produce inputs for the multinational, it may be denied
payments due to some contingency, resulting in a renegotiation scenario. If
its investment is specific to the relationship with the multinational and has
no alternative use, its bargaining position will be weak. The point is that
this fear may lead the third party to invest in a sub-optimal manner ex ante,
which causes inefficiency that reduces the multinational’s return from out-
sourcing. Second, there are potential costs resulting from the depreciation
of firm-specific assets.176 If the knowledge embodied in intangibles that are
exploited in local production is vulnerable to appropriation, an unrelated
licensee may choose to compete with the multinational after having gained
access to the information and experience in applying it.177 This problem is
pronounced for manufacturing intangibles (e.g. technical know-how), but
173. On this topic, see, in particular, Francescucci (2004a), at p. 57. On multinational

enterprise (MNE) value chains, see also Roberge (2013), at p. 225.
174. See Navaretti et al. (2004), at p. 35.
175. See Navaretti et al. (2004), at p. 36; and Schmitz (2001).
177. To prevent this, the multinational has two alternatives: it can either (i) share
some of its super profits with the licensee by charging a low royalty rate; or (ii) keep its
activities internal through FDI instead of outsourcing.
50
To stay home (outsource) or to go out (FDI)?
also relevant for marketing intangibles (e.g. trademarks).178 Third, there

may be informational asymmetries due to the geographical distance be-
tween the head office of the multinational and local third parties, giving
rise to agency costs connected to monitoring and motivating the employees
of the local parties, as their actions cannot be perfectly observed.179 The
enterprise may avoid these three types of costs if it instead structures its
foreign business through FDI.
In conclusion, and particularly relevant for the purposes of this book, an

enterprise will risk reducing the value of its unique intangibles if it out-
sources its activities to third parties, regardless of whether the outsourcing
is horizontal or vertical. The local third parties may appropriate the knowl-
edge embodied in manufacturing intangibles and start to compete (imita-
tion problem) or dilute the value of marketing intangibles (free riding).180
If the enterprise in light of this chooses FDI, the major costs it will incur
will be due to it not using the comparative advantages of a local specialized
agent (input supplier, assembler or distributor), resulting in an increase in
the average costs for supplying the final product.181 Thus, multinationals
that choose FDI are likely to own unique and valuable intangibles, from
which firm-level economies of scale originate, i.e. that generate residual
profits.182 Further, a key reason as to why enterprises choose to carry out
FDI is to control and protect such intangibles, which constitute their com-
petitive advantages.183
The implication of the above is that there are pronounced business reasons
as to why an enterprise should choose to become a multinational, apart
from tax planning. This is an important recognition in the current BEPS
178. A multinational may incur costs in the latter context if the foreign licensee lacks
sufficient incentives to maintain the multinational’s goodwill. The licensee will be-
come a “free rider” if it reaps the benefits of the multinational’s trademarks without
contributing to the maintenance of its value. Again, the multinational is faced with
the same two alternatives: either (i) share some of its super profits with the unrelated
licensee; or (ii) internalize local activities through FDI.
180. In this direction, see also Schön (2010a), at p. 246; and Parekh (2015), at p. 302.
On the economics of franchising, see Blair et al. (2005).
181. Establishing a subsidiary will also entail fixed costs, which can be avoided
through third-party outsourcing.
182. See also, in this direction, Purnell (1992); and Mogle (2001).
183. It should also be noted that some of the largest multinationals are extractive
companies (oil, gas and other natural resources) and carry out FDI to benefit from
locational advantages. They must go where the resources are geographically located.
They reap super profits due to their monopolistic access to resources, not due to their
intangibles.
51
Structures
climate. Tax incentives are important, but their significance should not be
exaggerated. The costs and benefits of FDI discussed above are, however,
only relevant where the enterprise actually carries out substantive opera-
tional value chain activities abroad. Conversely, if a multinational simply
establishes a foreign “post box” entity without business substance (e.g. to
hold legal ownership of group intangibles), and perhaps even “cash box”
entities (e.g. for research and development (R&D) financing purposes), the
costs and benefits discussed above are not really relevant. In these highly
practical scenarios, it is, in the author’s view, realistic to assume that the
structure of the value chain is tax motivated.
2.4. The centralized principal model for profit allocation
It should generally be assumed that multinationals, as rational economic

actors, seek to minimize their global effective tax rates.184 They may do
so by allocating their operating profits to jurisdictions where the profits
will be subject to the least amount of taxation. A model for attaining
this result is the so-called “centralized principal model”, which has been
popular at least since the wave of international business restructurings
began in the mid-1990s.185 The author will refer to the model throughout
the book.
184. Empirical evidence supports this assumption; see, in particular, OECD, Measur-
ing and Monitoring BEPS – Action 11: 2015 Final Report (OECD 2015), International
Organizations’ Documentation IBFD. See also, e.g. Bauer et al. (2013).
185. See also the discussion of the principal model in the 2010 Joint Committee on
Taxation Report, JCX-37-10, at pp. 16-17, which finds that US taxpayers, already prior
to the 1986 tax reform, were engaged in business restructurings aimed at foreign, cen-
tralized ownership of intangible property (IP), typically trough cost sharing arrange-
ments (CSAs). See also Brauner (2014a), at p. 97; Zollo (2011), at p. 776; Benshalom
(2013), at p. 445; Pankiv (2016), at p. 470, Schön (2014), at p. 4; and Fedusiv (2016), at
p. 483 on the centralization of IP rights; as well as Wu (2005); Huibregtse et al. (2011);
Zammit (2015), at p. 542; Wilkie (2016), at p. 66; and Pankiv (2017), at p. 158. See
Musselli et al. (2008a); and Musselli et al. (2008b) on risk-stripping of local entities;
as well as Bullen (2010), at p. 624. For further information on business restructurings,
see Lemein (2005); Hill et al. (2007); Hill et al. (2008); Schatan (2008); Ihli (2008);
Bakker (2009); Morgan et al. (2009); Ainsworth et al. (2010); Ainsworth et al. (2011);
IFA (2011); Andrew et al. (2012); and Chakravarty et al. (2013). On centralized “hub”
structures, see Bilaney (2017). From the German perspective, see Beck (2008). With
respect to tax structures from a US perspective, see, in particular, Kleinbard (2011);
Kleinbard (2012a); Kleinbard (2012b); as well as Eden (1998). On value chain struc-
turing, see Brem et al. (2005). On the relationship between IP holding structures and
intellectual property rights (IPR) law, see, e.g. Cowan et al. (2013); and Quiquerez
(2013). For an interdisciplinary analysis of firm theory and international tax law, see
52
The centralized principal model for profit allocation
The model is important for understanding the BEPS challenges underlying

the design of the current US and OECD profit allocation rules, in particular
with respect to outbound transfers of R&D-based intangibles developed in
high-tax jurisdictions. It may be illustrated as shown in figure 2.1.
Figure 2.1
Normal return remuneration (cost+resale/TNMM basis)
Routine entity Routine entity Routine entity Routine entity Routine entity
(manufacturing, (manufacturing, (manufacturing, (manufacturing, (manufacturing,
distribution, etc.) distribution, etc.) distribution, etc.) distribution, etc.) distribution, etc.)
Extraction of
residual profits
from source
through transfer
pricing Intangible
ownership
Principal Residual profits
Extraction of
residual profits
from source
through transfer
pricing Routine entity Routine entity Routine entity Routine entity Routine entity
(manufacturing, (manufacturing, (manufacturing, (manufacturing, (manufacturing,
distribution, etc.) distribution, etc.) distribution, etc.) distribution, etc.) distribution, etc.)
Normal return remuneration (cost+resale/TNMM basis)
Figure 2.1 shows a group structure where the unique intangibles, entitled
to the residual profits, are assigned to a single, centralized group entity
(the principal), normally resident in a jurisdiction with no or low tax on
IP income. All other group entities, including manufacturing, R&D and
distribution entities are treated as routine value chain input providers, re-
munerated with a normal market return. In other words, all residual profits
are extracted from the jurisdictions where R&D, manufacturing and sales
take place and injected into the residence jurisdiction of the principal. The
profit allocation profile of the model will only be compatible with the US
and OECD profit allocation provisions (and thus only yield the desired
allocation result for the multinational) if there are no unique intangibles
outside of the principal jurisdiction, as the other group (routine) entities
may then be remunerated as the tested party under a one-sided method
(cost-plus, resale price or transactional net margin method (TNMM)) with
a normal market return.186
Tavares (2016). For IP value chains and the relationship with cost sharing, see Ben-
shalom (2007), at p. 646. For an analysis of Apple’s tax structure and comments on
non-taxation, see Ting (2014).
186. An issue here is the extent to which operating profits can be extracted from local
taxation in source countries through the application of the one-sided transfer pricing
53
Structures
This extraction of residual profits to a low-tax principal jurisdiction is

only possible to achieve if multinationals are able to migrate the unique
intangibles responsible for the residual profits from the high-tax jurisdic-
tions where they were created without triggering taxation of their full
values.187 If done successfully, (at least parts of) the residual profits will
be allocated pursuant to the multinational’s choice as opposed to where
the intangible value was created. As will be discussed extensively later in
the book, the current US and OECD rules seek to prevent this practice, an
ambition that lies at the core of the new US cost-sharing regulations and
the 2017 OECD Transfer Pricing Guidelines (TPG) on intangible owner-
ship.
2.5. IP regimes and the 2015 OECD nexus approach
The US and OECD profit allocation provisions analysed in this book fun-
damentally seek to allocate residual profits from unique intangibles to the
jurisdictions where the intangible value was created. They say nothing
about whether the profits in the end actually should be taxed, but only
which jurisdiction is entitled to tax. Jurisdictions are provided with a piece
of the cake, but are not required to actually eat it. These transfer pricing
provisions are continually revised to prevent BEPS caused by the tax prac-
tices of multinationals and have, in recent years, been supplemented by a
widespread introduction of thin capitalization rules in the domestic laws of
methods, e.g. the transactional net margin method (TNMM). In other words, how low
can the “normal market return” profit allocation to source states can be set, indirectly
determining the size of the residual profits to be extracted from source? This problem
encompasses several key transfer pricing issues, such as the identification of local mar-
keting intangibles and the allocation of incremental operating profits due to local mar-
ket characteristics, cost savings and synergies, which the author will revert to later in
the book: see, in particular, ch. 23 and ch. 24 with respect to the treatment of internally
developed marketing IP under US law and the OECD TPG, respectively; and ch. 10 on
the allocation of profits from local market characteristics, cost savings and synergies
under the OECD TPG.
187. See, e.g. Ballentine (2016), where buy-in pricing under the 1995 US CSA regula-
tions is attributed blame for much of the migration of US developed R&D-based IP at
low prices, resulting in non-arm’s length profit allocations (on this, see also fn. 1276).
See Pankiv (2016), at p. 470, where a “freeze” strategy is suggested as a migration
technique, pursuant to which existing IP is licensed to the centralized IP company
while the future IP is developed at the level of the IP company. In the author’s view, this
technique will be more difficult to apply under the important functions doctrine of the
2017 OECD TPG (see the discussion in sec. 22.3.3.3.) than what was the case under the
1995/2010 TPG.
54
IP regimes and the 2015 OECD nexus approach
many jurisdictions, limiting interest deductibility on multinationals’ intra-

group (and in some cases, also external) debt.188
A range of jurisdictions have, parallel to this development, introduced pref-

erential tax regimes for IP profits, ensuring that residual profits are either
taxed at very low rates or not at all (often called “IP boxes”, “patent boxes”,
etc.).189 Thus, even though the transfer pricing and thin capitalization rules
provide a jurisdiction with cake, it may, pursuant to its IP regime, refuse
to eat it. The aim of such regimes is to attract foreign-based multinationals
with the hope that they will bring some of their economic activities with
them or at least use local services (legal and tax, accounting, financing,
etc.). There is tax competition among jurisdictions to offer the most attrac-
tive IP regimes with the lowest tax rates. If this “race to the bottom” were
to continue without restrictions, one could, in theory, end up with residual
profits not being taxed at all.
The OECD does not advocate the harmonization of income taxes or tax
rates within or outside the OECD. It accepts non-distortive tax competi-
tion.190 Tax competition for geographically mobile activities (e.g. IP-driven
value chains), however, may result in an unfair eroding of the tax base of
other countries (i.e. where the intangible value was created), distort the
location of capital and functions and shift parts of the tax burden to less
mobile tax bases, including labour, property and consumption.191 This par-
ticular form of tax competition is deemed undesirable by the OECD.
188. See IFA (2012) for an overview. Some jurisdictions limit interest deductions
by setting an allowable debt-to-equity ratio (e.g. 2:1 in Italy’s prior thin capitalization
regime, 5:1 in Belgium and 3:1 in several jurisdictions), while others limit interest
deductions by reference to the earnings of the issuer; e.g. Germany’s “interest deduc-
tion ceiling” rule limits interest deductions to 30% of the EBITDA of the issuer, Italy
limits deduction for net interest expenses to 30% of the EBITDA of the issuer, Spain
and Norway adopted similar approaches in 2012 and 2014 respectively and the United
States has thin capitalization rules as a part of their “earnings stripping” provisions;
the allowable interest deduction is calculated by reference to net interest expense to
income.
189. For example, the Netherlands (5% taxation of intangible income), Belgium
(6.8%), Luxembourg (5.72%), Switzerland (8.4%-12%), Ireland (2.5%-12.5%), Malta
(0%-10%), Hungary (5%-9.5%), Cyprus (2%), France (15%) and the United Kingdom
(10%). See also fn. 1934.
190. OECD, Harmful Tax Competition: An emerging global issue, paras. 26-27 and
41 (OECD 1998) [hereinafter the 1998 Report]; and OECD, Agreement on a Modified
Nexus Approach for IP Regimes – Action 5: 2015 Final Report, para. 3 (OECD 2015),
International Organizations’ Documentation IBFD [hereinafter BEPS Action 5: 2015
Final Report].
191. BEPS Action 5: 2015 Final Report, at para. 3.
55
Structures
The 1998 OECD Report on Harmful Tax Competition introduced four

key factors to determine whether a preferential tax regime was potentially
harmful,192 requiring that (i) no or low effective tax rates are imposed on in-
come from geographically mobile activities;193 (ii) the regime is ring-fenced
from the domestic economy;194 (iii) the regime lacks transparency;195 and
(iv) the regime has no effective exchange of information.196 In determining
whether these criteria are fulfilled, it should, inter alia,197 be taken into ac-
count whether the regime encourages operations or arrangements that are
purely tax-driven and involve no substantial activities.198 The report offers
little guidance as to how this latter factor is to be applied.199
Action 5 of the 2013 OECD BEPS Action Plan called for a revision of
the guidance on substantial activity for IP regimes in order to better align
taxation with substance to avoid artificial shifting of profits out of jurisdic-
tions where value is created.200 This resulted in the 2015 nexus approach
consensus among OECD member countries and the G20 countries for the
determination of whether there are substantial activities.201
The nexus approach represents a balancing of the conflicting acknowledge-

ments that preferential IP regimes inherently trigger BEPS concerns, but
also that IP-intensive enterprises facilitate economic growth and employ-
192. OECD, Report on Harmful Tax Competition (OECD 1998) [hereinafter 1998
OECD Report].
193. 1998 OECD Report, para. 61.
197. The OECD 1998 Report lists eight additional factors in paras. 68-79. These are:
(i) an artificial definition of the tax base; (ii) failure to adhere to international transfer
pricing principles; (iii) foreign-source income being exempt from residence country
taxation; (iv) negotiable tax rate or tax base; (v) existence of secrecy provisions; (vi)
access to a wide network of tax treaties; (vii) the regime being promoted as a tax mini-
mization vehicle; and (viii) the regime encouraging operations or arrangements that are
purely tax-driven and involve no substantial activities.
198. For retrospective comments on the 1998 OECD Report, see Avi-Yonah (2009b).
200. See OECD, Action Plan on Base Erosion and Profit Shifting, p. 13 (OECD 2013),
International Organizations’ Documentation IBFD [hereinafter BEPS Action Plan].
201. BEPS Action 5: 2015 Final Report, at para. 28. The nexus approach is inspired
by R&D credits and similar “front-end” tax regimes to incentivize R&D. The OECD
also considered two alternative approaches, namely (i) the value creation approach;
and (ii) the transfer pricing approach, but these did not gain consensus support (para.
27). No new entrants are permitted in any existing IP regime that is not consistent with
the nexus approach as of 30 June 2016. For an interesting analysis of the OECD nexus
approach geared towards the relationship with EU law, see Faulhaber (2017). See also
Zammit (2015), at p. 545, with respect to the EU State aid aspect of IP regimes.
56
ment and that jurisdictions should remain free to offer tax incentives for
R&D, provided that such incentives do not entail harmful effects for other
countries.202 The aim is to ensure that the tax competition with respect to
IP regimes is non-distortive.
The nexus approach establishes the outer boundaries for the permissible
material content of IP regimes. It says that a preferential IP regime may
give favourable tax treatment to the income from an intangible in propor-
tion to the extent to which the taxpayer itself has incurred R&D expenses
to develop it in the jurisdiction where the IP regime is located. The logic is
that the taxpayer’s R&D expenditures indicate to which extent it carries out
substantial activities adding real value in the IP regime jurisdiction. Thus,
R&D expenditures are used as a proxy for substantial activities.
If the taxpayer,203 for instance, has only one intangible and has itself in-
curred all of the expenses to develop it, an IP regime is allowed to bestow
favourable treatment on the entire IP income.204 If the taxpayer instead
has acquired the intangible or outsourced its development to related par-
ties, a portion of the IP income will be denied favourable treatment.205 The
intention is to prevent capital contributions or related-party R&D funding
from qualifying for benefits under an IP regime, as such expenses are not
deemed to be indicative of substantial activity.
The nexus approach is a fraction (nexus ratio) that, when multiplied with
the relevant IP profits,206 yields the amount of profits that may be encom-
passed by the IP regime, as follows:207
a+b
a+b+c+d
202. See BEPS Action 5: 2015 Final Report, at para. 26. There is an interesting eco-
nomic study on the impact of corporate taxes on R&D and patent holdings; see Riedel
et al. (2015).
203. Qualifying taxpayers include resident companies, domestic permanent estab-
lishments (PEs) and foreign PEs of resident companies that are subject to tax in the
jurisdiction providing benefits. See BEPS Action 5: 2015 Final Report, at para. 33.
206. The income to be included is limited to income derived from the IP asset (roy-
alties, capital gains and embedded IP income from the sale of products or the use of
processes directly related to the IP asset). Thus, other income that is unrelated to IP
must be excluded (e.g. manufacturing and marketing returns.) The OECD suggests that
separation could be based on transfer pricing principles.
57
Structures
The elements in the numerator are jointly referred to as “qualifying

expenditures”.208 They are:
(a) R&D expenditures incurred by the taxpayer itself. These must be di-
rectly connected to the IP asset and incurred for the purpose of actual
R&D activities; 209 and
(b) expenditures for unrelated-party outsourcing. Taxpayers that engage
in such outsourcing must nevertheless carry out substantial activities
themselves. The OECD rejects the notion that it will be realistic for
an enterprise to outsource its core R&D to unrelated parties, as that
would entail the relinquishment of a competitive advantage (informa-
tion and value inherent in R&D).210 Ergo, allowing only third-party
outsourcing expenditures as qualifying expenditures ensures that only
income derived from substantive R&D activities carried out by the
taxpayer itself will be entitled to favourable treatment under an IP re-
gime.
The elements in the denominator are jointly referred to as “overall expen-

ditures”. In addition to elements (a) and (b), this includes:
(c) the acquisition costs for IP purchased from unrelated and related
parties; 211 and
(d) the expenditures for related-party outsourcing. Income from acquired
IP and from IP developed by a related party is excluded from favour-
able treatment because the taxpayer has not carried out the R&D itself.
This is necessary in order to prevent distortive effects from an IP re-
gime. It will therefore, for instance, not be possible to “sell in” IP to a
208. Jurisdictions are allowed to permit taxpayers to apply a 30% “uplift” to qualify-
ing expenditures; see BEPS Action 5: 2015 Final Report, at para. 40. The uplift may,
however, only increase the qualifying expenses to the extent that the taxpayer has ac-
quisition or related-party R&D-outsourcing expenditures (the increased amount of ex-
penditures due to the uplift may not exceed these costs). The point of this uplift is sim-
ply to not penalize taxpayers too much for acquiring IP or outsourcing R&D activities
to related parties. The uplift option has been necessary in order to achieve a consensus
of the nexus approach. The overall picture is then that the nexus approach ensures that
taxpayers only receive benefits if they themselves undertook R&D activities and at the
same time acknowledges that taxpayers that acquired IP or outsourced a portion of
the R&D to a related party may still be responsible for much of the value creation that
contributed to the intangible income. See the comments on this conflicting logic below
in this section.
209. Interest expenses, building costs, acquisition costs or any costs that cannot be
directly linked to a specific IP asset are excluded.
211. R&D expenses connected to an acquired IP asset that is carried out subsequent
to the acquisition are deemed ordinary R&D expenses and are thus encompassed by
(a); see BEPS Action 5: 2015 Final Report, at para. 52.
58
central group IP-holding entity,212 regardless of whether the considera-

tion is at arm’s length, or to use a group “cash box” entity to fund R&D
and at the same time achieve favourable treatment of the subsequent
income from the acquired or group-developed IP.
It can be observed that the only way in which the nexus ratio – and thereby
the proportion of IP income entitled to favourable treatment – can be de-
creased from 100% is if the taxpayer acquires the IP or outsources the
R&D to related parties.
A key feature of this OECD approach to IP regimes is the restriction that

they may only encompass income from patents and functionally equivalent
intangibles (qualifying IP assets).213 Preferential IP regimes that grant ben-
efits to income from other assets will not be deemed to meet the substantial
activity requirement and are thus not allowed. This significant restriction
entails that the super profits from a wide range of practically important
unique intangibles (marketing intangibles, e.g. trademarks), goodwill,
know-how, workforce in place, etc.) can never be subject to favourable tax
treatment under a preferential IP regime, but must be taxed in the ordinary
tax regime of the relevant jurisdiction.214 The potential “race to the bottom”
in the form of preferential IP regimes will therefore be limited to the profits
from a highly significant, but ultimately narrow, group of intangibles. This
is aligned with the view that IP regimes may be used to incentivize R&D
activities, a goal that is not realized if the IP regime includes intangibles
that are not based on R&D.
One of the cornerstones of the 2017 OECD TPG on intangible ownership

is to ensure that residual profits from R&D-based intangibles are allocated
to the jurisdictions where the intangible value was created. The OECD
position on preferential IP regimes allows beneficial tax treatment of prof-
its from exactly this group of intangibles. As the author will revert to in
detail,215 the ownership guidance allows, to a certain extent, residual profits
212. There is an incentive to undervalue controlled transfers of IP to a group hold-

ing entity resident in a jurisdiction with an IP regime in order to maximize the pro-
portion of qualifying expenditures. Tax authorities should review such pricing criti-
cally.
213. Such assets must be both legally protected and subject to similar approval and
registration processes where such processes are relevant. These assets include (i) pat-
ents, defined broadly; (ii) copyrighted software; and (iii) in certain circumstances,
other IP assets that are non-obvious, useful and novel.
215. See ch. 22 on the allocation of profits from internally developed manufacturing
IP under the OECD TPG.
59
Structures
to be allocated to a group entity that has outsourced its R&D to related

entities, as long as it remains in control of important R&D functions.216
The nexus approach, however, excludes the profits yielded by intangibles
developed under such controlled agreements from preferential taxation, re-
gardless of whether the outsourcing entity is in control.
Thus, there is a “conflict” of logic behind the OECD approach to intangible

ownership and preferential IP regimes.217 While the ownership guidance
asserts that the outsourcing entity is indeed the entity that creates the in-
tangible value and therefore entitled to the residual profits, as long as it
remains in control of the related-party outsourced functions,218 the nexus
approach rejects the same logic under the view that third parties simply
would not outsource any “core R&D”.219
This conflict resulted in the compromise that the taxpayer, pursuant to

the nexus approach, nevertheless is entitled to an “uplift” of expenses in
the calculation of the nexus ratio.220 The taxpayer is treated as if parts of
its income from (acquired IP and) IP created through related-party R&D
qualify for preferential treatment. The end result is that the lion’s share of
residual profits generated by an intangible created through related-party
R&D outsourcing allocated to a group entity pursuant to the ownership
guidance will be excluded from beneficial treatment under a preferential
IP regime and must therefore be taxed in the ordinary tax regime in the ju-
risdiction where the ownership entity is resident. A smaller portion of these
residual profits, however, may nevertheless be included in the preferential
IP regime due to the “uplift”. The end rationale behind the nexus approach
is therefore a hybrid between the reasoning applied in the ownership guid-
ance and the core idea behind the nexus approach that the qualifying entity
itself must carry out substantial activities in the IP-regime jurisdiction.
216. See the analysis of the OECD important functions doctrine in sec. 22.3.2.
217. Of course, the OECD intangible ownership rules pertain to the allocation of
residual profits to jurisdictions (how the cake is split), while the OECD nexus approach
deals with the question of whether a jurisdiction may establish a preferential regime
for the taxation of the residual profits (whether the piece of cake must be eaten). The
“conflict” of logic pertains to the disparate views taken by these two sets of rules with
respect to the underlying factual (or casual) relationship between related-party out-
sourcing and intangible value creation. Even though the two sets of rules deal with
different issues, there does not seem to be any convincing reason as to why they should
have conflicting views on the same underlying causal relationships.
218. OECD TPG, at paras. 6.56 and 6.58.
220. See the ambiguous justification for the uplift in BEPS Action 5: 2015 Final Re-
port, at para. 41. See also the author’s comments in fn. 208.
60
The author will add that there seems to be potential for practical challenges
in implementing the nexus approach.221 It remains to be seen whether it is
viable.
In conclusion, the nexus approach allows preferential treatment of IP in-

come where real R&D activities are performed within the relevant juris-
diction. It will likely become onerous for multinationals to locate suitable
“principal jurisdictions” that can be used as the residence jurisdiction for
group IP-holding companies for R&D-based intangibles.222 This, in con-
cert with the 2017 OECD TPG on intangible ownership and transfer pric-
ing of intangibles, will likely make it more difficult for multinationals to at-
tain low effective tax rates through structuring their value chains in models
akin to the centralized principal model.
Further, the 2017 OECD TPG on intangible ownership and transfer pricing
of intangibles will allocate super profits to the jurisdiction(s) where the
intangible value was created. These jurisdictions are, under the nexus ap-
proach, allowed to have IP regimes for only a narrow group of R&D-based
intangibles. With respect to the profits from such intangibles, these juris-
dictions may, in theory, lower the rate of taxation to 0%. The super profits
from all other types of unique intangibles, however, must be taxed in the
ordinary tax regimes of these jurisdictions. Thus, for the latter category of
super profits, it seems unlikely that the rate of taxation will end up being
anywhere near as low as the IP regime rate. It therefore seems likely that
a significant portion of the total super profits generated through the intan-
gible value chains of multinationals in the future will be taxed effectively
where the intangible value is created. The severe restrictions on IP regimes
imposed by the nexus approach may, however, incentivize jurisdictions to
drop IP regimes and instead go for ordinary tax regimes with relatively low
tax rates on all business income, akin to that of Ireland.
221. Among the challenges is that qualifying expenditures are to be included in the
nexus ratio calculation at the time at which they occur; see BEPS Action 5: 2015 Final
Report, at para. 45. The expenses must be treated cumulatively, which may be difficult
to keep track of. It may also be challenging to track the relationship between expenses
and the relevant intangible (or derivative product) (see para. 55, and to determine the
scope of expenses to be included in the calculation, with respect to unsuccessful prod-
ucts when a portfolio approach is appropriate to apply (see para. 44).
222. See BEPS Action 5: 2015 Final Report, at p. 63, where there is a table listing the
IP regimes of 16 countries (Belgium, China, Colombia, France, Hungary, Israel, Italy,
Luxembourg, the Netherlands, Portugal, Spain, Spain – Basque Country, Spain – Na-
varra, Switzerland – Canton of Nidwalden, Turkey and the United Kingdom). These
regimes do not fulfil the nexus approach requirements. The respective countries are
now in the process of reviewing their regimes for the purpose of complying with the
nexus approach.
61
Chapter 3
Controlled Intangibles Transfers
3.1. Introduction
The arm’s length transfer pricing standard seeks to ensure parity in the
taxation of related and unrelated parties. If rights to an intangible are trans-
ferred among group entities and third parties would be willing to pay for
such rights, arm’s length compensation must be rendered from the receiv-
ing to the transferring group entity to comply with the standard.
It may, however, be that a jurisdiction exempts certain controlled intangible

transfers from taxation to promote other policies than transfer pricing. Tak-
ing into account that intra-group intangible transfers may be carried out in
different legal forms, spanning contracts such as licence agreements, cost
sharing agreements and business transfer agreements (where the assets
transferred include intangibles), it may be that there are direct exemptions
for specific IP transfers that supersede the transfer pricing requirement, for
instance, in connection with business reorganizations. Further, IP trans-
fers may pertain to a wide variety of different types of intangibles, such
as patents, trademarks, know-how, customer lists, workforce in place and
goodwill. There may be indirect transfer pricing exemptions in the form of
provisions that limit the scope of application of the transfer pricing meth-
odology (the transfer pricing methods and the intangible ownership provi-
sions), typically by way of linking the methodology to a definition that sets
out which intangibles are subject to a transfer pricing charge.223
Tax exemptions for intra-group asset transfers in general, and for valu-
able intangibles in particular, represent roadblocks on the way to arm’s
length pricing. This blocking effect comes into play if the transfer pric-
ing methodology requires arm’s length compensation for an IP transfer but
the transfer nevertheless goes tax-free under direct or indirect exemptions.
This may create incentive effects with respect to the classification and val-
uation of intangibles. If some types of intangibles and forms of IP transfers
are exempt from transfer pricing, there will be pressure from taxpayers to
classify transferred intangibles as tax-exempt and maximize their value, as
223. Most jurisdictions (with the apparent exceptions of China, India, Malaysia,
South Africa and the United States) do not have a definition of intangibles tailored for
transfer pricing purposes; see Rocha (2017), at p. 212.
63
Chapter 3 - Controlled Intangibles Transfers
well as to minimize the value of taxable intangible transfers. Thus, there

will not only be a lack of arm’s length pricing for the exempt intangibles;
taxpayers may also be enticed to misclassify taxable intangibles as non-
taxable, thereby also preventing arm’s length pricing of what, in reality, is
taxable intangible value.
Direct and indirect exemptions from transfer pricing for IP transfers give
rise to lack of neutrality and depart fundamentally from the arm’s length
transfer pricing standard. For the standard to work fully, there must be
absolute neutrality: all intangibles transfers, irrespective of the type of in-
tangible transferred and the legal form in which the transfer occurred, must
be subject to transfer pricing.
The purpose of this chapter is to explore the extent to which the US and
OECD transfer pricing systems are limited by intangibles definitions that
set out the types of intangibles encompassed by the transfer pricing rules,
as well as provisions that regulate the tax treatment of specific transac-
tion types through which intangibles are transferred. The main focus will
be put on the US rules, as these are the most problematic and have raised
significant issues in practice. This approach is also useful with respect to
highlighting the differences between the US and OECD solutions.
The author discusses the US and OECD transfer pricing definitions of in-
tangibles in sections 3.2. and 3.3., respectively. Categorizations of intan-
gibles that are useful for the later analysis of the US and OECD transfer
pricing methodology are discussed in section 3.4. The US and OECD rules
for specific controlled intangible transaction types are discussed in sec-
tions 3.5. and 3.6., respectively, before concluding comments are provided
in section 3.7.
3.2. The US intangibles definition
3.2.1. Introduction
The definition of intangibles that is applied for US transfer pricing purpos-

es is contained in US Internal Revenue Code (IRC) section 936, which gov-
erns Puerto Rico and possession tax credit (the section 936 definition).224
The commensurate-with-income standard in IRC section 482 and section
367(d) refers to this definition.
224. IRC sec. 936(h)(3)(B).
64
The US intangibles definition
An important issue is that the section 936 definition was amended in the
December 2017 US tax reform.225 The amendment does not change the
overall structure or content of the provision, but expands it to encompass
more intangible value. The amendment can be regarded as an add-on to the
pre-2018 version of the intangible property (IP) definition. The amendment
has effect from and including the income year 2018. Thus, all US intra-
group IP transfers carried out up to and including 2017 will be assessed
under the pre-2018 version of the definition, while all transfers carried
out after 2017 will be assessed under the new 2018 version. The pre-2018
version has caused significant interpretation problems in practice and has
been – and currently is – a challenging issue in some of the most notable
transfer pricing cases in US history, e.g. the 2009 Veritas and 2017 Ama-
zon.com Tax Court cases (the latter of which is now appealed to the Ninth
Circuit).226
It must be acknowledged that interpretation of the pre-2018 version of the

US IP definition will continue to be highly relevant in US transfer pric-
ing practice for years to come until all pre-2018 intra-group IP transfers
have been assessed and finalized. As an example of the considerable lag in
the treatment of transfer pricing cases, it can be mentioned that the 2017
Amazon.com case pertains to the income years 2005-2006. Due to the
relevance of the pre-2018 version of the US IP definition for years going
forward, the author will provide an analysis of its interpretation in section
3.2.2. This analysis will also serve as a discussion of the historical develop-
ment of the US IP definition, as well as a contextualization of how the IP
definition relates to the overall US IRC section 482 transfer pricing frame-
work. The amendments made to the IP definition in the 2017 tax reform
will be discussed thereafter, in section 3.2.3.
3.2.2. The pre-2018 version of the US IP definition
3.2.2.1. Introductory comments
The section 936 definition contains 28 specific intangible items across

five categories: “(i) patent, invention, formula, process, design, pattern, or
know-how; (ii) copyright, literary, musical, or artistic composition; (iii)
trademark, trade name, or brand name; (iv) franchise, license, or contract;
225. The content of the amendment was in line with earlier proposals for revision of
the sec. 936 definition. On this, see Brauner (2017), at sec. 2.3.1.2.
226. See the analyses of the rulings in secs. 14.2.4. and 14.2.5., respectively.
65
(v) method, program, system, procedure, campaign, survey, study, forecast,

estimate, customer list, or technical data”.227
In addition to the positively listed items, the pre-2018 version of the defini-
tion also encompasses “any similar item”. Each item must have “substantial
value independent of the services of any individual”. The listed items are
not defined or further elaborated on in IRC section 936. The common and
ordinary use of the terms must therefore apply.228
The definition provides historical continuity. It is virtually identical to the

intangibles definition in the 1968 version of the IRC section 482 regula-
tions.229 The current section 482 regulations also contain a definition of
intangibles, which lists the same items as the section 936 definition.230
The section 936 definition was introduced as part of the Tax Equity and
Fiscal Responsibility Act of 1982231 to ensure that intangibles that gave rise
to US tax deductions for research and development (R&D) expenses dur-
ing their development phases could not be transferred out of the US after
the development had been completed without triggering a tax charge.232
The legislative motivation was thus to match income and deductions in the
United States for the same intangible.
There is a tension between this matching purpose and the broad design of
the section 936 definition. The creation of R&D-based intangibles (e.g. pat-
ents) will normally entail US tax deductions for R&D expenses throughout
the development phase. However, it may be challenging to distinguish be-
tween development and exploitation for some marketing intangibles (e.g.
customer lists and trademarks). If it is not possible to identify a link be-
tween the development phase of US tax deductions and the subsequent
227. The US Internal Revenue Service (IRS) has made clear that contractual rights
to use intangible assets, e.g. a licence agreement, are themselves intangibles; see Treas.
Regs. § 1.482-4(f)(3) and the discussion in sec. 20.2.3.
228. See Amoco Production Co. v. Southern Ute Indian Tribe, 119 S.Ct. 1719 (1999).
229. 33 FR 5848, § 1.482-2(d)(3). The 1968 definition contains 27 of the 28 specific
intangibles listed in the sec. 936 definition (it does not contain the item “know-how”),
spread across the same five categories.
230. Treas. Regs. § 1.482-4(b). The current definition first surfaced in proposed form
in the 1992 regulations, 57 FR 3571-01, § 1.482-2(d)(1)(ii)(A)(8) without – as far as the
author can see – any specific justification. The 1992 definition contained the add-on “any
interests in any such items”, which was removed in the temporary regulations version;
see 58 FR 5263-02, § 1.482-4T(b). On this, see also Odintz et al. (2017), at sec. 2.2.1.3.
231. 96 Stat. 324 (1982), at sec. 213.
232. On the Congressional motivation, see S. Rep. No. 494, 97th Cong., 2d Sess., vol.
1, at pp. 158-159 (1982). For a historical discussion, see Driscoll (1988), at pp. 166-168.
66
exploitation phase of income for marketing intangibles, it is, in principle,

not clear that such intangibles should be encompassed by the definition, as
they may not facilitate the matching of US income and deductions.
The author will not discuss each of the listed items in the 936 definition
separately.233 Such general discussions would not facilitate the later analy-
sis of the profit allocation rules. The reason for this is that the section 936
definition does not inform about the profit allocation that follows from ap-
plying the transfer pricing methods; it may only restrict this allocation.
The question with respect to the intangibles definition is simply whether
the transferred intangible falls within the definition. If so, it must be priced
(and vice versa), as IRC section 482 links its arm’s length transfer pric-
ing requirement for intangibles to the 936 definition. If the intangible is
a patent, for instance, it will fall within the definition and must therefore
be subject to a charge upon transfer. The more particular legal aspects of
the patent will be highly relevant when it later comes to how much should
be charged for its transfer. For instance, the remaining duration of legal
protection may inform the amount of future profits can be expected from
the patent, and may affect the choice of comparables and comparability ad-
justments under the comparable uncontrolled transaction (CUT) method.
However, such discussions belong under the analysis of the transfer pricing
methods, not the 936 intangibles definition. Bluntly put, the question un-
der the transfer pricing methods regards the extent to which an intangible
transfer should be priced, while the question under the 936 definition is
simply whether the transfer should be priced.
This is not to say that there has not been controversy surrounding the inter-
pretation of the pre-2018 version of the 936 definition. The interpretation
issues have not, however, pertained to the particular legal aspects of the
listed 936 intangibles, but rather to whether goodwill, going concern value
233. For discussions of sec. 936 intangibles, the author refers to Andrus (2007), at
pp. 632-634; and Wittendorff (2010a), at pp. 596-610. See also DHL Corp. v. CIR,
T.C. Memo. 1998-461 (Tax Ct., 1998), affirmed in part, reversed in part by 285 F.3d
1210 (9th Cir., 2002), where the infrastructure and operating know-how of affiliates in
a package delivery business were deemed to be intangibles. In Hospital Corporation
of America v. CIR, 81 T.C. No. 31 (Tax Ct., 1983), nonacquiescence recommended by
AOD- 1987-22 (IRS AOD, 1987) and Nonacq. 1987 WL 857897 (IRS ACQ, 1987), the
court found that the mere opportunity to enter into a contract did not constitute a sec.
936 intangible. In Merck & Co., Inc. v. U.S., 24 Cl.Ct. 73 (Cl. Ct., 1991), it was found
that the power to determine which entity in a controlled group will earn income does
not qualify as a sec. 936 intangible. In line with the views expressed in these cases, see
also the 2009 OECD guidance on business restructurings in the 2010 OECD TPG, at
para. 9.65.
67
and workforce in place are encompassed by the definition, and thereby also
the commensurate-with-income standard in IRC sections 482 and 367(d),
even though they are not positively listed in the definition. This latter inter-
pretation problem illustrates how the US intangibles definition may serve
as a roadblock for arm’s length pricing. The author analyses the issue in
section 3.2.2.3. In order to properly frame the discussion, however, he will
first further elaborate on the relationship between the 936 definition and
profit allocation in section 3.2.2.2.
3.2.2.2. The relationship between the 936 definition and

profit allocation
The 936 definition is, in principle, decisive as to which controlled inbound

and outbound intangible transfers are subject to a tax charge under US
transfer pricing law. The reason for this is that, as mentioned in section
3.2.1., both IRC section 482 and 367 refer to the definition. Arm’s length
compensation for intra-group intangibles transfers therefore cannot con-
tain compensation for any elements that do not qualify as intangibles under
the definition.
Say, for instance, that five different intangibles are transferred intra-
group together as a package. Assume that the true market value of each
individual intangible is 10, but that the package as a whole – determined
in the aggregate under an applicable transfer pricing method – has an
arm’s length value of 60. If the taxpayer prices the transfer at 50, the tax
authorities may only reassess the pricing to 60 if the residual value of
10 (typically goodwill) qualifies as an intangible under the section 936
definition.
Thus, even though the transfer pricing methods may say that the arm’s
length value indeed should be 60, the definition may restrict the pricing
outcome to 50, given that the 10 in goodwill does not meet the definition.
This restrictive roadblock effect of the intangibles definition has been a
problem for the US tax authorities, particularly in two contexts.
First, problems in this regard have occurred in cost sharing transactions in

which a US group entity makes existing intangibles together with the ex-
perience and skills of an existing team of researchers (workforce in place)
available to a foreign group entity for the purpose of further development
of the existing intangible. In these cases, the value of the package of re-
sources made available by the US entity may far exceed the value of the
68
existing intangible. Nevertheless, in order for IRC sections 482 and 367 to
capture the exceeding value (workforce in place), the intangible must, in
principle, qualify as an intangible under the section 936 definition.
The US Internal Revenue Service (IRS) has asserted that workforce in

place is an intangible under the pre-2018 version of the 936 definition and
thus taxable when transferred in a section 367 or 482 transaction.234 In the
2009 Veritas ruling, however, the US Tax Court rejected the notion that
workforce in place was a compensable intangible in the context of a cost
sharing arrangement (CSA) under the previous generation of cost sharing
regulations.235 Essentially, the same position was taken in the 2017 Ama-
zon.com.236 As the author will come back to, the IRS, in the current 2009
cost sharing regulations and the 2016 section 367 regulations, has, to some
extent, “bypassed” the 936 definition. The cost sharing regulations require
valuation of items such as workforce in place, while the relevance of the
936 definition in the context of section 367 now has been removed. The
result is that residual value elements, such as workforce in place, in practice
will be taxed under the current regulations.
Second, the roadblock effect has been a problem in the context of so-called
“936 exits”, where US-based multinationals with business operations in
Puerto Rico that have benefited from the section 936 tax credit (subject to a
10-year phase-out) reorganize the operations into a controlled foreign cor-
poration (CFC) for US tax purposes, effectively to replace the section 936
234. The IRS position may be influenced by Ithaca Industries Inc. v. CIR (97 T.C.
No. 16 [Tax Ct., 1991], affirmed by 17 F.3d 684 [4th Cir., 1994], certiorari denied
by 513 U.S. 821 [1994]), in which the court found that workforce in place was an in-
tangible under pre-IRC sec. 197 tax law. See also the 1996 IRS issue paper on the
amortization of assembled workforces (96 TNT 49-27), which explained that the fact-
sensitive approach of the US Supreme Court in Newark Morning Ledger Co. v. US
(734 F.Supp. 176 [D.N.J., 1990], reversed by 945 F.2d 555 [3rd Cir., 1991], reversed by
507 U.S. 546 [1993]; see the comments in sec. 3.2.2.4. of this chapter) and Ithaca In-
dustries, Inc. v. CIR (97 T.C. No. 16 [Tax Ct., 1991], affirmed by 17 F.3d 684 [4th Cir.,
1994], certiorari denied by 513 U.S. 821 [1994]) rendered inappropriate the prior IRS
rejection of workforce in place as non-amortizable, reversing the IRS position taken in
a 1991 issue paper (91 TNT 90-35) that workforce in place was part of going concern
value. Further, with respect to goodwill, the IRS has seemingly indirectly admitted that
it is not encompassed by the sec. 936 definition, as the 1992 proposed regulations (58
FR 5310, 5312) asked for comments on whether the definition should be expanded to
also encompass goodwill. On this point, see also Odintz et al. (2017), at sec. 2.2.2.2.
235. 133 T.C. No. 14 (U.S. Tax Ct. 2009, IRS nonacquiescence in AOD-2010-05); see
footnote 31 in the ruling, as well as the author’s analysis of the ruling in sec. 14.2.4. See
also Zollo (2010), at p. 73.
236. 148 TC No. 8.
69
tax credit with deferral benefits under the subpart F rules.237 In these con-
versions, the Puerto Rican business operations are contributed to a CFC in
exchange for shares in the CFC or reincorporated as CFCs in non-taxable
reorganizations.238
The transfer pricing issue in 936 exit cases is the determination of the arm’s
length compensation required to be paid by the CFC for the transferred
assets.239 Taxpayers have generally structured these conversions to fall
within the so-called active trade or business (ATB) exception in section
367 (which is discussed in section 3.5.4.). A key design of these structures
is to classify the transferred intangibles as goodwill or going concern value
under the assertion that such items do not qualify as 936 intangibles in
order to escape US taxation.
The 936 exit cases have provided the IRS with an opportunity to relitigate
some of the same pricing issues tried in Veritas outside of the CSA context.
The core of the IRS argument in these cases is that in the end, the CFC
should only earn a routine return on its activities, akin to the profit alloca-
tion under the comparable profits method (CPM), with the residual profits
allocable to the US transferor under IRC section 482 or 367(d).240 So far,
the IRS has been unsuccessful in this litigation. The 2016 Tax Court ruling
in Medtronic v. CIR rejected the IRS claim that the Puerto Rican posses-
sions company had transferred intangibles to the CFC in the conversion.241
The 2017 Tax Court ruling in Eaton Corp v. CIR rejected the assertion
237. See the 2007 Industry Director Directive No. 1 on section 936 Exit Strategies
(LMSB-04-01-07-002) and audit guidelines on exit strategies (Notice 2005-21). The
2007 directive supplements the IRS position on aggregated valuation of cost-sharing
contributions under the pre-2009 regulations, as described in the 2007 Coordinated
Issue Paper, which provides guidance on buy-in issues (LMSB-04-0907-62, since with-
drawn). A second directive was issued in 2008 (LMSB-04-0108-001), emphasizing
that intangible transfers in sec. 936 exits may be taxable under sec. 367(d) and that
workforce in place should not be treated as part of foreign goodwill or going concern
value.
238. See IRC sec. 368(a)(1)(F).
239. As taxpayers may claim foreign tax credit for Puerto Rican withholding tax on
royalties, the IRS reassessments are generally based on price adjustments on goods.
240. This stand is similar to the contract manufacturer theory professed by the IRS
in the 1970s and 1980s (see the analyses of Bausch & Lomb Inc. v. CIR., 92 T.C. No.
33 [Tax Ct., 1989], affirmed by 933 F.2d 1084 [2nd Cir., 1991] and Sundstrand Corp.
v. CIR, 96 T.C. 226 [Tax Ct., 1991] in secs. 5.2.5.2. and 5.2.5.3., respectively). Earlier
case law is, however, distinguished due to the elevated position that the realistically
available alternatives principle has gained in pricing methodology under the current
regulations, as well as the commensurate-with-income standard.
241. T.C. Memo. 2016-112. See Gupta (2016) for comments in relation to the IRS’s
attempt to apply the commensurate-with-income standard in the case.
70
that US intangibles had been transferred to two Cayman Islands-registered

group entities with branch operations in Puerto Rico, in spite of the fact
that the two foreign entities earned super profits that indicated ownership
of unique intangibles (as opposed to mere licensing positions that should
earn only a normal CPM market return).242 It remains to be seen whether
the IRS will convince the court in other pending 936 exit cases.243
The IRS’s concern is that taxpayers have not been sufficiently thorough
in identifying and valuing all section 936 intangibles transferred to for-
eign corporations in section 367 transactions. The same concern applies
for CSA transactions under the pre-2009 cost sharing regulations. This is
the reason why the current cost sharing regulations require buy-in pay-
ments for all intangible development contributions, including workforce
in place. These concerns are closely connected to the IRS’s dissatisfaction
with the “bottom-up” valuation approaches normally taken by taxpayers,
under which transferred intangibles are valued separately. This stands in
contrast to the “top-down” approach adopted by the IRS (codified in the
last sentence of IRC section 482 in the 2017 tax reform),244 pursuant to
which transferred assets shall be valued in the aggregate and may include a
residual amount (goodwill, going concern value and workforce in place).245
3.2.2.3. Are goodwill, going concern value and workforce in place

encompassed by the pre-2018 version of the 936 definition?
Rights in intangibles are often transferred intra-group on a stand-alone ba-

sis, e.g. licensing or sales transactions pertaining to a specific patent or
242. TC Memo 2017-147.

243. Other sec. 936 exit cases include Eaton Corp. v. CIR (Tax Court Docket No.
5576-12); Boston Scientific Corporation v. CIR (Tax Court Docket No. 26876-11); Gui-
dant LLC v. CIR (Tax Court Docket No. 5989-11); Cardiac Pacemakers, Inc. v. CIR
(Tax Court Docket No. 5990-11); and Medtronic, Inc. v. CIR (Tax Court Docket No.
6944-11).
244. As a way of background, see JCX-51-17, at p. 236.
245. The IRS valuation approach is described in the 2007 CIP (LMSB-04-0907-62).
The CIP was withdrawn, but the IRS nevertheless remained dedicated to the described
valuation approach for pre-2009 cases (as amply illustrated in Veritas v. CIR (133 TC
No. 14) and Amazon.com Inc. v. CIR (148 TC No. 8)) and as codified in the new 2015
regulations clarifying the coordination of the application of the arm’s length standard
and the best-method rule under IRC sec. 482 with other IRC provision 80 FR 55538-01
(see, in particular, § 1.482-1T(a)(i)(A)), requiring that all value provided in controlled
transactions be compensated at arm’s length, and § 1.482-1T(a)(i)(B) on aggregation,
requiring that the combined effect of two or more transactions (including under multi-
ple IRC or regulations provisions) must be considered if that would yield a more reliable
arm’s length profit allocation).
71
trademark. In these cases, it will normally be clear that the transferred

intangible is encompassed by the 936 definition and therefore available for
an arm’s length charge under IRC section 482 or 367(d).
In some cases, however, it may be that intangibles are transferred as part of

a larger package of assets. Practical examples include the transfer of ongo-
ing R&D, patents and workforce in place as contributions to a section 482
cost sharing agreement, or intangibles transferred as part of a larger section
367 business restructuring. In these cases, the pricing should be founded
on an aggregated valuation of the entire package of transferred assets on
a going concern basis.246 This may yield a higher price than the sum of
the individual asset values transferred, resulting in a residual amount. The
residual may be caused by a range of conditions, such as the utility to the
buyer of integrating the purchased asset package into his existing business,
economies of scale of using the assets as a group, or customer preferences
that can be exploited only if the assets are used together. The residual will
typically be classified either as goodwill, going concern value or the value
of a workforce in place.
The treatment of such residuals in US outbound transfers has been subject

to considerable debate.247 Multinationals have historically claimed that the
residual falls outside the pre-2018 version of the 936 definition, thus escap-
ing a tax charge under IRC section 482 or 367, while the IRS has asserted
that the value should be allocated to one of the specified section 936 intan-
gibles, and therefore subject to tax.248
The point of departure is that goodwill, going concern value and work-
force in place are not listed in the pre-2018 version of the 936 definition
as separate items. The US tax code recognized goodwill, going concern
value and workforce in place as concepts for other tax purposes (such as
depreciation) long before the 936 definition was introduced.249 The absence
246. See, in particular, the author’s comments on the 2015 revisions of the regulations
clarifying the interaction between the arm’s length standard and the best-method rule
with other IRC provisions (80 FR 55538-01) in sec. 6.4. The application of aggregated
DCF valuations has also been debated in the context of CSAs; see the analyses of the
2007 CIP (LMSB-04-0907-62) in sec. 14.2.3. and the income method in sec. 14.2.8.3.
247. See, e.g. Zollo (2010).
248. See the author’s comments on the 2015 revision of the IRC sec. 367 regulations
in sec. 3.5.4.
249. See, e.g. Metropolitan Nat. Bank v. St. Louis Dispatch Co., 36 F. 722 (C.C.E.D.Mo.,
1888), affirmed by 149 U.S. 436 (S.Ct., 1893), where goodwill is described as an intangi-
ble of an ongoing publishing business; and U.S. Indus. Alcohol Co. (West Virginia) v.
C.I.R., 42 B.T.A. 1323 (B.T.A., 1940), affirmed in part, reversed in part by 137 F.2d 511
72
of these items as positively listed intangibles in the pre-2018 version of

the 936 definition may thus be read as an indication that Congress did not
intend to tax the value transferred through their migration. Also, Congress
has amended section 367(d) on two occasions since the 1993 enactment of
the depreciation provision in section 197 (which encompasses goodwill) 250
without changing the reference to the 936 definition or the definition it-
self.251 These arguments carry some persuasive power, but it still cannot
be ruled out that the pre-2018 version of the 936 definition encompasses
goodwill, going concern value and workforce in place, as it includes intan-
gibles “similar” to the ones listed. The question is therefore whether these
residual intangible values can be seen as “any similar item”.
As the 28 items listed in the 936 definition span a wide range of different
types of intangibles, it is unclear which specific intangible features should
be guiding for the “similar” determination.252 It has been argued in the lit-
erature that the listed items share two main aspects, i.e. they are susceptible
to (i) separate valuation; and (ii) transfer.253 The author agrees, in principle,
with this interpretation. The question is whether goodwill, going concern
value and workforce in place can be transferred and valued separately.
For the purpose of this discussion, the author will refer to going concern
value and workforce in place as “goodwill”, as these three items in practice
may be comprehensively difficult to separate from each other. The author
therefore sees no convincing reason as to why they should be treated sepa-
rately for the purpose of discussing whether they are encompassed by the
pre-2018 version of the 936 definition.254
(C.C.A.2, 1943), where going concern value is described as the value of the ability to
operate a business without interruption. See also IRC sec. 993(c)(2)(B); and the former
IRC sec. 927(a)(2)(B), which was repealed in 2000 (Pub.L. 106-519, § 2). See Zollo
(2010), at p. 73.
250. Treas. Regs. § 1.197-2(b)(1) defines goodwill as the “value of a trade or business
attributable to the expectancy of continued customer patronage”.
251. Zollo (2010), at p. 73. However, it should be noted, as discussed in sec. 3.5.4., that
the 2015 revision of the IRC sec. 367 regulations has removed the relevance of the sec.
936 definition in the context of IRC sec. 367 transfers.
252. Zollo (2010) argues that, due to the positive nature of the list of specified items in
the sec. 936 definition, any expansion of the list based on similarity to the items listed
should be narrow, with the burden being on the party advocating the expansion to dem-
onstrate that the expansion is correct, with reference to US v. Merrill, 258 F.R.D. 302
(E.D.N.C., 2009), in which the position was that a catch-all provision should be read
narrowly. See also Sinclair (1985).
253. See Blessing (2010a), Part D, 2, a).
254. The long-standing position of the IRS to consider workforce in place as part of
the going concern value of a business was withdrawn in the 2007 Industry Director Di-
73
The primary issue is whether goodwill should be understood to have a

substantive meaning,255 aligned with the “continued customer patronage”,
meaning developed through case law for depreciation purposes.256 If so,
goodwill will at least not necessarily fall outside the scope of the pre-2018
version of the 936 definition, as it may be possible to value and transfer on
a separate basis, and could therefore, in principle, be deemed “similar” to,
for instance, customer lists or trade names, as opposed to simply being the
residual value that surfaces when the total value of the transferred assets
is larger than the aggregated sum of the single asset values. If goodwill is
understood as a residual value, it will fall outside the scope of the pre-2018
version of the 936 definition, as it will not be susceptible to separate valu-
ation and transfer. A residual is, by its very nature, an amount that cannot
be allocated to any of the positively listed intangibles, and is therefore a
negatively-defined concept.
The narrow, substantive IRC section 197 meaning of goodwill developed

for depreciation purposes would not promote the purpose of IRC sections
482 and 367,257 as it would likely restrict the potentially encompassed value
of goodwill to a lesser amount than would typically be included in a re-
sidual goodwill value. This may conflict with the arm’s length allocation of
operating profits under both provisions, as operationalized in the transfer
pricing methods of the section 482 regulations. It is an undisputable fact
rective No. 1 on Section 936 Exit Strategies (LMSB-04-01-07-002). See also the 2005
audit guidelines on sec. 936 exits in Notice 2005-21.
255. See, e.g. Newark Morning Ledger Co. v. US, 734 F.Supp. 176 (D.N.J., 1990),
reversed by 945 F.2d 555 (3rd Cir., 1991), reversed by 507 U.S. 546 (1993) (discussed
below in this section).
256. Early depreciation cases on this issue include Northern Natural Gas Co. v. U.S.,
470 F.2d 1107 (8th Cir., 1973), certiorari denied by 412 U.S. 939 (S.Ct., 1973); Black
Industries v. CIR, 38 T.C.M. 242 (Tax Ct., 1979); Philip Morris Inc. v. CIR, 96 T.C. No.
23 (Tax Ct., 1991), affirmed by 970 F.2d 897 (2nd Cir., 1992); and VGS Corp. v. CIR,
68 T.C. 563 (Tax Ct., 1977), acquiescence recommended by AOD-1978-186 (IRS AOD,
1978) and acq. in 1979 WL 194041 (IRS ACQ, 1979).
257. The 1993 enactment of IRC sec. 197 (as part of the Omnibus Budget Recon-
ciliation Act of 1993 Pub. L. No. 103-66, § 13261[g], 107 Stat. 312 [1993]) made the
distinction between goodwill, going concern value and workforce in place and other
identifiable intangibles, which is irrelevant for depreciation purposes, as both groups of
intangibles (i.e.(i) residual” intangibles (negatively defined IP such as goodwill, going
concern value and workforce in place); and (ii) identifiable intangibles (positively de-
fined intangibles, such as patents, trademarks, etc.) now are depreciable. IRC sec. 197
was enacted to end the stream of disputes pertaining to whether acquired intangibles
could be depreciated. The categories of intangibles qualifying for depreciation were
significantly expanded, but all intangibles (including short-lived) now had to be depre-
ciated over 15 years. Prior to the enactment, intangibles could be depreciated over the
period of their useful lives.
74
that third-party business acquisitions often result in purchase price alloca-

tions where considerable parts of the total purchase price are allocated
to goodwill. In these third-party contexts, goodwill is the residual value
above the sum of the values of the transferred assets. The author sees no
sensible reason as to why goodwill should not have a residual meaning
also in the context of transfer pricing, where the overarching purpose of
the section 482 system is to provide parity in the taxation of related and
unrelated parties.
This chain of reasoning, where the substantive meaning of goodwill is

rejected based on its lack of usefulness in the specific legal context, is
aligned with the methodological approach taken by the US Supreme Court
in the 1993 Supreme Court ruling in Newark Morning Ledger Co. v. US
(Newark).258 At issue was whether parts of the purchase price paid for the
shares in a newspaper publishing company could be allocated to the list
of paid subscribers in the purchased company and depreciated by the pur-
chasing taxpayer on a straight-line basis.259 The majority of the Newark
court found the pre-section 197 substantive definition of goodwill as the
“expectancy of continued patronage” to be of little value for a taxpayer
trying to evaluate whether his intangibles qualify for depreciation deduc-
tions260 and concluded that the value of the customer lists over time dimin-
ished and that these costs should be matched against income.
Also, while there are no general legislative or regulatory clarifications of

what is included in goodwill in the context of section 482 or 367, for the
specific context of section 367(d) transactions, foreign goodwill or going
concern value is defined as “the residual value of a business operation con-
ducted outside of the United States after all other tangible and intangible
assets have been identified and valued”.261
258. Newark Morning Ledger Co. v. US, 734 F.Supp. 176 (D.N.J., 1990), reversed by
945 F.2d 555 (3rd Cir., 1991), reversed by 507 U.S. 546 (1993).
259. IRC sec. 167(a) allows “as a depreciation deduction a reasonable allowance for
the exhaustion, wear and tear (including a reasonable allowance for obsolescence)”.
The regulations allowed depreciation of intangibles if they were used in the business
for only a limited period (see Treas. Regs. § 1.167(a)-(3)). Depreciation was not avail-
able for goodwill. The IRS, since 1927, had consistently held that goodwill was non-
depreciable; see the 1927 IRS Treasury Decision on the amendment of Art. 163 of
Regulations 45 (1920 edition) in T.D. 4055 (IRS TD).
260. See Boe v. CIR, 35 T.C. 720 (Tax Ct., 1961), affirmed by 307 F.2d 339 (9th Cir.,
1962).
261. Treas. Regs. § 1.367(a)-1T(d)(5)(iii). The value of the right to use a corporate
name outside of the United States has, in practice, been treated as foreign goodwill
and going concern value. The definition is now irrelevant, due to the elimination of
75
The author cannot see any clear reason as to why the basic understanding
of the notion of goodwill as a residual value should not also be applicable
in a general interpretation of the pre-2018 version of the 936 definition.
In light of the above, he concludes that goodwill should be seen as the

residual value that remains after all other identifiable section 936 intangi-
bles have been allocated at arm’s length values. Goodwill falls outside the
scope of the pre-2018 version of the 936 definition.262
As mentioned, this roadblock effect of the US intangibles definition has

been problematic for the IRS in the context of CSAs and outbound busi-
ness transfers under section 367. With respect to CSAs, the 2009 cost
sharing regulations effectively override the pre-2018 version of the 936
definition’s exclusion of workforce in place through its “platform contribu-
tions” concept for the determination of the tax charge (buy-in amount).263
These platform contributions are contributions of “any resource, capabil-
ity, or right … that is reasonably anticipated to contribute to developing
cost shared intangibles”, for which an arm’s length consideration must
be charged. The regulations clarify that making available “the exper-
tise and existing integration of [a] research team is a unique resource or
capability”.264
Thus, the regulations effectively require compensation for outbound trans-

fers of workforce in place, even though such intangibles are not encom-
passed by the pre-2018 version of the 936 definition. In defence of this
regulatory “override” of the IRC, it can be argued that it is necessary to
secure an arm’s length profit allocation through the calculation of the buy-
the foreign goodwill exception in the 2015 revision of the sec. 367 regulations; see the
discussions in secs. 3.5.4. and 3.5.5.
262. Some commentators have concluded otherwise; see, e.g. Bowen (2008); and
Zollo (2010). Others, however, have concluded that foreign goodwill and going con-
cern value (in the context of IRC sec. 367) fall within IRC sec. 936(h)(3)(B) property
and are therefore subject to IRC sec. 367(d), not IRC sec. 367(a); see Collins et al.
(2013). Recently issued regulations clarify that all property is subject either to IRC sec.
367(d) or IRC sect. 367(a); see Treas. Regs. § 1.367(a)-7(f)(10) and (11). See also the
2013 regulations on certain outbound property transfers by domestic corporations and
certain stock distributions by domestic corporations (78 FR 17024-01).
263. See Treas. Regs. § 1.482-7(c)(1). Goodwill and going concern value are seldom
or never relevant in the context of a cost-sharing agreement because only resources that
contribute to the development of a particular joint R&D effort are priced through buy-
ins. The relevant “residual” in cost-sharing agreements is therefore captured through
the inclusion of the requirement to price workforce in place. Of course, workforce in
place may be difficult to separate from goodwill and going concern value.
264. Treas. Regs. § 1.482-7(c)(5), Example 2.
76
in amount.265 The arm’s length charge requirement is embodied in IRC

section 482 (and, by reference, also section 367), and is therefore of equal
legal rank as the 936 definition. It may therefore be asserted that the posi-
tion of the regulations is loyal to the overarching legislative policy inten-
tion in that the allocation of profits among related parties should mirror
that between unrelated parties, and that the regulations have done little
more than harmonize two fundamentally conflicting provisions of the IRC.
Second, the IRS has “overridden” the pre-2018 version of the 936 definition
in the context of tax-free non-recognition transactions carried out under
IRC section 367. Such business transfers are typically valued as a going
concern through a discounted cash flow (DCF) valuation, and therefore
often result in goodwill. Under the pre-September 2015 version of the IRC
section 367 regulations, the transfer of foreign goodwill and going concern
value could be carried out tax-free.266 The foreign goodwill exception is
now removed from the section 367 regulations. Outbound transfers of such
residual intangible value will thus now be taxed in spite of section 367(d)
being limited to 936-definition intangibles and of the historical legisla-
tive intention for section 367(a) to permit tax-free conversion of a foreign
branch.267
Even though goodwill transfers now will be taxed under the current regula-
tions, the author finds the exclusion of goodwill from the pre-2018 version of
the 936 definition to be a comprehensively unsatisfactory solution in principle.
First, the development of US goodwill will likely have resulted in concurrent

US tax deductions in the form of normal business costs (e.g. salaries, custom-
er support and marketing campaigns). While it has been asserted that good-
will is not an asset that gives rise to US tax deductions prior to an outbound
265. This is also the IRS’s rationale. The preamble to the final 2011 cost-sharing
regulations (76 FR 80082-01) states that “these regulations do not turn on whether
a given transaction in connection with a CSA involves intangible property within the
meaning of section 936(h)(3)(B), or whether such item has been transferred, licensed,
or retained. Rather, if a controlled participant devotes, in whole or part, any existing
resource, capability, or right to intangible development for the benefit of another con-
trolled participant, whether by transfer or license to the other controlled participant, or
by leveraging such resource, capability, or right within the context of the CSA, then the
regulations require an arm’s length charge for such platform contribution, in addition to
the funding of intangible development costs”.
266. In the author’s view, this was more than was intended. The development ex-
penses for US goodwill and going concern value will likely have been deducted against
current US income. Thus, a tax-free transfer will result in a mismatch in the allocation
of expenses and income connected to the goodwill in the United States.
267. See H.R. Rep. No. 98-432, Pt. 2 (1984), at p. 1320.
77
transfer,268 it is the author’s view that the residual connected to outbound

transfers of US-developed intangibles should be seen as an asset, the devel-
opment costs of which have been subject to deduction in the United States
in the form of ordinary business costs. Goodwill is closely connected to, and
often virtually inseparable from, a range of the intangibles positively listed in
the 936 definition, such as trademarks, trade names and customer lists. This
is amply illustrated by the pre-section 197 depreciation cases. The author
would assert that if goodwill does not give rise to US tax deductions, neither
do these intangibles. To the extent that his observation is justified, US match-
ing of deductions and income will not be achieved if outbound transfers of
goodwill are not encompassed by the 936 definition, contrary to its purpose.
Second, a fundamental problem with the exclusion of goodwill from the

pre-2018 version of the 936 definition – and therefore, in principle, from the
commensurate-with-income standard in IRC sections 482 and 367 – is that
it conflicts with the requirement that there shall be an arm’s length profit al-
location among the related US transferor and the foreign group entity trans-
feree. As third parties clearly pay for goodwill, it seems inappropriate that
the same should not apply to related parties under the arm’s length standard.
Third, the exclusion of goodwill from the pre-2018 version of the 936 defi-
nition introduces significant potential for controversy with respect to clas-
sification and valuation of intangible value, resembling the pre-section 197
depreciation disputes, as well as the historical principal purpose standard
litigation under IRC section 367. Recent taxpayer applications of section
367 have amply underlined this point (see the discussion in section 3.5.4.).
In light of the above, the author finds that the exclusion of goodwill (includ-
ing going concern value and workforce in place) from the pre-2018 version
of the 936 definition is a serious impediment to the US transfer pricing
system, facilitating base erosion.
The Obama administration suggested, in its budget proposals, to include

these items.269 The proposal stated that to “prevent inappropriate shifting
268. Joint Committee on Taxation, General Explanation of the Revenue Provisions of

the Deficit Reduction Act of 1984 (1984 Blue Book), at pp. 428-434. See also Aksakal
et al. (2013), at p. 11; and Blessing (2010a), Part D, 2, a).
269. US Department of the Treasury, General explanations of the Administration’s
Fiscal Year 2010 Revenue Proposals (FY 2010 Greenbook), at p. 32. See also JCX-37-
10, FY 2014 Greenbook, at p. 49, which contains a proposal to tax currently excess
returns associated with transfers of intangibles offshore; and p. 51, which contains a
proposal to limit shifting of income through IP transfers.
78
of income outside the United States … would clarify the definition of in-
tangible property for purposes of sections 367(d) and 482 to include work-
force in place, goodwill and going concern value”.
The proposal is angled from the perspective that the revision will only
clarify existing law, not change it. This is important for the IRS, as it then
would be possible to tax outbound transfers of residual values carried out
in income periods prior to the revision. This would go a long way towards
reversing the effects of the 2009 ruling in Veritas and the 2017 ruling in
Amazon.com, where the Tax Court rejected the assertion that residual in-
tangible value (workforce in place, goodwill and going concern value) was
a taxable intangible under the pre-2018 version of the 936 definition.270
It is the author’s view that a revision of the pre-2018 version of the 936 defini-
tion to also include residual intangibles clearly should be seen as a change to
the law, not a clarification. The amendment will therefore not be relevant for
past income periods.271 Nevertheless, in light of the fact that the 2009 cost
sharing regulations and the 2016 revision of the 367 regulations effectively
render the pre-2018 version of the 936 definition largely irrelevant in CSA
and IRC section 367 transactions, respectively, it must be admitted that an
amendment, in practice, will be a de facto clarification of the law as it now
exists in the mentioned regulations. More precisely, the IRC will then have
“caught up” with the IRS override in the CSA and section 367 regulations.
3.2.2.4. Is goodwill distinguishable from synergy value attributable

to a group of identifiable 936-definition intangibles valued
in the aggregate?
In section 3.2.2.3., the author concluded that goodwill, arising from an

aggregated valuation of intangibles transferred either as a package or as
part of a larger business transfer, should be understood as a residual value
that falls outside the scope of the pre-2018 version of the 936 definition.
Taxpayers thus have an incentive to allocate a minimum amount of the
total transaction value to specified section 936 intangibles and a maximum
amount to goodwill, and the other way around for the IRS.
270. Veritas Software Corporation & Subsidiaries v. CIR (133 T.C. No. 14 [U.S.Tax
Ct. 2009], IRS nonacquiescence in AOD-2010-05; see the author’s analysis of the rul-
ing in sec. 14.2.4.), footnote 31 of the ruling.
271. It is clear that the 2017 tax reform amendment of the sec. 936 definition (dis-
cussed in sec. 3.2.3.) has effect only from and including 2018; see JCX-51-17, at p. 237;
and Pub. L. 115–97, title I, §  14221(c).
79
Likely in an effort to circumvent taxpayer goodwill classification and valu-

ation, the IRS, in a Technical Advice Memorandum (TAM), concluded that
the residual value that is the difference between the total value of a group
of contracts when valued in the aggregate and the sum of the value of each
individual contract should not be seen as goodwill, but rather as a synergis-
tic value allocable to the specified 936-definition intangibles (the portfolio
of contracts).272 The author will tie some comments to this issue, as the
problem is relevant to more general interpretations of the 936 definition.
The taxpayer valuation in the specific case addressed by the TAM allocated
97% of the total value to goodwill based on the assertion that each contract
by itself purportedly had little value. Any incremental value on top of the
sum of the individual contract values was not due to the contracts, but to the
relationship between the distribution agents, constituting goodwill or, alter-
natively, going concern value. The reasoning behind the opposing IRS “syn-
ergy” view was that the transferred contracts (i) prior to the transfer were
used by a single entity as an integrated network; (ii) were transferred as a
single network to a single legal entity; (iii) would, at arm’s length, be valued
as an integrated whole; and (iv) were used as a single, integrated asset.
The IRS claimed support for its view in pre-section 197 depreciation case
law. For instance,273 in Computing and Software, Inc. v. CIR,274 the tax-
payer acquired the assets of three credit-reporting companies, including
files with credit information. Depreciation deductions for the files were
272. See the technical advice memorandum in TAM 200907024, the factual pattern
of which is akin to that in First Data Corp. v. CIR (Tax Court Docket No. 007042-09
[T.C. petition filed 20 Mar. 2009, case settled]), where the taxpayer argued that the IRS
was wrong in deeming a network of foreign agent relationships as specified intangibles
separate from goodwill and going concern value. The taxpayer argument was that the
network had been aggregated into an ongoing business separate and distinct from the
value of the contracts themselves and that the value attributable to the network was
foreign goodwill or going concern value not subject to IRC sec. 367(d).
273. The IRS also claimed support in Kraft Foods Company v. CIR, 21 TC 513 (Tax
Ct., 1954), where the Tax Court held that a group of 31 related patents should have been
valued as a group for depreciation purposes. The case is, in the author’s view, unsuit-
able to serve as an argument for classifying a residual value as “synergy” added to the
sum of the individual values of transferred specified intangibles, as the patents in the
case were not susceptible to stand-alone valuation. Further, the IRS claimed support
in Massey-Ferguson, Inc. v. CIR, 59 T.C. 220 (Tax Ct., 1972), acq., 1973 WL 157476
(IRS ACQ, 1973), a case that did not pertain to the classification of a residual value,
but rather to the treatment of a group of distribution contracts as a single “system”
intangible with respect to the issue of whether the taxpayer was entitled to deductions
for abandonment loss of the distribution contracts in the year in which the last contract
expired. The relevance of the case is therefore, in the author’s view, severely limited.
274. 64 T.C. 223 (Tax Ct., 1975), acq., 1976 WL 175506 (IRS ACQ, 1976).
80
claimed based on an allocation of a substantial part of each purchase price

to their cost basis. The court agreed with this allocation, as it found that
the value allocated to the files was separable from the goodwill and going
concern value of the acquired businesses. Each individual credit file was of
little value. This case, is in the author’s view, is a rather good argument for
classifying a residual value as an enhanced synergy value, as the issue in
the case is analogous to the classification problem under the 936 definition.
Further, the IRS found that the mismatch result of the taxpayer treatment,
i.e. that 97% of the transferred value escaped US taxation through clas-
sification as goodwill, was contrary to the policy intention of IRC section
367(d). The significance of this purpose is pronounced here, as the link
between current US deductions and the development of the distribution
contracts will likely in general be clearer than for goodwill.
The author nevertheless finds the IRS’s arguments flawed. The taxpayer
business valuation resulted in a large goodwill value, but similar purchase
price allocations are not unusual among third parties. Also, the IRS did
not seem to dispute that the value of the individual contracts was modest.
In this lies a crucial implication: that goodwill is the residual value above
the sum of the individual asset values transferred. Thus, when the IRS
characterized the residual value in this case as synergy value attributable
to the portfolio of contracts, this violated the core system of the 936 defini-
tion, which is to only include positively defined intangibles. Goodwill is a
negatively defined intangible. It is, per definition, the amount that cannot
be allocated to specified 936 intangibles.
The logical flaw in the IRS’s argument is precisely that it was possible to
ascertain the value of each individual contract. Had this not been the case
and the contracts could only be valued as a whole, the situation would have
been different. It should then, in the author’s view, be possible to allocate
the entire synergy value to the portfolio of contracts, provided that this
value indeed was determined to stem from the contracts as opposed to
general goodwill. Such an allocation would not violate the system of the
936 definition and would be more aligned with the pre-section 197 depre-
ciation case law.
The fact that, inter alia, the contracts prior to transfer were used by a single
entity as an integrated network are only arguments as to why the busi-
ness transfer should be valued in the aggregate. They provide no direction
with respect to whether the residual amount resulting from such a valuation
should be classified as a synergy value or goodwill. The IRS’s interpreta-
81
tion is illustrative of the intricate and rather artificial constructions that

may arise when taxpayers and the IRS have such pronouncedly different
incentives to classify residual values, in particular for customer-based in-
tangibles that are difficult to distinguish from goodwill.275
In conclusion, there is, in the author’s view, no legal basis to classify a

residual amount from an aggregated valuation of intangibles as a synergy
value.276 The residual amount is, by definition, goodwill, and thus falls out-
side the pre-2018 version of the 936 definition.277
3.2.2.5. Concluding comments on the pre-2018 version of

the US IP definition
Residual intangibles fall outside the pre-2018 version of the 936 defini-
tion. Thus, when goodwill, going concern value or workforce in place
is transferred outbound, there is, in principle, no requirement under the
commensurate-with-income standard in IRC section 482 or 367(d) for a
tax charge. This is detrimental to an arm’s length profit allocation among
related parties under US law, as the transfer of such values undoubtedly
would be compensated by third parties. There is a basic conflict between
275. See also, in this direction, Aksakal et al. (2013), who admits that, in cases in
which grouped intangibles are transferred along with significant tangible or business
assets, there will be tension between valuing the separate sec. 936 intangibles in the
aggregate and the policy intention to exclude foreign goodwill or going concern value
from IRC sec. 367(d).
276. See, conversely, Aksakal (2013), at p. 11. The argument is that excluding going
concern value from the value of a collection of intangibles is not consistent with the
purpose of IRC sec. 367(d) to the extent that the underlying intangibles themselves give
rise to mismatches and that it would be inappropriate to treat the value attributable
to the combination of intangibles differently than the intangibles themselves. It could
perhaps be claimed that Veritas (133 T.C. No. 14), where the court found that the valu-
ation of pre-existing software should focus on the software itself and not on synergies
between the software and other assets (separate valuation approach), stands in contrast
to this aggregated valuation approach. In the author’s view, that is not the case, as the
issue at hand is the classification of the residual value that results from a valuation as
either an add-on value to identifiable sec. 936 intangibles (synergy) or as goodwill, go-
ing concern value or workforce in place.
277. Veritas dealt with the aggregated going concern valuation of a buy-in payment;
see the analysis in sec. 14.2.4. The classification of the residual as a specified intangible
or as goodwill was not an issue in Veritas. Nevertheless, the court noted that the IRS
valuation took into account the value of a workforce in place (described as “access to
research and development team” and “access to marketing team”) when calculating the
buy-in amount and that workforce in place did not qualify as a sec. 936 intangible (see
footnote 31 of the ruling). Had the IRS valuation approach been accepted by the court,
this classification issue would likely have been brought to the forefront.
82
the arm’s length transfer pricing requirement and the fact that the two key
IRC provisions that govern related-party profit allocation, i.e. sections 482
and 367(d), are contingent on the 936 definition.
The author applauds the IRS for alleviating the negative consequences
of this contradictory legislation in the context of CSAs through the 2009
requirement to compensate platform contributions (and thereby also the
value of workforce in place). Further, the DCF-based aggregated cost-
sharing transfer pricing methods are influential in determining the pric-
ing of other section 482 valuation-based transactions through the ap-
plication of unspecified methods. Thus, the cost sharing regulations, in
practice, ensure that the relevant residual intangibles in the context of
section 482 transfers are taken into account and compensated. This IRS
“vigilantism” has also fixed the problem with respect to section 367(d)
transfers. The 2015 narrowing of the scope of the section 367(a) ATB
exception and the removal of the foreign goodwill exception (see the dis-
cussion in section 3.5.4.) has ensured that transfers of residual intangibles
in the context of outbound business transfers must now be compensated
at arm’s length.
It is a problem that these regulations conflict with the pre-2018 version of

the 936 definition. They represent the IRS’s interpretation of the IRC. Tax-
payers could challenge the validity of the 2009 cost sharing regulations and
the 2015 revision of the section 367 regulations before the courts, based
on the assertion that they require compensation that is irreconcilable with
the pre-2018 version of the 936 definition referred to in the commensurate-
with-income standard in IRC section 482 and the first sentence of IRC
section 367(d). The courts would then need to harmonize the IRC section
482 arm’s length profit allocation requirement with the 936 definition. In
the author’s view, the arm’s length requirement should prevail, as there
otherwise would be a significant disparity between the taxation of related
and unrelated parties, which seems comprehensively unlikely to have been
intended by Congress. The compensation of intangible value transferred
intra-group prior to 2018 hinges on the legal formality that the pre-2018
version of the 936 definition must be met, when the focus instead should
have been on whether a transfer of intangible value has occurred that third
parties would have required compensation for. It is unfortunate that the
material profit allocation provisions alone, in the form of the US transfer
pricing methods and intangible ownership rules, do not fully govern the al-
location of operating profits among related parties, but are cluttered by the
constraints of the rather arbitrary 936 definition.
83
The above discussions pertaining to the exclusion of residual intangibles un-

der the pre-2018 version of the US IP definition will not be relevant for intra-
group IP transfers carried out from and including 2018, as the law has now
been amended (see the discussion in section 3.2.3.). As mentioned in section
3.2.1., however, the reality is that this problem will continue to be relevant for
years to come, until the assessment of all pre-2018 controlled IP transactions
have been carried out and finalized. Depending on the outcome of the Ninth
Circuit’s review of the 2017 Tax Court ruling in Amazon.com, there may be
more cases on pre-2018 valuations of outbound IP transfers in which inter-
pretation of the pre-2018 version of the IP definition may become a key issue.
3.2.3. The 2018 version of the US IP definition (the 2017

tax reform amendment)
The 2017 US tax reform finally amended the 936 definition to include resid-
ual intangibles. As mentioned in section 3.2.1., the amendment should best
be seen as an add-on to the pre-2018 version of the definition, as the revision
“does not modify the basic approach of the existing transfer pricing rules
with regard to income from intangible property”.278 The amendment leaves
clauses (i)-(v) of the pre-2018 version of the 936 definition (which lists the 28
specific IP items) untouched,279 but replaces clause (vi) with the following:
(vi) any goodwill, going concern value, or workforce in place (including its
composition and terms and conditions (contractual or otherwise) of its em-
ployment); or
(vii) any other item the value or potential value of which is not attributable to
tangible property or the services of any individual.
The targeted inclusion of residual intangibles in the new clause (vi) brings
an effective end to the long-running discussions with respect to whether the
936 definition encompasses residual intangible value in the form of good-
will, going concern value and workforce in place. Intra-group transfers of
such value may now, from and including 2018, be taxed under IRC sections
367 and 482, with a clear basis in the law.
The revision goes further, however, by also including other items of value
that are not attributable to either tangibles or services. Residual value (i.e.
the total value of the transfer minus the value of tangibles and services)
278. JCX-51-17, at p. 236.

279. This part of the provision is quoted above under sec. 3.2.2.1.
84
The OECD intangibles concept
is, in other words, regarded as value from IP. On this point, the revised
US IP definition is similar to the OECD definition (see the discussion in
section 3.3.). An interesting aspect of this new clause is that it simply
refers to “value” without requiring that this value stems from something
that can be owned or controlled by the transferring group entity, nor that
the value is something that can be benchmarked against comparable un-
controlled transactions (CUTs). The fact that the provision shall be in-
terpreted in this manner must be regarded as relatively certain, in light
of the 2017 codification of the aggregated valuation and realistic alterna-
tives principles in the last sentence of IRC section 482.280 Thus, any intra-
group transfer of value that is not value from tangibles or services will be
encompassed by the commensurate-with-income standard and subject to
aggregated pricing.
Going forward, this legislative development will likely entail that disa-
greements between the IRS and taxpayers with respect to intra-group IP
transfers will go from the traditional classification debates (discussed in
section 3.5.4.) to more pure valuation debates, as there will no longer be
any question as to whether the IP must be priced. The only issue will be the
correct valuation of the IP transfer pricing charge. When viewed together
with the codification in IRC section 482 of the IRS’s aggregated valuation
approach based on the realistic-alternatives pricing principle, it is difficult
to imagine that the IRS in the future will face defeat akin to that which it
incurred in the Tax Court rulings in Veritas and Amazon.com.
3.3. The OECD intangibles concept
Paragraph 6.6 of the 2017 OECD Transfer Pricing Guidelines for Multina-
tional Enterprises and Tax Administrations (OECD TPG) defines an intan-
gible negatively, as follows:
[S]omething which is not a physical asset or a financial asset, which is capable
of being owned or controlled for use in commercial activities, and whose use
or transfer would be compensated had it occurred in a transaction between
independent parties in comparable circumstances.
The broad design of this definition is intended to ensure arm’s length profit
allocation in intra-group intangibles transactions, as well as to avoid the
typical problems associated with defining intangibles for transfer pricing
280. See JCX-51-17, at p. 237, for background on the codification.
85
purposes, in particular misclassification.281 The definition encompasses all

intangibles relevant for transfer pricing. This includes patents,282 know-how
and trade secrets,283 trademarks and trade names,284 rights under contracts
and government licences285 and (other) licences.286 Residual intangible val-
ues, such as goodwill, going concern value and workforce in place, are
also encompassed, even though such items cannot be transferred or valued
separately.287
Further, the definition is disconnected from the concept of royalties under

article 12 of the OECD Model Tax Convention on Income and on Capital
(OECD MTC).288 An intra-group transfer of, for instance, goodwill or go-
ing concern value may require arm’s length compensation for transfer pric-
ing purposes that is nevertheless not regarded as a royalty under article 12
of the OECD MTC. The definition is also not linked to financial account-
ing classifications.289 For instance, even if the R&D costs connected to a
particular intangible are expensed for financial accounting purposes and
therefore not recognized as a balance sheet item, the resulting intangible
may be compensable for transfer pricing purposes.
The most important aspect of the OECD intangibles definition is that it

does not impose any restrictions on the profit allocation that follows from
applying the OECD transfer pricing methodology and the intangible own-
ership rules (analysed in parts 3 and 4 of the book, respectively). Thus, as
opposed to the pre-2018 version of the US definition, it has no roadblock
effect on transfer pricing. This is made clear in paragraph 6.2 of the OECD
TPG, which states that “[t]o the extent that an item or activity conveys eco-
nomic value, it should be taken into account in the determination of arm’s
length prices whether or not it constitutes an intangible within the meaning
of paragraph 6.6”.290
281. See also OECD TPG, at para. 6.5. See the discussion in sec. 3.5.4. for examples
of misclassification under the US regime. On the OECD IP concept, see also Wilkie
(2016), at p. 76. See also Brauner (2016), at p. 105, who notes that the IP definition
pertains to the scope of the OECD IP transfer pricing definitions, but also admits that
the definition does not seem to exclude much in practice.
282. OECD TPG, at para. 6.19.
284. OECD TPG, at paras. 6.21-6.22.
285. OECD TPG, at para. 6.24. These must be distinguished from company registra-
tion obligations, which do not qualify as intangibles under the OECD definition.
287. OECD TPG, at paras. 6.28 and 1.156.
86
The OECD intangibles concept
Thus, the reality is that the definition merely serves as a reference to the
underlying profit allocation rules that operationalize the arm’s length prin-
ciple, i.e. the intangible ownership rules and the transfer pricing methods.
If these profit allocation rules identify an intangible item that must be
compensated to ensure arm’s length profit distribution, the definition will
never intervene. This approach removes any distortion effects that could be
caused by legal formalities.
The broad OECD definition is combined with the requirement that taxpay-
ers and tax authorities must clearly identify the intangibles that purportedly
are transferred and compensable.291 If there is an intangible that is difficult
to classify (e.g. local marketing know-how), taxpayers or tax authorities
could simply “default” to a goodwill classification. As both know-how and
goodwill are encompassed by the definition and, even more importantly,
would be compensated if transferred between independent parties, such
classification issues will not impact profit allocation. The identification re-
quirement may, however, serve to protect taxpayers from reassessments
based on loosely founded claims that vague intangibles have been trans-
ferred, such as local marketing intangibles, without further justification.
The OECD intangibles concept must be distinguished from local market

characteristics (e.g., cost savings and local purchasing power) and group
synergies.292 Such characteristics and synergies may – similarly to unique
intangibles – create incremental profits. However, such characteristics and
synergies are not intangibles and cannot be owned by any particular group
entity. Thus, for profit allocation purposes, the incremental operating prof-
its from such characteristics and synergies must be kept separate from re-
sidual profits generated by unique intangibles and then allocated based on
other principles than those that govern the allocation of residual profits
from unique intangibles. This “segregation” of profits yields an important
result: incremental profits that are caused by local market characteristics
and group synergies cannot automatically be allocated to the group entity
that owns unique intangibles.
This is a new and important recognition by the OECD that, at least in

principle, should affect international profit allocation. The reason for this
is that, under past transfer pricing practices (often based on the TNMM),
incremental operating profits from local market characteristics and syner-
gies, more or less automatically, have been “blended” with the residual

292. OECD TPG, at paras. 6.9, 6.30 and 6.31.
87
profits generated by unique intangibles and then allocated to the foreign

principal group entity that held ownership of the unique intangibles. This
ownership entity – and thereby, its residence jurisdiction – thus received
more profits than it was entitled to in reality. The source jurisdictions,
where the incremental profits arose due to local market characteristics, re-
ceived correspondingly less profits. This “blending” of profits was made
possible due to a lack of a more developed profit allocation doctrine. The
historical practice has been a major concern for source jurisdictions. The
new OECD rules on local market characteristics and synergies are a step
in the right direction towards addressing this concern.293
3.4. Useful distinctions on the intangibles concept
3.4.1. Introduction
Above, the author discussed whether and how the US and OECD IP defini-
tions may restrict arm’s length profit allocation. In this section, the author
will turn the perspective and briefly draw some useful distinctions with
respect to the concept of intangibles that may facilitate the application of
the US and OECD transfer pricing methodologies.
3.4.2. Manufacturing and marketing intangibles
The distinction between manufacturing (or trade) and marketing intan-

gibles represents established transfer pricing terminology.294 There is no
legal importance attached to the use of these labels under US law or the
OECD TPG.295 However, due to their ability to convey relevant transfer
pricing differences among groups of intangibles, the author will employ
them extensively in his later discussions throughout the book.296 The dis-
tinction is also used to structure his analysis of the intangible ownership
293. See further analysis of this topic in ch. 10.

294. See, e.g. Vann (2003), at p. 154. See also Lagarden (2014), at p. 334, for a broader
analysis on the classification of intangibles. On the categorization of IP, see also Wilkie
(2016), at p. 76.
296. Further, the term “hard intangibles” is sometimes used to refer to the most typi-
cal intangibles in the context of transfer pricing, such as patents, trademarks and know-
how, while the term “soft intangibles” is used for goodwill, going concern value and
workforce in place. The author does not find this terminology particularly useful and
will not employ it often in this book.
88
Useful distinctions on the intangibles concept
provisions in part 3 of the book. The author is not aware of any prior legal
analysis that has used this framework to analyse the US and OECD IP
ownership provisions.
Manufacturing intangibles include patents, know-how and similar techni-

cal knowledge that is used in the process of producing goods or services.297
Marketing intangibles encompass trademarks, trade names, customer lists,
etc. that are used in the distribution and sale of products and services.298
There are relevant differences between these two groups of intangibles.299

Manufacturing intangibles tend to be developed through risky, expensive
and often lengthy R&D, reliant on the talents of skilled researchers, and
are typically performed at R&D facilities located in only one or a few ju-
risdictions. The value of marketing intangibles tends to be mainly driven
by the amount of expenses incurred to create and carry out, for example,
marketing campaigns. Successful development of a marketing intangible is
also normally (to some extent at least) contingent on the quality of the un-
derlying product being marketed, and therefore also on the manufacturing
intangible that the product is based on. Marketing strategies and designs
tend to be developed at a centralized level within the group, but the value
of a marketing intangible must be developed in each market in which the
product is sold. Thus, as opposed to manufacturing intangibles, marketing
intangible development will not be contained within a few jurisdictions,
but will rather be carried out in all jurisdictions where the product is mar-
keted and sold.
Another relevant difference is that it will, as touched upon above, normally

be straightforward to distinguish between the development and exploita-
tion of a manufacturing intangible, while this is often not true for market-
297. The OECD TPG glossary defines “manufacturing intangibles” (there denoted as
“trade intangibles”) negatively, as commercial intangibles other than marketing intan-
gibles.
298. The OECD TPG glossary defines “marketing intangibles” as intangibles within
the meaning of “intangibles” in para. 6.6, which relate to marketing activities, aid in the
commercial exploitation of a product or service and/or have an important promotional
value for the product concerned. Depending on the context, marketing intangibles may
include trademarks, trade names, customer lists, customer relationships and proprietary
market and customer data that is used or aids in marketing and selling goods or services
to customers. On marketing intangibles, see, in particular, Roberge (2013), at p. 213;
and Levey et al. (2006).
299. See, e.g. the informed discussion in Brauner (2014a), at p. 100, where he criticiz-
es the OECD for (in the run-up to the 2017 OECD TPG) not putting enough effort into
distinguishing between different groups of IP based on how the value of the respective
IP is created.
89
ing intangibles.300 For instance, a marketing campaign will build the value
of the trademark being marketed while at the same time exploiting it for the
purpose of attaining sales.
Further, marketing intangibles may, in some cases, be difficult to distin-

guish from residual intangibles (e.g. goodwill, going concern value and
workforce in place) and local market characteristics (e.g. local customer
preferences). Certain manufacturing intangibles, such as know-how, may
be difficult to distinguish from the services of, for example, technical per-
sonnel.
3.4.3. Unique and non-unique value chain contributions
Another useful distinction, which both the US and OECD transfer pric-
ing rules fundamentally rely on, is between unique (non-routine) and non-
unique (routine) value chain contributions.
Routine contributions will typically be generic and relatively straightfor-

ward functions (e.g. contract manufacturing, distribution and marketing),
but may also be intangibles. Such routine IP encompasses widely avail-
able technology, production, programming and marketing know-how, etc.
Competition among enterprises that all have access to such technology
will drive prices down to a level at which routine intangibles only will
be capable of generating normal market returns. Thus, a group entity that
contributes a routine intangible to the value chain only contributes to a
normal return. Arm’s length compensation to this entity will therefore also
be limited to a normal return.
In contrast, an enterprise that owns unique intangibles may reap super prof-
its by exploiting them, because competing enterprises will not have access
to equivalent input factors. Therefore, an entity that contributes a unique
intangible to the value chain may contribute to the creation of residual prof-
its. This will, however, only be the case if the unique intangible actually
has contributed to such residual profits. This will not always be the case,
as not all unique intangibles are commercially successful.301 Provided that
there is causality between the unique intangible and residual profits, arm’s
length compensation to the group entity that contributed the unique intan-
gible to the value chain must include the residual profits.
300. On this issue, see, in particular, Roberge (2013), at p. 212.

301. On this point, see also Lagarden (2014), at p. 334.
90
Even though the US and OECD transfer pricing methodologies provide a

lot of leeway to adjust profit margins, it can be said that residual profits in
general represent a significantly greater amount of operating profits than
normal market returns. This means that the distinction has a significant
impact on international profit allocation. A group entity that is deemed to
contribute a unique intangible to the value chain may receive “big” profits,
while a group entity that contributes a routine intangible will only receive
“small” profits. This impact may be described as a threshold or “cliff”
effect, as the issue of whether a group entity shall be allocated residual
profits depends on whether it contributes unique intangibles to the value
chain. The distinction is therefore capable of providing clear direction with
respect to how profits shall be allocated among the entities within a multi-
national, and thus lies at the heart of modern transfer pricing.
The distinction comes into play in two main settings. The first setting is
the process of allocating profits between group entities that contribute dif-
ferent value chain inputs. The distinction has direct effect on the choice
of transfer pricing method to be applied. As the author will come back
to in more detail (see sections 8.3. and 9.2.), the one-sided methods (the
resale price and cost plus methods, as well as the TNMM) can directly
only be used to price non-unique contributions,302 leaving only the CUT
method and the profit split method as available methodologies to directly
price unique contributions.
A basic example of the cliff effect in this setting is as follows: A subsidiary

in France sells a pharmaceutical product that generates 1,000 in profits.
The subsidiary provides local manufacturing, distribution and sales func-
tions. The product is produced under a patent that is licensed from the UK
parent. The unique value chain contribution is the UK patent, while the
non-unique contributions are the French manufacturing, distribution and
sales functions.
Let us say that, based on, for instance, the cost plus method, the profits
attributable to the non-unique contributions are 200. The remaining 800
are super profits caused by the unique value chain contribution, i.e. the UK
302. See the Canadian Tax Court ruling in Alberta Printed Circuits Ltd. v. The Queen
(2011 TCC 232), where the Court rejected a reassessment by the Canadian tax authori-
ties denying deductions claimed by a Canadian company for royalty payments made to
its subsidiary in Barbados based on the view that the Barbados subsidiary was a routine
entity that did not own any unique and valuable IP and thus could be allocated profits
based on the transactional net margin method (TNMM). See also infra n. 917 on this
case.
91
patent. As the French subsidiary is deemed to contribute only routine value

chain contributions, the profits allocable to it are limited to a normal mar-
ket return. The 800 are extracted from taxation in France and allocable to
the United Kingdom, thereby creating a “cliff” compared to the relatively
small profits allocable to France.
The second setting in which the distinction comes into play is in the deter-
mination of intangible ownership, i.e. the process of allocating the residual
profits between group entities that contributed to the development of the
relevant unique intangible from which the residual profits stem. The pro-
vision of relatively unique and high-skill R&D functions, as opposed to
more generic and administrative functions, is generally determinative as
to which group entity shall be assigned the right to residual profits from an
intangible that it has contributed to the development of under the OECD re-
gime, as well as under the transfer pricing methods of the US cost-sharing
regulations, in particular, the income method.303
Imagine that the UK patent in the above example was developed by R&D
staff of a Norwegian subsidiary. The development process was, however,
financed by the UK parent. As the author will revert to in detail later in the
book (see section 22.3.2.), the OECD intangible ownership provisions will
allocate the bulk of the residual profits to the group entity that provided
the most unique and valuable R&D inputs, which, in this example, was
the Norwegian subsidiary. The idea is that such inputs are relatively more
important for the creation of intangible value than more generic functions
and inputs, and thus should attract relatively more profits.
The potentially significant difference in profit allocation, depending on

whether the relevant value chain contribution is deemed to be unique or
non-unique, triggers the question of what constitutes unique contributions
entitled to residual profits versus non-unique contributions only entitled to
normal market returns.
The US profit split method regulations define a non-routine contribu-

tion negatively, as “a contribution that is not accounted for as a routine
contribution”.304 A routine contribution, on the other hand, is described by
focusing on the normal market return profits allocable to it, which will not
include residual profits.
303. Treas. Regs. § 1.482-7(g)(4). Relatively unique and high-level R&D functions
are often provided in combination with the contribution of a pre-existing unique intan-
gible to the CSA for the purpose of further research.
304. Treas. Regs. § 1.482-6(c)(3)(i)(B).
92
The 2017 OECD TPG define “unique and valuable intangibles” as those
“(i) that are not comparable to intangibles used by or available to parties to
potentially comparable transactions, and (ii) whose use in business opera-
tions (e.g. manufacturing, provision of services, marketing, sales or admin-
istration) is expected to yield greater future economic benefits than would
be expected in the absence of the intangible”.305
Similarly to US law, this OECD definition is closely linked to the scope of

application for the profit split method. The definition requirement that there is
no comparable intangible will render the CUT method unavailable to allocate
residual profits, leaving the profit split method as the only available method-
ology capable of directly allocating residual profits. The link between the
distinction and the profit split method is further underlined by the fact that
the OECD introduced this definition alongside its new 2017 preference for
the profit split method as the go-to method for allocating residual profits.306
Both the US and OECD definitions point towards the same, but they start at
opposite ends. The US definition indirectly describes unique contributions
as those that are able to generate residual profits. The OECD definition
describes unique intangibles as those that are not available to competitors
and therefore generate higher profits.
The core point is the same under both US and OECD law: in order for an
intangible to be entitled to residual profits, (i) it must be unavailable for
competing enterprises; and (ii) there must be either actual or likely causal-
ity between the intangible and the creation of residual profits, as not all
exclusive intangibles generate residual profits; only those that are commer-
cially successful do. Only intangibles that satisfy both (i) and (ii) should be
seen as unique value chain contributions under US and OECD law.
The border between routine and non-routine contributions will, in most

cases, likely be clear, as it was in the above example with the UK patent.
Other cases may be less clear. These have the potential to generate signifi-
cant controversy. The stakes are high; the issue is whether a jurisdiction is
entitled to tax residual profits or not.
305. OECD TPG, at para. 6.17. The second part of the definition contains a nonsensi-
cal statement, as the operating profits of any business will be less if one removes a busi-
ness asset and its associated profits. The point must rather be that the business profits
would be less if the intangible instead had also been available to competitors. These
competitors could then offer the same products, resulting in price competition, driving
down prices and eliminating the residual profits.
306. OECD TPG, at paras. 6.138, 6.145 and 6.146.
93
Imagine, for instance, that the French subsidiary in the example above over
time has developed a valuable network of customers and a strong reputa-
tion for delivering quality products. These factors are no doubt exclusive
to the French subsidiary, and it does not seem unreasonable to assume that
they may have contributed to the residual profits from local sales of the
product. French tax authorities may then argue that the marketing network
and reputation of the subsidiary should be seen as unique intangibles and
then apply the profit split method to split the residual profits of 800 between
the UK patent and the French marketing contributions. Say, for the sake of
simplicity, that the split is set at 50/50. Because a unique value chain con-
tribution has been identified on the side of the French subsidiary, it will go
from receiving only 200 in profit as compensation for its routine contribu-
tions to receiving a profit of 600, which also compensates it for its unique
marketing intangibles. The UK tax authorities may oppose this and assert
that the patent is the only unique value chain contribution and claim taxing
rights over the entire residual profits of 800.
Given the potentially significant profit allocation effect of identifying

unique value chain contributions, it is decisive that the functional and fac-
tual analysis of value chain contributions is carried out in a prudent and
thorough manner. On the one side, there must be a relatively high threshold
for classifying a value chain contribution as non-routine, as the economic
rationale underlying the transfer pricing methodology is that such contribu-
tions represent monopoly positions that enable the owner to accrue super
profits. The widespread application of the one-sided transfer pricing meth-
ods, however – which are fundamentally based on the tested party only
contributing routine inputs – seems to indicate that transfer pricing prac-
tice has adopted a too-narrow definition of non-routine contributions.307
The author’s impression of the US and OECD rules is that the “unique
intangible” threshold – and thereby also the threshold for attracting re-
sidual profits – is, in practice, set too high, in particular in the context of
recognizing local marketing intangibles.
Given the profound significance of the distinction between unique and non-
unique value chain contributions, it is rather striking that it has not been de-
veloped further in US and OECD law. Both systems have instead put effort
307. It can, however, be observed that the 2009 US service regulations stretch the
non-routine input concept far, as the mere obligation to perform under a service agree-
ment qualifies; see Treas. Regs. § 1.482-9(g)(2), Example 2. This regulatory position
(which, in the author’s view, is too relaxed with respect to what can qualify as unique)
can be seen as a reaction to current transfer pricing practices (which, in the author’s
view, are too restrictive with respect to what can qualify as unique).
94
Controlled intangibles transfers subject to transfer pricing under US law
into simply defining what intangibles are. However, the simple fact that an
intangible is present in the value chain says very little about how transfer
pricing should be carried out. The truly relevant question is whether that
intangible qualifies as a unique value chain contribution or not, because it
will be decisive for the allocation of residual profits. Further work should
be done within the two regimes to develop the distinction.

transfer pricing under US law
3.5.1. The taxation of US inbound and outbound intangibles

transfers
US inbound intangibles transfers occur when a foreign group entity owns

an intangible and transfers rights to exploit it in the United States to a local
group entity. This may be relevant for both manufacturing and marketing
intangibles, such as patents and trademarks, respectively.308 The allocation
of profits from inbound transfers is governed solely by IRC section 482.
The allocation of profits from outbound transfers of US-developed intangi-

bles may be governed by either IRC section 482 or 367, depending on the
circumstances. A US taxpayer can transfer intangibles to a foreign related
entity in five main ways:
(1) through a sale;
(2) through a licence agreement;309
308. Particularly, inbound transfers of marketing intangibles have caused challenging

allocation issues, for instance, where a local US entity incurs substantial marketing
costs to build the local value of a foreign-owned trademark. The question in these cases
is whether the US entity should be allocated residual profits from the intangible due to
its development contribution to the local value of the trademark. The author will dis-
cuss the US profit allocation rules in ch. 23.
309. The distinction between a sale and a licensing transaction lies in whether the
transferor has transferred all substantial rights to the intangible and may affect the
character of the income (e.g. capital or ordinary income), the source of the income (US
or foreign), “basketing” for foreign tax credit purposes of the transferor’s gain (active
or passive income) and the recovery of the transferee’s costs. On the distinction, see
Treas. Regs. § 1.1235-1(a), pertaining to the determination of capital gains and losses
with respect to the sale or exchange of patents. See also Waterman v. McKenzie, 29 F.
316 (1886), affirmed by 138 U.S. 252 (1891); Hooker Chem. & Plastics Corp. v. US,
1978 WL 21534 (Ct.Cl. Trial Div., 1978), affirmed by 219 Ct.Cl. 161 (Ct.Cl., 1979), sup-
plemented by 221 Ct.Cl. 988 (Ct.Cl., 1979); and Bell Intercontinental Corp. v. US, 1967
WL 156523 (Ct.Cl., 1967), report and recommendation adopted by 180 Ct.Cl. 1071 (Ct.
Cl., 1967). See also Zollo (2011), at p. 768.
95
(3) through a CSA;

(4) through a contribution of capital; or
(5) through a reorganization.
Transaction types (1)-(3) are governed by IRC section 482, while transac-
tion types (4)-(5) are governed by section 367. Transactions carried out un-
der section 367 pertain to the transfer of an ongoing business in exchange
for stock or securities in the foreign transferee, organized as a non-recog-
nition (tax-free) transaction. Here, intangibles are transferred as part of a
larger business transfer, in contrast to the typical section 482 transaction
types (1)-(3).
Both IRC sections 482 and 367 seek to ensure that intangible value devel-
oped in the United States is not migrated without taxation. The US trans-
feror will normally have deducted development-phase R&D expenses in
the United States. Taxation of the full value of the intangible when trans-
ferred abroad is therefore necessary to match income and deductions con-
nected to the same intangible in the United States.
It is up to the taxpayer to choose whether to structure a transaction as a

section 482 or section 367 transfer. Section 482 will likely be the natural
choice when only intangibles are transferred, while section 367 will be
chosen when there is a business transfer that includes intangibles.
The author will discuss transaction types (1)-(3) (licence, sale and CSAs)
in connection with his analysis of the US and OECD transfer pricing meth-
odologies in part 2 of the book. Focus in sections 3.5.2.-3.5.8. will be on
section 367 transaction types (4)-(5).
Section 367 has important interactions with section 482. If taxation is trig-
gered under section 367, the value of the transferred intangibles must be
calculated pursuant to the material profit allocation rules under section
482. Thus, the section 482 transfer pricing methodology that is analysed in
part 2 of the book also governs profit allocation under section 367. The aim
of the discussions in this chapter on section 367 is to contextualize how this
provision fits into the larger section 482-driven US framework for related
outbound IP transfers and to pinpoint the most significant nuances of the
relationship between sections 482 and 367.
96
3.5.2. The context in which IRC section 367 applies:

Non-recognition transactions
The federal US tax code affords non-recognition treatment for certain

transactions (including business reorganization asset transfers).310 A trans-
fer that qualifies under a non-recognition provision generally entitles the
transferee to earn the income associated with the transferred property.
Non-recognition transactions that are carried out among related parties
will trigger a conflict between section 482 – which requires arm’s length
profit allocation when value is transferred among related parties – and the
tax-free treatment granted under the non-recognition provision applied to
the transaction.
The IRS has, in general, reserved the right to perform section 482 realloca-
tions notwithstanding the application of non-recognition provisions.311 The
section 482 regulations indicate a rather narrow window for applying sec-
tion 482 in the context of non-recognition transfers.312 This reflects the fact
that the IRS has only been successful in reallocating income by overriding
non-recognition provisions when the transferred property was disposed of
by the transferee quickly after the contribution.
In more sophisticated cases, when the transferee has used the transferred
intangible in its business to generate profits, the IRS has lost. The prime
example of this is Eli Lilly v. CIR, where the Tax Court found that the
residual profits earned by a Puerto Rican subsidiary from exploiting a US-
developed patent transferred to it in a section 351 non-recognition transac-
tion could not be reallocated to the US parent under section 482.313
310. On US non-recognition treatment of corporate reorganizations in general, see, in

particular, Brauner (2004).
311. Treas. Regs. § 1.482-1(f)(1)(iii).
312. See the example in Treas. Regs. § 1.482-1(f)(1)(iii), pertaining to a tax-free trans-
fer of shares to a subsidiary in exchange for capital stock, which sells the shares to
a third party the year after, realizing a loss. The regulations state, with reference to
National Securities Corp. v. CIR, 46 B.T.A. 562 (B.T.A., 1942), affirmed by 137 F.2d
600 (C.C.A.3, 1943), certiorari denied by 320 U.S. 794 (S.Ct., 1943), that the IRS may
disallow the loss on the basis that it was incurred by the transferor. See also Southern
Bancorporation v. CIR, 67 T.C. 1022 (Tax Ct., 1977); and Northwestern Nat. Bank of
Minneapolis v. U.S., 1976 WL 1016, (D.Minn. 1976), affirmed by 556 F.2d 889 (8th
Cir., 1977). Similarly, in Bank of America v. U.S., 1978 WL 1257 (N.D. Cal., 1979), the
court found that the IRS could not apply sec. 482 to disregard the tax consequences of
a transaction governed by the former sec. 311.
313. 84 TC No. 65. The intangible transfer in Lilly occurred prior to the enactment of
sec. 367(d). See the analysis of the case in sec. 5.2.4.3.
97
In the context of section 367 transactions, however, non-recognition treat-

ment that follows from other sections of the IRC (typically, IRC sections
351 and 361) is overridden, unless an exception is applicable. In other
words, section 367 in itself ensures that a non-taxable transaction becomes
taxable.
3.5.3. The historical background of IRC section 367
The predecessors of the current IRC section 367 were section 112(k) of
the Revenue Act of 1932 and section 367 of the Internal Revenue Code of
1954.314 The motivation behind the introduction of the 1932 provision was
to close a “serious loophole for avoidance of taxes”, i.e. that a US transferor
could transfer appreciated property to a foreign subsidiary in a reorganiza-
tion under section 351 without triggering taxes and then defer or avoid tax
upon the subsequent sale of the transferred property by the foreign subsidi-
ary.315 The 1932 provision and the subsequent 1954 section 367 determined
that a section 351 or 332 exchange would be immediately taxable unless it
was established that the exchange was not in pursuance of a plan having
the avoidance of tax as one of its principal purposes (the principal purpose
standard).316
A transfer of assets used in an active foreign trade or business was gener-

ally not deemed to have tax avoidance as its principal purpose, as opposed
to transfers of passive income generating intangibles or intangibles relat-
ing to income from the manufacture or sale of goods in the United States
(roundtrip transactions). Congress observed that the principal purpose
standard was interpreted narrowly by courts and difficult to administer.317
314. Revenue Act of 1932, Pub. L. No. 72-154, 46 Stat. 169, § 112(k). The initial focus
of the rule was to trigger tax on built-in gains.
315. H.R. Rep. No. 708, 72d Cong., 2nd Sess. (1932), 1939-1 (Part 2) C.B. 471.
316. There was comprehensive IRS ruling practice on this standard. See the IRS rul-
ing guidelines, originally set out in Revenue Procedure 68-23. See also Rev. Proc. 75-
29; Rev. Proc. 76-20; Rev. Proc. 77-17; Rev. Proc. 80-14; and Rev. Rul. 2003-99.
317. In the early 1980s, the IRS was faced with two significant Tax Court reversals
of its interpretation of the principal purpose standard, with the result that high-value
intangibles were migrated without taxation. The first case, Dittler Brothers, Inc. v. CIR,
72 T.C. 896 (Tax Ct., 1979), affirmed by 642 F.2d 1211 (5th Cir.(Ga.), 1981), pertained
to the transfer of scratch-off lottery ticket technology to a foreign joint venture. The
Tax Court rejected the IRS assertion that there was a tax avoidance purpose because
the technology was to be used in mainly passive activities of the foreign joint venture,
with the result that the transfer qualified for tax-free non-recognition treatment. In the
second case, Hershey Foods Corp. v. CIR, 76 T.C. 312 (Tax Ct., 1981), declined to fol-
low by Private Letter Ruling (IRS PLR, 1982), 1982 WL 204647, the taxpayer wanted
98
The standard was replaced in 1984 with the section 367(a)(1) requirement
that gains be recognized on any transfer by a US taxpayer to a foreign cor-
poration otherwise qualifying for tax-free treatment, unless an exception
applies.
While the current version of section 367 contains exceptions to the non-
recognition override, these do not apply for transfers of intangibles. Thus,
tax will now always be triggered when intangibles are migrated as part
of a section 367 transaction. The question now is only with regard to how
the income from the intangibles transfer shall be recognized. The answer
to this depends on whether section 367(a) or (d) shall be applied. The
author discusses these provisions in the following two sections (3.5.4.-
3.5.5.).
3.5.4. Current gain recognition under IRC section 367(a)
When property is transferred outbound from a US taxpayer to a foreign

corporation in the form of a non-recognition exchange described in sec-
tions 332, 351 (incorporation transfers), 354, 356 or 361 (reorganization
transfers),318 IRC section 367(a)(1) requires that the gain realized is subject
to immediate recognition and taxation for the US transferor.319
There are some important exceptions from this non-recognition override.

Particularly relevant for the purposes of this book is an exception pertain-
ing to transfers of property to a foreign corporation for use in the active
conduct of a trade or business outside the United States, i.e. the mentioned
ATB exception (the predecessor of which was the principal purpose stand-
to incorporate the assets of two Canadian branches, one of which was loss-making, at
a time when it could be expected to become profitable. The Tax Court found that IRC
sec. 367 was not intended to recapture past losses when a branch is incorporated in a
foreign country.
318. The non-recognition rules apply only to the transfer of property. See, e.g. E.I.
Du Pont de Nemours and Co. v. US, 296 F.Supp. 823, (D.Del., 1969), affirmed in part,
reversed in part by 432 F.2d 1052 (3rd Cir., 1970), where the court observed that US
tax law contains a “generous definition of property” for sec. 351 purposes. See also
Hospital Corp. of America v. CIR, 81 T.C. No. 31 (Tax Ct., 1983), nonacquiescence
recommended by AOD- 1987-22 (IRS AOD, 1987) and Nonacq. 1987 WL 857897 (IRS
ACQ, 1987), where the rendering of negotiation and other support pertaining to a busi-
ness opportunity was deemed to be a service and not a transfer of property under sec.
351.
319. IRC sec. 367(a)(1); and Treas. Regs. § 1.367(a)-1T(b)(1). Taxation is a conse-
quence of the foreign corporation not being considered a corporation.
99
ard, referred to in section 3.5.3.).320 Outbound transfers of property that

qualify for this exception are tax free, i.e. the non-recognition treatment for
these transfers is not overridden.
Until autumn 2015, the ATB exception was designed so that it encom-
passed all types of property unless specifically excluded.321 Section 936
intangibles were excluded, and thus taxable. This meant that intangibles
that were not encompassed by the pre-2018 version of the section 936 defi-
nition fell within the wording of the ATB exception in IRC section 367(a)
(3)(A) as “property transferred to a foreign corporation”, and could thus be
migrated without triggering taxation.
The legislative history behind the ATB exception documents that the origi-
nal intention was to facilitate tax-free conversion of an active foreign busi-
ness from branch to subsidiary form.322 Foreign goodwill and going con-
cern value transferred in such conversions was seen as the result of genuine
non-US development activities. Non-taxation of these items was therefore
not deemed to have any potential for abuse of the US tax system. This
point was highlighted by Congress in the 1984 revision of section 367:323
“The committee does not anticipate that the transfer of goodwill or going
concern value developed by a foreign branch to a newly organized foreign
corporation will result in abuse of the U.S. tax system.”324
Congress envisioned that the transfer of foreign goodwill or going concern

value would be assessed under the ATB exception, as opposed to a separate
320. IRC sec. 367(a)(3)(A); and Treas. Regs. § 1.367(a)-2T(a)(1). US taxpayers may
structure their foreign active business operations in branch form or through deferral
vehicles (controlled foreign corporations, CFCs). The trade-off is that deferral of US
tax may be achieved through the use of deferral vehicles, but not if the branch form
is chosen. On the other side, the losses of a foreign branch may be offset against US
income, as opposed to the losses incurred in a CFC. The active foreign trade exception
lies at the cross-section of these two regimes, governing incorporations of active for-
eign branches and transfers of other active foreign business assets to deferral vehicles.
US outbound transfers to a foreign branch of a US enterprise is not taxable; see IRC
secs. 351 and 367(b). The branch will remain subject to current US taxation, entitled
to credit for foreign income taxes paid; see IRC sec. 901. However, the transfer of an
active foreign business to a foreign parent corporation is a fully taxable transaction; see
IRC secs. 311(b) and 367(e)(2).
321. This exclusion was operationalized through the so-called “tainted asset rule” in
IRC sec. 367(a)(3)(B)), which encompassed, in addition to sec. 936(h)(3)(B) intangi-
bles, outbound transfers of inventory, copyrights, accounts receivable, foreign currency
and property of which the transferor was the lessor at the time of transfer.
322. See Zollo (2010), at p. 73.
323. On this, see, in particular, Driscoll (1988), at p. 171.
324. H.R. Rep. No. 98-432 (1984), Pt. 2, at pp. 1317-1319.
100
rule for intangibles. The Treasury Department and the IRS followed up
on this and released temporary section 367 regulations in 1986.325 These
stated, in line with the Congressional intention, that the “transfer of foreign
goodwill or going concern value” was exempt from section 367(d) and thus
fell under the ATB exception (given that foreign goodwill was not a sec-
tion 936 intangible), with the result being that such transfers would not be
subject to tax.326
This regulatory interpretation would give rise to abusive classification and

valuation practice where multinationals asserted that US outbound trans-
fers of intangible value (i) should not be classified as 936 intangibles, but
rather as foreign goodwill or going concern value qualifying under the
ATB exception; and (ii) minimized the value of the transferred 936 intan-
gibles for which income inclusion was required and maximized the value
of the foreign goodwill exempt from tax.
This practice relied on creative interpretations of foreign goodwill, where it

was even argued that US marketing activities and value from business op-
erated by US employees where income was earned remotely from foreign
customers should be seen as foreign goodwill.327 Such interpretations were
combined with unreliable valuation approaches, using standalone item-by-
item valuations, even where aggregated valuations of the entire package of
transferred assets would yield more reliable value estimates.
These practices became more widespread and aggressive over time, as tax-
payer incentives for structuring US outbound intangibles transfers in the
form of section 367 (as opposed to section 482) transfers grew. Such incen-
tives included the 1997 removal of the treatment of section 367(d) deemed
royalties as US income in relation to the foreign tax credit limitation,328 the
combined effect of the 1996 “check-the-box” regulations and the 2005 sub-
325. TD 8087, 51 FR 17936.

326. Treas. Regs. § 1.367(d)-1T(b). See the definition of foreign goodwill/going con-
cern value in Treas. Regs. § 1.367(a)-1T(d)(5)(iii).
327. See Zollo (2010), at p. 84, on the IRS position that taxpayer classifications of for-
eign goodwill could also encompass “US goodwill”. Zollo is critical of the IRS position
on the basis that foreign goodwill was defined in the regulations as a residual amount
of the value of the foreign business, implying that the entire residual should be deemed
foreign.
328. Taxpayer Relief Act of 1997, Public Law 105-34, 111 Stat. 788. The 1997 amend-
ment to sec. 367(d)(2)(C) secured that deemed royalty amounts received under sec.
367(d) are treated as ordinary income that is sourced in the same manner as royalties,
and therefore potentially as from sources outside of the United States. Prior to 1997,
deemed royalties received under sec. 367(d) were seen as income from US sources (and
thus not eligible for credit).
101
part F “look-thru” rule, making it easier to move funds earned in a foreign

group entity intra-group without incurring US tax and for foreign group
entities to license intangibles intra-group without triggering subpart F in-
come.329 The likely strongest taxpayer incentive to structure transactions
under section 367 came in 2009, when the new temporary cost-sharing
regulations significantly strengthened the valuation methods under IRC
section 482 for buy-in transactions, ensuring “full value” taxation of in-
tangibles transferred abroad through CSAs.330 This development occurred
over decades in which the importance and value of intangible assets for the
businesses of large multinationals had just increased.331 The result of these
practices was that significant US-developed intangibles were not taxed at
their full value – or even escaped taxation altogether – when transferred
within the group to a foreign jurisdiction.
The IRS issued new proposed section 367 regulations in September 2015
to put an end to these BEPS practices.332 These regulations were finalized
in December 2016.333 The revised ATB exception is significantly narrowed
and represents a clear departure from previous, 3-decades-long practice.
The ATB exception now only applies to positively listed (so-called “eligi-
ble”) property. The list of eligible property encompasses tangible property,
working interests in oil and gas property and certain financial assets.334
Intangibles do not qualify as eligible property.
329. For the “check-the-box” regulations of § 301.7701–3 TD 8697 (61 FR 66584), see
TD 8697 (61 FR 66584). For the subpart F “look-thru” rule, see sec. 954(c)(6) of the
Tax Increase Prevention and Reconciliation Act of 2005, Public Law 109-222, 120 Stat.
345.
330. TD 9441 (74 FR 340), finalized in TD 9568 (76 FR 80082). It was, in particu-
lar, the introduction by these regulations of the income method to value buy-ins that
discouraged taxpayers from structuring outbound IP transfers under sec. 482. See the
analysis of the income method in sec. 14.2.8.3.
331. See the discussion in the preamble to the 2016 regulations, TD 9803.
332. The revision was made effective immediately, in proposed form, without allow-
ing for notice and comment before making the rule change; see 80 FR 55568-01. This
is unusual and emphasizes the urgent need to address the profit shifting practices of
multinationals under IRC sec. 367(a) and (d). It should be noted that the 2015 revision
of the active trade or business (ATB) exception also intended to make the information
more easily accessible. The text of the pre-2015 ATB exception was scattered through
a range of different regulations (Treas. Regs. § 1.367(a)-2; § 1.367(a)-2T; § 1.367(a)-4;
§ 1.367(a)-4T; § 1.367(a)-5; and § 1.367(a)-5T). The revised 2015 exception combines
these regulations into a single provision, i.e. Treas. Regs. § 1.367(a)-2, published in 80
FR 55568-01. The revised provision does not include the depreciation loss recapture
rule, which is now moved from § 1.367(a)-4T to 1.367(a)-4) (see 80 FR 55568-01).
333. TD 9803.
334. Treas. Regs. § 1.367(a)-2(c) (see 80 FR 55568-01) lists four categories of in-
eligible property. These are (i) inventory; (ii) instalment obligations, etc.; (iii) foreign
currency, etc.; and (iv) certain leased tangible property.
102
The exception for foreign goodwill and going concern value was re-
moved.335 The IRS assessed whether it could be retained in a modified form
that ensured that it would not be abused. Both the definition of a foreign
branch and a “cap” on the foreign goodwill that could be exempted was
considered,336 but the IRS ultimately found that such a modified exemption
would be difficult to administer.
The scrapping of the foreign goodwill exception was a controversial move

by the IRS,337 as Congress clearly envisioned in 1984 that there should be
an exception for the transfer of foreign goodwill and going concern val-
ue.338 The author nevertheless finds the IRS reasoning for the removal con-
vincing. In 1984, when the ATB exception was introduced – coming from
a situation with severe difficulties in administering the principal purpose
standard – it was clearly also the congressional intent that (i) the foreign
goodwill exception should not entail abusive profit shifting practices; and
(ii) outbound transfers of US-developed 936 intangibles should be taxable.
Given the aggressive taxpayer practices of classifying outbound transfers
of (what in reality, in many cases, must have been 936) intangibles as for-
eign goodwill and maximizing the value of such wrongly classified intan-
gibles at the expense of 936 intangibles transferred in the same business
transfer using standalone item-by-item valuations, it is indeed difficult to
see how the exception for foreign goodwill could be upheld while at the
same time protecting the integrity of the federal US tax system.339 Even
though the 2015 revision of the ATB exception certainly came abruptly,
it was, in the author’s view, a necessary modification of the section 367
regulations. It will undoubtedly facilitate arm’s length profit allocations in
the context of controlled intangibles transfers going forward.
Thus, under the current regulations (as revised in 2015 and finalized in
2016), all outbound transfers of intangibles under IRC section 367 – includ-
ing foreign goodwill and going concern value – will be taxable for the US
transferor. The intangibles transfers shall be valued using the transfer pri
335. Thus, outbound transfers of foreign goodwill or going concern value are subject
to either current gain recognition under IRC sec. 367(a) or gain deferral under sec.
367(d), but are nevertheless taxable.
336. See Velarde (2015).
337. The removal of the foreign goodwill exception was met with considerable criti-
cism from taxpayers. The IRS discussed the objections thoroughly in the preamble to
the final regulations, but upheld its position.
338. See H.R. Rep. No. 98-432, Pt. 2 (1984), at p. 1320.
339. Reinstatement of an exception for active trade or business is, however, now under
consideration with respect to cases that pose little risk for abuse and administrative
challenges; see TNT Doc-2017-72131. See also JCX-51-17, at p. 195.
103
cing methodology under IRC section 482 so that the full value is captured
under US taxation. However, the current gain recognition treatment under
section 367(a) cannot be applied to transfers of 936 intangibles,340 as such
intangibles must be taken into account under section 367(d) and its com-
mensurate-with-income deemed royalty construction (discussed in section
3.5.5.). The scope of application for section 367(a), in the context of intan-
gibles transfers, is therefore narrow. Only non-936 intangibles can now be
treated to immediate gain recognition under section 367(a). However, as
the author will elaborate on in section 3.5.5., taxpayers are now given the
choice to elect treatment under section 367(d) for such non-936 intangibles.
3.5.5. Deemed royalty inclusions under IRC section 367(d)
3.5.5.1. Historical background
In the 1970s and early 1980s, US pharmaceutical companies had a practice

of transferring ownership of US-developed, high-profit-potential manufac-
turing intangibles to Puerto Rico as well as other US possessions through
non-recognition transactions, often at the point in time when the intangi-
bles were expected to become profitable. This was profit shifting, as the
subsequent profits generated by the intangibles would then not be allocated
to the United States, in spite of the fact that US tax deductions had been
claimed for R&D expenses throughout the development phase.
IRC section 936(h) was introduced in 1982 as a reaction to these practices,

requiring US shareholders of possession companies to include in their cur-
rent income the profits generated by the intangibles owned by the pos-
session companies.341 Congress acknowledged that some multinationals
would, as a result, transfer intangibles from Puerto Rico to foreign juris-
dictions. This was the background for the 1982 introduction of the initial
version of section 367(d), which deemed intangibles transfers from a pos-
session corporation to a foreign corporation as having tax avoidance as the
principal purpose and being thus ineligible for non-recognition treatment
under section 367. This first version of section 367(d) required recognition
of the full realized gain at the time of the transfer.342
340. Treas. Regs. § 1.367(a)-(1), in 80 FR 55568-01.

341. See the Tax Equity and Fiscal Responsibility Act of 1982, Pub. L. No. 97-248, 96
Stat. 324 (1982).
342. See Driscoll (1988), at p. 177.
104
Section 367(d) was amended in 1984 to apply more generally to outbound

intangible transfers. Even though it no longer applied solely to possession
corporations, the provision retained its original cross-reference to the 936
intangibles definition. The requirement for current and definitive recogni-
tion of gains at the time of transfer was replaced with a deemed royalty
construction, aimed at better reflecting the actual profits generated by the
transferred intangible.
In 1986, the first IRS interpretation of section 367(d) was released in the
form of the temporary IRC section 367(d) regulations.343 That same year
also saw the incorporation of the new IRC section 482 commensurate-with-
income standard into IRC section 367(d).344 In 2010, the IRS announced
a project to revise the 1986 temporary section 367(d) regulations,345 which
resulted in the 2015 proposed regulations (finalized in 2016).346
3.5.5.2. The material content of IRC section 367(d): Sale of

contingent payments
The system under the section 367 regulations is that transfers of 936 in-
tangibles must be subject to deemed royalty recognition under section
367(d) (a so-called “transfer pursuant to a sale of contingent payments”),
and thereby the commensurate-with-income standard. There are four main
scenarios for the application of section 367(d):
(1) Intangible transfers to a foreign corporation under section 351 or 361:

The US transferor will, in these cases, be treated as having transferred
the intangible in exchange for annual royalty payments, contingent on
the productivity or use of the property, subject to the commensurate-
with-income standard.347
(2) Subsequent disposal by the foreign transferee of the intangible to a

third party: The US transferor must, in these cases, recognize any gain
343. See 51 Fed. Reg. 17936-01.

344. Tax Reform Act of 1986, Pub. L. No. 99-514, 100 Stat. 2085 (1986), at sec.
1231(e)(1).
345. For informed recommendations pertaining to the regulation project, see Bless-
ing (2010a).
346. 80 FR 55568 (proposed 2015 regulations); and TD 9803 (final 2016 regulations).
347. Treas. Regs. § 1.367(d)-1T(c). The initial version of IRC sec. 367(d), as proposed
by the House Committee on Ways and Means, described the IRC sec. 367(d) payments
as an exclusive licensing agreement. The Senate version, as adopted in the final legisla-
tion, used the contingent sale approach.
105
that comes from deducting its former adjusted tax basis for the intan-
gible from the value of the intangible as of the date of the transferee’s
disposition.348
(3) Subsequent transfers of the stock of the foreign transferee corporation

to a related US person:349 The new related US stock owner will, in
these cases, be treated as having received a proportional right to the
annual royalty payments that otherwise would have accrued to the US
transferor.350
(4) Subsequent transfers of the stock of the foreign transferee corporation

to a third party: The US transferor will, in these cases, be treated as
having sold the intangible and must recognize any gain that comes
from deducting its former adjusted tax basis for the intangible from the
value of the intangible at the date of the stock disposition.351
The author will focus on the first scenario, as this is the scenario that is rel-
evant for direct transfers of intangibles out of the United States (as part of
a larger business transfer). In this scenario, the US transferor is, under sec-
tion 367(d), treated as having sold the intangible in exchange for payments
contingent upon the productivity, use or disposition of the intangible for
amounts that would have been received annually in the form of such pay-
ments over the useful life of the intangible.352 Thus, section 367(d) sees the
intangible transfer as separate from the larger business transfer of which
it is a part.
The sale for contingent consideration approach (1) entails that the income
is deferred over the life of the intangible, and the contingent amount of
gain is taxable as ordinary income. The useful life of the intangible – and
thus the period of the potential income stream – was set to 20 years under
348. Treas. Regs. § 1.367(d)-1T(f)(1).

349. If the related entity that acquires the stock is foreign, the US transferor will con-
tinue to include in its income the annual deemed payments under IRC sec. 367(d) as if
the subsequent stock disposition had not occurred.
350. Treas. Regs. § 1.367(d)-1T(e).
351. Treas. Regs. § 1.367(d)-1T(d). The calculation will be reduced by the gain that
the US transferor has recognized with respect to the deemed sale of the intangible un-
der IRC sec. 367(d).
352. IRC sec. 367(d)(2)(A). A US taxpayer may, as the point of departure, transfer
intangibles to a foreign partnership without recognizing gains on the transfer (see IRC
sec. 721). The US transferor will, however, be subject to tax on its share of the partner-
ship’s income subsequent to the transfer (see IRC secs. 701 and 702). Regulations were
issued in February 2017 that extend the deemed sale treatment to transfers to foreign
partnerships under IRC sec. 367(d)(3); see TD 9814.
106
the 1986 regulations.353 The revised 2015 regulations broadened the rule to
encompass the entire period during which the exploitation of the intangible
was reasonably anticipated to occur, including the use of the intangible for
the purpose of further R&D, as well as the use in subsequently developed
intangibles akin to CSAs.354 The 20-year limitation was removed due to the
concern that it could result in less than all of the income from the trans-
ferred intangible being recognized as income by the US transferor. This
2015 rule was criticized by taxpayers for being difficult to administer.
The IRS therefore modified the rule in the final 2016 regulations so that
taxpayers may, in the year of the transfer, choose to take into account
deemed royalties under section 367(d) for only 20 years.355 However, this
modification was combined with the requirement that the 20-year limi-
tation should not affect the present value of all amounts included by the
taxpayer under section 367(d). The result is essentially that the “full value”
of the intangible (as captured by the 2015 rule) can now be spread over
“only” 20 years, should the taxpayer choose the 20-year limitation. While
this modified rule will also undoubtedly be challenging to administer, it
seems necessary to capture the entire income from long-lived intangibles.
It has been asserted that the purpose of section 367(d) is merely to “claw
back” the US tax advantages gained through deducting R&D costs related
to the transferred assets or from operating in branch form, as opposed to
establishing a general exit tax on US persons that incorporate their foreign
branch operations or transfer foreign business assets outbound.356
The author finds this assertion comprehensively difficult to reconcile with

the methodology for calculating income under section 367(d), which is a
construction for deemed royalty inclusion based on the section 482 com-
mensurate-with-income standard, designed to capture the actual income
generated through exploitation of the transferred intangibles, unrelated to
development phase costs.
Nevertheless, the literature argues that this is not incompatible with a mere
claw-back intention, as the commensurate-with-income standard purport-
edly may better address mismatch concerns due to the difficulty in iden-
353. The 20-year rule was found in Treas. Regs. § 1.367(d)-1T(c)(3).

354. This is consistent with the cost-sharing regulations; see Treas. Regs. § 1.482-7(g)
(2)(ii)(A) and § 1.482-7(g)(4)(viii), Examples 1 and 7.
355. From and including the first year in which the US transferor takes into account
such income.
356. Aksakal et al. (2013), at p. 13.
107
tifying R&D costs for those specific intangibles that are successfully de-
veloped.357 While the author finds this argument somewhat alluring, it is
not convincing. Section 367(d) goes further than simply clawing back US
R&D deductions; it captures the entire intangible value transferred.
3.5.6. Income recognition under section 367(a) or (d) for

intangible transfers?
The US transferor may now apply IRC section 367(d) to a transfer of intan-
gibles that otherwise would be subject to IRC section 367(a).358 The tax-
payer is, in other words, given a choice between immediate income recog-
nition under section 367(a) and deferred income recognition under section
367(d). This is new, as non-936 intangibles could not be assessed under
section 367(d) under the previous regulations. To implement this taxpayer
choice (as well as the removal of the foreign goodwill exception), the defi-
nition of intangibles that applies for the purpose of section 367(a) and (d)
was revised in 2015. The new definition encompasses both 936 intangibles
and property to which a US person applies IRC section 367(d) in lieu of
applying IRC section 367(a).359 The IRS reasoning for this election is that
it does not have the authority to force taxpayers to apply IRC section 367(d)
(and thereby the commensurate-with-income standard) to intangibles that
fall outside the 936 definition.
In connection with the finalization of the 2016 regulations, taxpayers re-

quested that the regulations provide certainty that taxpayers would be per-
mitted to treat goodwill and going concern value as non-936 intangibles,
subject to section 367(a) and not 367(d). The IRS refused to take a clear
position on this issue. The answer is, in the author’s view, clear. Section
367(d) applies only to 936 intangibles. Residual intangible value, such as,
inter alia, foreign goodwill and going concern value, is not encompassed
by the pre-2018 version of the 936 definition. There is no legal basis for ob-
jecting to taxpayer treatment of non-936 intangibles under section 367(a).
357. Ibid., at p. 14.

358. Treas. Regs. § 1.367(a)-1(b)(5), in 80 FR 55568-01. The IRC sec. 367(d) treat-
ment must be applied consistently for all controlled transfers with respect to such items
of property.
359. Treas. Regs. § 1.367(a)-1(d)(5), in 80 FR 55568-01. A similar definition will be
used for the purpose of implementing Notice 2012-39 (IRB 2012-31) on outbound trans-
fers of intangibles in certain asset reorganizations. Odintz et al. (2017), at sec. 2.2.2.2,
are under the impression that the definition may exceed the IRS’s authority, as it may
require compensation for intangible items that fall outside the sec. 936 definition.
108
Whether taxpayers will choose income recognition under section 367(a)

or (d) will likely depend on the circumstances. The entire value of the
residual intangible must be recognized immediately upon transfer under
section 367(a). A valuation must be performed to estimate this value,
based on the probable future income generated by the intangible. This
may lead to high values and corresponding high immediate taxation.
The IRS will likely take a critical view towards taxpayer valuations. An
upside for taxpayers will be that the income recognition under section
367(a) is a one-off event with no future follow-up obligations. Section
367(d), on the other hand, offers taxpayers a significant deferral benefit,
entailing a lower present value of the tax obligation as compared to the
situation under section 367(a). However, the commensurate-with-income
standard applies under section 367(d), requiring that the income recog-
nized must correspond to the actual income generated by the intangible
over the 20-year period, triggering risk of subsequent periodic adjust-
ments.360 As taxpayers must report income over the 20-year period, sec-
tion 367(d) also entails considerable follow-up obligations subsequent to
the transfer.
3.5.7. The further relationship between profit allocation under

sections 482 and 367
The material content of the profit allocation rules used to determine the
amount of taxable income for all outbound intangible transaction types are,
in principle, the same under US law, regardless of whether the transfer is
structured as an actual sale, licence or CSA directly subject to section 482,
or as a section 351 or 361 exchange subject to section 367. Thus, the fair
market value necessary to calculate any gain under section 367(a), as well
as the determination of the annual deemed royalty amounts under section
367(d), relies on the transfer pricing methods under section 482.361 This
“catch all” application of the arm’s length standard is further accentuated
by the 2015 regulations clarifying the application of the arm’s length stand-
ard and the best-method rule with other IRC provisions (including IRC
section 367), where the overarching principle is that arm’s length compen-
sation must be provided for all intra-group value transfers independent of
the form or character of the controlled transaction.362 There are, however,
360. See ch. 16 for a discussion of the US periodic adjustment authority.

361. See Treas. Regs. § 1.367(d)-1T(c)(1) for royalty payments and § 1.367(d)-1T(g)(5)
with respect to the determination of the fair market value of the transferred intangible.
362. Treas. Regs. § 1.482-1T(f)(2)(i)(A).
109
some nuances with respect to the application of the section 482 profit al-
location rules in the context of section 367, which deserve to be mentioned.
First, the form of payment rules are flexible under section 482, allow-
ing royalty, lump sum, instalment and contingent payments. In contrast,
section 367(d) restricts the form of payment to periodic deemed royalty
payments over a 20-year period (reflecting in total the full value of the
intangible) or to immediate recognition of the entire gain if the intangible
is transferred out of the group. These differences do not affect the profit al-
location as such (i.e. the amount of income to be recognized), but only how
and when this amount shall be recognized. The deemed royalty under sec-
tion 367(d) will, as mentioned, give the taxpayer a beneficial present value
effect. Under section 482, it must be assumed that the taxpayer will often
opt for deferred treatment in the form of royalty, instalment or contingent
payments. This will “smooth the difference” between the form-of-payment
rules of sections 482 and 367(d) in practice.
Second, none of the section 482 exemptions from the periodic adjustment
authority under the commensurate-with-income standard are specifical-
ly mentioned in the section 367(d) regulations.363 It has been questioned
whether they are applicable under section 367(d).364 The asserted relevance
of the question is due to the observation that the US transferor remains
taxable on the appreciated value if he disposes of the intangible, regardless
of whether the gain is due to factors that could not reasonably have been
anticipated at the time of the transfer.
The author doubts whether the question is justified. The deemed section 367
considerations are treated as royalty-like contingent payments based on ac-
tual profits, typically determined either through the CPM or the profit split
method.365 The recognized income will not be restricted to the profits that
were possible to estimate at the time of transfer. The exceptions to the com-
mensurate-with-income standard, which generally turn on the degree of dis-
crepancy between the profits that could reasonably be estimated at the time of
the transfer and the subsequent actual profits,366 are therefore not meaningful
363. For an analysis of the US periodic adjustment authority, see sec. 16.3.
364. See Blessing (2010a), Part D, 1, a).
365. It has been suggested that the “contingent upon the productivity” language ap-
pearing in the original statute in 1982 and 1984 was rendered duplicative, or at least
unnecessary, when the commensurate-with-income standard was added to IRC secs.
367(d) and 482 in 1986.
366. See the analysis of the exceptions from the US periodic adjustment authority in
sec. 16.3.3.
110
in the context of section 367. Had section 367 instead prescribed a gain cal-
culation based on the estimated value of the intangible at the time of transfer
and the subsequent actual profits had indicated that the transfer price was set
too low, the periodic adjustment authority could have been applied in a sensi-
ble way. That, however, is not the case. The author therefore fails to see that
the exceptions to the standard are relevant in the context of section 367(d).
Third, there are some nuances between section 482 and section 367(d) with
respect to basis recovery. Under section 482, the basis will be deductible
against the amount realized upon disposition.367 The deemed royalty construc-
tion under section 367(d) does not allow the US transferor to deduct his basis
in the transferred intangible,368 as the full amount of the deemed payment
received must be included in the gross income.369 Thus, if the taxpayer has a
basis in the transferred intangible, the result of the section 367(d) deemed sale
construction will be less favourable than a section 482 sale. The author does
not find any indications in the legislative history that this inconsistency was
intended.370 He cannot see any good reason as to why the transferor should
not be able to recover his basis over the useful life of the intangible.
Fourth, section 367(d) gains are not subject to section 453A interest charg-
es, resulting in a potential deferral benefit under section 367(d) compared
to the treatment under section 482.
Fifth, as touched upon above, prior to 1997, there was a disadvantage con-
nected to section 367(d) transactions (compared to licence transactions
under section 482) in the sense that the deemed royalty inclusions were
ineligible for foreign tax credits under the then-current sourcing rule. The
1997 revision removed the requirement to treat the deemed royalty income
as US-source income.371 The royalty is now considered foreign-source to
the extent that an actual payment made by the foreign corporation under
367. IRC sec. 1001(a). This is subject to the instalment method rule in IRC sec. 453,
under which the income recognized for any taxable year from a disposition is the pro-
portion of the payments received in that year that the gross profit bears to the total
contract price; see IRC sec. 453(c).
368. The only circumstances in which the transferor recovers his original basis in the
intangible are where (i) he disposes of his stock in the transferee corporation to a third
party (Treas. Regs. § 1.367(d)-1T(d)(1)); and (ii) the transferee corporation disposes of
the transferred intangible to a third party (Treas. Regs. § 1.367(d)-1T(f)(1)(i)).
369. Treas. Regs. § 1.367(d)-1T(c).
370. See also Blessing (2010a), at part D, 4, b).
371. See the Taxpayer Relief Act of 1997 (P.L. 105-34), which amended subparagraph
(c) of IRC sec. 367(d)(2) to read as follows: “[…] any amount included in gross income
by reason of this subsection shall be treated as ordinary income.”
111
a licence or sale agreement would be treated as foreign-source income.372

Prior to this amendment, taxpayers generally chose section 482 licence
structures over outbound section 367(d) transfers.373 The main problem
for taxpayers with section 367(d) transactions is the likelihood that the
deemed royalty payments will not be deductible under the domestic law of
the foreign jurisdiction. The author sees this as more of a consequence of
the relevant domestic law than a necessary feature of section 367(d).
In light of the above, it should be acknowledged that the most important

differences between the tax treatment of intangible transactions governed
by sections 482 and 367 lie not in the profit allocation rules themselves, but
rather in subordinate consequences pertaining to the treatment of the recog-
nized income, in particular the differences in basis deductibility when cal-
culating US income and the probable lack of local deductibility for deemed
section 367 royalty inclusions when calculating foreign income. These dif-
ferences may, seen in isolation, incentivize taxpayers to structure intangi-
bles transfers as section 482 transactions, i.e. as a licence, sale or CSA. As
discussed above (in section 3.5.4.), however, there has been a trend towards
structuring outbound intangibles transactions as section 367 transfers,
geared at classifying the transferred intangibles as residual intangible value
(goodwill) to escape taxation under the ATB exception. Now, subsequent
to the 2016 narrowing of the ATB exception and the removal of the foreign
goodwill exception, it remains to be seen whether multinationals will retain
their historical interest in migrating US intangibles by way of section 367.
3.5.8. The relationship between profit allocation under

the section 482 cost-sharing regulations and
section 367(d) 374
Qualified CSAs have historically been popular legal vehicles for migrating
valuable US-developed intangibles using “below full value” buy-in valua-
tions, thereby escaping US taxation on significant parts of the true intan-
372. IRC sec. 367(d)(2)(C); see also ibid.

373. In terms of the context of this revision, the pre-1997 rule was intended to dis-
courage outbound intangible transfers, but the enhanced information reporting require-
ments included in the Taxpayer Relief Act of 1997 made it unnecessary to uphold this
rule. See the discussion in Congressional Record (107th Congress), V. 147, P.T. 20, 19
Dec. 2001-3 Jan. 2002.
374. It should be noted that the author’s comments in this section refer to the termi-
nology of the US cost-sharing regulations in Treas. Regs. § 1.482-7. He refers to his
analysis in sec. 14.2.7. (with further references) for an elaboration of these concepts.
112
gible value. The current 2009 section 482 cost-sharing regulations,375 with
their highly developed pricing methods, based on the realistic alternatives
available to the controlled parties, seem to be effective in curtailing these
practices.376 As a result, US-based multinationals– at least until September
2015, when the new proposed section 367 regulations were issued – have
seemed to focus their efforts on migration of US-developed intangibles
through, inter alia, section 367 transactions in an attempt to avoid the valu-
ation principles of the CSA regulations.377
Put plainly, the CSA pricing methodology, where the go-to methods (the
income method and the RPSM) rely on aggregated DCF valuations, will
generally yield a higher price for a transferred group of intangibles than
stand-alone pricing of each transferred intangible based on the ordinary
specified pricing methods (which undoubtedly apply under section 367(d)).
Thus, in theory, it could be possible for a multinational to attain a “valua-
tion rebate” by avoiding the CSA pricing methodology through a section
367(d) transfer.378
The author will analyse the CSA pricing methods later in the book.379 The
question at this point is whether they are applicable to value section 367(d)
intangibles transfers that are connected or similar to a CSA. This is relevant
when intangibles are transferred for the purpose of serving as the basis for
further research, e.g. when a US multinational transfers its R&D branch to
a foreign corporation or transfers its intangibles in a section 367(d) transfer
to subsequently use them as platform contributions via the foreign transfer-
ee to a CSA. The economic substance of such transactions may lie closely
to that of CSAs, in the sense that ownership to – and thus entitlement to
residual profits from – US-developed intangibles are migrated to foreign
group entities. It is clear that a section 367(d) intangibles transfer as such
is not a CSA and that the pricing methods of the CSA regulations therefore
cannot be used (section 367(d) intangibles transfers fall outside the scope
of the CSA regulations).380 The issue is whether the CSA pricing methods
may be applied analogically.
375. Treas. Regs. § 1.482-7.

376. See Sheppard (2011), who assumes that CSAs are no longer being used for profit
shifting purposes due to the new rules.
377. Another popular migration method is the contribution of intangibles to partner-
ships with foreign related entitles as partners; see the comments infra n. 388 and n. 390.
378. See also Reams et al. (2005), at sec. V.
379. See ch. 14 on CSAs.
380. The IRC sec. 367(a) regulations state that “a person’s entering into a cost-sharing
arrangement under § 1.482-7 or acquiring rights to intangible property under such an
113
The regulations were clarified on relevant points in September 2015, when

the IRS and the Treasury, concurrently with the issuance of the new section
367 regulations, issued new temporary and final regulations clarifying the
application of the arm’s length standard and the best-method rule under
section 482.381
Particularly relevant in the current context is the requirement that all value
provided in controlled transactions must be compensated at arm’s length
without regard to the form or character of the transaction.382 For this pur-
pose, the entire arrangement must be considered, as determined by the
contractual terms, whether written or imputed in accordance with the eco-
nomic substance of the arrangement. Further, there is a new requirement
for a coordinated best-method analysis of two or more controlled transac-
tions to which one or more provisions of the IRC or regulations apply.
The point is to ensure that that the overall value transferred, including any
synergies, is properly taken into account.383
The new 2015 rules target, inter alia, speculative controlled transactions that
seek to reduce the arm’s length compensation amount through the use of sec-
tion 482 and section 367 to different parts of what is, in economic substance,
the same transaction. A key point of these new rules is that outbound section
367 transfers of limited rights in US-developed intangibles combined with
a subsequent R&D arrangement between the US and foreign group entities
(e.g. contract R&D) – where the substance of the entire set of transactions
resembles a CSA – must be valued in the aggregate based on the best real-
istic alternatives of the controlled parties so that all value transferred is re-
flected in the consideration (as opposed to separate pricing of the IP transfer
and the subsequent R&D arrangement), regardless of the legal form.384
A notable aspect of the 2015 rules is that they only address the requirement
for arm’s length pricing as such and not the particular pricing methodology
that should be applied. The CSA pricing methods should be applied to de-
termine the valuation of section 367(d) transactions, where their economic
substance is akin to CSAs.
arrangement shall not be considered a transfer of property described in section 367(a)

(1)”; see Treas. Regs. § 1.367(a)-1(d)(3). It follows from the cost-sharing regulations
that shares in the transferee cannot be used as a buy-in payment.
381. 80 FR 55538-01. See the author’s comments on these provisions in the context
of (i) the best-method rule in sec. 6.4.; (ii) aggregated valuation in sec. 6.7.2.; and (iii)
intra-group R&D services (contract R&D arrangements) in sec. 21.4.
382. Treas. Regs. § 1.482-1T(a)(i)(A), as added by 80 FR 55538-01.
383. Treas. Regs. § 1.482-1T(a)(i)(C), as added by 80 FR 55538-01.
384. See Treas. Regs. § 1.482-1T(a)(i)(E), Examples 6, 7 and 8.
114
Nevertheless, it is a problem that valuations under IRC section 367(d) in

principle should be restricted to taking into account only section 936 in-
tangibles.385 For US intra-group IP transactions carried out before 2018 and
thus assessed under the pre-2018 version of the US IP definition, this will
typically exclude residual intangibles, such as workforce in place, which
tends to be a part of the intangible value transferred through CSA transac-
tions. However, the new 2015 section 367 regulations, which narrowed the
application of the ATB exception and removed the foreign goodwill ex-
ception, effectively “override” the section 936 definition restriction in the
sense that all intangible transfers under section 367 now are taxable either
on a current gain recognition basis under section 367(a) or on a deferred
deemed royalty basis under section 367(d), leaving residual intangibles
open for taxation in the context of section 367 transfers. Likewise, the 2009
cost-sharing regulations circumvent the section 936 definition restriction
in the context of CSAs through the design of the calculation of the buy-in
amount.
Nevertheless, the author finds that an analogical application of the CSA

pricing methodology will likely be the only option that will yield an arm’s
length profit allocation in transactions combining sections 482 and 367.
These methods are specifically designed to capture the value transferred
when intangibles are made available to related parties for the purpose of
further research. Also, the income method,386 the hallmark of which is to
allocate the entire residual profits to the US transferor, will, in the author’s
view, generally provide the most reliable profit allocation result in these
cases, as the non-routine development contributions come from the US en-
tity alone.
There is also a clear link between the emphasis of the realistic alternatives
available for pricing purposes in the new 2015 regulations on the applica-
tion of the arm’s length standard and the income method of the 2009 cost-
sharing regulations.387 This also aligns with the recent IRS approach to
the valuation of transfers of US-developed intangibles by a US partner to
385. See the discussion of the US intangibles definition applicable under IRC secs.
482 and 367 in sec. 3.2.
386. Treas. Regs. § 1.482-7(g)(4). See the analysis in sec. 14.2.8.3.
387. The author’s view correlates with that of the IRS; see IRB 2012-12. Further,
the preamble to the final cost-sharing regulations envisions that the aggregated valu-
ation approach of the buy-in pricing methods should be applied to reflect the “com-
bined effect of multiple contributions”, including controlled transactions “outside of the
CSA (for example, make-or-sell licenses, or intangible transfers governed by section
367(d))”, if an aggregated valuation would yield the most reliable measure of an arm’s
length result. See 76 FR 80082-01, Explanation of provisions, sec. A.
115
a partnership with foreign controlled partners (an alternative tax planning

approach to section 367 transfers that is currently popular among US-based
multinationals), which is linked to the CSA methodology, in particular,
the income method.388 New regulations were issued in 2017 on transfers of
certain property by US persons to partnerships with related foreign part-
ners.389 The IRS is of the view that section 482 applies to controlled trans-
actions that fall under the new regulations, and will likely apply the income
method valuation approach for such transfers.390
In light of the above, the author concludes that the CSA pricing methods
should be applicable by analogy to pricing section 367(d) transfers if the
economic substance of the controlled transaction is akin to that of a CSA.
This conclusion is strengthened by the 2017 tax reform codification (con-
tained in the last sentence of IRC section 482) of the aggregated valuation
approach based on the best realistic alternatives of the controlled parties.391

transfer pricing under the OECD TPG
The transfer pricing methods and the intangible ownership rules set out in
the 2017 OECD TPG will allocate profits to any group entity that, in con-
trolled transactions, has contributed functions, assets or risks to the devel-
opment, enhancement, maintenance, protection or exploitation (DEMPE)
of intangibles.392 These OECD allocation rules are triggered if intangible
value is transferred intra-group, regardless of how the transfer was struc-
tured contractually. Thus, the same material allocation norms govern profit
allocation in:393
− transactions in which an intangible is sold;394
− transactions in which an intangible is licensed;395
388. This overrides the non-recognition provision in IRC sec. 721; see 2015-34 I.R.B.
210. See also the comments in 2014-42 I.R.B. 712 on inversions and related transac-
tions.
389. TD 9814.
390. See the preamble to TD 9814. See also IRB 2015-34; Tax Notes Intl., 24 Aug.
2015, at p. 646; Tax Notes Intl., 14 Sept. 2015, at p. 942; and Sheppard (2015).
391. As a way of background for the codification, see JCX-51-17, at pp. 236-237.
393. See Lagarden (2014), at p. 333, for a discussion of different types of IP transac-
tions.
395. Ibid.
116
Concluding comments
− cost-sharing transactions;396 and

− business transfers that include intangibles.397
This makes the delineation of the scope of the OECD guidance much
clearer – and far less problematic – than the intricate interplay between
sections 482 and 367(d) for outbound intangible transfers under US law.
For instance, even though the guidance on cost-sharing agreements and
valuations is contained in different chapters of the OECD TPG, the mate-
rial profit allocation turns on the same principles. A taxpayer will therefore
gain nothing by structuring an intra-group IP transfer as a cost-sharing
transaction as opposed to a sale. The OECD rules are neutral with respect
to transaction forms.
3.7. Concluding comments
The purpose of this chapter was to explore the extent to which the US and
OECD transfer pricing rules are limited by intangibles definitions and pro-
visions that regulate specific intra-group transaction types through which
intangibles are transferred.
The US and OECD rules represent opposite approaches. The US intangibles

definition and specific IP transaction provisions are complex and, in princi-
ple, restrict the application of the transfer pricing methodology, a problem
which is significant under the pre-2018 version of the US IP definition. Such
restrictions do not facilitate a neutral tax system, as they put pressure on
the classification and valuation of intra-group intangible transfers. While
US law still links the application of the transfer pricing methods and the
intangible ownership provisions to a positive definition of intangibles, the
2017 tax reform has amended the definition to include residual intangibles.
This will ensure taxation of such IP when transferred intra-group from and
including 2018. This amendment of the law is most welcome, but it would
likely have been even better to remove the reference to the intangibles defi-
nition from IRC sections 367 and 482, thereby entirely freeing the transfer
pricing methodology of any potential unnecessary constraints and enabling
arm’s length profit allocation for all intangible value.
The US approach departs significantly from the relatively straightforward

(and neutral) approach taken by the OECD, which requires that all intra-
396. OECD TPG, at ch. 8.

397. OECD TPG, at paras. 6.92-6.103.
117
group IP value transfers are priced as unrelated parties would have done.
The OECD approach is not cluttered by arbitrary constraints in the form of
a restrictive intangibles definition or provisions that require special treat-
ment for certain forms of IP transfers. This is commendable and should
serve as a blueprint for the design of domestic transfer pricing law solu-
tions.
The problems illustrated by the pre-2018 version of the US intangibles defi-

nition raise the question of whether it is at all sensible to link transfer pric-
ing methodology to a restrictive intangibles definition. In the author’s view,
it is not. All intra-group transfers of intangible value should be subject to
arm’s length pricing, regardless of whether the transferred value qualifies
under an ultimately arbitrary definition. Of course, an intangibles defini-
tion may possibly aid in the identification of transferred intangible value,
but the same will be sufficiently ensured by a thorough delineation of the
actual transaction. The benefits of linking a definition of intangibles to the
transfer pricing methodology are unclear, but the dangers are not. As long
as such a link exists, taxpayers will have an incentive to classify intangibles
as falling outside the definition, thereby escaping transfer pricing. This can
be avoided by doing as the OECD has done: removing the link.
Both the United States and the OECD should, however, go a step further.
Focus should be set on developing the distinction between routine and non-
routine value chain contributions, as it indeed plays a truly significant role
in ensuring arm’s length profit allocation.
118
Chapter 4
Introduction to Part 2
Operating profits from IP value chains are allocated in two steps under
Internal Revenue Code (IRC) section 482 in US law and article 9 of the
OECD Model Tax Convention (OECD MTC). The operating profits must
first be allocated among the different value chain inputs. This process is
governed by transfer pricing methodology, which will be analysed in this
part of the book. For instance, say that a Norwegian subsidiary of a US-
based multinational performs routine distribution functions and licenses
a patent from an Irish group entity in connection with its local sales of a
blockbuster drug. The role of the transfer pricing method here is to allocate
the profits from the Norwegian sales among the value chain inputs, i.e. the
Norwegian routine distribution functions and the Irish-owned patent. As-
sume that the Norwegian profits are 100 and that the transfer pricing meth-
ods indicate that an arm’s length return to the subsidiary is 20. The residual
profits of 80 are allocable to the Irish patent. The next step is to allocate the
residual profits assigned to the patent among group entities. This second
step is governed by the IP ownership provisions in the US regulations and
the OECD Transfer Pricing Guidelines (OECD TPG), which are analysed
in part 3 of the book.
The author’s analysis of the transfer pricing methodology should be seen in

the context of the principal model.398 The model is contingent on successful
migration of ownership to unique intangibles developed in high-tax coun-
tries to a group entity resident in a jurisdiction with low or no taxation (or a
favourable intangible property (IP) regime), in the sense that the full value
of the IP is not taxed on the way out of the jurisdiction in which it was cre-
ated. Otherwise, there would be no point in the structure. The 2017 OECD
TPG on contractual allocations of risk,399 valuation techniques400 and hard-
to-value intangibles,401 as well as the 2009 US cost-sharing regulations,402
are responses that seek to ensure that the allocation of operating profits

399. See sec. 6.6.5.5.
400. See ch. 13.
401. See sec. 13.3.5. with respect to the consequences of unreliable valuation parame-
ters, as well as the discussion of the new 2015 OECD guidance on periodic adjustments
in sec. 16.5.
402. See ch. 14.
119
Chapter 4 - Introduction to Part 2
from IP is aligned with value creation by requiring taxation of the full IP

value on the way out.
Further, the principal model creates tension between local market jurisdic-
tions in which sales are made and routine functions are carried out, on the
one hand, and the jurisdiction in which IP ownership is located, on the oth-
er. The reason is that the residual profits will be extracted from the former
jurisdictions and allocated to the latter. The question here is of how low the
normal “routine” return allocable to the market jurisdictions can be set un-
der the current pricing methodology (in particular, the comparable profits
method (CPM) and transactional net margin method (TNMM)). The lower
this return is, the larger the residual profits (which may be extracted from
source) will be. Market jurisdictions now object to the sparse level of oper-
ating income afforded to them. Their view is that they should be entitled to
something more. Even though the OECD has issued new guidance on the
allocation of incremental operating profits generated by location savings,
local market characteristics and multinational enterprise (MNE) synergies,
it may be challenging to accommodate these jurisdictions under the pre-
vailing transfer pricing methodology, given that no unique IP (know-how,
customer relationships, etc.) can be identified in these jurisdictions that
may attract a portion of the residual profits.
The author will begin his analysis with a discussion of the historical de-
velopment of the so-called “profit-based” pricing methodology.403 Due to
increased difficulties in addressing controlled IP transactions (that resulted
in profit shifting) under the traditional, transaction-based pricing meth-
ods of the 1968 US regulations in the 1970s and 1980s, a 1986 tax reform
amending IRC section 482 was carried out, with subsequent studies form-
ing the platform for new and more effective pricing methods. This eventu-
ally resulted in the 1994 US regulations, introducing two specified profit-
based pricing methods: (i) the CPM; and (ii) the profit split method (PSM).
These methods were quickly incorporated, with some modifications, into
the OECD TPG.
In light of this historical discussion, the author analyses the current US and
OECD transfer pricing methodology relevant to IP value chains. He begins
by analysing metaconcepts underlying the US and OECD transfer pric-
ing methodologies to facilitate the subsequent discussions of the particular
methods.404 He discusses the relationship between the concept of operating
403. See ch. 5.

404. See ch. 6.
120
profits and the transfer pricing methods, the relationship between gross and
net profit methods, the selection of an appropriate transfer pricing method-
ology, the concept of the arm’s length range, the concept of comparability
and the aggregation of controlled transactions.
He then moves on to discuss the core workings of the current US and

OECD transfer pricing methodology. His focus is on how the amount of
residual profits allocable for unique and valuable IP employed in the value
chain should be determined. First, he discusses how this residual profit
amount can be determined directly under the traditional, transaction-based
US and OECD comparable uncontrolled transaction (CUT) method.405 Af-
ter this, he discusses how the residual profit amount alternatively can be
determined indirectly through the one-sided US and OECD profit-based
methods, i.e. the US CPM and its OECD counterpart, the TNMM.406 His
key focus here will be on comparability problems. Next, he discusses how
the residual profit amount allocable to a unique IP alternatively can be de-
termined directly under the two-sided profit-based US and OECD PSM.407
The 2017 OECD guidance on the allocation of incremental operating prof-
its due to location savings, local market characteristics and MNE synergies
and on the application of the transfer pricing methods in the “post-BEPS”
era will then be analysed.408 Further, he discusses the US and OECD rules
governing the use of unspecified methods.409
The rest of part 2 of the book is dedicated to analysing particular applica-

tions of the transfer pricing methodology. The author discusses the 2017
OECD TPG on IP valuation,410 the 2009 US regulations and the 2017
OECD TPG on cost-sharing arrangements (CSAs),411 as well as taxpayer
initiated compensating adjustments and periodic adjustments under US
law and the OECD TPG.412 The allocation of IP profits under article 7 of
the OECD MTC to a permanent establishment that contributes intangible
value chain inputs will be discussed in chapter 17.
405. See ch. 7.

406. See ch. 8.
407. See ch. 9.
408. See ch. 10 and ch. 11, respectively.
409. See ch. 12.
410. See ch. 13.
411. See ch. 14.
412. See ch. 15 and ch. 16, respectively.
121
Part 2
Chapter 5
The Historical Development of

Profit-Based Transfer Pricing Methodology
5.1. Introduction
The question in part 2 of the book is how to determine the amount of op-
erating profits allocable for unique intangible property (IP). As such, an
asset will generally not produce profits on a stand-alone basis, but will be
one of several inputs to an intangible value chain designed by the multina-
tional to bring products or services to the global marketplace. The ques-
tion of the transfer pricing of intangibles is really a question of how the
operating profits from the relevant value chain (i.e. from the sale of an end
product in different countries) shall be distributed among the group enti-
ties that contribute the different value chain inputs, including the unique
intangibles. The transfer pricing methodology under US and OECD law
governs – and restricts – this allocation process, thereby distributing the
profits among the contributing group entities, and thus also among the
jurisdictions in which these entities are resident for tax purposes. The ma-
terial content of the transfer pricing methodology therefore has significant
impact on how the profits of multinationals are allocated among jurisdic-
tions.
In this chapter, the author will discuss how the US and OECD transfer
pricing methodology has evolved over the last decades. The chapter will
serve as a lead-in to, and will provide context for, the analysis of the current
transfer pricing methodologies in chapters 7-16.
The common theme throughout this chapter is that historical experience

has demonstrated that there are severe practical problems associated with
applying the three traditional transfer pricing methods (the comparable
uncontrolled transaction (CUT) method, the resale price method and the
cost-plus method) used for the purpose of allocating operating profits from
intangible value chains. In the United States, these problems lead to the de-
velopment of alternative – and more substance-based – profit-based meth-
odologies (methods that allocate net profits to a group entity based on its
contributions of functions, assets and risks). Through case law, the profit
split method (PSM) was developed, and the US Internal Revenue Service
(IRS) designed the so-called “contract manufacturer theory” for profit al-
125
Chapter 5 - The Historical Development of Profit-Based Transfer Pricing
Methodology
location to “routine” group entities (this development is discussed in sec-

tion 5.2.). These two methodologies would, over time, go on to become
highly influential methods for allocating profits to intangibles in interna-
tional transfer pricing. A step along this path was when the methodolo-
gies were implemented into US transfer pricing legislation. The 1986 US
tax reform and the process towards implementing the methodology into
the 1994 Treasury Regulations are discussed in section 5.3. The methodol-
ogy was hesitantly incorporated into the OECD Transfer Pricing Guide-
lines (OECD TPG) in its 1995 revision, and has since gained a central
role among the transfer pricing methods accepted by the OECD. This is
discussed in section 5.4.
5.2. Development of the US PSM and the contract

manufacturer theory through case law
5.2.1. Introduction
In this section, the author discusses early US transfer pricing case law on
intangibles. These seminal cases were decided under the 1968 regulations
and formed the basis for the development of the profit-based transfer pric-
ing methods, i.e. the comparable profits method (CPM) and PSM, which
today enjoy a dominant position in international transfer pricing. Essen-
tial methodological concepts, in particular the idea that routine value
chain inputs should only be allocated a normal market return and that
residual profits could be allocated among unique value chain contribu-
tions according to their relative values, were developed here. This case
law remains relevant, as it illustrates the application of a profit allocation
methodology significantly similar to that in the current regulations. It also
provides useful analogies for the resolution of current profit allocation
problems. The author will refer extensively to these cases throughout his
analysis of the current regulations. He has systematized the case law into
three groups.
First, cases pertaining to inbound intangibles transfers (in which the IP

ownership resides with a foreign group entity) are discussed in section
5.2.3. Second, cases pertaining to outbound intangibles transfers (in which
IP ownership resides with a US group entity) are discussed. The first group
of outbound cases pertains to the transfer of ownership to valuable US-
developed intangibles to foreign subsidiaries under non-recognition pro-
visions from the late 1960s to the early 1980s. These are discussed in
126
Development of the US PSM and the contract manufacturer theory
through case law
section 5.2.4. The second group of outbound cases pertains to “roundtrip

transactions”.413 These are discussed in section 5.2.5. Third, there are cases
pertaining to the allocation of intangible profits through services and sales
contracts. Even though these cases were single instances and not part of an
identifiable tax-planning trend, they are illustrative. A brief discussion is
provided in section 5.2.6.
Seen together, the cases that the author has selected paint a picture of
courts struggling with the application of the CUT-based pricing methods
of the 1968 regulations, thereby resorting to an alternative profit-based
technique, i.e. the PSM. The cases contribute significantly to understand-
ing the background of the 1986 US tax reform, in which the commen-
surate-with-income standard was introduced and which would eventually
result in the 1994 final regulations, introducing the CPM and the PSM as
specified transfer pricing methods. Before beginning the case law analysis,
it is necessary to tie some comments to the profit allocation rules under the
1968 regulations.
5.2.2. The 1968 regulations and their background414
For more than 30 years prior to the issuance of the 1968 regulations, no
significant changes were made to the US transfer pricing rules.415 In the
early 1960s, concern had spread as to whether the rules were being prop-
erly enforced.416 The US Congress initiated a process to investigate how
the foreign business activities of US-based multinationals were taxed, both
through their direct presence abroad through permanent establishments
413. Roundtrips were arrangements in which foreign subsidiaries were established

to manufacture products based on US-developed intangibles to sell them back to US
entities at high market prices for resale to the end customers in the United States. The
residual profits were shifted from the United States to the foreign subsidiaries through
low royalty payments for the licensed intangibles.
414. On the 1968 US regulations, see, in particular, Culbertson et al. (2003). See also
Avi-Yonah (1995); Bischel (1973); McCawley (1990); and Duerr (1972).
415. On the history of US Internal Revenue Code (IRC) sec. 482, see supra n. 95.
416. See Miyatake et al. (1994) for a comprehensive discussion of the historical de-
velopment. The 1960s saw attention directed at US-based multinationals that operated
within the oil industry. The US Internal Revenue Service’s (IRS’s) focus was on the
profit allocation of these multinationals to low-tax countries. In the 1970s, the focus
shifted to the question of whether these multinationals were entitled to foreign tax
credits, which resulted in foreign tax credit regulations. See Exxon Corp. v. CIR, T.C.
Memo. 1993-616 (Tax Ct., 1993), affirmed by 98 F.3d 825 (5th Cir., 1996), certiorari
denied by 520 U.S. 1185 (S.Ct., 1997), pertaining to an IRC sec. 482 reallocation of
profits from crude oil sales.
127
Methodology
(PEs) and through foreign subsidiaries.417 In 1966, the Treasury, in coop-

eration with the IRS, issued proposed regulations.418 The final regulations
were issued in 1968.419
The standard to be applied to determine an arm’s length profit allocation

under the 1968 regulations was the amount that would have been paid by
an unrelated party for the same IP under the same circumstances.420 Thus,
the 1968 regulations required exact or inexact CUTs for the pricing of con-
trolled intangibles transfers, at least as the point of departure. If no CUTs
were available, the 1968 regulations prescribed a broad pricing assessment
in which a list of 12 factors should be taken into consideration “in arriv-
ing at the amount of the arm’s length consideration”.421 This provision was
often referred to as the “fourth method” under the 1968 regulations, or
simply as the “12-factor list”.
The factors were as follows:

(1) prevailing industry rates;
(2) offers of competing transferers;
(3) terms of transfer;
(4) uniqueness of the property;
(5) degree and duration of protection afforded to the property;
(6) value of services rendered by the transferer to the transferee in connec-
tion with the transfer;
417. In 1962, the US Congress deliberated on a proposal to introduce formulary

methods to allocate income from transactions in tangible property among related en-
terprises where no comparable uncontrolled prices were available; see H.R. Rep. No.
87-1447 (1962), at p. 28 (reprinted in 1962-3 CB 405). The proposal was rejected, but
the conference committee recommended that the US Treasury develop new transfer
pricing regulations that provided additional guidance; see H.R. Rep. No. 87-2508, at
18-19 (1962), reprinted in 1962-3 CB 1129, 1146. In 1963, the United States addressed
location savings reaped by US-based multinationals that moved manufacturing activi-
ties to Puerto Rico to take advantage of low labour costs and tax holidays; see Rev.
Proc. 63-10. This was one of the first US attempts to tackle the profit shifting of multi-
nationals; see the discussion in sec. 34.1.
418. The chain of regulatory development was as follows: the 1965 notice of proposed
rulemaking (30 FR 4256); the 1966 proposed regulations (31 FR 10394); and the 1968
final regulations (33 FR 5848).
419. 33 FR 5848. Pankiv (2017), at p. 33, asserts that the 1968 US regulations “are an
example of how domestic law and regulations can depart from the general arm’s length
principle”. The author finds this assertion to be unclear, as profit allocations carried
out under the pricing methodology of the 1968 regulations obviously would qualify as
arm’s length.
420. 1968 Treas. Regs. (33 FR 5848) § 1.482-2(d)(2)(ii).
421. 1968 Treas. Regs. (33 FR 5848) § 1.482-2(d)(2)(iii).
128
through case law
(7) prospective profits to be realized by the transferee through its use or

transfer of the property;
(8) capital investment and start-up expenses required of the transferee;
(9) availability of substitutes for the property transferred;
(10) arm’s length rates and prices paid by unrelated parties when the prop-
erty was resold or sublicensed;
(11) costs incurred in the transfer or in developing the property; and
(12) any other fact or circumstance that unrelated parties would have been
likely to consider in determining an arm’s length price.
The IRS and the courts made active use of this 12-factor list, as it enabled
effective allocation of arm’s length profits for intangibles in the absence of
CUTs. The list only provided high-level guidance, resulting in leeway with
respect to its application. This provided fertile breeding ground for the
courts to develop new transfer pricing methodologies, as they indeed did
with the PSM. The downside to the list was that it did not ensure uniform
profit allocations, leading to unpredictable transfer pricing outcomes. The
12-factor list has a modern-day parallel in the “unspecified transfer pricing
methods” under current US regulations and the OECD TPG,422 which al-
low discretionary profit allocation solutions as long as they align with the
overarching arm’s length principle.
5.2.3. Three inbound cases: Nestlé, French and Ciba
5.2.3.1. Introduction
In sections 5.2.3.2.-5.2.3.4., the author discusses three cases pertaining to

inbound intangibles transfers. Either directly or indirectly, the question
was whether the royalties paid by US group entities for the use of foreign-
owned intangibles were at arm’s length under IRC section 482. The cases
discussed are Nestlé Co., Inc. v. CIR in section 5.2.3.2.,423 R. T. French
Co. v. CIR in section 5.2.3.3.424 and Ciba-Geigy Corp. v. CIR in section
5.2.3.4.425
422. See the analysis of unspecified transfer pricing methods in ch. 12.
423. Nestlé Co., Inc. v. CIR, T.C. Memo. 1963-14 (Tax Ct., 1963).
424. R. T. French Co. v. CIR, 60 T.C. 836 (Tax Ct., 1973), distinguished by 1977 WL
191022 (IRS TAM, 1977), 1979 WL 56002 (IRS TAM, 1979) and 1992 WL 1354859
(IRS FSA, 1992). For recent comments on the ruling, see, e.g. Navarro (2017), at p. 244.
425. Ciba-Geigy Corp. v. CIR, 85 T.C. No. 11 (Tax Ct., 1985), acquiescence recom-
mended by AOD-1987-21 (IRS AOD, 1987) and acq. 1987 WL 857882 (IRS ACQ, 1987).
129
Methodology
5.2.3.2. Nestlé (1963)
At issue in Nestlé was whether royalties paid by a US licensee to foreign

group licensers for manufacturing and marketing intangibles (patents,
know-how and trade name) were deductible as ordinary business expenses
under IRC section 23(a)(1)(A).426 Thus, Nestlé did not pertain to a section
482 reallocation, but the theme for assessment was significantly similar to
that under IRC section 482.427
The intangibles required for the manufacturing and sales of Nescafé and
related products were owned by other group entities, mainly resident in
Switzerland, where research and development (R&D) was carried out. For
the period under review (the income years 1947-1952), the US licensee
paid annual royalties ranging from 7% to 10% on its US sales, limited to a
maximum of one third of its net profits. The royalties paid equalled a profit
split of two thirds to the US licensee and one third to the foreign licensers.
During the late 1930s and the early 1940s, the US licensee incurred signifi-
cant marketing expenditures to build the local value of the Nescafé brand
through marketing. The marketing expenses incurred by the US licensee
in 1947-1952 were approximately 33% larger than the royalties paid by the
US licensee in the same period.
The IRS disregarded the licence agreement and treated the payments as
non-deductible dividend distributions from the US licensee.428 The court
looked closely at the reasonableness of the outbound royalty payments.
The highly profitable US sales of Nescafé grew rapidly in the early 1940s,
mainly as a result of the then-unique quality of the product and the mar-
keting efforts of the US licensee. When significant sales materialized, the
group revised the US licence agreement for Nescafé. In 1941, it was de-
cided that the US licensee was to pay 5% royalties on its US Nescafé sales.
426. Now IRC sec. 162(a).

427. The agreements would be valid and enforceable if they were fair and reasonably
judged by the standards of a transaction entered into by parties dealing at arm’s length;
see Granberg Equipment, Inc. v. CIR, 11 T.C. 70479 (Tax Ct., 1948); and Differential
Steel Car Co. v. CIR, 16 T.C. 41388 (Tax Ct., 1951).
428. The reallocation was based on several arguments: (i) the controlled licence
agreements lacked clauses commonly found in third-party licence agreements; (ii)
there was an incentive to choose royalty over dividend payments because of the income
tax deductibility of the former type of payment; (iii) group entities usually did not enter
into licence agreements; (iv) the agreements were lengthy (over the remaining 14-year
life of the Nescafé patent); and (v) the US licensee (in the late 1940s, when the technol-
ogy, encompassed by the licence agreement, had become common knowledge) should
have repudiated the licence agreements and entered into the business on its own.
130
through case law
Only a year later, it was agreed to increase the royalty substantially, to

between 7-10% of sales, limited to a total of one third of the US net profits.
The US Tax Court reviewed whether the revision was warranted. It found
the profitability of US Nescafé sales largely to be a result of the continuing
R&D efforts made by the foreign licenser and that this development war-
ranted a higher royalty. The Court stated that “as long as the amount of the
royalty paid was commensurate with the value of the benefits received and
was reasonable we would not be inclined to, nor do we think we would be
justified to, conclude that the increased royalty was something other than
what it purported to be”.429
The author is not aware of any earlier transfer pricing case law that estab-
lished a principle of proportionality between actual profits related to an
intangible and royalties to be paid. In this respect, Nestlé is the seminal
case. The decision did not, however, provide much further guidance on
the content of the principle. The only legal guidance to be deduced from
the ruling is that an increase in profits should generally be followed by
an increase in royalties. Significant room was thus left for the taxpayer to
adapt its transfer pricing to changes in profits. Further, the author finds it
curious that the significance of the US marketing expenditures was not
assessed more closely, as they undoubtedly increased the local value of
the manufacturing and marketing intangibles. It could be argued that at
least part of the expenditures should have been seen as outbound royalty
payments or dividend distributions. Alternatively, it could be argued that
the US distribution entity had developed local marketing intangibles that
entitled it to residual profits.430
5.2.3.3. French (1963)
At issue in French was whether royalties paid by a US licensee to a related

UK licenser were at arm’s length under IRC section 482. The IRS denied
deductions for outbound royalty payments for the income years 1963 and
429. The Court found that the value of the patents licensed to Nestlé was evidenced
by the US profitability (which went from sales of USD 16 million in 1938 to USD 111
million in 1952, and accumulated profits of USD 31 million after royalties in the period
of 1939-1952), of which the US subsidiary retained approximately two thirds, despite
the increase in royalties. Based on a broad assessment, the Court found the royalties
paid to be reasonable for the tested income years.
430. See the analysis of the allocation of profits from intra-group-developed market-
ing intangible property (IP) under US law in ch. 23.
131
Methodology
1964, based on the view that the US licensee did not receive any benefits in
return for its royalty payments.431
The factual pattern was that in 1946, a process (including patents and tech-
nical know-how) for making mashed potatoes was licensed from a UK
licenser to a US licensee. Under the agreement, the US licensee received
the exclusive make-sell rights for mashed potatoes in the United States for
the 21-year duration of the patent protection, as well as rights to techni-
cal advice and know-how transfers from the United Kingdom. Originally,
the royalty rate was set to 3% of the US sales.432 In the early phase of the
licence agreement, all R&D was carried out in the United Kingdom, and
the US production process followed the patent description closely. The UK
licenser shared information relating to developments in the production pro-
cess with the US licensee. Gradually, the process was improved through
US production experience, and the US licensee began performing its own
R&D.
It was not disputed that the royalty was at arm’s length at the time at which
the agreement was entered into.433 The IRS’s argument was that the US
licensee did not receive any significant benefits in return for its royalty
payments to the UK licenser in the 17th and 18th income years covered by
the agreement. The IRS claimed that the process described in the origi-
nal 1946 agreement was now widely understood in the food industry and
therefore had little value left. It also claimed that the US licensee had, by
1963-1964, developed its own research capability and made the results of
431. A peculiar point is that French had little fiscal importance, as almost all US
profits were repatriated to the United Kingdom through dividend payments. The royalty
payments were insignificant. The author assumes that the case was followed through
Tax Court litigation because of the potential precedential value of the question at issue.
432. The licence agreement was altered in 1956 and 1960, when the royalty rate was
lowered to 2%.
433. First, a third party owned a significant portion of the UK licenser company at
that time. This in itself made profit shifting from the wholly owned US company to
the partially owned UK company unlikely. In DHL Corp. & Subsidiaries v. CIR, T.C.
Memo. 1998-461 (Tax Ct., 1998), affirmed in part, reversed in part by 285 F.3d 1210
(9th Cir., 2002), the taxpayer also argued that the presence of a third party in price
negotiations ensured the arm’s length character of the price. DHL is distinguishable
from French in that the third party in DHL was only concerned about the total price for
a comprehensive package of assets (shares and rights in a trademark). The third party
in DHL was indifferent as to how the taxpayer allocated the total package price among
the purchased assets. Second, the UK licenser in French also licensed the same mashed
potato process to an unrelated company in France, under a non-exclusive licence for a
royalty of 5% of local sales. The unrelated company thus paid more for less, compared
to the US licensee. Neither factor left much doubt that the agreement was at arm’s
length at the outset.
132
through case law
this process available to the UK licenser without receiving any royalties in

return.
The most interesting aspect of the case is how alterations in the balance of
the controlled parties’ value chain contributions subsequent to the forma-
tion of a long-term licence agreement should be treated for transfer pricing
purposes. The court found it inappropriate to view the income years of
1963 and 1964 in isolation, as the royalties paid then were not necessarily
compensation for value received in those particular years.434 This observa-
tion was based on comparing the royalty payments with the value of the
parties’ contributions over the duration of the agreement as a whole as op-
posed to viewing each income period separately with respect to the value
exchanged between the parties. The logic applied by the court seems to be
that the UK licenser provided most of the value in the first part of the dura-
tion of the agreement by making available the patent and know-how, train-
ing US employees and supervising newly established factory processes in
the United States, which eventually enabled the US licensee to perform its
own research and make improvements to the mashed potato process. In the
later stages, the US licensee performed most of the functions, incurred the
most risk (research) and controlled the significant asset in the relationship
(know-how for the improved process).
The shift in functions and risks between the parties seemed mainly ena-
bled by the platform established by the contributions of the UK licenser in
the early stages of the agreement. Over the course of 14 years, from 1948 to
1961, the yearly royalty rate paid by the US entity was between 2-3%. Thus,
the royalty rate charged by the UK licenser did not vary much depending
on the activity performed by him. In the author’s view, it is, to some degree,
natural that the relative lack of compensation in stages in which the UK
licenser performed the greatest part of the functions was made up for by
relative overcompensation in periods in which the UK licenser performed
fewer functions.
It is, however, not obvious that the holistic reasoning of the Tax Court was
correct. First, it could not possibly be clear at the time at which the agree-
ment was entered into that there would be a change of functions performed
and risks assumed by the contracting parties over the duration of the agree-
ment. This must particularly be the case for the core performance elements
434. Further, the Tax Court found no reason to believe that an unrelated party in the
UK licenser’s position would have allowed the US licensee to avoid its obligation to pay
royalties prior to the expiration date of the agreement.
133
Methodology
of the agreement. After all, the agreement was that the UK licenser was
to supply the US licensee with the process, not the other way around. The
best estimate of the activity to be rendered by the contracting parties over
the life of the agreement, foreseeable at the outset, was presumably the
required performance as codified in the agreement itself. As royalties were
to be paid yearly, it is tempting to deduce from this that the royalty pay-
ments were to correspond to a yearly provision of performance by the UK
licenser.
Second, given that the US licensee performed its own research, which
eventually led to a process different from the one described in the licensed
patent, it seems likely that the improved process was a new intangible, dis-
tinguishable from the licensed process. The new intangible, even if based
on the old process, would likely not be covered by the existing agreement,
since it was technically different from the patented process and developed
and funded by the US licensee. Income should then be allocated to the US
licensee pursuant to section 482 as consideration for making the new asset
available to the UK licenser.435 This would be more in line with legal and
economic realities.436
Third, the view of the Tax Court was one-sided, as it stated that “there is no
reason to believe that an unrelated party would have permitted petitioner to
avoid its contractual obligations at any time prior to the expiration date of
the 1946 and 1960 agreements”. The focus was on whether the agreement
turned out to be beneficial for the UK licenser and whether a third party
would willingly surrender such a beneficial legal position. The obvious
response from the perspective of the US entity would be to ask if an unre-
lated party would be willing to share its manufacturing intangibles free of
charge, which seemed to be exactly what the US licensee did in the later
stages of the agreement. The author finds that unlikely.437
Fourth, when the actual conduct of unrelated contracting parties over time
deviates from their codified agreement, the likely explanation must be that
435. This line of reasoning would presumably be more consistent with the intellec-
tual property law treatment of the US process pursuant to the third amendment of the
agreement, where reference to the original patent had been removed because the actual
process deviated from the patent description.
436. The new process would likely qualify as a separate intangible. See the discussion
of the US IP definition in sec. 3.2.
437. The fact that the new process represented value for the UK entity was acknowl-
edged by the Tax Court, which stated that “in later years, to be sure, MPP (through
Chivers) derived substantial benefits from petitioner’s own considerable research ef-
forts”.
134
through case law
they have entered into a modified agreement consistent with their actual
performance. The court did not assess this rather obvious question, likely
due to the one-sided perspective of the review.
The core question in French, i.e. whether a transfer pricing review of a

long-term licence agreement with fixed pricing should consider each pe-
riod within the duration of the agreement separately or as a whole, was
fundamental. The perhaps somewhat formalistic stand assumed by the
Court offered rich tax-planning opportunities for multinational enterprises
(MNEs). In essence, French would shield non-arm’s length transfers of
intangibles from reassessments, as long as the taxpayer could prove that the
agreement was at arm’s length at the outset.438
5.2.3.4. Ciba (1985)
The issue of limiting US income tax deductions for outbound royalty pay-
ments also arose in Ciba with respect to payments from a wholly-owned
US licensee to a Swiss parent licenser for US make-sell rights to agricul-
tural products.439 The IRS denied deductions in full and reclassified the
payments as deemed dividend payments subject to US withholding tax.440
The principal assertion of the IRS was to impute a cost-sharing agreement

(CSA) between the US licensee and the Swiss licenser, under which the
former was entitled to residual profits, based on the observation that the US
licensee had been involved in a joint R&D effort with the parent company.
The court rejected this argument, as there was no evidence supporting any
significant R&D contribution from the US subsidiary, nor had it agreed to
share any costs or benefits in relation to the R&D activities. Further, the
IRS claimed that the agreed royalty rate of 10% had to be adjusted down-
wards to 6%.441 As no CUTs were available, the court rejected the assertion
438. The 1986 US tax reform specifically addressed and rejected the result of French
by introducing the commensurate-with-income standard, under which, periodic adjust-
ments are to be carried out in each income year to ensure that the income from intangi-
bles is commensurate with the profits generated by them; see sec. 5.3.3. The author also
refers to the discussion of periodic adjustments in ch. 16.
mended by AOD-1987-21 (IRS AOD, 1987) and acq. 1987 WL 857882 (IRS ACQ,
1987). For a discussion of the case, see, e.g. Wittendorff (2010a), at p. 645; and Brauner
(2008), at p. 134.
440. At least to the author’s knowledge, at the time at which the case was filed with
the Tax Court in August 1985, it was the largest transfer pricing reallocation to be tried
before a US court.
441. 1968 Treas. Regs. (33 FR 5848) § 1.482-2(d)(2).
135
Methodology
based on the 12-factor list in the 1968 regulations,442 taking into account
royalty offers that the parent company had received from competing licen-
sees, prevailing rates in the industry and required capital investments and
start-up costs for the licensee.443
5.2.4. Three outbound cases: Eli Lilly, Searle and Merck
In sections 5.2.4.2.-5.2.4.5., the author discusses three transfer pricing cas-

es litigated in the 1980s and early 1990s involving significant reallocations
of residual profits. The cases pertained to structures where high-value, US-
developed intangibles had been transferred tax-free under non-recognition
provisions to related Puerto Rican subsidiaries.444 Eli Lilly and Company
and Subsidiaries v. CIR is discussed in section 5.2.4.3.,445 G.D. Searle &
Co. v. CIR is discussed in section 5.2.4.4.446 and Merck & Co., Inc. v. CIR
is discussed in section 5.2.4.5.447
The first two cases, in particular, Eli Lilly, are relevant to the interpreta-
tion of the profit split method of the current US regulations. The third case
(Merck) did not result in a reallocation of residual income, but the author
includes a modest discussion of it, as it, in his view, should have resulted in
a profit split, and is therefore relevant as an analogy in discussions of the
profit split method.
Before beginning the discussion, the author will provide a brief overview
of relevant background law pertaining to US investments in Puerto Rico
in section 5.2.4.2.

443. Thus, the entire reallocation of approximately USD 23 million fell.
444. The transfers were carried out under IRC sec. 351 as non-taxable contributions
to the capital of the involved subsidiaries. For an informed analysis of transfer pricing
issues in this context, see O’Brien et al. (2007).
445. Eli Lilly and Company and Subsidiaries v. CIR, 84 T.C. No. 65 (Tax Ct., 1985),
affirmed in part, reversed in part by 856 F.2d 855 (7th Cir., 1988).
446. G.D. Searle & Co. v. CIR, 88 T.C. 252 (Tax Ct., 1987).
447. Merck & Co., Inc. v. U.S., 24 Cl.Ct. 73 (Cl.Ct., 1991). The Court of Claims deci-
sion was rendered in 1991, but the case pertained to the income years 1975 and 1976
and was decided under the 1968 Treas. Regs. (33 FR 5848). Hence, the author finds the
case relevant for the discussion of case law prior to the 1986 US tax reform.
136
through case law
5.2.4.2. The historical tax treatment of investments in

US possessions448
US multinationals with business activities in US possessions were original-

ly subject both to federal corporate income tax and taxes levied by the US
possession.449 To encourage US investments into possessions and to remove
the competitive disadvantage of such investments relative to foreign direct
investment (FDI) in the United States by foreign companies,450 Congress
exempted business income earned by US companies in a possession from
US federal income tax by enacting IRC section 931.451 Local taxation in
possessions was also generally beneficial.452 The result was zero taxation of
the profits earned by possession subsidiaries, as long as they were not dis-
tributed to the US parent as dividends.453 Accumulated profits were often
reinvested abroad by the possessions subsidiaries.454
5.2.4.3.
Eli Lilly (1985)
Eli Lilly is a landmark transfer pricing case for several reasons.455 First,
the court in Eli Lilly developed an allocation pattern for the profit split
method, in which the residual profits were divided among the controlled
448. The term “US possession” refers to a specific geographic territory under the
jurisdiction of the United States for which the US Congress has decided that the US
Constitution is to be applied to the territory’s local government and inhabitants in the
same manner as it applies to the local governments and residents of the states of the
United States (examples include Puerto Rico and the US Virgin Islands).
449. See sec. II of the Tariff Act of 1913 (Ch. 16, 38 Stat. 166); and the Revenue Act
of 1918 (Ch. 18, 40 Stat. 1058).
450. See H. Rept. No. 350, 67th Cong., 1st Sess. 1 (1921), 1939-1 C.B. (Part 2) 168,
174.
451. IRC sec. 931 had its origin in the Revenue Act of 1921, which exempted income
from US companies operating within a US possession; see the Revenue Act of 1921,
sec. 262, Ch. 136, 42 Stat. 271. The provision was re-enacted without material change
in 1928 and 1954 (see, respectively, sec. 251 of the Revenue Act of 1928, Ch. 852, 45
Stat. 850; and sec. 931 of the IRC of 1954, Ch. 736, 68A Stat. 162). The first material
change of the provision was made in 1976, when Congress substituted it with an equally
favourable tax credit regime; see IRC sec. 936; and the Tax Reform Act of 1976, Pub.
L. 94-455, 90 Stat. 1643.
452. In particular, in Puerto Rico, where special investment incentive programmes
were put into place.
453. See IRC sec. 246(a)(2)(B).
454. Congress altered the dividends-received deduction in 1976 in order to encourage
the multinationals to invest their possession profits in the United States.
affirmed in part, reversed in part by 856 F.2d 855 (7th Cir., 1988). On Eli Lilly, see, in
particular, Brauner (2008), at p. 136. See also Wittendorff (2010a), at p. 754.
137
Methodology
parties according to the relative value of their intangible value chain con-
tributions. This method would later become highly influential in both US
and international transfer pricing jurisprudence, and is now, alongside the
US CPM and the OECD TNMM, the de facto profit allocation method for
residual profits.
Second, Eli Lilly marked the beginning of the IRS’s “contract manufac-
turer” litigation stance, pursuant to which a group licensee that merely
contributed routine inputs to the value chain should not be allocated any
residual profits, but receive a normal market return. Also, the contract
manufacturer return theory would later become highly influential in US
transfer pricing law through the proposed basic arm’s length return method
(BARLM) of the 1988 White Paper and the CPM of the 1994 regulations,
as well as in international transfer pricing law through the TNMM of the
OECD TPG.
Third, Eli Lilly was the first of three transfer pricing cases involving Puerto
Rican subsidiaries to be litigated in the 1980s. In its time, with a realloca-
tion of USD 71 million, it was the largest transfer pricing case litigated
before a US court. The case supposedly also marked the start of a new ap-
proach by the IRS in developing big transfer pricing cases for reassessment
and litigation, involving large multidisciplinary audit teams.
In Eli Lilly, the patents and technical know-how connected to a top-selling

drug in the 1960s US pharmaceutical market (the painkiller Darvon and
its successor, Darvon-N) were transferred from a US parent company to
a wholly owned Puerto Rican subsidiary. These intangibles represented
the main value in the Darvon value chain, as the patent protected Eli Lilly
from direct competition. The marketing intangibles connected to Darvon
(the trade name and so forth) were not transferred. The US parent had de-
veloped the drug through its own costly R&D.
The idea behind the restructuring was to transfer ownership of the manu-
facturing intangibles to the Puerto Rican subsidiary. The residual profits
allocable for the manufacturing intangible would thus be taxable in Puer-
to Rico, not the United States. This profit allocation was operationalized
through the prices charged for the finished drug from the Puerto Rican
subsidiary to the US parent, which would then sell the drug to unrelated
customers in the United States. The price was set so that the parent could
recover its costs on marketing functions and earn a profit of 90-100% on
those expenses. The residual profits allocated to the subsidiary were not
subject to tax. The return position of the taxpayer was therefore that the
138
through case law
residual profits from the US-owned marketing intangibles were also allo-
cable to the Puerto Rican subsidiary. The reassessment allocated the entire
residual profit (from both manufacturing and marketing intangibles) to the
US parent. The Puerto Rican subsidiary was treated as a mere contract
manufacturer and remunerated on a cost-plus basis. Thus, the taxpayer and
IRS positions were extreme opposites.
The principal IRS argument was that the IRC section 351 non-recognition
transfer of the manufacturing intangibles was carried out to avoid federal
income tax and resulted in a distortion of income for the US parent.456 The
court found that IRC section 482 in principle could be used to disregard
section 351 transfers in narrow circumstances, but that the factual pattern
in Eli Lilly did not qualify for such treatment.457 Even as the Court was un-
willing to disregard the intangibles transfer in and of itself, it did find that
a section 482 reallocation could be made with respect to the subsequent
pricing of goods. The view of the Court was that the controlled pricing of
sales from the Puerto Rican subsidiary to the US parent did not allocate
sufficient profit to the parent so that it could fund its R&D expenses. The
IRS reallocation was, however, set aside because it did not afford any re-
sidual profits to the Puerto Rican subsidiary.
The CUP, resale price and cost-plus methods of the 1968 regulations were
found inapplicable due to a lack of CUTs. The court based its profit al-
location on the PSM, which it, perhaps questionably, claimed support
for in the 1968 regulations,458 as well as in case law, in particular PPG
Industries Inc. v. CIR,459 Eli Lilly & Co. v. U.S.460 and Lufkin Found-
ry and Mach. Co. v. CIR.461 Distinguishing the Fifth Circuit reversal of
456. The IRS further argued that there was no good reason for the US parent company
to transfer its highly valuable intangibles to a wholly owned subsidiary in Puerto Rico.
The court rejected this “realistic alternatives” argument on the basis that transfer pri
cing should be based on the transactions actually carried out by the parties.
457. See prior case law: National Securities Corp. v. CIR, 46 B.T.A. 562 (B.T.A.,
1942), affirmed by 137 F.2d 600 (C.C.A.3, 1943), certiorari denied by 320 U.S. 794
(S.Ct., 1943); Southern Bancorporation v. CIR, 67 T.C. 1022 (Tax Ct., 1977); and
Northwestern Nat. Bank of Minneapolis v. U.S., 1976 WL 1016, (D.Minn. 1976), af-
firmed by 556 F.2d 889 (8th Cir., 1977).
458. 1968 Treas. Regs. (33 FR 5848) § 1.482-2(e)(1)(iii).
459. PPG Industries Inc. v. CIR, 55 T.C. 928 (Tax Ct., 1970).
460. Eli Lilly & Co. v. U.S., 372 F.2d 990 (Cl.Ct, 1967).
461. Lufkin Foundry and Mach. Co. v. CIR, T.C. Memo. 1971-101 (Tax Ct., 1971).
In the first two cases, PPG Industries Inc. v. CIR and Eli Lilly & Co. v. U.S., the
profit split method was only used to support a profit allocation carried out under
other transfer pricing methods. In PPG, the court upheld the taxpayer’s comparable
uncontrolled price (CUP)-based pricing of sales of from a US parent to a Swiss resale
139
Methodology
Lufkin, the Court found the transactional methods of the 1968 regulations
“clearly inapplicable due to a lack of comparable or similar uncontrolled
transactions” and that “some fourth method not only is more appropri-
ate, but is inescapable”. The Court then went on to apply the PSM, in
which the profits from sales of the Darvon drug for 2 of the income years
under review were divided between the US parent and the Puerto Rican
subsidiary.462
The PSM applied in Eli Lilly essentially consisted of two elements. First,
each controlled party was allocated a normal market return on its func-
tions. Thus, the US parent was allocated a return on its marketing func-
tions, and the Puerto Rican subsidiary a return on its manufacturing ac-
tivities. Due to the extreme profit associated with the Darvon value chain,
these “normal” returns were generous.463 Second, the residual profits were
allocated to the intangibles exploited in the Darvon value chain, i.e. the
US-owned marketing intangibles and the Puerto Rican-owned manufac-
turing intangibles. Based on a broad assessment of the relative contribu-
tions of these intangibles to the profits from the Darvon drug, the court
found that while the manufacturing intangibles were responsible for most
of the profits, the marketing intangibles also were central to the value crea-
tion. On this basis, 45% of the residual profits were allocated to the US
parent.
subsidiary, which split the profit with 55% to the United States and 45% to Switzer-
land. In Eli Lilly, the court upheld the IRS’s reallocation of profits from a distribu-
tion subsidiary to a parent company. Transfer prices on sales from the parent to the
subsidiary were set low in order for the subsidiary to realize the bulk of the profits
from final sales. In Lufkin, the IRS reallocated income from two export sales and
service subsidiaries to a manufacturing parent company. The Tax Court rejected the
comparable advocated by the IRS due to functional differences. The taxpayer pricing
was accepted based on an overall assessment, the key factor of which was a profit split
argument. On appeal, the Fifth Circuit rejected – in the author’s view, on a poorly
founded basis – the Tax Court’s decision in Lufkin, holding that the transaction-based
methods of the 1968 regulations required “evidence of transactions between uncon-
trolled” taxpayers.
462. For these 2 years, the patent protection for Darvon applied, and thus there were
no comparables available. For the third year on review, the patent protection had ex-
pired and competitors had entered the market. Comparables were available for that
year, and the CUP method was applied.
463. The parent company was afforded a 25% mark-up on its marketing functions.
The Puerto Rican subsidiary received a 100% mark-up on its manufacturing functions.
Location savings were exclusively allocated to the subsidiary, as the Court viewed the
specific benefits of localization in Puerto Rico (low labour costs, tax holidays and so
forth) allocable to the subsidiary only.
140
through case law
In the author’s view, the profit split between the United States and Puerto
Rico was surprisingly even (45/55), given that the manufacturing intangi-
bles were the most valuable assets in the Darvon value chain for 2 of the
3 income years under review. The value of the marketing intangibles was,
to some degree, contingent on the continued access to the Darvon patent.
Regardless of this, the Court essentially divided the residual profits equally
between the two groups of intangibles. The reasoning was not elaborated
on. The logical allocation would have been to split the residual profit in
proportion to the relative values of the marketing and manufacturing in-
tangibles, which likely would have skewed the profit split further to the
benefit of Puerto Rico.
Eli Lilly was appealed, but the Court of Appeals upheld the allocation of
the Tax Court with only minor adjustments. The most significant element
in the Court of Appeals ruling was that it altered the Tax Court’s allocation
of R&D expenses, which was an allocation that did not have a correspond-
ing effect on the Puerto Rican subsidiary. The Court of Appeals seemed to
have misunderstood this, as it stated:
[T]he only cost increase for Lilly P.R. to which Lilly specifically objects on
appeal is a charge to Lilly P.R. for general research and development expens-
es. Because we reject the Tax Court’s basis for allocating general research
and development expenses to Lilly P.R., we uphold this objection. In all other
respects we affirm the Tax Court’s cost adjustments.
Further, the Court of Appeals stated that “the Tax Court declined to re-
solve this issue in light of its substantial reallocation of general research
and development costs to the subsidiary”.
Both statements show that the Court of Appeals was under the impression
that the Tax Court reallocated R&D expenses from the parent company
to the subsidiary. That is incorrect. The Tax Court simply charged a por-
tion of the general R&D expenses of the parent company to the Darvon
profits of the parent. This reduced the combined net income of the two
companies, and therefore ultimately also the profits allocated to the parent
company under the PSM; the profits to be split became smaller as a result
of the charging of general R&D costs. The consequence of the Court of
Appeals’ position was, as far as the author can see, that the charging of
general R&D costs to the profits of the parent had to be removed, increas-
ing the total net profits of the two companies, and therefore ultimately
also the amount of profits to be allocated to the parent company under the
PSM.
141
Methodology
5.2.4.4. Searle (1987)
In Searle, a US pharmaceutical parent transferred self-developed manu-

facturing and marketing intangibles connected to established, high-profit
products to a wholly owned Puerto Rican subsidiary.464 The factual pat-
tern in Searle was distinguishable from Eli Lilly in that the subsidiary did
not sell the finished drugs back to the parent, but instead sold directly to
unrelated US customers.465 The question that Searle had in common with
Eli Lilly was whether the Puerto Rican subsidiary should be regarded as
the owner of the transferred intangibles for IRC section 482 purposes and
thereby allocated the residual profits.
The IRS argued that there was no business purpose behind the transfer
and that an adjustment was necessary to correct the distortion of income.
The IRS in essence disregarded the subsidiary’s ownership position and
treated it as a contract manufacturer for profit allocation purposes. The
Tax Court, Judge Wiles presiding,466 regarded the subsidiary as both the
legal and economic owner of the intangibles, but found that unrelated par-
ties, for the transfer of similar intangibles, would have demanded a lump-
sum payment or a substantial share of the profits of the transferee or a
royalty.
The Court held that the transferred intangibles were of little value to
the subsidiary without the marketing, administrative and regulatory
services provided by the parent. Using its best judgement, the Court
found that 25% of the subsidiary’s total net sales in 1974 and 1975 were
to be allocated to the US parent. The result was arguably both suitable
for the particular factual pattern in the case and reasonable. The Court,
however, did not offer much in the way of reasoning for the chosen
profit split.
Also, the result seems incompatible with the fact that the court did find the
subsidiary to be the legal and economic owner of all relevant intangibles.
In light of Eli Lilly, where the profit split was decided by weighing the rela-
tive economic contributions and significance of intangibles owned by each
contracting party, there should have been specific language addressing why
the parent company was legally entitled to a share of the residual profits
464. G.D. Searle & Co. v. CIR, 88 T.C. 252 (Tax Ct., 1987). On Searle, see Brauner
(2008), at p. 140. See also Wittendorff (2010a), at p. 646.
465. However, the parent company did provide sales and marketing services under a
separate service agreement.
466. Wiles also wrote the Tax Court decision in Eli Lilly.
142
through case law
when the company did not have any ownership interests in the intangibles
in the income years under review. A key point in the profit allocation was
what the Court deemed should have been done, not the actual transaction.
On this point, the Court went against its own clear statements in Eli Lilly
on basing transfer pricing on actual transactions. The ruling may, on this
point, be seen as applying the now-established principle that controlled
pricing must respect the pricing boundaries dictated by the realistic alter-
natives of the controlled parties.
5.2.4.5. Merck (1991)
In Merck, a US parent transferred manufacturing intangibles for estab-

lished pharmaceutical chemicals to a wholly owned Puerto Rican subsidi-
ary as contribution to capital in an IRC section 351 non-recognition trans-
fer.467 Prior to the transfer, the subsidiary licensed these intangibles from
the parent. All R&D for the transferred intangibles had been performed
in the US by the parent since the 1950s. The subsidiary manufactured and
sold the pharmaceutical product directly to foreign group companies for
resale in non-US markets.
Prior to the transfer, the subsidiary was dependent on three main con-
tributions from the parent: (i) the manufacturing intangibles; (ii) con-
current R&D; and (iii) a marketing programme. The IRS found that the
package of these three elements should be compensated with an annual
10% royalty payment from the subsidiary, pursuant to IRC section 482.
The twist in Merck was that the IRS, in light of its failures in Eli Lilly
and Searle, did not focus on the transferred manufacturing intangibles.
Instead, it argued that the parent’s “other contributions”, in particular
a marketing programme, were valuable enough to warrant a 7% royal-
ty.468
In essence, the parent’s marketing was nothing more than the vertical in-
tegration of the Merck group. The marketing advantage of this structure
was simply that every last bit of production by the Puerto Rican subsidiary
was purchased by other group entities for resale in their local markets.
467. Merck & Co., Inc. v. U.S., 24 Cl.Ct. 73 (Cl.Ct., 1991). The value of the trans-
ferred intangibles was estimated to be USD 235 million in 1975.
468. The IRS viewed the transfer of ownership to the manufacturing intangibles
largely as a formality, but the reassessment was based on the argument that the parent
company supplied the subsidiary with marketing services that were separate from the
transferred intangibles.
143
Methodology
Thus, the subsidiary, in reality, had no inventory or market risk. The sub-
sidiary benefited from the network of companies within the Merck group,
as well as the group’s planning and pricing mechanism. The IRS deemed
this organizational or synergy value as an intangible owned by the parent
company, which the subsidiary was obliged to pay for pursuant to IRC sec-
tion 482.469
The Court found that for an element to be compensated as an intangible un-

der the 1968 regulations, the element should have “substantial value as an
independent property”.470 No contracts were in place that legally required
other group entities to purchase products from the subsidiary. The fact that
other group entities actually purchased products from the subsidiary was
explained by the “group’s vertical integration and system of […] control”.
The Court dismissed the assertion that the organizational structure in itself
was an intangible for which compensation was required under IRC section
482, with the rationale that “the mere power to determine who in a con-
trolled group will earn income cannot justify a Section 482 allocation from
the entity that actually earned the income”.
The reallocation fell as a result of the Court’s rejection of the synergy

argument. Merck was therefore successful in shifting the entirety of the
residual profits from the transferred manufacturing intangibles away from
US taxation, resulting in zero taxation of that profit. The result of the rul-
ing stands in contrast to that of Eli Lilly and Searle, which also involved
non-recognition transfers of intangibles under IRC section 351. Of course,
here, the IRS’s profit split assertion rested solely on the synergy argument.
The author will not rule out that the Court may have been influenced by
the fact that Merck, as opposed to Eli Lilly and Searle, did not involve
sales to the United States of the product produced by the Puerto Rican
subsidiary. This might have contributed to giving the factual pattern less
of a roundtrip feel. Still, the result of Merck is pronouncedly difficult to
reconcile with, in particular, Searle. In both cases, the parent company
provided valuable services closely connected to the transferred intangi-
bles, but only in Searle were the services rewarded with a cut of the re-
sidual profits.
469. See the discussion of the US IP definition in sec. 3.2. and the discussion of the
allocation of incremental profits due to synergies under the OECD Model Tax Conven-
tion (OECD MTC) in sec. 10.7.3.
470. 1968 Treas. Regs. (33 FR 5848) § 1.482-2(d)(iii). See the analysis of the US IP
definition in sec. 3.2.
144
through case law
5.2.5. Four roundtrip cases: Bausch, Sundstrand, Perkin

and Seagate
By the mid-1970s, the IRS focus went beyond the typical Puerto Rican
structures to more elaborate structures involving US-based multinationals
with operations in Europe and Southeast Asia, where the manufacturing
of products intended for sale in the US market were migrated to low-cost
and tax-effective jurisdictions, typically Ireland or Singapore. The end
products were shipped back to the United States for sale. Due to the “full
circle” structure of these value chains, the transactions were referred to as
“roundtrips”.
These structures typically involved two legal instruments. First, a licence

agreement was used to deploy the required intangibles from the US parent
(IP owner) to a foreign manufacturing subsidiary. Second, a sales agree-
ment was used to set the terms for the sale of finished physical products
from the foreign subsidiary back to the US parent. Through the combina-
tion of low royalty payments for the outbound licensing and high prices
for the inbound product sales, these instruments effectively allocated the
residual profits from these value chains away from US taxation.
Multinationals that undertook this type of reorganization would normally

be indifferent towards the amount of profits allocated under each sepa-
rate instrument, as the point would be the net result of both instruments,
provided that each type of income was otherwise subject to the same tax
treatment (withholding taxes, tax credits, etc.). Reassessments of roundtrip
structures tended to hone in on the licensing agreement, where the issue
would be whether the royalty rate was at arm’s length. The reason for this
was that a multinational would normally be in a position where it could
defend its sales price for the physical end product from the low-tax juris-
diction subsidiary to the high-tax jurisdiction parent due to comparable
products. A high sales price into the United States would ensure that the
residual profit was placed in the jurisdiction of the selling subsidiary. It
would normally be more challenging to defend a royalty rate paid as com-
pensation for the exploitation of a unique intangible, as there likely would
be no CUTs available.
145
Methodology
In sections 5.2.5.2.-5.2.5.5., the author discusses four roundtrip cases liti-

gated in the late 1980s and early 1990s.471 All cases resulted in profit splits,
where the residual profits were split to provide a middle ground between
the extremes of the IRS’s contract manufacturer theory and the tax-mo-
tivated allocations of the multinationals. The profit splits in these cases
do not seem to follow any particular observable pattern, apart from, of
course, avoiding one-sided extremes. Bausch & Lomb Inc. v. CIR is dis-
cussed in section 5.2.5.2.,472 Sundstrand Corporation v. CIR is discussed
in section 5.2.5.3.,473 Perkin-Elmer Corporation v. CIR is discussed in sec-
tion 5.2.5.4.474 and Seagate Technology, Inc. v. CIR is discussed in section
5.2.5.5.475
5.2.5.2. Bausch (1989)
In Bausch,476 a US parent licensed self-developed manufacturing and mar-

keting intangibles for contact lenses to a wholly owned Irish subsidiary for
an annual royalty of 5% of the subsidiary’s sales. The subsidiary manu-
factured the lenses in Ireland and sold them to the US parent and foreign
group entities.477
The IRS deemed the Irish subsidiary as a contract manufacturer and at-
tacked the transfer pricing under both the inbound sales agreement and the
outbound royalty agreement. The argument was that the US parent would
not have been willing to pay an unrelated party more for contact lenses
than the cost of self-manufacturing. It was clear that the US parent could
have manufactured the contact lenses itself domestically at a cost of USD
1.5 per contact lens, but instead chose to purchase lenses from its Irish
subsidiary at a cost of USD 7.5. The IRS was indifferent as to whether addi-
471. Concerning income years from the late 1970s to the mid-1980s.
472. Bausch & Lomb Inc. v. CIR, 92 T.C. No. 33 (Tax Ct., 1989), affirmed by 933 F.2d
1084 (2nd Cir., 1991).
473. Sundstrand Corp. and Subsidiaries v. CIR, 96 T.C. No. 12 (Tax Ct., 1991).
474. Perkin-Elmer Corp. and Subsidiaries v. CIR, T.C. Memo. 1993-414 (Tax Ct.,
1993).
475. Seagate Technology, Inc. v. CIR, 102 T.C. No. 9 (Tax Ct., 1994), acquiescence in
result only recommended by AOD-1995-09 (IRS AOD, 1995) and acq., 1995-33 I.R.B.
4 (IRS ACQ, 1995).
1084 (2nd Cir., 1991). For discussions of the ruling, see Brauner (2008), at p. 142; An-
drus (2007), at pp. 639-641; Wittendorff (2010a), at pp. 646 and 653-655; and Navarro
(2017), at pp. 245 and 262.
477. Approximately 60% of the Irish production in the years under review was sold
back to the US parent.
146
through case law
tional profits were allocated to the United States through a decrease in the
outbound product purchase price or an increase in the royalty, as long as
the result was that the Irish subsidiary was allocated only a normal market
return for its routine contributions to the value chain.478
The first question was whether the USD 7.5 lens purchase price was com-
patible with IRC section 482. The taxpayer offered third-party agreements
supporting the USD 7.5 transfer price as evidence, which the court found
sufficiently comparable to apply the CUP method, thereby rejecting the
IRS’s argument that the price should be lowered to the USD 1.5 cost level,
at which the US parent could have self-manufactured the lenses.
The second question was whether the 5% royalty rate was sufficient. In
the absence of comparable licensing agreements, the court turned to the
12-factor list of the 1968 regulations,479 pursuant to which the “prospec-
tive profits to be realized […] by the transferee through its use […] of the
property” and “the capital investment and starting up expenses required of
the transferee” could be taken into account. The extreme profits of the Irish
subsidiary clearly made an impression. The Court deemed it necessary to
adjust the royalty rate upwards from 5% to 20%, in effect constituting a
split of the residual profits.480 Even after that adjustment, the Irish subsidi-
ary was enormously profitable, with an internal rate of return of 27%.
478. The IRS claimed that the subsidiary in reality was guaranteed to sell all of its
production intra-group and should thus not be entitled to high returns (as it had no price
or volume risk). This resembled the synergy argument professed by the IRS in Merck
& Co., Inc. v. U.S., 24 Cl.Ct. 73 (Cl.Ct., 1991). In Merck, however, the question was
whether the marketing advantage of a vertically integrated multinational enterprise
(MNE) structure was a separate intangible for which the US entity should receive ad-
ditional income. In Bausch, the argument was turned upside-down: the focus was now
on the foreign entity, and the question was whether it was entitled to abnormal profits,
given its de facto risk-free market position. The Court found that the subsidiary only
had “certain expectations as to the volume and price it could anticipate selling to”
group entities and that such expectations did not constitute a guarantee that effectively
insulated it from market risks. The Tax Court thus rejected the contract manufacturer
argument, as there was no legal requirement for the US parent to purchase contact
lenses from the Irish subsidiary.
479. Treas. Regs. (33 FR 5848) § 1.482-2(d)(2)(iii).
480. The Court distanced itself from the use of the actual profits in making its profit
allocation by stating that “such information would not have been available in 1980
to a potential licensee negotiating a license agreement which was entered on January
1, 1981. The arm’s-length nature of an agreement is determined by reference only to
facts in existence at the time of the agreement, R. T. French Co. v. Commissioner, 60
T.C. 836, 852 (1973)”. It relied on taxpayer estimates produced at the time at which the
licence agreement was entered into, showing that the Irish subsidiary was likely to be
highly profitable. Given this high level of earnings, the Court found that no independent
party would have been willing to accept a royalty that would “preclude any reasonable
147
Methodology
In its assessment, the Tax Court chose a middle ground between the con-
tract manufacturer assertion of the IRS, which placed all residual profits
in the United States, and the taxpayer assertion, which placed almost all
residual profits in Ireland. In particular, the Court placed weight on two
elements. First, the Irish subsidiary was only exposed to a “moderate level
of risk”, indicating entitlement to a smaller portion of the residual prof-
its. Second, in third-party contexts, each party is usually in possession of
some unique quality that makes a licensing relationship interesting for the
other party. In this case, the US entity was in possession of all resources
required, particularly as it owned both the manufacturing and marketing
intangibles.
The Irish subsidiary was, at first, not much more than a “cash box”, which
would have given it a weak bargaining position when negotiating a licens-
ing agreement at arm’s length. The Court found that an equal split of the re-
sidual profits between the US and Irish entities, equalling a royalty rate of
20%, left the Irish subsidiary with a generous cut of the intangible-related
profits, given its moderate bargaining position. Given the reasoning of the
Court, the author finds the result puzzling. It seems clear that all unique
value chain inputs were provided by the US parent. It should therefore have
been allocated the entire residual profits, in line with the IRS contract man-
ufacturer assertion. The Irish subsidiary, as a routine functions provider,
should have been allocated a normal market return on its contributions.
Such a profit allocation, in the author’s view, would have been the result
had the case been tried under the current rules, as the CPM then likely
would have been applied.481
5.2.5.3. Sundstrand (1991)
In Sundstrand,482 a US high-tech company licensed manufacturing and

marketing intangibles for an advanced aircraft component to a Singapore
subsidiary. The royalty rate was set to 2% of the subsidiary’s sales. The
parent company purchased all components manufactured by the subsidi-
expectation of earning a profit through use of the intangibles”. In the author’s view,
there was little reality in the Court’s reservation towards using actual profits for the
purpose of determining the profit split. The transferred intangibles were fully devel-
oped and proven, and the projected profits did not seem to vary noticeably from the
actual profits.
481. The same result would also follow from the transactional net margin method
(TNMM) in the context of the OECD MTC.
482. Sundstrand Corp. and Subsidiaries v. CIR, 96 T.C. No. 12 (Tax Ct., 1991). On
the ruling, see Brauner (2008), at p. 145. See also Wittendorff (2010a), at p. 647.
148
through case law
ary at the market catalogue price minus a 15% discount. The IRS argued
that the subsidiary should be regarded as a contract manufacturer and
remunerated on a cost-plus basis. The Tax Court dismissed the argument.
First, with respect to the sales agreement, the Court accepted the cata-
logue-based transfer price, with a minor adjustment of the discount. Sec-
ond, with respect to the royalty pricing, the Court found no CUTs,483 and
therefore went on to consider the 12-factor list of the 1968 regulations.484
It used licence agreements that the parent had entered into with third
parties as its point of departure, generally requiring a royalty around
6%. The Court adjusted the royalty upwards, due to specific benefits be-
stowed upon the subsidiary in the licensing contract, including the benefit
of entering a virtually risk-free market and using the parent’s market-
ing intangibles. Based on an overall assessment, the Court adjusted the
royalty rate from 2% to 10%, which was, in effect, a split of the residual
profits.485
5.2.5.4.
Perkin-Elmer (1993)
In Perkin-Elmer,486 a US parent in the high-tech industry established a

wholly owned subsidiary in Puerto Rico to manufacture specialized ana-
lytical instruments, accessories and lamps. The parent licensed its manu-
facturing and marketing intangibles to the subsidiary in exchange for a
3% royalty on the sales of the subsidiary.487 The subsidiary purchased the
components necessary for its production from the parent, and the parent
purchased the finished products from the subsidiary. The IRS reallocated
USD 30 million in income to the United States for the income years under
review, based on the contract manufacturer theory, leaving the subsidiary
with only a normal market return.488
With respect to the product pricing, the Court did not find the taxpayer’s
proposed benchmark transactions sufficiently comparable to apply the re-
483. The agreements differed in, inter alia, scope of rights licensed, differences in
territorial sales restrictions, cost allocation for technical assistance and extension fees.
485. Technical assistance provided by the parent company was remunerated on a
separate basis.
486. Perkin-Elmer Corp. and Subsidiaries v. CIR, T.C. Memo. 1993-414 (Tax Ct.,
1993). For a discussion of the ruling, see Brauner (2008), at p. 149.
487. The licensed intangibles included the right to manufacture the products in Puer-
to Rico, worldwide sales rights and the rights to all necessary information and know-
how on current and future licensed products.
488. The reallocation used the cost-plus method to allocate income to the subsidiary.
The contract manufacturer argument was abandoned by the Commissioner pre-trial.
149
Methodology
sale price method to the sale of finished products from the subsidiary to the
parent company.489 The transactions were, however, sufficient to convince
the Court that the prices charged fell within the arm’s length standard.490
With respect to the royalty pricing, the parties agreed that a licence agree-
ment that the parent company had entered into with a third party could be
used as a comparable under the 1968 regulations. The Court adjusted the
royalty rate up from 3% to 7.5%, increasing the profit split in favour of the
United States.491
5.2.5.5. Seagate (1994)
In Seagate,492 a US high-tech parent established a wholly owned subsidi-

ary in Singapore to manufacture and sell hard drives. The parent licensed
manufacturing (and some marketing) intangibles necessary for the produc-
tion and sale of the drives for a royalty of 1% of the subsidiary’s US sales.
The subsidiary sold finished drives and components back to the parent
company at cost plus 25%.493
The IRS reallocated USD 38 million in total taxable income to the parent
for the income years under review based on the assertion that the royalty
rate paid by the subsidiary should be adjusted upwards to 3%. Seagate ar-
gued that the subsidiary had low earnings, equalling approximately only a
profit margin of 20-30% on costs, and could therefore not sustain a higher
royalty payment. The essence of Seagate’s argument was that the licensed
489. See the resale price method in the 1968 Treas. Regs. (33 FR 5848) § 1.482-2(e)
(3).
490. See the discussion of the arm’s length range in sec. 6.6.1.
491. The adjustment only encompassed the sale of instruments; no adjustment was
made to the pricing of lamps.
result only recommended by AOD-1995-09 (IRS AOD, 1995) and acq., 1995-33 I.R.B.
4 (IRS ACQ, 1995). For comments, see also Wittendorff (2010a), at p. 648.
493. The structure also involved other agreements. In particular, there was market-
ing (a 5% commission fee for the subsidiary), procurement services (reimbursement
by the subsidiary of the parent’s procurement costs), cost sharing (50/50 split between
the parent and subsidiary for research and development (R&D) costs for the agreed
cost base) and warranty reimbursement agreements in place between the parent and
the subsidiary. The Court did not find the 50/50 cost-sharing split to be in accordance
with the 1968 Treas. Regs. (33 FR 5848) § 1-482-2(d)(4) on cost sharing because it was
expected that the subsidiary would likely benefit the most from the US R&D activities,
since most of the production was moved to Singapore (only high-performance, low-
cost, low-labour-intensive production would remain in the United States). The Court
discretionarily adjusted the cost-sharing split to 25% for the parent and 75% for the
subsidiary.
150
through case law
hard drive intangibles did not yield residual profits, as they did not provide
efficient protection from competitors. In order to make profits, the subsidi-
ary relied on cost-efficient manufacturing.
With respect to the pricing of the subsidiary’s sales of finished products

to the parent, the Court found no CUTs. In accordance with Sundstrand 494
and the 12-factor list of the 1968 regulations, the Court went on to find
a best estimate for the price. Even though the subsidiary’s sales transac-
tions with unrelated parties were not comparable, the Court used these
transactions as a baseline for further assessment, as it had previously done
in Bausch and U.S. Steel Corp. v. CIR.495 The subsidiary had been sub-
ject to pressure from its third-party customers to reduce prices, which it
had done. It had not, however, performed a corresponding reduction in
its sales price to the parent. Based on this, the Court concluded that the
arm’s length price for the sale of finished products to the parent was the
lowest of the prices actually charged and the average of prices to unrelated
customers.
Also in reviewing the royalty rate, the Court found no CUTs,496 and there-
fore based its allocation on the 12-factor list.497 First, the Court found
that a royalty rate higher than 1% was necessary in order for the parent
company to recover its R&D costs. Second, the parent had made valuable
marketing intangibles available to the subsidiary. Seagate was a world-
leading manufacturer of hard drives with established global markets for
its products. The subsidiary therefore had a market ready and waiting for
its products. The Court also found that some compensation was justified
for the loss of US sales due to direct sales from the subsidiary to former
US customers. Based on the transactions presented to the Court, it found
that a royalty in the 3-5% range was appropriate. Due to risks assumed by
the subsidiary, such as the age of the transferred intangibles, the volatility
of the hard drive market and the absence of cross-licensing provisions,
the Court discretionarily set the royalty rate at 3%, entailing an indirect
profit split.
494. Sundstrand Corp. and Subsidiaries v. CIR, 96 T.C. No. 12 (Tax Ct., 1991).
495. See Bausch & Lomb Inc. v. CIR, 92 T.C. No. 33 (Tax Ct., 1989), affirmed by 933
F.2d 1084 (2nd Cir., 1991); and U.S. Steel Corp. v. CIR, T.C. Memo. 1977-140 (Tax Ct.,
1977), reversed by 617 F.2d 942 (2nd Cir., 1980), and nonacquiescence recommended
by AOD-1980-179 (IRS AOD, 1980).
496. See 1968 Treas. Regs. (33 FR 5848) § 1.482-2(d)(2)(ii).
151
Methodology
5.2.6. Two cases on controlled services and sales contracts:

DuPont (1979) and Hospital Corporation
of America (1983)
E.I. Du Pont de Nemours and Co. v. the United States is fundamentally

a case on the allocation of intangible profits.498 It pertained to sales of
proprietary raw materials from a US chemical parent to its Swiss subsidi-
ary. R&D and manufacturing of the raw materials was carried out in the
United States. The Swiss subsidiary was used as a sales hub. Also, sales
to non-European regions were routed through Switzerland. The subsidi-
ary was to provide technical sales services and marketing functions. One
of the motivations behind the structure was to shift profits away from US
taxation by selling raw materials from the parent at a low price. The lion’s
share of the profit was allocated to Switzerland upon the resale of raw ma-
terials to unrelated customers around the world at the high market price.
The IRS performed a considerable reallocation, in total USD 18 million,
based on a discretionary profit split under the 12-factor list,499 allocating a
significant portion of the residual profits to the United States. The Court
accepted the IRS reallocation without much hesitation, no doubt in part
due to the fact that even subsequent to the reallocation, the profits of the
subsidiary were among the highest 4% of functionally comparable com-
panies.
In Hospital Corporation of America v. Commissioner,500 a successful US

parent, which owned and managed a range of domestic hospitals, entered
into a lucrative contract to manage a hospital in Saudi Arabia in the early
1970s via a Cayman Islands subsidiary. The Tax Court assessed the al-
location of profits from the management contract to the US parent based
on three alternative legal grounds. First, the IRS argued that the profits
should be allocated to the US parent directly by disregarding the subsidi-
ary’s ownership of the contract on the basis that it was a sham corpora-
tion. Secondly, the IRS argued that the management contract, as a taxable
value, had been transferred from the US parent company to the foreign
498. E.I. Du Pont de Nemours and Co. v. U.S., 1978 WL 3449 (Cl.Ct., 1978), adopted
by 221 Ct.Cl. 333 (Ct.Cl., 1979), certiorari denied by 445 U.S. 962 (S.Ct., 1980), and
judgment entered by 226 Ct.Cl. 720 (Ct.Cl., 1980).
500. Hospital Corporation of America v. CIR, 81 T.C. No. 31 (Tax Ct., 1983), nonac-
quiescence recommended by AOD- 1987-22 (IRS AOD, 1987) and Nonacq. 1987 WL
857897 (IRS ACQ, 1987). For a discussion of the ruling, see, e.g. Wittendorff (2010a),
at pp. 657-658 and 755.
152
US legislative and regulatory implementation of “profit-based” methods
subsidiary, thereby triggering a taxable charge. These arguments were re-

jected by the Court. Third, the IRS argued that the US parent had rendered
services and made intangibles (know-how, management systems, network
and goodwill) available to the foreign subsidiary, for which it should have
received arm’s length compensation. The Tax Court accepted this argu-
ment, as it found that the parent owned intangibles that indeed were made
available to the subsidiary. The result was that 75% of the residual profits
from the management contract were allocated to the United States through
a profit split.
5.3. US legislative and regulatory implementation of

“profit-based” methods
5.3.1. Introduction
In this section, the author comments on the legislative and regulatory

process that led to the 1994 section 482 regulations, which form the bulk
of the current rules relevant to the allocation of residual profits.501 This
will provide a significant background for the later analysis of the CPM
and PSM (in chapters 8 and 9, respectively). The author discusses the
relevant aspects of the 1986 tax reform in section 5.3.2., the 1988 White
Paper in section 5.3.3. and the development of the 1994 regulations in
section 5.3.4.
5.3.2. The 1986 tax reform
In 1985, the US House Committee on Ways and Means addressed the profit
shifting practices of multinationals, as documented in a separate report
(1985 House Report).502 Two main problems were identified. First, multina-
tionals had strong incentives to shift intangible profits away from US taxa-
tion to minimize their effective tax rates. One method of doing so was via
outbound transfers of US-developed manufacturing intangibles at an early
stage of R&D at a low price, where the transferring US group entity could
later claim that it was not possible to foresee the profit potential of the in-
501. On the development of US transfer pricing rules, see Rosenbloom (2006), at

p. 341.
502. H.R. Rep. No. 426, 99th Cong., 1st Sess. 424 (1985) (1985 House Report). For a
historical analysis of the run-up to the 1986 US tax reform, see Culbertson et al. (2003),
at p. 61.
153
Methodology
tangible at the time of transfer. Second, the lack of CUTs in general made
the application of the transactional pricing methods of the 1968 regulations
difficult.503 On this basis, the Committee concluded that the allocation of
residual profits should be commensurate with the income attributable to
the transferred intangibles. This concept had two distinct features:504 (i)
that the actual profit experience from an intangible should be taken directly
into account in the profit allocation; 505 and (ii) that the allocation should be
subject to periodic review.506
The conclusions of the 1985 House Report were supplemented in the 1986
Conference Committee Report from the House of Representatives (1986
Committee Report),507 introducing that the profit allocation should reflect
“the relative economic activity undertaken by each” controlled party, also
in the context of CSAs.508 The 1986 Committee Report recognized that
several transfer pricing issues pertaining to intangibles had been left un-
resolved. A comprehensive study of the profit allocation rules was ordered
from the IRS. On this basis, the commensurate-with-income concept was
added to IRC section 482 in the following wording:
[I]n the case of any transfer (or license) of intangible property (within the
meaning of section 936(h)(3)(B)), the income with respect to such transfer
or license shall be commensurate with the income attributable to the intan-
gible.509
503. The Committee was displeased with certain court interpretations of IRC sec.
482, particularly the use of comparables in U.S. Steel Corp. v. CIR, T.C. Memo.
1977-140 (Tax Ct., 1977), reversed by 617 F.2d 942 (2nd Cir., 1980) and nonacqui-
escence recommended by AOD-1980-179 (IRS AOD, 1980). In U.S. Steel, the IRS
reallocated costs charged to a US entity in a transfer pricing structure concerning the
shipping of iron ore from Venezuela to the United States. The Tax Court, in prin-
ciple, upheld – albeit substantially reducing – the reallocations. The Second Circuit
reversed the Tax Court’s decision on the basis that the evidence sufficiently showed
that, similarly enough, uncontrolled transactions supported the price paid by the US
entity as an arm’s length charge. The IRS recommended non-acquiescence (AOD-
1980-179).
504. See H.R. Rep. No. 426, 99th Cong., 1st Sess. 424 (1985), at pp. 425-426.
505. Thus, industry norms or “roughly similar CUTs” (comparable uncontrolled
transactions) should not provide safe harbours for the allocation of residual profits.
506. The author reverts to this issue in the discussion of periodic adjustments in ch.
16.
507. H.R. Conf. Rep. No. 841, 99th Cong., 2nd Sess. II-638 (1986).
508. Ibid., at p. 637.
509. Tax Reform Act of 1986, Pub. L. No. 99-514, Sec. 1231(e)(1), 100 Stat. 2085,
2562-63 (1986). The wording of the commensurate-with-income standard in the second
sentence of IRC sec. 482 has remained unchanged to this day.
154
5.3.3. The 1988 White Paper
The study of the profit allocation rules for intangibles ordered in the 1986
Committee Report was delivered in October 1988 in the report titled “A
study of intercompany pricing”, prepared by the Treasury Department and
the IRS (White Paper).510 The document provided an overview of section
482 practices under the 1968 regulations and discussed how the commen-
surate-with-income standard should be implemented. It delivered three
clear messages for the revision of the 1968 regulations. First, it was critical
towards the use of CUTs to allocate residual profits. There would normally
be no CUTs available, as plainly illustrated by high-profile transfer pricing
cases such as Eli Lilly, Searle, Hospital Corporation of America, Ciba and
DuPont. Further, if a purported CUT were found, there was a risk that it
could be misinterpreted.511 Second, as a response to this, and to cement and
further develop the case-law-driven profit split method, it introduced a new
profit allocation method: the basic arm’s length return method (BALRM).
Third, it signalled a stricter approach towards profit allocation in the con-
text of CSAs.
The impact of the White Paper on international transfer pricing jurispru-

dence can hardly be overstated. It remains relevant to this day, due to its
thorough analysis of the methodology used in the CPM, PSM and periodic
adjustments. The discussion at this point will be restricted to the BAL-
RM.512 The BALRM was to be applied if no third-party transactions satis-
fied the strict proposed CUT requirements. The method came in two ver-
sions: (i) the BALRM; and (ii) the BALRM with profit split.
The one-sided BALRM was designed to allocate a normal market return to

the controlled party that contributed only routine inputs to the value chain
(tested party).513 Thus, all operating profits, apart from the normal market
return, would be allocated to the other party to the controlled transaction,
which contributed unique value chain contributions, e.g. unique intangibles
and R&D. The logic behind the BALRM was aligned with economic rea-
soning. As there would be competition among providers of routine value
510. Notice 88-123. For insightful comments on the White Paper, see Clark (1993);
and Culbertson et al. (2003), at p. 65.
511. In particular, the White Paper viewed the CUT applied by the Second Circuit in
U.S. Steel as economically different from the controlled transaction.
512. For an analysis of the White Paper positions on cost-sharing agreements (CSAs)
and periodic adjustments, see the analysis in sec. 16.2.2.
513. For an analysis of the basic arm’s length return method (BALRM), see Clark
(1993), at p. 1184; and Culbertson et al. (2003), at p. 70.
155
Methodology
chain inputs, pressing prices down, no residual profits should be allocable

to the tested party.
Essentially, the BALRM was a profit split method in which 100% of the
residual profits were allocated to the party contributing unique functions
and assets. As opposed to the profit split method, the BALMR did not re-
quire discretionary assessments with respect to the allocation of residual
profits. It was a powerful response to the many tax-driven reorganizations
of MNEs involving unique intangibles that had plagued the United States
for decades. The problem with the BALMR was that its scope for applica-
tion was narrow. In cases in which the foreign subsidiary owned “some
type of intangible that is of major importance to the enterprise, and which
few unrelated parties possess”, the BALRM would be inapplicable,514 for
instance, where a foreign distribution entity had developed its own local
marketing intangibles.
In cases where the BALRM was inapplicable, the White Paper prescribed
a profit split addition to the method.515 This extension was meant for cases
in which the foreign subsidiary performed complex functions, owned sig-
nificant intangibles and was exposed to real risks. The extension was not as
simple as the clean-cut BALRM. However, as opposed to the BALRM, the
profit split addition ensured that both controlled parties would be allocated
residual profits. This result stood in stark contrast to that of the contract
manufacturer litigation position of the IRS during the 1970s-1980s, pur-
suant to which the intangibles owned516 or licensed517 by foreign manu-
facturing subsidiaries in effect were disregarded, and the subsidiary was
allocated only a normal market return on routine functions.
The BALRM with profit split complied with the intentions of the 1985 House
Report, which did not require the use of the contract manufacturer theory
for allocating profits to foreign related entities in all cases.518 Two alternative
profit allocation patterns were available to allocate residual profits under the
profit split extension: (i) the comparable profit split; and (ii) a profit split ac-
cording to the relative values of the unique value chain contributions.
514. Notice 88-123, at p. 99.

515. See Culbertson et al. (2003), at p. 73 for an insightful analysis of the BALRM
with profit split.
516. See the discussion on the possessions company string of cases (Eli Lilly, Searle,
Merck, etc.) in sec. 5.2.4.
517. See the discussion of the roundtrip case law in sec. 5.2.5.
518. H.R. Rep. No. 426, 99th Cong., 1st Sess. 424 (1985), at p. 426.
156
Under the comparative profit split allocation pattern, the residual profits
would be distributed using a fraction extrapolated from a CUT. The idea
was that one could apply the same profit split as unrelated parties had used,
provided that the functions they had performed in connection with the
CUT were broadly similar to those performed under the controlled trans-
action.519 It was apparent that this allocation pattern would not be practi-
cal. First, it would be practically impossible to identify and gain access to
CUTs for truly comparable functions in order to extrapolate the third-party
splits. Second, even if arm’s length profit splits for each of the relevant
functions could be ascertained, such third-party splits, on a stand-alone
basis, would not necessarily be meaningful. They could, for instance, be
determined in the context of a larger contractual package, where it was the
net result of several intertwined contracts that mattered for the involved
parties, as illustrated, for instance, in GSK-Canada520 and the roundtrip
case law.521 Thus, it could be misleading to rely on a single parameter to
allocate the residual profits.
The second allocation pattern was the significantly more relevant Eli Lilly-
inspired methodology, pursuant to which the residual profits would be dis-
tributed among the controlled parties according to the relative values of the
unique intangibles that they contributed to the value chain.522 This allocation
pattern mirrored the method that had gradually been developed through the
courts as a “fourth” method under the 1968 regulations and was therefore,
519. This was illustrated in the White Paper (Notice 88-123), Appendix E, Ex-
ample 10, where the R&D of a US parent had resulted in a high-tech patent. The
product required customization to be tailored to the needs of each local customer.
The parent had a European subsidiary that maintained an R&D staff and self-manu-
factured all of the devices that it sold. The R&D and design functions performed by
the subsidiary were viewed as complex and relatively unique, and thus warranted a
profit split. For these functions, the example identified CUTs purely based on broad
similarities.
520. GlaxoSmithKline Inc. v. R. (2012 SCC 52 [2012]), which affirmed 2010 CAF
201, F.C.A., [2010], which reversed 2008 TCC 324 [T.C.C., 2008]). See the analysis of
this case in sec. 6.7.4.
521. See the discussions in sec. 5.2.5.
522. The White Paper (Notice 88-123) contained an example in which this pattern
was applied (Appendix E, Example 11). The example pertained to a US parent that
licensed make-sell rights to a new toy to its European subsidiary, which incurred sig-
nificant marketing expenditures to develop a local trademark and reputation. The sub-
sidiary’s unique marketing intangible justified a split of the residual profits between
the US parent and the subsidiary. The split was determined on a discretionary basis
by weighing the estimated relative importance of the design and R&D activities and
the manufacturing intangibles of the parent against the marketing intangibles of the
subsidiary, akin to the profit split assessment in Eli Lilly. Due to the high threshold for
applying the comparable profit split, the White Paper, in reality, favoured profit splits
pursuant to this pattern.
157
Methodology
in principle, not a new concept in US transfer pricing jurisprudence. The

same cannot be said for the White Paper’s introduction of the BALRM.
The contract manufacturer theory of the IRS, in its different incarnations,
had repeatedly been rejected by the courts in high-profile cases such as Eli
Lilly, Searle and Bausch. In Sundstrand,523 the Tax Court rejected the IRS’s
proposition that the foreign manufacturing subsidiary should be remuner-
ated on a normal return basis, with the following wording:
[W]e are not satisfied from this record that the expert’s “normal” rate of re-
turn approximates in any way the proper rate of return for a company such as
SunPac or even that it is an acceptable criterion upon which to determine an
arm’s length price for SunPac parts.
The BALRM was therefore genuinely innovative, in the sense that it had
no basis in current law. The BALRM allocated actual profits, in line with
the legislative intentions behind the 1986 tax reform and the commensurate-
with-income standard. This aspect of the methodology was perceived as con-
troversial internationally.524 Nevertheless, the White Paper took the position
that the 1979 OECD report did not preclude the allocation of actual profits.
This stance is, in the author’s view, debatable. There were only brief and am-
biguous mentions of profit-based transfer pricing in the 1979 report.525 The
little wording there seemed to suggest that actual profits indeed could be tak-
en into consideration in setting transfer prices, but then primarily as a check
on the allocation results yielded by the traditional CUT-based methods.526
523. Sundstrand Corp. and Subsidiaries v. CIR, 96 T.C. No. 12 (Tax Ct., 1991). See
the discussion in sec. 5.2.5.3.
524. Also, other aspects were criticized, e.g. that the BALRM was difficult to apply,
that it was unfair to group companies that experienced returns that deviated from aver-
age returns and that it would allocate too much income to the United States. Japan was
concerned that if the BALRM was applied, US distribution subsidiaries of Japanese
MNEs could be compared to US distribution companies (thereby increasing their tax-
able income). Apparently, the view was that US distributors of US-manufactured goods
were able to reap higher profits than US distributors of Japanese goods due to the high
price at which the US distribution entity had to buy from the Japanese manufacturer.
See the example drawn up in Miyatake et al. (1994).
525. White Paper (Notice 88-123), at p. 61. See the 1979 OECD Report, at paras. 14
and 70-72.
526. Putting the issue of whether the allocated income should be based on projected
or actual profits aside, the White Paper’s positive take on the compatibility between the
BALRM and the 1979 OECD Report could arguably be defended for the BALRM with
profit split, but seemed more questionable with respect to the clean-cut BALRM that
operationalized the IRS’s contract manufacturer theory.
158
5.3.4. The 1994 US regulations
The White Paper eventually led to proposed regulations in 1992,527 op-

erationalizing most of the key ideas presented in the White Paper. The
BALRM from the White Paper was replaced with, or rather renamed as,
the comparable profits method (CPM). The criteria for qualifying compa-
rables under the CUT-based methods remained strict.528 Oddly, the profit
split method was not expressly presented as an independent pricing method
in the proposed regulations.529
The proposed regulations met a critical reception from international

commentators,530 in particular from the OECD. Shortly after the publica-
tion of the proposed regulations, the Committee on Fiscal Affairs of the
OECD established a special task force to provide the US tax administra-
tion with the collective views of other OECD member countries on the
proposed regulations. The findings of the task force were presented in a
report (hereafter the task force report),531 which was speedily approved at
the end of 1992. The Business and Industry Advisory Committee to the
OECD (BIAC) and the International Chamber of Commerce (ICC) were
among the business representatives that provided input to the task force
during its work on the report.532 The Committee on Fiscal Affairs also or-
dered its Working Party to urgently undertake a review of the 1979 OECD
report. In this review, the new US transfer pricing approach based on actual
profit experience was to be one of the points to be considered for a revised
guideline text.
527. 57 FR 3571-01.
528. The 1992 proposed regulations relaxed the requirements for inexact compara-
bles set out in the White Paper. This was, however, compensated for by requiring that
income allocation under the CUT, resale price and cost-plus methods that be based
on inexact comparables in order to be compatible with the comparable profits method
(CPM).
529. A limited provision pertaining to comparable profit splits in the context of profit
level indicators was, however, incorporated in § 1.482-2(f)(6)(iii)(C)(3) of the 1992
proposed regulations (57 FR 3571-01). Under the approach of the 1992 proposed regu-
lations for transfer pricing of intangibles, the comparable profit interval was to be used
when determining the pricing under the CPM in § 1.482-2(d)(5), or other methods in
§ 1.482-2(e)(1)(iv). See Langbein (2005), p. 1069, who found that the 1992 residual
profit split method (PSM) was so restricted that it offered little utility.
530. Mexico, for example, supposedly threatened to enact similar rules against the
United States.
531. OECD, Tax Aspects of Transfer Pricing Within Multinational Enterprises – The
United States Proposed Regulations, A report by the Committee on Fiscal Affairs on
the Proposed Regulations under Section 482 IRC (OECD 1993).
532. See task force report, at para 1.30.
159
Methodology
The task force report was highly critical of the proposed regulations.
The bulk of the critique was directed at the CPM (BALMR), but the
provisions on periodic adjustments and sound business judgement also
received negative comments. The author finds it somewhat ironic that
the profit split method was not subject to the same heavy criticism as the
CPM (as both methods use what is essentially the same methodology). In
fact, the report is somewhat contradictory on this point, as the profit split
method is given a rather favourable mention relative to the comments
on the CPM.533 The author will discuss the specific objections as part
of his later analysis of the CPM, the PSM and the periodic adjustment
authority.534
The temporary regulations were issued in the beginning of 1993535 and

largely carried over the approach for allocating residual profits unaltered
from the 1992 proposed regulations. However, the dominant position en-
joyed by the CPM in the proposed regulations was toned down, primar-
ily through the introduction of the “best method rule”.536 The profit split
method was introduced as an independent transfer pricing method. Never-
theless, due to strict requirements for applying both the CUT method and
the profit split method, the temporary regulations were, in reality, biased
towards the CPM with respect to the allocation of residual profits.
The OECD provided its comments in a second 1993 task force report,537
which was not as thorough as the first task force report, and essentially
reiterated the viewpoints of its predecessor. The apparent main aim was
to encourage the United States to restrict the scope of the CPM to abusive
cases and as a method of last resort.538
533. See task force report, at paras. 3.12-3.15.

534. In ch. 8, ch. 9 and ch. 16, respectively. See also sec. 5.4., which deals with the de-
velopment of the OECD transfer pricing methods leading up to the 2010 OECD Trans-
fer Pricing Guidelines (OECD TPG).
535. 58 FR 5263-02. For an insightful analysis of the 1993 temporary regulations, see
Clark (1993).
536. See the discussion of the best-method rule in sec. 6.4. Formally, the rule invited
flexibility, with the CUT method as the natural default transfer pricing method for in-
tangibles.
537. OECD, Intercompany Transfer Pricing Regulations under US Section 482 Tem-
porary and Proposed Regulations (OECD 1993).
538. See id., at para. 2.11. See also Clark (1993), at p. 1197, for a discussion of the
potential for conflict with foreign tax systems in light of the transfer pricing approaches
conveyed by the 1993 temporary regulations.
160
OECD implementation of “profit-based” transfer pricing methodology
The final regulations were issued in 1994, largely corresponding to the

temporary regulations.539 The changes that were made generally softened
the language without altering the core content of the commensurate-with-
income approach.
5.4. OECD implementation of “profit-based” transfer

pricing methodology
The author will comment in this section on the broad strokes of the develop-
ment of the OECD transfer pricing methodology for intangibles. He will go
more into depth in later chapters of the book in connection with his analysis
of the TNMM, the PSM and the OECD periodic adjustment authority.540
The OECD published a discussion draft in 1994 (1994 Draft) on the re-
vision of the 1979 OECD report,541 largely as a response to the new US
rules.542 The 1994 Draft approach was two-sided: while in principle ac-
cepting the CPM and the PSM, it also discouraged the use of them. The
position was that only where the CUT-based methods were inapplicable
should the CPM and the PSM be applied on their own to determine transfer
prices.543 A 1995 report dealt with some particular issues (1995 Draft), such
as the treatment of marketing intangibles, CSAs and periodic adjustments,
but provided no principal guidance on the pricing methods.544
Aligned with the 1994 Draft, the final 1995 OECD TPG did allow the use
of profit-based methods as stand-alone transfer pricing methodologies,545
539. 59 FR 34971-01.
540. See the discussion of the TNMM in ch. 8, the OECD PSM in ch. 9 and sec. 11.4.,
the OECD approach to periodic adjustments in sec. 16.5. and periodic adjustments in
the context of CSAs in sec. 16.7.
541. It has been asserted that the 1979 OECD report, being a response to the 1968 US
regulations (which applied to profit allocation among group entities), has an unclear
relationship with the pre-1979 efforts of the OECD (and before that, the League of
Nations) on international allocation of business profits, which were focused on branch
(permanent establishment) allocation. See Vann (2003), at p. 136 on this issue.
542. OECD, Transfer Pricing Guidelines for Multinational Enterprises and Tax Ad-
ministrations, Part I: Principles and methods, discussion draft (OECD 1994). For a
brief overview of the development of the OECD TPG, see also Pankiv (2017), at pp. 26-
30.
543. Pankiv (2017), at para. 173. See also para. 130 and Note to Readers, at p. 6.
544. OECD, Transfer Pricing Guidelines for Multinational Enterprises and Tax Ad-
ministrations, Draft text of Part II (OECD 1995).
545. 1995 OECD TPG, at paras. 3.50 and 3.56. See Culbertson et al. (2003), at p. 88
for comments on the historical context.
161
Methodology
albeit with hesitation, and preferably merely for the purpose of cross-check-
ing the results of transaction-based pricing methods. The 1995 OECD TPG
seemed to fear misapplication of the profit-based pricing methods by tax
authorities, in particular, the TNMM.546
The author has found three noteworthy aspects of the 1995 text. First,
it renamed the CPM to the TNMM.547 In addition, certain restrictions
were also added to the TNMM in order to separate it from the CPM,
making consensus on the final 1995 text possible.548 Second, the 1995
OECD TPG contained (in the author’s view, highly questionable) fac-
tual assertions pertaining to the profit-based methods.549 In particular,
it was stated that “in fact, enterprises rarely if ever use a transactional
profit method to establish their prices”.550 No empirical references or
support for this factual allegation were offered. The assertion stood in
stark contrast to the rapidly growing popularity of profit-based methods
in practice, in particular the CPM. Third, the preference for the transac-
tional methods was reiterated. Only in “exceptional cases”,551 in which
the transactional methods were inapplicable, and if all “safeguards”
set out in the 1995 OECD TPG were observed, should the profit-based
methods be applied, and even then only as a “last resort” (for instance,
where there was insufficient data on CUTs to apply the transactional
methods).552
Over the following decade, the use of profit-based pricing methods – in

particular, the TNMM – became so widespread among multinationals and
tax administrations that a disparity had become apparent with respect to
546. 1995 OECD TPG, at para. 3.56.

547. The 1994 and 1995 drafts applied the terms “profit-based methods”, “profit-split
method” and the “CPM”. The common term “profit-based methods” was altered to
“transactional profit methods”, and it was specified that the new term covered “methods
that examine the profits that arise from particular transactions among associated enter-
prises”; see 1995 OECD TPG, at para. 3.1. The profit split method, for the time being,
was not renamed.
548. See the discussion on comparability under the TNMM in sec. 8.6. See also inter-
esting comments in Lodin (2009), at pp. 427-437.
549. See also Schön et al. (2011), at p. 125, for critical reflections on factual assertions
made in the OECD TPG.
550. 1995 OECD TPG, at paras. 3.2 and 3.54.
551. It can be questioned whether it was useful for the 1995 OECD TPG to refer to
such cases as “exceptional”, given the importance of the transfer pricing of unique
intangibles for fiscal purposes and that transactional methods had proved to be practi-
cally inapplicable (leaving few other options than to apply profit-based methods). For a
discussion of the reservation, see Casley et al. (2003), at p. 164.
552. 1995 OECD TPG, at para. 3.50.
162
OECD implementation of “profit-based” transfer pricing methodology
the last-resort qualification of the 1995 OECD TPG. In 2008, the OECD re-
leased a discussion draft suggesting revised guidelines for the profit-based
methods (2008 Draft).553 It proposed to replace the last-resort qualification
with a general requirement to apply the most appropriate method for a
particular case. The traditional preference of the OECD TPG towards the
transactional methods was markedly toned down. The OECD still deemed
the transactional methods to have “intrinsic strengths” relative to the profit-
based methods.554
The 2010 revision of the OECD TPG removed the last-resort qualification.
The profit-based pricing methods were then the dominant pricing methods
in practice, but some countries still opposed the use of them in non-ex-
ceptional cases. Paragraph 2.2 stated the principle that the selection of the
most appropriate method for each particular case should take into account
the nature of the controlled transaction, the availability of CUTs and the
reliability of any comparability adjustments that had to be made. Then,
paragraph 2.3 modified this by stating that the transactional methods were
regarded as the most direct means of establishing whether conditions in
the commercial and financial relations between the related parties were at
arm’s length. Thus, if a transactional and a profit-based method could be
applied in an equally reliable manner, the former would be “preferable”.
In particular, the CUT method would be preferable over all other pricing
methods. Even though the paragraph 2.3 reservation was important as a
statement of principle, in the author’s opinion, it is unlikely that it would be
of any relevance for the transfer pricing of intangibles in practice.555
Further, the 2010 OECD TPG contained new guidance on the selection of
the appropriate transfer pricing method,556 as well as on the application of
the TNMM and the PSM.557 As this will be discussed later in the book as
part of the analysis of the current OECD transfer pricing methodology, the
author will not go into the guidance here.
553. OECD, Transactional profit methods, discussion draft for public comment
(OECD 2008).
554. See id., at Introduction, para. 6. The proposed revision of chs. I-III of the 1995
OECD TPG was released in 2009; see OECD, Proposed revision of Chapters I-III of
the Transfer Pricing Guidelines (OECD 2009). The proposed text contained updated
general comments on comparability in ch. I and detailed comparability guidance in ch.
III.
555. See the discussion of the new OECD guidance on the CUT method in sec. 12.4.
557. See OECD TPG, at paras. 2.68-2.107 for the TNMM; paras. 2.115-2.145 for the
PSM; as well as Annexes I-III to ch. II and the Annex to ch. VI.
163
Methodology
Also, for the first time, the OECD TPG acknowledged the concept of tax-
payer-initiated year-end compensating adjustments,558 or “true-ups”. The
OECD TPG refer to the concept as a procedure that allows a taxpayer to
report a transfer price for tax purposes that is an arm’s length price for the
controlled transaction in question, even though the price would differ from
the amount actually charged between the associated entities. The acknowl-
edgement of compensating adjustments should be seen as an additional
levying of the position of the profit-based pricing methods in the OECD
TPG.559
558. OECD TPG, at paras. 3.70-3.71 and 4.38-4.39.

559. See also Russo, in Weber et al. (2011), at p. 168. The author refers to his analysis
of year-end adjustments in ch. 15 for further discussion on this topic.
164
Chapter 6
Metaconcepts Underlying the US and OECD

Profit Allocation Rules
6.1. Introduction
In this chapter, the author will analyse a selection of six interconnected

transfer pricing metaconcepts that underlie the US and OECD profit al-
location rules. These discussions will form a backdrop for the analysis (in
chapters 7-16) of the US and OECD methodology for the transfer pricing
of intangibles, where active use is made of these concepts.
He will begin with a discussion of operating profits in section 6.2., as a

more comprehensive understanding of this concept (than what has been
necessary up to this point) will be needed for later chapters. The discussion
will be focused on how the concept relates to the different transfer pricing
methods, how the different methods test different components (or levels)
of operating profits and how the lack of publicly available profit data may
restrict the use of certain methods in practice.
He will then move on to develop the relationship between the transaction-

based gross profit methods (the comparable uncontrolled transaction (CUT)
method, resale price method and cost-plus method) and the profit-based net
profit methods (the transactional net margin method (TNMM) and profit
split method (PSM)) in section 6.3. He discusses how both groups of meth-
ods rely on the same basic methodology for allocating profits to group enti-
ties. Further, he discusses the traditional OECD point of departure that the
gross profit methods are generally more reliable than the net profit methods
and shows that the situation is not as clean-cut as the OECD might make
it out to be.
Next, he will discuss the US and OECD rules that govern which transfer
pricing method should be chosen to allocate operating profits in section
6.4. As will be shown, these rules are open-ended in the sense that they
generally do not dictate a specific method, but rather emphasize the impor-
tance of relative reliability. Nevertheless, as the transfer pricing of intan-
gibles has proven vulnerable to profit shifting practices, the US rules – in
the context of intangible property (IP) transfers – now emphasize that the
165
Chapter 6 - Metaconcepts Underlying the US and OECD Profit
Allocation Rules
selected method must take into account the need to perform aggregated
valuations to capture all intangible value transferred intra-group (irrespec-
tive of legal form), while the OECD – for many of the same reasons – now
leans away from the CUT method and towards the PSM.
Subsequently, the author will discuss the US and OECD rules that deter-
mine how much deviation is acceptable between controlled profit alloca-
tion, on the one side, and third-party profit allocation, on the other side, be-
fore a transfer pricing adjustment is triggered (the so-called “arm’s length
range”) in section 6.5.
He will then go on to discuss the concept of comparability in section 6.6.

The US and OECD profit allocation rules generally rely on comparing con-
trolled transactions with (or benchmarking them against) third-party trans-
actions to find acceptable results. The comparability provisions of these
rules have two sides. The first side pertains to the controlled transaction,
where legal form must give way to economic substance (in the case of
conflict) with respect to establishing the material content of the transac-
tion before it is benchmarked against third-party transactions. To put it in
the terminology of the OECD, only the actual controlled transaction as
“properly delineated” will be given effect. Economic substance is an es-
sential component in the comparability doctrine, not only for identifying
the actual controlled transaction that must be priced, but also for “peeling
away” artificial legal structures designed to shift profits (e.g. aggressive
contractual separations of risk from underlying functions). The second side
of the comparability provisions pertains to the uncontrolled benchmarking
transactions and is made up of rules that set the standard for how com-
parable these must be to the controlled transactions (with respect to geo-
graphical market, functions, risks, etc.) before they can be applied using an
applicable transfer pricing method.
The author will end the chapter with a discussion of the aggregation of
controlled transactions in section 6.7., which is a topic closely related to
the comparability doctrine. The core question here is to what extent the US
and OECD profit allocation rules allow that the transfer pricing methods
are applied to price several controlled transactions together as one pack-
age as opposed to pricing them on a transaction-by-transaction basis. The
underlying issue is that the transaction-by-transaction approach may fail
to capture all value transferred intra-group in a combination of controlled
transactions, thereby triggering a need for an aggregated valuation ap-
proach.
166
The relationship between operating profits and the transfer pricing methods
6.2. The relationship between operating profits and

the transfer pricing methods
6.2.1. Introduction
The transfer pricing methods allocate operating profits to the different value
chain inputs contributed by group entities. All of the five specified methods
in US and OECD transfer pricing law (the CUT method, resale price meth-
od, cost-plus method, comparable profits method (CPM)/TNMM and PSM)
– with the possible exception of the CUT method – draw upon a set of profit
concepts. The one-sided resale price and cost-plus methods build on the con-
cept of gross operating profit, while the CPM/TNMM and the PSM build on
the concept of net operating profit. Until now, the introduction of operating
profits provided in chapter 1 has sufficed, but for the purpose of the following
analysis of the transfer pricing methods, it will be necessary to have a more
developed understanding of the concept. The author discusses the concept of
operating profits in section 6.2.2., the components of operating profits in sec-
tion 6.2.2. and the accounting information on gross profits and transactional-
level profit information in sections 6.2.4. and 6.2.5., respectively.
6.2.2. The concept of operating profits
Operating profits are profits that accrue solely from business operations.
They do not include income and expense items that derive from the financ-
ing of business activities. Operating profits are thus profits before net in-
terest expenses and are similar to the financial accounting term “earnings
before interest and taxes” (EBIT).560 The traditional outline of operating
profits in transfer pricing is as follows:
Sales
− Cost of goods sold
= Gross profit
− Operating expenses
= Net profit = operating profit = EBIT
560. A measure closely related to earnings before interest and taxes (EBIT) is earn-
ings before interest, tax, depreciation and amortization (EBITDA), which is simply
EBIT with the twist that depreciation and amortization expenses have been added.
EBITDA is normally a better indicator than EBIT of the net cash flows generated by
the business (as depreciation and amortization expenses are accruals and do not have
cash flow effects) and therefore often used as an input in valuations.
167
Allocation Rules
It is not apparent how these elements should be delineated. A clarifying

analysis is necessary because much of the legal arguments that are in fa-
vour of or against the application of a particular transfer pricing method
draw upon opinions of the usefulness of these underlying income and ex-
pense items. This is particularly true for the debate on the utility of net
profit methods (i.e. the profit-based methods) relative to the gross profit
methods.
Neither the OECD Model Tax Convention (OECD MTC) nor the OECD
Transfer Pricing Guidelines (OECD TPG) directly define income and ex-
pense items.561 This is, however, to some extent, done in the CPM pro-
visions of the US regulations, but further clarification is needed.562 The
International Financial Reporting Standards (IFRS) represent the de facto
global standard for financial accounting. The author will use the relevant
IFRS as legal sources for delineating the elements in operating profits un-
der the US regulations and the OECD TPG.
The reasoning behind this methodological choice is that first, the termi-
nology in financial accounting is developed and goes – on many points
– deeper in delineating the various income and expense items than the US
regulations and the OECD TPG do. Second, the terminology applied in
both the US and OECD regimes must be presumed to build upon, and to be
coherent with, the terminology applied in authoritative financial account-
ing standards, as tax returns build on the data generated by the accounting
systems of business enterprises as presented in their audited financial state-
ments. Third, there is no sensible reason as to why the concept of operat-
ing profits should differ between the US regulations, the OECD TPG and
the authoritative financial accounting standards, as third-party accounting
data is used directly by the transfer pricing methods to extract the relevant
profit margins.
561. Further, art. 7 of the OECD Model Tax Convention on Income and on Capital
(OECD MTC) does not define “business income”. Para. 71 of the (2010) Commentaries
on the OECD MTC (art. 7, para. 4) states that it “should nevertheless be understood that
the term when used in this article and elsewhere in the Convention has a broad meaning
including all income derived in carrying on an enterprise. Such a broad meaning cor-
responds to the use of the term made in the tax laws of most OECD member countries”.
Thus, it should be rare that disagreements arise in practice as to whether an item of
income earned by an enterprise should be classified as business income.
562. Treas. Regs. § 1.482-5(d).
168
6.2.3. Delineating the components of operating profits
6.2.3.1. Sales
The first item in operating profits is sales. The US regulations define this as
the amount of the total receipts from the sales of goods and provision of ser-
vices, less returns and allowances.563 As far as the author can see, this is also
the understanding of the OECD TPG, even though there is no precise defini-
tion. The US definition is materially similar to the one applied in the IFRS.
The IFRS, however, use the term “revenue” instead of “sales”. Revenue,
according to the IFRS, is defined as the “gross inflow of economic benefits
during the period arising in the course of the ordinary activities of an entity
when those inflows result in increases in equity, other than increases relat-
ing to contributions”.564
This includes only inflows received and receivable by the entity on its own
account565 and encompasses income from the use by others of the entity’s
assets (i.e. interest, royalties and dividends).566 Income from interests and
dividends shall, however, not be included in the operating profits under
the US regulations or the OECD TPG.567 These exclusions are natural in a
transfer pricing context, as such income will not be connected to any value
chains for particular products or services (i.e. business activities), but are
the result of financing and investment activities, and should therefore not
be relevant under any of the transfer pricing methods.
563. Treas. Regs. § 1.482-5(d)(1).

564. International Accounting Standard (IAS) 18: Revenue, para. 7. The International
Financial Reporting Standards (IFRS) may require recognition of income and expense
items that have not been realized. Such accruals pertain to the timing of income and ex-
pense recognition. The author is only interested in what types of income and expenses
are recognized under the IFRS (in other words, classification issues, as opposed to tim-
ing issues).
565. E.g. in an agency relationship, revenue is only the amount of commission (IAS
18, para. 8). This was, for instance, the return position of the taxpayer in the Norwegian
Supreme Court ruling in Dell Products v. the State (Norwegian Supreme Court ruling
dated 2 Dec. 2011, Utv. 2012, s. 1, reversing Utv. 2011, s. 807). See the analysis of this
ruling in sec. 25.6.5.
566. IAS 18, para. 1c) and paras. 29-34.
567. Treas. Regs. § 1.482-5(d)(4); and OECD Transfer Pricing Guidelines for Multi-
national Enterprises and Tax Administrations (OECD TPG), para. 2.80. It should be
noted that para. 2.80 of the OECD TPG excludes “non-operating items” and exempli-
fies interest income and expenses. It does not, however, specifically mention dividend
income. In the author’s view, there is clearly no reason to assume that dividend income
should be included in operating profits under the OECD TPG.
169
Allocation Rules
6.2.3.2. Costs of goods sold
The second item in operating profits is the cost of goods sold (COGS).
Even though not explicitly defined in the US regulations or the OECD
TPG, it is clear that the term includes direct costs for producing the prod-
uct or service rendered.568 Under the IFRS (IAS 2: Inventories), the cost
of inventories comprises all costs of purchase (purchase price, import
duties, transportation, etc.),569 conversion (direct labour costs and sys-
tematic allocations of fixed and variable production overheads incurred
in converting materials into finished goods) 570 and “other” costs (costs
incurred in bringing the inventories to their present location and condi-
tion). The COGS for a manufacturing subsidiary, for instance, will typi-
cally consist of the costs for raw material and royalties paid for licensing
the necessary technology for manufacturing the product (patents, know-
how, etc.).571
6.2.3.3. Gross profit
The third item in operating profits is gross profit. This is the amount that
remains after the COGS are deducted from sales.572 Gross profit should be
interpreted as a preliminary measure of operating profit, as it only signals
the ability of the taxpayer to cover its operating expenses.573 The one-sided
resale price and cost-plus methods test the allocation of gross profits.
6.2.3.4. Operating expenses
The fourth item in operating profits is operating expenses. These encom-

pass all business expenses that are not included in the COGS.574 Both the
568. Such costs fall outside the other category of business costs (operating expenses)
in Treas. Regs. § 1.482-5(d)(3).
569. IAS 2, para. 11. Trade discounts, rebates, etc. are deducted in determining the
costs of purchase.
570. IAS 2, para. 12. Fixed production overheads are indirect costs of production that
remain relatively constant regardless of the volume of production (e.g. depreciation and
maintenance of factories). Variable production overheads are indirect costs of produc-
tion that vary directly (or nearly directly) with the volume of production (e.g. indirect
materials and indirect labour).
571. See Bhatnagar (2017), at pp. 26 for a discussion of costs of goods sold (COGS)
under applications of the Berry ratio (see sec. 8.4.2.) for distribution entities.
572. Treas. Regs. § 1.482-5(d)(2).
573. See sec. 8.4.2. on the Berry ratio.
574. Treas. Reg. § 1.482-5(d)(3).
170
US regulations and the OECD TPG include marketing and distribution

costs (e.g. advertising, warehousing, administration and salary costs con-
nected to such functions) in operating expenses.575 A reasonable allowance
for depreciation and amortization on plant and equipment will normally
also be included.576 This aligns with the core content of the term under the
IFRS. Distribution costs, administrative expenses and other expenses will
be reported as operating expenses (i.e. below gross profits in the profit-and-
loss statement) if the function-of-expense method is used.577
6.2.3.5. Net profit
The fifth and final item is net profit.578 This is the amount of income from
business operations left after all operating costs have been deducted from
sales.579 Income from activities not being tested by the selected transfer
pricing method580 or extraordinary gains and losses that do not relate to the
continuing operations of the tested party are not relevant for transfer pric-
ing purposes.581 As touched upon, net operating profits shall not contain
any financial elements, such as interest income or expenses,582 nor shall
foreign or domestic income taxes or any other expenses that are not related
to the operation of the relevant business activity included.583 The one-sided
CPM/TNMM (and one aspect of the profit-split method) 584 test the alloca-
tion of net profits.585
575. See OECD TPG, paras. 2.46-2.47, 2.87 and 2.102.

576. See Haugen (2005), at p. 226, who is critical with respect to the lack of guidance
provided in the OECD TPG on the issue of how depreciation and amortization costs
shall be allocated among COGS and operating expenses.
577. See the discussion in sec. 6.2.4. on the function-of-expense method. See also
Bhatnagar (2017), at pp. 26 for a discussion of operating expenses under applications of
the Berry ratio (see sec. 8.4.2.) for distribution entities.
578. On net profits, see also, e.g. Wittendorff (2010a), at p. 739.
579. Treas. Regs. § 1.482-5(d)(4); and OECD TPG, para. 2.77.
580. See the discussion of blended profits in sec. 8.6.
581. OECD TPG, para. 2.80.
583. Treas. Regs. § 1.482-5(d)(3).
584. The first step of normal return profit allocations to routine value chain inputs
under the resale price method before allocating residual profits. See sec. 9.3.
585. See the discussion of the comparable price method (CPM) and transactional net
margin method (TNMM) in ch. 8 and of the US and OECD profit split method (PSM)
in ch. 9.
171
Allocation Rules
6.2.4. Information on gross profits may be unavailable
Publicly available financial statements will not necessarily contain data on

gross profits. To do so, business expenses must be presented in the profit-
and-loss account according to the functions that the underlying costs serve
in the enterprise (function method).586 Such presentation may be resource-
demanding, as it will require that a range of different costs are allocated
to the COGS (including salary expenses, depreciation, amortization, etc.).
Such allocation may involve discretionary assessments (and arbitrary
allocations).587 As a minimum under this method, an entity must disclose
its cost of sales separately from other expenses. In that way, it will be pos-
sible to distinguish gross from net profits. International Accounting Stand-
ard (IAS) 1 illustrates cost classification pursuant to the function method
as follows:588
Revenue
− Cost of sales
= Gross profit
+ Other income
− Distribution costs
− Administrative
expenses
− Other expenses
= Net profit before
tax
IAS 1 requires the reporting entity to disclose information on the nature of

the expenses (see table below) if the function method is used,589 and there-
fore favours the nature method over the function method.
586. IAS 1, para. 103. The method is also known as the cost-of-sales method.
587. IAS 1, para. 103.
588. Authoritative financial accounting literature is critical towards the method; see
Van Breda et al. (1992), at p. 368. The argument is that the method will not put readers
of financial statements in a better position to make predictions about the profit-gener-
ating ability of the enterprise as a whole, nor will they be able to evaluate the contribu-
tions of the diverse functions.
589. See IAS 1, para. 104. The disclosure shall encompass additional information on
the nature of expenses, including depreciation and amortization expenses and employee
benefits expenses. This requirement comes from the fact that information on the nature
of the expenses is useful in predicting future cash flows; see IAS 1, para. 105.
172
In practice, with the nature method, business expenses are normally pre-
sented according to their nature (e.g. total salary costs or total depreciation
costs), with each item presented on a different line in the income state-
ment.590 Under this method, costs are not reallocated according to func-
tions within the entity. This makes the method simple to apply (and thus
cost-efficient) and free of discretionary assessments. IAS 1 illustrates clas-
sification using the nature method as follows:591
Revenue
+ Other income
− Changes in inventories of finished goods and
work in progress
− Raw materials and consumables used
− Employee benefits expense
− Other expenses
= Net profit before tax
In conclusion, if a reporting entity has classified its operating costs pur-

suant to the nature method, it will not be possible to extract gross profit
margins. This makes it difficult – or outright impossible – to apply the
resale and cost-plus methods.592 The only alternative for allocating a nor-
mal market return to a group entity that only contributes routine inputs to
the value chain will be to apply the CPM or the TNMM, which rely on net
profit margins.
6.2.5. Information on transaction-level profits may be

unavailable
The transfer pricing methods, at least at the outset, require accounting

information on a transactional level, as the gross or net operating prof-
it margins of a CUT must be compared to the margins of the controlled
transaction.593 In order to provide realistic legal analyses of key issues con-
nected to the transfer pricing methods, it is crucial to recognize that what
590. IAS 1, para. 102.

591. Id.
592. See also Haugen (2005), at p. 224; Casley et al. (2003), at p. 164; Hamaekers
(2001), at p. 35; Cools (1999), at p. 178; and Roberge (2013), at p. 232.
593. See the discussion of the aggregation of controlled and uncontrolled transactions
for transfer pricing purposes in secs. 6.7. and 8.6., respectively.
173
Allocation Rules
is offered in company-level financial statements is aggregated financial in-

formation on the total sales and total business expenses of an enterprise for
all products and services sold by it (i.e. not transaction-level profit data).594
The same, but on an even further aggregated level for the entire group, is
offered in consolidated financial statements. It is not the presentation of
aggregated transactions as such that is the problem from a transfer pric-
ing perspective, but rather that income and expense items contained in the
income statement blend different types of transactions that have different
profit characteristics. The use of blended profit margins in transfer pricing
poses a significant risk of non-arm’s length results. This issue is discussed
in depth in section 8.6.
6.3. The relationship between gross and net profit

methods
6.3.1. Introduction
The resale price and cost-plus methods that use data on third-party gross
profits to benchmark the controlled allocation of income, as well as the US
CPM and OECD TNMM that use data on third-party net profits, have the
same general purpose: to allocate a normal market return to group enti-
ties that contribute only routine inputs to the intangible value chain. As
they only allocate profits to one group entity (the tested party), they are
so-called “one-sided methods”, as opposed to the two-sided CUT method
and PSM.
In sections 6.3.1.-6.3.5., the author will discuss a selection of aspects per-

taining to the relationship between the gross and net profit methods. This
will form a backdrop for the later analyses of the CPM and the TNMM (in
chapter 8),595 and should be seen in connection with the discussions of the
best method rule and comparability.596 Most of the issues discussed below
are triggered by the OECD guidance on the TNMM. It should be kept in
mind that most of the current guidance on the TNMM is identical to the
1995 consensus text, which hesitantly introduced the methodology into the
OECD TPG. The rather negative historical attitude of the OECD towards
594. See also Haugen (2005), at p. 224; and Luckhaupt et al. (2011), at p. 102.
595. The author will not analyse the one-sided gross profit methods in depth, as they
are largely redundant with respect to profit allocation from intangible value chains due
to the availability of the CPM and TNMM.
596. See the discussions in secs. 6.4. and 6.6., respectively.
174
The relationship between gross and net profit methods
the TNMM may have influenced some of the assertions made in the OECD
TPG on the usefulness of the method relative to the traditional cost-plus
and resale price methods. The author doubts whether all of the assertions,
now 20 years later, are representative of the current attitude of the OECD
towards the same problems, but they are nevertheless a part of the current
OECD TPG and must therefore be taken at face value.597
The author discusses common methodological traits among the gross and
net profit methods in section 6.3.2., relevant parameters under the gross
and net profit methods in section 6.3.3., whether operating expenses are rel-
evant under the transactional pricing methods in section 6.3.4. and whether
comparability adjustments under the gross profit methods are more reliable
than under the net profit methods in section 6.3.5.
6.3.2. Common methodological traits among the gross and

net profit methods
The fundamental difference between the gross and net profit methods is
that operating expenses are deducted in the computation of net profits,
while in the computation of gross profits, they are not. Aside from this, the
methods are fundamentally similar with respect to the methodology used
to allocate income to the tested party. The resale price method, the cost-
plus method, the CPM and the TNMM all apply a two-step approach for
allocating income: (i) a profit margin is extracted from the accounting data
of comparable unrelated enterprises; and (ii) the extracted profit margin is
used to benchmark the corresponding profit margin of the tested party. The
author will develop this a bit further in this section.
Imagine a transfer pricing structure between a controlled manufacturing

entity and a distribution entity, where the latter purchases goods for re-
sale from the former. The question is whether the transfer pricing applied
between them is at arm’s length. In practice, only one of the parties to the
transaction will be selected as the tested party. In this example, however,
the author uses two parties simply to illustrate the application of both the
cost-plus method and the resale price method.598
597. The TNMM is still not in vogue with the OECD, but now for different reasons
than in 1995. See the analysis in ch. 8 and the comments on the 2017 OECD TPG re-
garding the TNMM in ch. 11.
598. This example draws inspiration from Culbertson (1995).
175
Allocation Rules
Manufacturer Distributor
(related) (related)
Sales - 200
COGS 150 -
Gross profit - -
Assume that the manufacturer’s COGS of 150 stem from purchases from
unrelated suppliers and that the distributor’s sales of 200 stem from unre-
lated customers. Both elements are therefore reliable. The cost-plus and
resale methods are selected to allocate income to the manufacturer and dis-
tributor, respectively. The first step in applying these methods is to extract
gross profit margins from CUTs. The selected unrelated enterprises display
the following accounting data:599 600
(unrelated) (unrelated)
Sales 600 400
COGS 500 360
Gross profit 100 40
Gross profit margin600 20% 10%
It is seen that the gross margins of the unrelated manufacturer and distribu-
tor are 20% and 10%, respectively, with the former calculated on the basis
of costs and the latter on the basis of sales.
The second step is to apply the extracted uncontrolled margins to the data
of the tested party. For the related manufacturer, the gross profits will be
20% of its COGS, which is 150 × 20% = 30. For the related distributor, the
gross profit will be 10% of its sales, which is 200 × 10% = 20. This will
result in a transfer price of 180 for the product, as follows:601
599. It is assumed, rather unrealistically, that the uncontrolled enterprises carry out
only one type of transaction and that the transaction is fully comparable to the con-
trolled transaction.
600. The manufacturer’s margin is calculated based on costs (100 ÷ 500 = 20%),
while the distributor’s margin is calculated based on sales (40 ÷ 400 = 10%).
601. The fact that the two transfer prices calculated here correspond is just by design
of the example. This would be highly unlikely to happen in practice by applying the
resale price method and the cost-plus method to two different related entities, even if
for the same value chain.
176
(related) (related)
Sales 180 200
COGS 150 180
Gross profit 30 20
Applicable margin Gross profit ÷ COGS Gross profit ÷ sales
Computation 30 ÷ 150 20 ÷ 200
Gross profit margin 20% 10%
Thus, on the basis of this example, it may be concluded that the gross profit
models allocate income to the tested party equal to the gross profit margin
it would have earned if its gross profit margin had been the same as that of
an uncontrolled taxpayer. The main problem with a gross profit analysis is
that it requires specified accounting data from reliable CUTs, which makes
it possible to ascertain the unrelated gross margin.602
The author will now illustrate the parallel methodology of the net profit
methods. Let us, for instance, suppose that the distributor from the example
above is selected as the tested party under the CPM and that the following
accounting data is available from a comparable uncontrolled enterprise:603
Distributor
(unrelated)
Sales 400
COGS 360
Gross profit 40
Operating expenses 5
Net profit 35
Applicable margin Net profit ÷ sales
Computation 35 ÷ 400
Net profit margin 8.75%
602. See the discussion in sec. 6.2.4.

603. Again, for the purpose of the example, it is assumed that the unrelated distribu-
tor carries out only one type of transaction, which is fully comparable to the controlled
transaction.
177
Allocation Rules
The first step, as under the gross profit methods, is to extract the relevant
profit margin from the CUTs. The net profit margin is 8.75%, calculated on
the basis of sales. When applied to the financial data of the tested party, the
extracted net profit margin will yield 17.5 in net profit for the party tested
under the CPM:
Distributor
(related)
Sales 200
Net profit 17.5
Thus, it may be concluded that the basic methodology of both the gross and
net-profit-based methods follows the same two-step approach to allocate
income to the tested party.
6.3.3. Relevant parameters under the gross and net profit

methods and their impact on reliability
The OECD TPG claim, as an argument against applying the TNMM, that
net profits are susceptible to influence by “some factors that would either
not have an effect, or have a less substantial or direct effect, on price or
gross margins between independent enterprises”.604
This assertion is ambiguous. Oddly enough, the OECD TPG do not offer
much in the way of reasoning to support it. Some vague indications of
which factors the assertion refers to, however, are provided. It is stated that
“factors other than products and functions can significantly influence net
profit”,605 and that net profit can be influenced by “the potential for varia-
tion of operating expenses across enterprises”. (Emphasis added) 606
Based on these indications, the core of the argument seems to be the obser-
vation that operating expenses are not included in gross profits, as well as the
OECD’s perception that this weighs in favour of the gross profit methods.
In this section, the author analyses to which extent the assertion should be
deemed justified. Before beginning this analysis, he would like to point out
604. OECD TPG, paras. 2.64 and 2.70.

178
the relevance of the basic motivation behind the transfer pricing methods.
The methods benchmark the controlled profit against the profit realized by
unrelated parties that carry out comparable transactions. If the controlled
profit lies outside the arm’s length range (see the discussion in section 6.5.),
there should be an adjustment so that the controlled allocation of income
aligns with the third-party profit data.
This logic does not hold up if the differences in profits realized by the
tested party and the uncontrolled comparables are caused by differing ac-
counting classifications or transfer pricing. First, differing accounting clas-
sification is a comparability problem that must be adjusted for, and will
typically pertain to the classification of business costs as COGS or operat-
ing expenses.607 Second, it should be ensured that third-party profit data are
not affected by any transfer pricing performed by the third party and its
related entities. Profit data extracted from such transactions will not reflect
profits realized between parties with genuinely conflicting interests, and
are thus not reliable. This may present a problem in practice, as a potential
comparable unrelated enterprise may be a member of a group.608
The author will now comment on the above assertion that “some factors
that would either not have an effect, or have a less substantial or direct ef-
fect, on price or gross margins between independent enterprises”.609
First, there seems to be an erroneous factual assumption underlying the as-

sertion. The quoted wording indicates that differences in functions are re-
flected in the COGS. That will normally not be the case. Differences in the
functions performed between enterprises will likely be mainly reflected in
the operating expenses.610
Second, the factors that affect gross profits logically also affect net profits.
Net profits are gross profits after reduction for operating expenses. How-
607. See sec. 6.2.4. and below in this section.

608. Also, with respect to the tested party, the OECD TPG state that the selected prof-
it indicator should be reasonably independent from controlled transactions; see OECD
TPG, para. 2.88. This guidance is not entirely clear. In the author’s view, it should, as
the point of departure, be irrelevant whether the profit indicator of the tested party is
affected by transfer pricing. In fact, that will presumably be the case, as well as the
reason why there is a transfer pricing assessment in the first place. In general, all profit
elements of the tested party (sales, COGS and operating expenses) will likely consist of
a mix of controlled and uncontrolled transactions.
610. See the discussion of operating expenses in sec. 6.2.4.
179
Allocation Rules
ever, the assertion of the OECD TPG pertains to the reverse causality: that
the net profits are influenced by factors that do not affect gross profits.
Of course, it should be recognized, from a purely factual point of view,

that there is merit to the argument, in the sense that operating expenses are
included in the net profits, while in the gross profits, they are not. It may
therefore be claimed that gross profit methods have an intrinsic advantage
relative to the net profit methods, in that they rely on fewer parameters.
This observation in and of itself, however, is not particularly meaningful.
The real issue is whether this difference is influential towards the reliabil-
ity of the results yielded by the gross and net profit methods.
Operating expenses come as an additional expense item that must be com-

pared to the corresponding expenses of an unrelated comparable enter-
prise, adding complexity and additional comparability concerns to a trans-
fer pricing assessment. One could therefore be tempted to assume that it
will be easier to achieve sufficient comparability under the gross profit
methods. If that were true, the cost-plus method and the resale price meth-
ods should generally be preferable over the CPM and the TNMM.
That, however, is not the case. On the contrary, there are, as mentioned,
limitations on the accounting data available on CUTs that may make a net
profit analysis more reliable than a gross profit analysis.611 The author will
illustrate this through an example in which the accounting data available
for the tested party and a comparable unrelated enterprise is as follows:
Distributor Profit Distributor Profit

(tested party) margins (unrelated) margins
Sales 1,000 1,000
COGS 650 600
Gross profit 350 35% 400 40%
Operating expenses 250 300
Net profit 100 10% 100 10%
The tested party earns a smaller gross profit margin than the comparable
enterprise, but the same net profit margin. If a gross profit method were
applied to test the pricing, for instance, the resale price method, the result
would be that the COGS should be adjusted downwards to 600 in order for
611. See also Culbertson (1995).
180
the tested party to earn a comparable gross profit margin.612 An effect of

this adjustment would be that the net profit of the tested party is increased
to 150, resulting in a higher net profit margin earned by the tested party
than the comparable unrelated enterprise.
If the tested party and the unrelated enterprise are closely comparable with
respect to the functions performed, assets used and risks assumed, it can-
not be ruled out that the deviation between their gross profit margins is due
to accounting differences with respect to the classification of expenses as
COGS or operating expenses. It could be, for instance, that the tested party
attributes certain depreciation or warranty costs to COGS, while the com-
parable enterprise does not. Had the third-party accounting specifications
been known, it would have been possible to perform an appropriate com-
parability adjustment at the gross profit level of the comparable enterprise.
In practice, however, it will normally not be possible to conclude with cer-

tainty whether the observed differences in gross profit margins are due to ac-
counting differences, as the third-party public financial statements will like-
ly not be sufficiently detailed. Further, the tested party and the tax authorities
will normally not have access to the internal management accounting data
of a comparable third-party enterprise. In this case, it will normally yield a
more reliable transfer pricing outcome if the income of the tested party is
tested at the level of net profits because the effect of accounting differences
will be eliminated at this level, as net profits include all business expenses.
Thus, had a net profit method been applied to the accounting data of the tested
party in the above example, no transfer pricing adjustment would have been
performed, as both the tested and third party have a net profit margin of 10%.
The author therefore concludes that it cannot be claimed without reserva-

tion that gross profit methods are intrinsically more reliable than net profit
methods simply because they rely on fewer parameters.
6.3.4. Are operating expenses relevant under the transactional

pricing methods (CUT, resale and cost-plus)?
The OECD TPG contrast the TNMM against the CUP, resale price and
cost-plus methods (the transactional pricing methods) for the purpose of
612. The sales of the tested party are to unrelated parties, making COGS the only
item in gross profits that are influenced by transfer pricing.
181
Allocation Rules
illustrating that net profits are influenced by factors other than those that
influence gross profits. The author will tie some comments to this.
As the difference between the calculation of gross and net profits is the
inclusion of operating expenses in net profits, a sensible way to assess the
assertion of the OECD TPG is presumably to analyse the relationship be-
tween operating expenses and the CUP, resale price and cost-plus methods.
If the results under these three methods are affected by operating expenses,
directly or indirectly, there would logically seem to be less merit to the as-
sertion. If so, the assertion should not be used as an argument against the
application of the TNMM.
The CUP method compares the price charged in a controlled transaction

to the price charged in an uncontrolled comparable transaction.613 The
method is particularly suited for the pricing of generic products with read-
ily available objective market prices.614 Over time, prices must be set so
that the seller reaps a net profit.615 That is, each product sold must, in the
long run, bear its share of the operating expenses of the selling enterprise.
Thus, logically, the price of a controlled transaction tested under the CUP
method must be influenced by both COGS and operating expenses.616
The resale price method allocates an arm’s length gross profit margin to
the tested party.617 More specifically, the resale price of the tested party is
reduced by a gross profit margin, extracted from CUTs, to find the price
that an unrelated distributor would require in order to cover its operating
expenses and provide it with a net profit. The view of the OECD is that a
low gross margin will be justified if the related distribution entity does not
carry on substantial business activities, but only “transfers the goods to a
third party”.618 Conversely, a higher margin is justified when the distributor
613. Treas. Regs. § 1.482-3(b); and OECD TPG, paras. 2.13-2.20. See also the analy-
sis of the US comparable uncontrolled transaction (CUT) method in sec. 7.2. and the
OECD CUT method in secs. 7.3. and 11.4.
614. See the unbranded coffee bean example in OECD TPG, para 2.18; identical
products apart from delivery terms in OECD TPG, para 2.19; and identical products
with only volume differences in OECD TPG, para. 2.20.
615. See the discussion in OECD TPG, para. 2.43.
616. For instance, para. 2.19 of the OECD TPG contains an example of an application
of the CUP method where an adjustment is made for differences in transportation and
delivery terms. Such costs may be classified as operating expenses.
617. Treas. Regs. § 1.482-3(c); and OECD TPG, paras. 2.21-2.38.
182
is in possession of particular marketing expertise or if it contributes sub-

stantially to the creation or maintenance of marketing intangibles.619
The reasoning behind this position is likely that a distributor that provides
more value, through the incurrence of more operating expenses, should
be compensated with a larger gross profit margin than a distributor that
provides relatively less value and incurs relatively less operating expenses.
In other words, as the operating expenses (and risks) of the distributor in-
crease, so should its gross profit margin. The reasoning of the OECD on
this point is therefore, in fact, a net profit view.620
The cost-plus method calculates the return due to a supplier as a gross mar-
gin mark-up on an appropriate cost base.621 The method is typically suit-
able for determining the transfer price for semi-finished products or for the
provision of services, either by using internal or external comparables. The
OECD recognizes that it may be problematic to determine an appropriate
cost base in order to apply the cost-plus method.622
There may be differences between the tested party and the unrelated com-
parables with respect to the level of business expenses incurred, as well as
the classification of expenses as COGS or operating expenses. Of particular
interest is the recognition by the OECD that it may be necessary in order to
establish a cost base for applying the cost-plus method to also include cer-
tain operating expenses in order to achieve consistency and comparability,
619. Id. However, if the distribution subsidiary is the owner of unique intangibles, it
will be highly unlikely that the TNMM is a suitable transfer pricing method to allocate
income to the distribution subsidiary in the first place.
620. If the accounting classifications used in the controlled and uncontrolled transac-
tions differ, adjustments should be made to the accounting data in order to properly
apply the resale price method; see OECD TPG, para. 2.35. For instance, it may be that
R&D costs are included in the COGS, and thus in the gross margin of the tested party,
while classified as operating expenses in the financial statements of the comparable
unrelated enterprises. Other practical scenarios where adjustments are required may,
for instance, be where the tested party and the comparable unrelated enterprises have
classified warranty costs differently; see OECD TPG, para. 2.37.
621. Treas. Regs. § 1.482-3(d); and OECD TPG, paras. 2.39-2.55.
622. See OECD TPG, paras. 2.42-2.43. The author’s focus here is on the classification
of business expenses as either COGS or operating expenses, but there are also other
relevant classification problems connected to the cost-plus method. For instance, as
the method is one-sided, it is limited to taking into account the costs of the controlled
supplier, not the controlled buyer. This represents a potential problem, as there will be
an incentive to allocate costs to the buyer side of the transaction in order to reduce the
basis for allocation of income to the supplier, which is the tested party under the cost-
plus method.
183
Allocation Rules
and that “in these circumstances the cost plus method starts to approach a
net rather than gross profit analysis”.623
Further, if the tested party is a pure contract-manufacturing subsidiary, the

basis for applying the cost-plus method will be all of the costs contractually
connected to the assembling activities.624 The OECD TPG take the same
stance with respect to contract research and development (R&D) agree-
ments.625 All costs agreed between the controlled parties for the provision
of the R&D services shall be used as the basis for the cost-plus method
to remunerate the research provider. Such contractually agreed costs may,
and likely will, encompass expense items that, on a stand-alone basis,
would be classified as operating expenses.
This touches on the greater issue that there is no bright line in all cases
between the COGS and operating expenses.626 The OECD also recognizes
that variations in accounting standards among countries may render it dif-
ficult to draw precise lines between the COGS and operating expenses and
that the application of the cost-plus method may include the consideration
of some operating expenses.627 Even still, the OECD TPG take the rather
theoretical stance that these problems “do not alter the basic practical dis-
tinction between the gross and net profit approaches”.628
6.3.5. Are comparability adjustments under the gross profit

methods more reliable than under the net profit methods?
The OECD TPG assert, in somewhat ambiguous language, that gross and
net profits may be influenced by some of the same factors, “but the effect
of these factors may not be as readily eliminated” under the TNMM as
under the gross profit methods.629 Such factors may include the threat of
new entrants, competitive positions, management efficiency and individual
strategies, the threat of substitute products, varying cost structures (includ-
ing age of plant and equipment), differences in the cost of capital and the
degree of business experience.630 The author takes issue with this assertion.

626. See also Haugen (2005), at p. 225 on this issue.
184
First, the only way in which a comparability factor may influence both gross
profit and net profit is if the factors influence either sales or the COGS.
Clearly, most of the factors listed by the OECD TPG will have a direct effect
on, first and foremost, the sales of the enterprise. This includes the com-
petitive position, strategy and substitute products. Other factors, such as cost
structures, will clearly impact both the COGS and the operating expenses.631
The author fails to understand why it should be more onerous to adjust for
the same factors under the TNMM than under the resale price or cost-plus
methods. No justification for this assertion is offered by the OECD TPG.
Of course, it may be difficult to perform reliable comparability adjustments
in and of themselves, but that does not weigh on the relative difficulty of
adjustment between the gross and net profit methods.
Second, the underlying motivation behind the assertion seems to be the per-
ception that the effects may be eliminated under the gross profit methods by
way of stricter comparability requirements relative to the TNMM.632 In the
author’s view, this is an unrealistic and unfounded point of view. The main
comparability problem associated with the use of the TNMM is the extrac-
tion of blended profit margins from comparable unrelated enterprises.633 It
is crucial that this issue is not associated with the TNMM in particular.634
Both the resale price and the cost-plus methods require data on individual
CUTs. Such data is largely unavailable. Both the gross profit margin under
the resale price method and the mark-up on costs under the cost-plus meth-
od are therefore, in practice, extracted from the financial statements of se-
lected comparable enterprises. It is the case for the net profit methods that it
will normally not be possible to separate the transactions of the comparable
unrelated enterprise that are relevant for transfer pricing purposes from the
transactions that are not. Gross profit methods are susceptible to the same
measurement errors from blended profits as the net profit methods are.
Third, the OECD TPG support the OECD’s assertion with another asser-
tion, i.e. that net profit indicators can be less sensitive than gross margins to
631. The cost of capital is irrelevant in this context. Neither the gross nor net profit
methods include financial income or expense items. Cost of capital should therefore be
disregarded. It must be due to an error that the factor is even mentioned in OECD TPG,
para. 2.71 as a comparability factor that may influence both gross and net profits.
633. See the discussion of the use of aggregated third-party accounting data under the
TNMM and blended profits in sec. 8.6.
634. Id.
185
Allocation Rules
differences in the extent and complexity of functions, as well as to differ-

ences in risk.635 The 2010 OECD TPG contain three examples that seek to
illustrate the relative sensitivity of gross and net profit indicators in certain
scenarios.636 More specifically, the examples pertain to whether unadjusted
comparability differences will impact gross or net profits more.
The author observes that these examples are highly specific and rely fully
on the assumptions made. They therefore have little value as general guid-
ance. In the author’s view, the takeaway from the examples should not be
more than the fact that unadjusted comparability issues will likely affect the
results regardless of whether a gross or net method is used and that the im-
pact of the unadjusted differences may influence gross and net profits dif-
ferently. The choice of an appropriate transfer pricing method should take
this into account. Further, he also finds that it is somewhat inappropriate to
directly compare gross and net margins in the way that the examples do.
When both margins are calculated based on the same denominator, it is, of
course, obvious that gross margins will tend to be larger than net margins.
This will again result in the likely scenario, as reflected in the first and sec-
ond example, that the difference between the controlled and uncontrolled
gross and net margins will tend to be larger for the former than the latter,
with the exception of cases in which net profits are zero or negative, as the
situation was in the third example. For these reasons, the author rejects the
OECD assertion that that the effects of the mentioned factors may not be
as readily eliminated under the TNMM as under the gross profit methods.
The US regulations – opposite to the OECD’s perspective discussed above

– take the stance that their net profit method, the CPM, will generally pro-
vide a more reliable result than the gross profit methods in cases where
the available CUTs contain significant product or functional differences
relative to the controlled transaction. It is stated that as “operating profit
usually is less sensitive than gross profit to product differences, reliability
under the comparable profits method is not as dependent on product simi-
larity as the resale price or cost plus method”.637 Further, it is stated that the
“degree of functional comparability required to obtain a reliable result un-
der the comparable profits method […] is generally less than that required
under the resale or cost plus methods”.638

636. OECD TPG, annex 1 to ch. 2.
637. Treas. Regs. § 1.482-5(c)(2)(iii).
638. Treas. Regs. § 1.482-5(c)(2)(ii).
186
Which transfer pricing method should govern the profit allocation
among value chain inputs?
The reasoning behind this position is that differences in functions are often
reflected in operating expenses, and therefore that taxpayers performing
different functions may have very different gross profit margins, but earn
similar net profit margins.639 It is not entirely easy to follow this. If differ-
ences in functions primarily are reflected in operating expenses, would the
gross profits of the tested party and the comparables not precisely remain
unaffected by the functional differences? After all, operating expenses in-
fluence net, not gross, profits.
A possible interpretation is that net profit methods may be more reliable in

cases where it is clear that there are functional differences, but it is unclear,
due to a lack of sufficiently detailed accounting data, whether the differ-
ences in costs have been classified as COGS or as operating expenses. If
this interpretation is valid, the US position is, in reality, a variant of the
argument that net profit methods eliminate distortions from differing ac-
counting classifications.
The point of departure is that the best method rule determines which pric-
ing method is appropriate.640 However, most comparables will likely suffer
from a lack of both product and functional comparability. In light of the
factors discussed in this chapter, in particular that the effects of differing
accounting classifications are eliminated under the net profit methods, the
author agrees that there does not seem to be any convincing reason as to
why the CPM should not be the preferred method in these cases.
6.4. Which transfer pricing method should govern

the profit allocation among value chain inputs?
With respect to the important question of which transfer pricing methodol-
ogy should govern the allocation of operating profits among value chain
contributions, the approach under both the US and OECD regimes is that
the most appropriate method for the individual case should be selected, i.e.
the so-called “best-method rule”.641
639. Id.
640. Treas. Regs. § 1.482-1(c)(1).
641. For general comments on the best method rule, see, e.g. Wittendorff (2010a), at
pp. 702-711. See also Luckhaupt et al. (2011), at p. 116, where it is emphasized that there
is a large degree of discretionary assessment associated with (selecting and) applying
the OECD transfer pricing methods.
187
Allocation Rules
This rule first surfaced in the 1993 US temporary regulations.642 The US

version of the rule states that the arm’s length result of a controlled trans-
action shall be determined under the transfer pricing method that provides
the most reliable measure of an arm’s length result.643 The arm’s length
result may be determined under any method without establishing the inap-
plicability of another method. However, if another method subsequently is
shown to produce a more reliable measure of an arm’s length result, that
other method must be used.
The US regulations recognize that data from uncontrolled comparable

transactions provide the most objective basis for determining whether the
results of a controlled transaction are at arm’s length.644 In determining
which transfer pricing method is the most appropriate for a particular case,
the primary factors to take into account are the degree of comparability
between the controlled and uncontrolled transactions, as well as the qual-
ity of the data and the assumptions used in the analysis. For the purpose
of allocating operating profits to intangible value chain inputs, the strict
comparability requirements for applying the CUT method will generally
ensure that the CPM (or the income method in the context of cost-sharing
arrangements (CSAs)) is the preferred method in cases where only one
party to the controlled transaction provides unique value chain contribu-
tions, and the PSM where more than one party renders such value chain
inputs. The author will revert to this in detail in the subsequent analysis of
the transfer pricing methods (in chapters 7-16).
As touched upon in chapter 3, new temporary and final regulations were

issued in September 2015, clarifying the application of the arm’s length
standard and the best-method rule with other tax code provisions.645 These
aim to ensure that there is an arm’s length charge – based on an aggregated
valuation approach – for all intangible value transferred among related en-
tities, regardless of the legal form in which the transaction is carried out.646
In particular, controlled transactions combining sections 367 and 482
642. See Higinbotham et al. (1993) for a discussion.

643. Treas. Regs. § 1.482-1(c)(1). For an overview of the development behind the US
best-method rule, see Ahmadov (2011), at pp. 195-196. See also Pichhadze (2015), at
sec. 3.1.1.
644. Treas. Regs. § 1.482-1(c)(2). It may also be relevant to consider whether the re-
sults of an analysis are consistent with the results of an analysis under another method.
645. 80 FR 55538-01. See the comments on these provisions in the context of the
relationship between US Internal Revenue Code (IRC) secs. 367 and 482 in sec. 3.5.8.
of this book, on aggregated valuation in sec. 6.7.2. and on intra-group R&D services
(contract R&D arrangements) in sec. 21.4.
646. See Treas. Regs. (80 FR 55538-01) § 1.482-1T(a)(i)(A) and (C).
188
Which transfer pricing method should govern the profit allocation
among value chain inputs?
transfers are targeted.647 While no particular method is emphasized as the

best for dealing with such (typically outbound) transfers, the requirement
for an aggregated valuation will normally entail that the income method
(or an unspecified method of similar content) is selected.648 The residual
profits will then be allocated to the group entity that provides unique inputs
to the intangible development (R&D, workforce in place, know-how, etc.)
or exploitation (unique IP), which, in practice, tends to be the US group
entity, leaving only a normal market return to the foreign group entities
involved in the controlled transaction.
With respect to the OECD TPG, there has been – as touched upon in chap-
ter 5 –a clear historical preference for the traditional transactional trans-
fer pricing methods (the CUT, resale price and cost-plus methods).649 The
OECD blatantly disregarded the fact that the CUT method would generally
be difficult (if not impossible) to apply in a reliable manner to allocate op-
erating profits from intangible value chains.650 While the 2010 OECD TPG
abandoned the sharp divide between the traditional transactional transfer
pricing methods and the profit-based methods (the TNMM and PSM), the
former were still regarded as the most direct means of establishing arm’s
length pricing.651 The OECD has now rid itself of the historical bias to-
wards the CUT method.652 While the 2017 OECD TPG on intangibles do
not alter the general (2010) OECD approach for selecting pricing methods,
significant supplementary guidance is provided on the selection of pricing
methodology in the context of IP value chains. The new guidance expresses
scepticism towards the CUT method, combined with a pronounced prefer-
ence for the PSM.653 This new supplementary guidance on the choice of
647. See the analysis of combined sec. 482 and 367 transfers in sec. 3.5.8. of this
book.
648. The income method is discussed in sec. 14.2.8.3.
649. For an instructive discussion of the development of the OECD best-method rule,
see Ahmadov (2011), at pp. 184-193.
650. 1995 OECD TPG, paras. 3.50 and 3.54. See also Luckhaupt et al. (2011), at
p. 104 on the relationship between the available third-party profit data and the selection
of an appropriate transfer pricing method.
652. For a different approach (that the author does not agree with), see the views
expressed in Kotarba (2009), at p. 154 and p. 170 pertaining to the best-method rule.
Kotarba argues, much in light of his analysis on p. 160 of the Australian ruling in Roche
Products Pty Ltd. v. Commissioner of Taxation ((2008) 70 A.T.R. 703 (Austl.)), that the
use of profit-based methods should be restricted to cases in which the traditional trans-
actional methods cannot be employed. See also Cauwenbergh (1997), at pp. 139-141
for a useful (but now dated) comparative overview of the approach taken by different
jurisdictions to the priority of transfer pricing methods. See also Markham (2004) in
this respect.
653. OECD TPG, paras. 6.131-6.132.
189
Allocation Rules
methodology is part of the more sophisticated approach introduced in the

2017 OECD TPG for determining ownership of unique intangibles based on
the important functions doctrine,654 revised guidance on the application of
the PSM,655 as well as new rules on the allocation of incremental operating
profits due to cost savings, location-specific advantages and synergies.656
The author will revert to the 2017 OECD TPG on the selection of pricing
methodology to allocate profits from intangible value chains in chapter 11.
6.5. The arm’s length range
6.5.1. Introduction
The application of a transfer pricing method to a set of CUTs will normally

not yield one single arm’s length price or gross or net profit margin, but sever-
al.657 The arm’s length range is a spectrum of arm’s length results. If the con-
trolled price (CUT method), gross profit margin (resale price and cost-plus
methods) or net profit margin (CPM/TNMM) fall within the arm’s length
range, there will be no adjustment under US law or under the OECD MTC.658
The US regulations do not require that the Internal Revenue Service (IRS)
determine an arm’s length range prior to making a reassessment.659 For
instance, the IRS may propose an allocation based on a single uncontrolled
price if the CUT method is applied. If the taxpayer subsequently demon-
strates that his return position falls within the arm’s length range produced
by equally reliable CUTs, the IRS will not be authorized to reassess.660 The
OECD takes a similar stance in that only if a taxpayer is unable to establish
that the controlled transaction falls within the arm’s length range, the tax
administration will be entitled to reassess. In this respect, the arm’s length
range works, under both the US and OECD regimes, as a safe harbour for
the taxpayer. Conversely, as seen from the point of view of the tax authori-
ties, the arm’s length range represents a reassessment limitation.
654. See the analysis of the “important functions doctrine” in sec. 22.3.2.
655. See the analysis of the OECD PSM in ch. 9 and the relative role of the method in
the post-BEPS OECD regime in sec. 11.4.
656. See the analysis of the allocation of incremental operating profits from local
market characteristics, location savings and synergies in ch. 10.
657. For a recent discussion of the arm’s length range (under transfer pricing laws in
Denmark, Hungary and the United States), see Koue et al. (2017).
658. Treas. Regs. § 1.482-1(e)(1); and OECD TPG, para. 3.60.
659. Treas. Regs. § 1.482-1(e).
660. Id.
190
The arm’s length range
In the context of allocating operating profits from intangible value chains

(where unique IP is used as input), the arm’s length range will, in prac-
tice, be relevant mainly for the one-sided pricing methods (in particular the
CPM and TNMM), or more specifically, where the purpose is to allocate
a normal market return to the routine value chain contributions rendered
by the tested party (e.g. compensation of a distribution subsidiary that li-
censes a trademark from another group entity).661 The close relationship
between the arm’s length range and the CPM (which is the key US method
for allocating normal market returns to routine entities) is accentuated by
the fact that the arm’s length range provision was introduced into the US
regulations alongside the CPM.662
The author will comment on the positions taken by the United States and the
OECD on use of the arm’s length range in sections 6.5.2.-6.5.3. The main fo-
cus will be on the US regulations, as the OECD TPG are brief on this issue.
6.5.2. The level of comparability required to include

an uncontrolled transaction in the arm’s length range
The selection of CUTs must comply with the comparability criteria set out
by the particular pricing method applied and must be sufficiently similar to
the controlled transaction in order to provide a reliable measure of an arm’s
length result.663 Under the 1993 temporary regulations, all valid applica-
tions of transfer pricing methods were included in the arm’s length range.664
The final 1994 regulations amended this rule to reflect the possible use of
inexact comparables. The idea behind the amendment was that it would
be inappropriate to derive the arm’s length range from a mix of exact and
661. See Markham (2005), at pp. 309-310, who argues that the arm’s length range may
“counteract” many of the problems associated with applying the arm’s length principle
to intangible transactions. This is, in the author’s view, an unrealistic assertion, as the
arm’s length range only pertains to the determination of a normal market return to
routine value chain contributions. It does not, in and of itself, aid in the allocation of
residual profits from unique intangibles.
662. See the preamble to the 1993 temp. Treas. Regs. (58 FR 5263-02) § 1.482-1T(d)
(2). The CPM was favoured in a range of arm’s length results rather than a single arm’s
length price. Even though primarily motivated by the CPM, the 1993 temporary regula-
tions extended the use of the arm’s length range to all of the pricing methods. However,
to form a part of the range, each application of the applicable pricing method must
independently satisfy the criteria for the application of that method and the general
comparability standards under temp. Treas. Reg. § 1.482-1(c)(2).
663. Treas. Regs. § 1.482-1(e)(2)(ii).
664. 1993 Temp. Treas. Regs. (58 FR 5263-02) § 1.482-5T(d)(2)(i).
191
Allocation Rules
inexact comparables, as that would assign the same weight to results with
potentially widely varying degrees of reliability. The rule under the final
regulations is therefore that, as the point of departure, the arm’s length
range can only include uncontrolled transactions of “similar comparability
and reliability”.665 Third-party transactions with significantly lower levels
of comparability and reliability must therefore be discarded.666 This leaves
two possible alternatives for determining the arm’s length range.
The first alternative would be to include all uncontrolled transactions in the

arm’s length range.667 This requires that the information available on the
CUTs is “sufficiently complete”, i.e. that it is likely that every material dif-
ference between the controlled and each uncontrolled transaction has been
“identified” and that each difference has a “definitive and reasonably ascer-
tainable effect” on price, mark-up or profit, and that an adjustment is made
to eliminate the effect of such differences. If these requirements are ful-
filled, all of the uncontrolled transactions are included in the range.668 The
OECD takes a similar stance as the US regulations on this point. Under the
OECD TPG, uncontrolled transactions that have a lesser degree of compa-
rability than the other CUTs included in the range should be discarded.669
The second alternative for determining the arm’s length range is to use the
so-called “interquartile range”.670 This will be relevant where the above-
mentioned comparability standard (identification and adjustment of price-
relevant differences in the CUTs) is not met. In this situation (which is highly
practical), the arm’s length range will contain uncontrolled transactions of
varying degrees of comparability and reliability. Had the whole range been
used, that would imply that results with varying degrees of comparability
and reliability would be assigned equal weight. This also would violate the
rule that uncontrolled comparables with significantly lower levels of com-
parability and reliability must be discarded from the arm’s length range.671
Because it would be impossible to directly identify and quantify the mate-
rial differences between the controlled and each uncontrolled transaction,
the US regulations require that the differences be taken into account indi-
rectly through the use of a statistical range. When it is highly probable that
665. Treas. Regs. § 1.482-1(e)(2)(i).

666. Treas. Regs. § 1.482-1(e)(2)(ii).
667. Treas. Regs. § 1.482-1(e)(2)(iii)(A).
668. See Treas. Regs. § 1.482-1(e)(5), Example 1.
670. Treas. Regs. § 1.482-1(e)(2)(iii)(B). See Culbertson et al. (2003), at p. 108; and
Casley et al. (2003), at p. 167 on the interquartile range.
671. Treas. Regs. § 1.482-1(e)(2)(ii).
192
The arm’s length range
the uncontrolled comparables are not of roughly equal comparability, it is

reasonable to assume that the results diverging significantly from the norm
are not comparable to the controlled transaction.
The interquartile range exception (from using only CUTs that satisfy the
above-mentioned comparability requirements) allows CUTs that contain
material differences (that may affect pricing) relative to the controlled
transaction. To compensate for the reduced reliability caused by the use of
such comparables, the arm’s length range itself must be adjusted through
the application of a valid statistical method so that there will be a 75%
probability of a result falling above the lower end of the range and a 75%
probability of a result falling below the upper end of the range. The idea
behind the interquartile range is to narrow the range of results, thus exclud-
ing the lowest and highest prices provided by the selected uncontrolled
transactions, in order to increase the reliability of the data by excluding the
extremes at both ends of the data range.672
The position taken in the US regulations is that the interquartile range will
ordinarily provide an acceptable measure of the arm’s length range (even if
the range includes somewhat “incomparable” third-party transactions).673
The OECD takes the same position.674
The interquartile range is the range from the 25th percentile to the 75th
percentile of the results from the uncontrolled comparables. For instance,
if seven uncontrolled transactions have been identified and the pricing ap-
plied in them are 1, 2, 3, 4, 5, 6 and 7, the middle value of the prices, i.e. the
median, will be 4. The interquartile range is the range of the middle 50%
of the data. The median of the lower half of the data, 2, is the first quartile,
while the median of the upper half of the data, 6, is the third quartile. The
difference between the first and third quartiles is the interquartile range.
Thus, if the controlled pricing of the tested party lies between 2 and 6,
there will be no adjustment, as that pricing will be within the interquartile
range. However, if the controlled pricing is either below 2 or above 6, there
will be an adjustment.
672. The US regulations contain an example pertaining to an outbound structure be-

tween a US manufacturing parent and a foreign subsidiary in which the US Internal
Revenue Service (IRS) considers applying the CPM; see Treas. Regs. § 1.482-1(e)(5),
Example 4.
673. Treas. Regs. § 1.482-1(e)(2)(iii)(B).
674. OECD TPG, para. 3.57. See Koue et al. (2017), where it is argued that there
should be additional guidance on how to apply the interquartile range in practice.
193
Allocation Rules
6.5.3. On which point within the arm’s length range may

a reassessment be based?
If it is clear that the price or profit result of the controlled transaction lies
outside of the arm’s length range, there will be an adjustment. The question
is then as to which point inside the arm’s length range may the tax authori-
ties reallocate income.
The US regulations allow the IRS to reassess the taxpayer’s result at any
point within the arm’s length range.675 The OECD takes the same posi-
tion.676 Nevertheless, if the interquartile range is applied, the taxpayer’s
result will normally be adjusted to the median of the interquartile range.677
6.6. Comparability
6.6.1. Introductory comments
Comparability is fundamental in transfer pricing. The pricing of a con-

trolled transaction will satisfy the arm’s length standard if it results in the
same allocation of profits that would have been realized if unrelated enter-
prises had carried out the same transaction under the same conditions. Yet,
because it will normally not be possible to identify uncontrolled transac-
tions that are identical to the controlled transaction in all respects, the US
and OECD transfer pricing methods only require third-party benchmark
transactions to be comparable.678
Comparability requirements are specific to the transfer pricing method be-

ing applied. As the pricing methods test different parameters, comparabil-
ity problems will vary from method to method. Thus, it should be recog-
nized that transfer pricing comparability requirements form a heterogene-
675. Treas. Regs. § 1.482-1(e)(3).

677. Treas. Regs. § 1.482-1(e)(3). The median of a set of data is the middle-most num-
ber in the set. For instance, in the dataset 1, 2, 3, 4 and 5, the median is 3. The OECD
TPG find it “appropriate to use measures of central tendency to determine this point
… in order to minimise the risk of error due to unknown or unquantifiable remaining
comparability defects”; see OECD TPG, para. 3.62.
678. On comparability, see an overview of significant issues in practice in Haugen
(2005), at p. 227. See also Higinbotham et al. (1998); Fris et al. (2010); and Luckhaupt
et al. (2011), at p. 100.
194
Comparability
ous set of rules, not one coherent norm; there are as many comparability
norms as there are pricing methods.
For instance, the CUT method applied to allocate profits to unique IP seeks
to identify a third-party royalty rate that can be used to benchmark the
rate used in the controlled transaction. The CPM/TNMM benchmark the
normal return net profit margin allocable to the tested party’s routine value
chain contributions against the net profits of comparable third parties. The
need to attain transactional comparability is particularly pronounced un-
der the CUT method, while functional comparability is more relevant un-
der the CPM/TNMM.679
Due in part to the manner in which multinationals organize their trans-

actions and business structures through centralized ownership of unique
value chain inputs and dispersed performance of routine functions through
specialized local group entities, it has, in practice, proven genuinely prob-
lematic to achieve the ideal comparability goals of the arm’s length stand-
ard as set out by the pricing methods.680 This problem runs deep.
First, the comparability problem has historically been the main catalyst for
the development of new transfer pricing methods. From its origin in the
transaction-oriented CUT method,681 the development has gone through
the more functionally-geared CPM/TNMM,682 while the core application
of the PSM is to allocate residual profits based not on comparables, but
on concrete assessments of the relative values of the unique intangibles
contributed to the value chain.683 The latest addition to this development
is the income method of the US cost-sharing regulations,684 which require
intra-group profit allocations to be aligned with the realistic alternatives of
the controlled parties.685
679. OECD TPG, para. 1.109. See also the discussion in sec. 8.6. of this book, with
further references.
680. See Francescucci (2004a), at p. 68 for a discussion of how the transactions of
group entities may differ from those of third parties. See also Schön et al. (2011), at
pp. 102-103, for reflections pertaining to the lack of comparables in transfer pricing.
There are also other problematic aspects pertaining to comparables, such as the use by
tax authorities of secret comparables; see, e.g. Przysuski et al. (2003).
681. See the discussion of the US CUT method in sec. 7.2. and of the OECD CUT
method in secs. 7.3. and 11.4.
682. See ch. 8 on the CPM and TNMM.
683. See ch. 9 on the US and OECD profit split methods.
684. See the discussion of the income method in sec. 14.2.8.3.
685. The “realistic alternatives available” pricing principle has also had significant
influence on the new OECD guidance; see, e.g. OECD TPG, paras. 1.38, 1.40, 1.122
and 6.139. See also the discussions of the 2017 OECD TPG on valuation in ch. 13.
195
Allocation Rules
Second, the comparability problem has resulted in tension between the per-
ceived need of some jurisdictions to maintain transfer pricing rules based
on ideal comparability requirements, on the one side, and the need to adapt
the methods sufficiently to the real-world problems encountered in the field
by multinationals and tax authorities – revolving in particular around the
lack of reliable disaggregated third-party accounting data – on the other
side.686 A consequence of this tension has been that the current OECD con-
sensus on rather fundamental comparability requirements is ambiguous.
The author imagines that comparability problems will remain a relevant

issue in transfer pricing also in the future, likely necessitating revised or
new pricing methods, perhaps with an even stronger emphasis on the real-
istic alternatives available to the controlled parties, moving transfer pric-
ing (even) further away from traditional, transaction-based comparability
assessments.
Because comparability is inherently linked to the transfer pricing methods,

the author will analyse the specific comparability issues in connection with
his discussions of the individual transfer pricing methods.687 Of course, this
link makes it difficult to present a generalized overview of comparability
without the discussion becoming either too bland or even redundant in the
sense that the topics discussed must nevertheless be reverted to when the
particular pricing methods are later analysed. Nevertheless, he finds it nec-
essary at this stage of the book to at least introduce some broad information
on comparability in this chapter that can form the backdrop for the more
specific discussions that will follow.
The standard of comparability under the US regulations and the OECD

TPG is discussed in section 6.6.2.; the question of whether the required
degree of comparability varies between the pricing methods is discussed
in section 6.6.3.; some comments on the relationship between comparabil-
ity and the rules for determining ownership to self-developed intangibles
are provided in section 6.6.4.; and the comparability factors under the US
regulations and the OECD TPG are discussed in section 6.6.5.
686. See the discussion of the use of aggregated third-party financial profit data as
comparables under the TNMM in sec. 8.6.
687. Comparability problems under the US CUT method are discussed in secs. 7.2.2.-
7.2.3.; under the OECD CUT method in secs. 7.3.2.-7.3.3. and 11.4.; under the CPM and
TNMM in secs. 8.5. and 8.6., respectively; and under the US and OECD PSMs in ch.
9. Buy-in pricing under the CUT method of the US cost-sharing arrangement (CSA)
regulations is discussed in sec. 14.2.8.2.
196
Comparability
6.6.2. The standard of comparability
The US standard of comparability is that an uncontrolled transaction must

be “sufficiently similar that it provides a reliable measure of an arm’s length
result”688 and, further, that “all factors that could affect prices or profits in
arm’s length dealings (comparability factors)” must be taken into account
when determining the degree of comparability.689
If there are material differences between the controlled and uncontrolled

transactions, comparability adjustments must be made, provided that the
effect of the differences on the prices or profits can be ascertained with
sufficient accuracy. The adjustments must be made based on commercial
practices, economic principles or statistical analyses. If such adjustments
cannot be made, the comparable may still be used, but the reliability of the
analysis will be reduced. The interquartile range (see section 6.5.2.) may
then be used to limit the potential errors caused by the reduced degree of
comparability.
The OECD comparability standard is largely similar.690 The OECD TPG

state:
[B]ecause the separate entity approach treats the members of an MNE group
as if they were independent entities, attention is focused on the nature of the
transactions between those members and on whether the conditions thereof
differ from the conditions that would be obtained in comparable uncontrolled
transactions.691
Similarly to the US regulations, the OECD TPG also accept comparables

when it is not possible to achieve ideal comparability.692 Also here, the
interquartile range will normally be used to lessen the potential errors of
using uncontrolled transactions with limited comparability.
688. Treas. Regs. § 1.482-1(d)(2). See the comments in Wittendorff (2010a), at p. 393.
689. Treas. Regs. § 1.482-1(d)(1).
690. See the comments in Wittendorff (2010a), at p. 396. See Bullen (2010), at p. 215
on comparability adjustments.
197
Allocation Rules
6.6.3. Does the degree to which comparability is required

vary among the pricing methods?
It seems to be suggested from time to time that the degree to which com-
parability is required varies depending on the method being applied. A
typical assertion may be that the CUT method demands the strictest degree
of comparability, while the other methods require a lesser degree of com-
parability. If this assertion were true, it would entail qualitative differences
between the profit allocation results yielded by the different methods.
The pricing methods work in different ways. The CUT method is wholly
reliant on finding an uncontrolled transaction from which the third-party
royalty rate is extracted and used to benchmark the rate used in the con-
trolled transaction. The method can only be applied in a meaningful way
to allocate profits to unique IP if there is an extreme degree of comparabil-
ity between the intangibles transferred in the controlled and uncontrolled
transactions (in particular, with respect to profit potential).693
The CUT method directly allocates residual profits by way of applying

the royalty rate extracted from the third-party transaction to benchmark
the controlled profit allocation. Due to the large amount of income direct-
ly allocated (residual profits may be larger than normal market returns),
shortcomings in comparability may have profound effects on the ultimate
allocation of profits among controlled parties. Conversely, one-sided meth-
ods directly allocate only a normal market return profit margin to the rou-
tine inputs (e.g. contract manufacturing and sales functions) provided by
the tested party. Such returns are easier to benchmark, and comparability
shortcomings may not have the same impact on the ultimate profit alloca-
tion between the controlled parties.694
Thus, the CUT method is most vulnerable to comparability shortcomings

when applied to allocate profits to unique IP. This, however, does not en-
tail that its comparability requirements as such are stricter than those of
other methods. There must, of course, be sufficient comparability under all
methods so that the results yielded by them are reliable. The point is, rath-
er, that the CUT method and the one-sided methods test different classes of
third-party transactions. Achieving comparability under the CUT method
693. See also, in this direction, OECD TPG, para. 1.40, which states that “the method
becomes a less reliable substitute for arm’s length transactions if not all the charac-
teristics of these uncontrolled transactions that significantly affect the price charged
between independent enterprises are comparable”.
694. See the analysis of the aggregation of uncontrolled transactions in sec. 8.6.
198
Comparability
– when applied to allocate profits to unique IP – is reliant on finding third-

party transfers of IP comparable to the unique IP transferred in the con-
trolled transaction (which, of course, will be difficult), while the one-sided
methods rely on third-party comparables for routine inputs. Simply due to
this difference in parameters tested under the respective methods, it will,
in general, be more likely that sufficient comparability can be attained un-
der the one-sided methods than under the CUT method.695
The assertion that the one-sided methods require a lesser degree of com-
parability than the CUT method must therefore be rejected, as the com-
parability requirements cannot meaningfully be compared directly. They
pertain to different classes of third-party comparables. The only common
factor between the comparability requirements of the one-sided methods
and those of the CUT method is that the ultimate profit allocation, irre-
spective of the method applied, must be sufficiently reliable so as to reflect
what unrelated enterprises would have agreed to in a comparable scenario.
6.6.4. The relationship between comparability and the rules

for determining ownership of intra-group-developed
intangibles
As discussed, comparability requirements stem from the transfer pricing

methods. The rules for determining intangible ownership are – and have
historically been deemed to be – independent of the transfer pricing rules
at the outset, and therefore also independent of the comparability require-
ments.696 The 2017 OECD TPG on the determination of IP ownership,
however, draw upon transfer pricing principles, in particular, the profit split
method (e.g. that the relative importance of each group entity’s IP develop-
ment contributions shall be influential for profit allocation purposes).697
Because of this connection, it is no longer possible to operate with a bright
line between rules on IP ownership and transfer pricing rules. Likewise,
it cannot categorically be stated that comparability considerations are re-
695. In this direction, see also Luckhaupt et al. (2011), at p. 104 on the relationship
between the third-party profit data available and the selection of an appropriate transfer
pricing method.
696. The US and OECD intangible ownership provisions are analysed in part 3 of this
book.
697. See OECD TPG, para. 6.56. The author refers to his discussions of important
functions for determining ownership of co-developed intangibles under the OECD
TPG in sec. 22.3.2.; the remuneration of intangible development funding in sec. 22.4.;
and the relationship between the rules on intangible ownership and the PSM in sec.
11.4. See also the discussion of risk as a comparability factor in sec. 6.6.5.5.
199
Allocation Rules
served for the transfer pricing stage only. The impact of some compara-
bility factors, in particular the allocation of risk, are perhaps even more
pronounced in the context of determining IP ownership than in the context
of transfer pricing.
6.6.5. Comparability factors
The ultimate goal of transfer pricing is to ensure that controlled profit al-
locations reflect arm’s length profit allocations. In order to provide this
outcome, all factors having influence on the third-party profit allocation
must be taken into account and be comparable to the corresponding factors
in the controlled scenario. These factors have traditionally been referred to
as “comparability factors” in the US regulations and the OECD TPG. The
2017 OECD TPG, however, go far by replacing the term with “economi-
cally relevant characteristics”.698 Neither regime operates with exhaustive
lists of comparability factors. The point is rather to ensure that all factors
that possibly could affect pricing are properly taken into account. If a rel-
evant factor is disregarded, the application of a pricing method could yield
a non-arm’s length result.
Due to the comprehensiveness of a proper comparability analysis, there is

no bright line separating it from a functional analysis. As a point of depar-
ture, a functional analysis pertains to an analysis of the functions, assets
and risks contributed to the relevant value chain (and linking these contri-
butions to specific group entities) for the purpose of gaining an overview
of the multinational business,699 while a comparability analysis refers to an
assessment upon the application of a specific pricing method for the pur-
pose of allocating operating profits (i.e. whether a third-party transaction
is sufficiently similar to the controlled transaction so that it can be used for
benchmarking purposes).700 A functional analysis is thus carried out prior
to the actual pricing and comparability process (the application of a trans-
fer pricing method). Comparability analyses are so intensive that every
relevant aspect of a value chain must be clarified, with the result that a
typical comparability analysis may resemble a general functional analysis.

699. OECD TPG, para, 1.34
200
Comparability
Both the US regulations and the OECD TPG contain general guidance on
comparability.701 The key comparability factors under both regimes are (i)
contractual terms; (ii) functions; (iii) economic conditions; and (iv) risks.
These will be commented on in sections 6.6.5.2.-6.6.5.5.
6.6.5.2. Contractual terms
6.6.5.2.1. Introduction
A comparability analysis will normally start with the terms of the con-
trolled transaction (see section 6.6.5.2.2.),702 at least when applying the
CUT method for pricing intangibles. The contractual terms of the con-
trolled transaction must, however, align with the economic substance of
the actual behaviour of the controlled parties in order to be given effect
under the US and OECD regimes. Thus, if there is a discrepancy between
the written terms of the controlled agreement and the actual behaviour of
the controlled parties, the terms that are indicated by the actual behaviour
will prevail over the written terms. Economic substance and the notion of
non-recognition are discussed in section 6.6.5.2.3.
6.6.5.2.2. Comparability of contractual terms
All terms contained in the controlled agreement are, in principle, relevant

for the comparability analysis, of course apart from the provisions per-
taining to compensation (e.g. royalties).703 Price provisions are subject to
testing under the pricing methods once comparability has been established
for other terms. Significant contractual terms include the form of payment,
701. Treas. Regs. § 1.482-1(d); and OECD TPG, ch. 1, sec. D.

702. On contractual terms as a comparability factor, see also Wittendorff (2010a), at
pp. 419-421.
703. See Pichhadze (2015a); and Pichhadze (2015b) for a discussion of the signifi-
cance of contractual interpretation (US) law for determining the content of the con-
trolled transaction. Pichhadze fears that unless the contractual terms of the actual con-
trolled transaction are interpreted correctly, there will be a risk that comparables will
be collected based on misconstrued contractual terms, with the result that income may
be allocated incorrectly. While Pichhadze’s views are interesting, this author’s impres-
sion is that Pichhadze assigns too much weight to controlled contractual terms. There
are several other comparability factors than such terms that must be taken into account
for profit allocation purposes, and the current transfer pricing methodology ensures that
profit is allocated according to the controlled parties’ contributions of functions, assets
and risks.
201
Allocation Rules
volume of sales,704 scope and terms of warranties, rights to updates and

revisions, duration of a licence agreement, termination and re-negotiation
rights.705
Information on the terms of uncontrolled agreements may, in practice,

be limited, or simply unavailable. That may influence the selection of the
best pricing method. If the controlled transaction, for instance, is a licence
agreement, a lack of information on controlled terms may make it difficult
to apply the CUT method (i.e. it will be unclear as to what to compare). If
the PSM is applied, however, there may be no particular need for such in-
formation (as the residual profits may be allocated based on discretionary
assessments of relative value and not comparables).
6.6.5.2.3. Economic substance and non-recognition
As mentioned, the terms of a written controlled agreement will be given

effect (i.e. will form the basis for applying a transfer pricing methodology)
under US and OECD transfer pricing law only if they are consistent with
the economic substance of the actual behaviour of the controlled parties.706
Thus, if there is a discrepancy between the written terms and the actual
behaviour, the terms that can be deduced from the actual behaviour will
prevail and form the basis for subsequent comparability and transfer pri
cing assessments.707
The 2017 OECD TPG refer to this “reality check” as the “delineation of
the actual transaction”.708 It may, for instance, be that the written controlled
agreement reflects that the parent has licensed IP to its foreign subsidiary
and agreed to provide relevant technical support in return for royalty pay-
ments. The actual conduct of the parties, however, shows that the parent
negotiates with the subsidiary’s clients and assists the subsidiary in per-
forming its obligations towards these clients. The subsidiary is not capable
of fulfilling its contracts without help from the parent, and it does not inde-
pendently develop its own business and or act as a licensee. The controlled
704. The US regulations include two examples on comparability adjustments for dif-
ferences in volumes, undoubtedly motivated by U.S. Steel Corp. v. C.I.R., T.C. Memo.
1977-140 (1977). See Treas. Regs. § 1.482-1(d)(3)(ii)(C), Examples 1 and 2.
705. Treas. Regs. § 1.482-1(d)(3)(ii).
706. Treas. Regs. § 1.482-1(d)(3)(ii)(B)(1); and OECD TPG, para. 1.45. See Bullen
(2010), at pp. 433-441; and Monsenego (2014) on economic substance.
707. On this issue, see, e.g. Navarro (2017), at p. 130; and Odintz et al. (2017), at sec.
2.2.3.1.
202
Comparability
agreement should therefore not be deemed a licence agreement, but rather

a service agreement.709 If there is no written agreement in place between
the controlled parties, all aspects of the controlled transaction must be de-
duced from the actual conduct of the parties.710
In a world apart from delineating the actual controlled transaction lies so-
called “non-recognition”.711 Under the 2017 OECD TPG, the actual deline-
ated transaction may be recast:
[…] where the arrangements made in relation to the transaction, viewed in
their totality, differ from those which would have been adopted by independ-
ent enterprises behaving in a commercially rational manner in comparable
circumstances, thereby preventing determination of a price that would be ac-
ceptable to both of the parties taking into account their respective perspec-
tives and the options realistically available to each of them at the time of
entering into the transaction.712
If these criteria are fulfilled and the controlled transaction is recast, it

should be made to comport as much as possible with the facts of the actual
transaction while achieving a commercially rational result that would have
enabled the parties to arrive at a price acceptable to both of them.713
The OECD non-recognition doctrine is reserved for “exceptional

circumstances”.714 Non-recognition is generally inapplicable if it is possible
to price the transaction715 and will, in any case, be inapplicable where the
controlled pricing is based on a genuine CUT.716 The system of the OECD

710. OECD TPG, para. 1.47. On this issue, see Navarro (2017), at p. 137.
711. On non-recognition under the 2017 OECD TPG, see Navarro (2017). For a com-
prehensive analysis of this subject, based on the pre-2015 BEPS revision language of
the OECD TPG, see also Bullen (2010), pp. 215-219 on the distinction between compa-
rability and structural adjustments. See also Pankiv (2017), at p. 114.
712. OECD TPG, para. 1.122. The “previous generation” text in para. 1.65 of the
2010 OECD TPG stated that if the controlled transaction differed from what would
have been adopted by rational independent enterprises and the actual structure of the
controlled transaction “practically impedes the tax administration from determining
an appropriate transfer price”, tax authorities were entitled to restructure the controlled
transaction. On the concept of options realistically available, see Parekh (2015), in
particular at pp. 297-298, where Parekh finds that the concept has application in three
transfer pricing contexts apart from recharacterization, namely (i) comparability analy-
sis; (ii) pricing of IP transactions; and (iii) pricing of compensation triggered by cross-
border business restructurings.
716. Id.
203
Allocation Rules
TPG is that the controlled transaction, as delineated in accordance with

economic substance, shall be analysed using the rules on intangible owner-
ship and the transfer pricing methods in order to provide an arm’s length
profit allocation. Given the scope, depth and sophistication of these rules,
there is little left for the non-recognition doctrine to deal with.
Comparatively, the US regulations use the realistic alternatives of the con-

trolled parties as a central component in the pricing methods (which is also
the case under the OECD TPG),717 but will not restructure the controlled
agreement as if the alternative had been adopted by the taxpayer.718
The author will elaborate a bit on why the non-recognition doctrine plays
such a modest role in transfer pricing practice. First, as stated by the OECD,
“nonrecognition can be contentious and a source of double taxation”.719
The guidance on intangible ownership and the transfer pricing methods
seek to ensure uniform profit allocations. Even if the current transfer pric-
ing methodology certainly provides some leeway with respect to the profit
allocation patterns that are acceptable, the spectrum of allowed solutions
will normally be reasonably limited (e.g. the arm’s length range lists the
acceptable profit margins that can be used to allocate a normal market
return to the tested party). The hope is that the methodology is applied in
the same manner by both treaty jurisdictions, thereby yielding the same
pricing outcomes and avoiding double taxation.
The non-recognition doctrine triggers a more significant risk of dishar-

monized profit allocations, and thereby double taxation. If the criteria for
applying the doctrine are not sufficiently restrictive, there is a risk that
a controlled transaction may be subject to non-recognition, even if it is
susceptible to arm’s length pricing. This would be detrimental to the inter-
national transfer pricing system, as controlled transactions priced at arm’s
length provide parity in the taxation between related and unrelated parties,
and should be respected. As expressed by the OECD:
[N]onrecognition of a transaction that possesses the commercial rationality
of an arm’s length arrangement is not an appropriate application of the arm’s
length principle. Restructuring of legitimate business transactions would be
a wholly arbitrary exercise the inequity of which could be compounded by
717. See, e.g. the income method in the new cost-sharing regulations, as discussed in
sec. 14.2.8.3.
718. Treas. Regs. § 1.482-1(f)(2)(ii).
719. Id.
204
Comparability
double taxation created where the other tax administration does not share the
same views as to how the transaction should be structured.720
Closely related to this is that the profit allocation effect of non-recognition

is elusive and unpredictable, and may first be determined after the pricing
methods have been applied to price the recast controlled transaction. Given
the amount of freedom that tax administrations have with respect to re-
structuring the controlled transaction (provided that the criteria for apply-
ing the doctrine are fulfilled), it does not seem far-fetched to assume that
there is a risk that the transaction may be recast in a manner that maxi-
mizes profit allocation to the relevant jurisdiction,721 creating an obvious
potential for double taxation.
Second, there is a negative correlation between the relevance of the non-

recognition doctrine, on the one side, and the effectiveness of the rules on
intangible ownership and the transfer pricing methods, on the other. The
more developed the latter rules are, the more redundant a non-recognition
doctrine will be. Non-recognition has traditionally been discussed in the
context of pronouncedly BEPS-motivated transactions. It will normally be
possible to price the actual delineated transaction in most of these cases, as
the controlled parties tend to provide some form of routine or non-routine
value chain contributions that are susceptible to arm’s length pricing.722
The 2014 proposed OECD guidance on non-recognition provides an ex-

ample that illustrates this point.723 It pertains to a subsidiary, resident in
jurisdiction P, which owns a valuable, self-developed trademark crucial
to its business and carries out extensive local marketing. The trademark is
sold to a foreign group entity resident in a low-tax environment for a lump
sum. The foreign entity then licenses rights to the marketing intangible
back to the subsidiary for a fixed royalty rate, combined with a service
agreement pursuant to which the subsidiary is to provide local marketing.

721. The degree of freedom is limited by the requirement that the recast transaction,
once priced, yields a profit allocation outcome that is aligned with the realistic alterna-
tives of the controlled parties. Nevertheless, it will likely, in most cases, be possible to
accommodate a range of different “recast alternatives”, offering considerable leeway
for profit allocation.
722. For instance, the main reason why the IRS argued non-recognition in the Puerto
Rican cases discussed in sec. 5.2.4. was that the domestic rules at the time allowed the
transfer of valuable intangibles without triggering a taxable event (tax continuity), ren-
dering the transfer pricing methods inapplicable. In normal contexts, however, transfer
pricing will be triggered when intangibles are transferred among group entities.
723. OECD, Guidance on Transfer Pricing Aspects of Intangibles, Action 8: 2014
Deliverable (OECD, 2014) [hereafter 2014D, D4], paras. 90-93.
205
Allocation Rules
The transaction is not recognized, on the basis that it does not enhance or
protect the subsidiary’s commercial and financial position.724 The ultimate
logic is that the subsidiary would be better off had it not entered into the
controlled transaction.
The author finds it clear that it was inappropriate to apply non-recognition

in the example. The first question should have been: What is the conse-
quence of not recognizing the actual delineated transaction? The answer
is that the subsidiary would be treated as the owner of the intangible, re-
sulting in allocation of the residual profits from the trademark to jurisdic-
tion P.725 The next question should have been: What is the consequence of
pricing the actual delineated transaction? The subsidiary developed the
marketing intangible, and should therefore be allocated all residual profits
from it.726 There would likely be no existing CUT on which to base the
pricing, probably resulting in an application of the TNMM in the context
of a net present value (NVP) calculation.727
Key in this assessment are the realistic alternatives available to the con-
trolled parties.728 For the subsidiary, this would, of course, be to not sell
the intangible, but to continue its ownership and marketing efforts as be-
fore. The NPV of the operating profits allocable to it under this alternative
forms the lowest acceptable price for the marketing intangible. No rational
economic actor would sell below this amount. However – as is the author’s
point – if the transfer price equals or exceeds this amount, the subsidiary
will either be neutral or better off by selling the intangible.
Thus, the question of whether the actual controlled transaction “enhances

or protects” the subsidiary’s “commercial or financial position” is entirely
dependent on the pricing of the transaction. The example seems to disre-
gard this fundamental point. Also, if the transfer price of the marketing
intangible lies below the NPV of the subsidiary’s best realistic alternative,
this would constitute normal mispricing, which should be corrected in the
proper way, i.e. through transfer pricing, not through non-recognition.
724. 2014D, D4, para. 91.

725. 2014D, D4, para. 92.
726. See the analysis of the OECD intangible ownership guidance on “important
functions” in sec. 22.3.2.
727. See the discussion of the OECD guidance on valuation in ch. 13. This applica-
tion of the TNMM would largely be parallel to an analogical application of the income
method under the US cost-sharing regulations; see sec. 14.2.8.3.
728. See the discussion of the significance of the alternatives realistically available in
the context of valuation in section 13.5.
206
Comparability
Fortunately, this example was scrapped in the final 2017 consensus text
and replaced with a new example, which indicates a significantly high-
er threshold for applying the non-recognition doctrine.729 It pertains to a
group entity that self-develops intangibles through its own R&D and trans-
fers unlimited rights to all intangibles that may arise from its future work
over the next 2 decades for a lump-sum payment to a foreign group entity.
The example deems the controlled transaction to be commercially irration-
al for both parties. Neither company has any reliable means to determine
whether the payment reflects an arm’s length valuation, as it is uncertain
as to what future R&D activities will be carried out, making a valuation of
the potential outcomes entirely speculative.730
In light of the omitted 2014 draft example, there should be no doubt that the
final 2017 example is intended to establish a significantly narrow scope of
application for the doctrine. The point of the latter example is that the con-
trolled transaction is so comprehensively nonsensical (essentially a transfer
of unspecified rights) that an application of the transfer pricing methodol-
ogy would not be meaningful, distinguishing the scenario from normal
cases of valuation uncertainty and mispricing in which non-recognition
clearly is inapplicable. In the author’s view, the steep threshold indicated
in the 2015 example for application of the non-recognition doctrine is both
useful and necessary to preserve the arm’s length transfer pricing system,
while at the same time ensuring that there is an “emergency” mechanism
in place to adjust absurd transactions so that it may be possible to attain an
arm’s length profit allocation in such outlier scenarios as well.731
6.6.5.3. Functions
The US regulations require that the degree of comparability between the

economically significant functions performed by the controlled and un-
controlled taxpayers must be ascertained.732 Such functions include R&D,
729. OECD TPG, at para. 1.128. An additional example is provided, also indicating a
comprehensively high threshold for non-recognition; see OECD TPG, para. 1.126.
730. The example suggests that the controlled transaction should be recast in accord-
ance with the economically relevant characteristics, e.g. as the provision of financing
by company S2 or as the provision of research services by company S1, or if specific
intangibles can be identified, as a licence with contingent payment terms.
731. See also Navarro (2017), at p. 237, where it is concluded that non-recognition
should never not be applied to deal with the assignment of IP ownership, as this issue
shall be dealt with through remuneration of the involved group entities (transfer pri
cing). The author agrees with this.
732. Treas. Regs. § 1.482-1(d)(3)(i). For critical comments on the role of the func-
tional analysis in transfer pricing, see Roberge (2013), at p. 226.
207
Allocation Rules
product design and engineering, manufacturing, production and process

engineering, product fabrication, procurement, marketing and distribution
functions, transportation and warehousing and management services. The
analysis must take into account the resources (e.g. unique intangibles) em-
ployed in connection with the functions performed.
The message of the 2017 OECD TPG is essentially the same.733 The new
2017 text elaborates on “capabilities” and “fragmented functions”.734 It is
stated that the actual contributions and capabilities of a controlled party
will affect its realistic alternatives. Further, it is stated that multinationals
have the ability to fragment functions into highly specialized group enti-
ties that, as a whole, are interdependent of each other, because the func-
tions refer to the same value chains. Sufficient functional comparability is
particularly important under the CPM and the TNMM.735 The guidance on
fragmented functions is related to the revised text on “highly integrated
operations” under the OECD PSM (see section 9.2.3.).736
6.6.5.4. Economic conditions
6.6.5.4.1. Introduction
The US regulations and the OECD TPG require that significant economic
conditions that may affect profits be taken into account when determin-
ing the degree of comparability.737 This includes the similarity between
geographical markets (relative size, extent of economic development and
competition, whether expanding or contracting, etc.), at which level of the
value chain transactions take place (wholesale or retail), product market
shares, location-specific costs and the alternatives realistically available to
the buyer and the seller.738
733. OECD TPG, paras. 1.51-1.55.

734. OECD TPG, paras. 1.52 (capabilities) and 1.55 (fragmented functions).
735. See the analysis of the CPM and TNMM in ch. 8.
736. See ch. 9.
737. Treas. Regs. § 1.482-1(d)(3)(iv); and OECD TPG, paras. 1.110-1.113. Vann
(2003), at p. 167, argues that there is a “clash” between the focus of the (1995 version
of the) OECD TPG on the identification of economic value drivers in the form of func-
tions, assets and risks, on the one side, and the transactional focus of the OECD TPG,
on the other. Vann seems to admit, however, that the introduction of the profit-based
pricing methods into the OECD TPG in reality relaxed the transactional paradigm; see
Vann (2003), at p. 168.
738. The OECD TPG additionally mention the availability of substitute goods and
services, the degree of local consumer purchasing power and the extent of government
regulation; see OECD TPG, para. 1.110.
208
Comparability
The author will revert to some of these factors when analysing the transfer
pricing methods (in chapters 7-16). Here, he will just briefly comment on
the US and OECD positions on the use of comparables from other markets,
location savings and temporary pricing strategies, as these issues are not
necessarily particular to any specific pricing method.
6.6.5.4.2. Use of comparables from other markets
The point of departure under both the US regulations and the OECD TPG
with respect to the use of comparables from other markets is that CUTs
normally should be chosen from the same market in which the tested party
operates, as there may be significant differences in the economic condi-
tions prevailing in different markets.739 Nevertheless, if there are no CUTs
available in the same market, comparables from other markets may be
used, provided that proper comparability adjustments are carried out. Even
if there is not enough information available to carry out such adjustments,
the CUTs may still be used, but unadjusted differences will affect the reli-
ability of the pricing method applied.740
6.6.5.4.3. Location savings
The stance of the US regulations on how incremental operating profits that

are due to location savings shall be allocated among the controlled parties
is that comparability adjustments must be made for significant differences
in costs if the tested party and the unrelated comparable (benchmark) en-
739. Treas. Regs. § 1.482-1(d)(4)(ii)(A); and OECD TPG, para. 1.110. See, e.g. the
Stanley Black & Decker Norway AS v. Skatt Øst ruling by the Oslo City Court (Utv.
2016/1143), where the Court rejected the taxpayer’s assertion that comparables from a
range of European countries should be used under the TNMM to determine the profits
allocable to a Norwegian group distribution entity (the Norwegian tax authorities had
reassessed based on Norwegian comparables). Further, it should be noted that there
may also be a wide spread of profit margins on comparables that are drawn from within
the same market, and such “third-party” comparables may also be group entities (mak-
ing reliance on their profit data problematic); see Rozek et al. (2003).
740. The US regulations contain an example in which a foreign subsidiary manu-
factures products for sale to its US parent; see Treas. Regs. § 1.482-1(d)(4)(ii)(B). In
the absence of CUTs, the IRS considers applying the cost-plus method or the CPM
to remunerate the subsidiary. However, information is not available on uncontrolled
taxpayers performing comparable functions under comparable circumstances in the
same geographical market. Thus, data from US manufacturers, adjusted for differences
between the US and the foreign market, are considered in order to apply the cost-plus
method.
209
Allocation Rules
terprises operate in different geographical markets.741 In other words, the

allocation of profits to the tested party should be based on third-party com-
parables from the same market. For instance, if the total costs of operat-
ing in the controlled manufacturer’s market are less than the total costs
of operating in other markets, that will normally justify higher profits to
the manufacturer only if the cost differences would increase the profits of
comparable uncontrolled manufacturers operating at arm’s length in the
same market.
This position will normally entail that incremental profits due to cost
savings are extracted from local taxation. For instance, imagine that
a US clothing company contracts a foreign subsidiary to manufacture
clothes.742 The cost of sewing in the foreign jurisdiction is significantly
lower than in the United States. Several competitors in geographic mar-
kets similar to that of the subsidiary offer the same services. The US reg-
ulations assume the position that the fact that the production is less costly
in the country of the subsidiary does not in and of itself justify allocating
additional profits to it, as the operating profits of the local benchmark
competitors indicate that arm’s length the profits would not be retained
by the subsidiary.
The OECD has now assumed essentially the same position (see the discus-
sion in chapter 10). Where local comparables are available, these should
be used to benchmark the profits allocable to the tested party.743 Where lo-
cal comparables are unavailable, the bargaining position of the controlled
parties will determine the allocation of the incremental operating profits
from cost savings.744 Where the tested party is a routine input provider, as
normally is the case, this will likely entail that the profits are extracted
from taxation at source.
6.6.5.4.4. Temporary pricing strategies
To increase the market share of a particular product in an existing market

or to enter an entirely new market, unrelated taxpayers may adopt strategies
involving temporarily increased operating costs or reduced sales prices.745
741. Treas. Regs. § 1.482-1(d)(4)(ii)(C). On this, see Andrus (2007), at p. 641; and
Allen (2004).
742. See the example in Treas. Regs. § 1.482-1(d)(4)(ii)(D).
744. OECD TPG, para. 1.143. See also paras. 9.148-9.153.
745. Treas. Regs. § 1.482-1(d)(4)(i); and OECD TPG, para. 1.115. From a business
point of view, market share improvement strategies, in addition to temporary pricing
210
Comparability
The question is to what extent such temporary prices are acceptable also in
controlled transactions.
The US regulations will only give effect to such pricing strategies if four
criteria are satisfied:
(1) it must be shown that an unrelated taxpayer was engaged in a compara-
ble strategy under comparable circumstances for a comparable period
of time;746
(2) the costs incurred for implementing the market strategy must be borne
by the same taxpayer that will obtain the potential future profits from
the strategy, and there must be a reasonable likelihood that the strategy
will result in future profits that reflect an appropriate return relative to
the costs incurred for implementing it;747
(3) the market share strategy must be pursued only for a reasonable period
of time;748 and
(4) the controlled agreement that allocates current costs and future prof-
its must have been established before the strategy was implement-
ed.749
The OECD position seems to rest on the same causality-based approach as

adopted in the US regulations.750 The question is whether there is a plausi-
ble expectation that (i) the business strategy will produce a return sufficient
to justify its costs within a reasonable timeframe; and (ii) an unrelated
party at arm’s length would be willing to sacrifice profits temporarily in
expectation of a greater future reward.
6.6.5.5. Risks
6.6.5.5.1. Introductory comments on risk
The notion of risk is problematic in transfer pricing. Economic logic dic-

tates that there is a positive correlation between risks and return.751 The
connected to penetration strategies, also encompass market maintenance and market

expansion strategies; see Przysuski et al. (2004b), at p. 631.
746. Treas. Regs. § 1.482-1(d)(4)(i).
747. Treas. Regs. § 1.482-1(d)(4)(i)(A).
748. Treas. Regs. § 1.482-1(d)(4)(i)(B).
749. Treas. Regs. § 1.482-1(d)(4)(i)(C).
750. OECD TPG, paras. 1.116-1.117.
751. See, in particular, Schön (2014), at p. 17. See also, e.g. Penelle (2014), at p. 239;
and Hafkenscheid (2017), at p. 19. For a recent comparative overview of the treatment
of risk in different (IFA branch) jurisdictions, see Rocha (2017), at p. 223.
211
Allocation Rules
riskier the investment is, the higher the return a rational economic ac-
tor will require. Applied to transfer pricing, this rationale entails that a
group entity that bears a particular risk should be allocated the profits
associated with it, and the higher the risk, the more profits it should be
allocated.
The problem with this logical axiom is that it may incentivize multina-
tionals to contractually allocate risks to group entities resident in low-tax
jurisdictions, even if the operational business activities that generate the
risks (R&D, manufacturing, etc.) are performed by other group entities
resident in high-tax jurisdictions, thereby facilitating profit shifting.752
Contractual risk allocations (separation of risks from underlying busi-
ness activities) have traditionally been given effect for transfer pricing
purposes, and have therefore become classic taxpayer strategies. The risk
separation practices that have garnered the most attention in recent years
include:753
– contractually stripping local routine value chain input providers (con-
tract manufacturers and low-risk distributors) of as much risk (inven-
tory, raw material, exchange rate, obsolescence, capacity risks, etc.) as
possible in order to allocate a minimum amount of operating profits
(typically under the TNMM) to such entities (and thereby to the local
jurisdictions where they are resident);754 and
– contractually stripping ownership of self-developed IP from the group
entity that carries out R&D and actually creates intangible value (typi-
cally resident in a high-tax jurisdiction) and assigning the IP owner-
ship to a group entity resident in a low-tax jurisdiction (typically an IP
holding or “cash box” entity) in order to facilitate low taxation of the
intangible profits.
752. On the role of contractual risk (stripping) in transfer pricing in general, see, in
particular, the seminal article in Schön (2014). See also Andrus et al. (2017), at p. 90;
Vann (2003), at p. 153; Navarro (2017), at p. 219; Pankiv (2017), at p. 95; Musselli
et al. (2008a); Musselli et al. (2008b); Kane (2006); Vann (2010); Reyneveld et al.
(2012); Durst (2012c); Kofler (2013); Weisbach (2004); Glaize et al. (2011); Osborn et
al. (2017), at sec. I.A-C; Pantelidaki et al. (2013); a seminal economic study in Hines
(1990); Musselli et al. (2007b); Musselli (2006); Heggmair (2017); Gonnet (2016), at
p. 40; and Koomen (2015b), at p. 238. With respect to contractual risk stripping in the
context of contract R&D agreements in particular, see Musselli et al. (2017); and the
analysis in sec. 22.3.3. of this book.
753. These strategies should be seen in the context of the centralized principal model;
see sec. 2.4. On stripping manufacturing entities, see Vann (2003), at p. 153; and Mus-
selli (2008b).
754. E.g. the conversion of local fully-fledged manufacturers or distributors to con-
tract manufacturers and low-risk distributors in connection with business restructur-
ings.
212
Comparability
The fundamental issue is to what extent such controlled contractual

risk allocations should be given effect for transfer pricing purposes.755
The author’s ambition in this chapter is to provide some general com-
ments on key points756 in order to form a backdrop for his later discus-
sions.
6.6.5.5.2. Contractual risk allocation among group entities
Both the US regulations and the OECD TPG require that the profit-relevant
risks in the tested and uncontrolled transactions are comparable before the
uncontrolled transaction can be used under an applicable transfer pricing
method to benchmark the pricing of the controlled transaction.757 Relevant
risks include (i) market risks (fluctuations in cost, demand, pricing and
inventory levels); (ii) R&D risks (success or failure); (iii) financial risks
(fluctuations in foreign currency rates or exchange and interest rates); (iv)
credit and collection risks; (v) product liability risks; and (vi) general busi-
ness risks.
The position taken in both the US regulations and the OECD TPG is that
contractual risk allocation will be respected only if it is consistent with the
economic substance of the actual transaction.758 This is not a comparability
criterion in and of itself, but rather a premise for accepting the controlled
risk allocation for transfer pricing purposes (which thereafter will be com-
pared to the risk allocation found in CUTs). In considering the economic
substance, three main factors are relevant: (i) whether the actual behaviour
of the controlled taxpayer over time is consistent with the contractual allo-
cation of risk;759 (ii) whether the group entity that is contractually assigned
755. Vann (2010) argues for non-recognition of such provisions. On the targeted non-
recognition approach expressed in Schön (2014), see infra n. 773.
756. On risk as a comparability factor, see also, e.g. Wittendorff (2010a), at pp. 407-
411.
757. Treas. Regs. § 1.482-1(d)(iii); and OECD TPG, para. 1.73. The nature and de-
gree of risks incurred may be indicative of the level of profits that should be allocated
to the tested party; e.g. a distributor that is reimbursed for marketing costs will gen-
erally command a lower return than a distributor that must bear its own marketing
costs.
758. Treas. Regs. § 1.482-1(d)(iii)(B); and OECD TPG, para. 1.98. Controlled al-
locations of risk after the outcome of the risk is known or reasonably knowable lack
economic substance. Osborn et al. (2017), under sec. II, argue that the 2017 OECD
TPG on control allow considerably less deference for contracts and thus rely even more
on economic substance considerations to allocate risks, with the result that the US and
OECD provisions are not consistent.
759. Treas. Regs. § 1.482-1(d)(3)(iii)(B)(1); and OECD TPG, paras. 1.86 and 1.88.
213
Allocation Rules
the risk has the financial capacity to bear the losses that will occur if the
risk materializes;760 and (iii) whether the group entity that is contractually
assigned the risk is in control of it.761
The 2017 OECD TPG contain voluminous information on risk,762 which

significantly influences key parts of the new intangibles guidance, in
particular the provisions on intangible ownership (which are analysed in
part 3 of the book). Risk is defined broadly as “the effect of uncertainty
on the objectives of the business”,763 encompassing all risks relevant for
transfer pricing purposes (e.g. strategic, financial and hazard risks). The
centre of attention of the new guidance, however, is on operational busi-
ness risks.764
The core assumption underlying the OECD approach to risks is that third
parties generally will not be willing to assume risks that they do not have
control over.765 For instance, if a supplier determines which products a
distributor should sell and at which volumes and prices, as well as which
marketing campaigns to implement, the supplier directly affects the risks
associated with the distributor’s business. At arm’s length, the distributor
would likely find it difficult to determine the additional return that it would
require from the supplier if it were to assume stock obsolescence risk, as
that risk would be determined by the supplier. Moral hazard considerations
would likely also be relevant, as the supplier could be inclined to maximize
its returns at the expense of the distributor.
760. Treas. Regs. § 1.482-1(d)(3)(iii)(B)(2) and OECD TPG, para. 1.64. See Treas.
Regs. § 1.482-1(d)(3)(iii)(C), Example 1 for an illustration of contractual allocation of
market risk and financial capacity.
761. Treas. Regs. § 1.482-1(d)(3)(iii)(B)(3); and OECD TPG, para. 1.65. The US
regulations contain examples on currency and product liability risk; see Treas. Regs.
§ 1.482-1(d)(3)(iii)(C), Examples 3 and 4. See also Bullen (2010), at pp. 496-506, on
control of risk (with respect to the previous generation [1995/2010] OECD TPG text on
risk).
762. OECD TPG, paras. 1.56-1.106. The treatment of risk has received heightened
attention in the 2017 OECD TPG compared to the previous 1995/2010 generation ver-
sions.
765. On this, see, in particular, Schön (2014), at p. 18. For critical comments on the
2017 OECD “functional” approach to risk allocation, see Hafkenscheid (2017), at p. 20
and p. 23, where he argues that the new language on risk may be subject to diverging
interpretations and thus raises a potential for tax planning and controversy.
214
Comparability
The 2017 OECD guidance defines “control”, largely based on the 2009
business restructuring wording,766 as the “capability to make decisions to
take on … and … whether and how to respond to the risks”.767
In line with this,768 the new guidance distinguishes between “control” and
“risk management”. The latter concept encompasses control, but is broader
in that it also includes the capability to take measures that affect risk out-
comes, i.e. so-called “risk mitigation”.769 The distinction is not clear-cut,
as the capability to decide how to respond to risks (the second element
of the control concept) seems rather similar to risk mitigation. The point
of the distinction seems to be that a group entity may retain control and,
therefore, entitlement to the returns associated with the controlled risk,
even if it outsources risk mitigation to another group entity.770 However, the
entity that outsources the risk mitigation function must, in order to retain
the return associated with the risk it purports to control, have the capacity
to (and actually):
determine the objectives of the outsourced activities, to decide to hire the
provider of the risk mitigation functions, to assess whether the objectives are
being adequately met, and, where necessary, to decide to adapt or terminate
the contract with that provider, together with the performance of such assess-
ment and decision-making.771
766. See OECD TPG, paras. 9.23 and 9.28. On control of risk (under the 2010 OECD
TPG), see Monsenego (2014), at p. 14; Huibregtse et al. (2011); Musselli et al. (2009);
Rosalem (2010); Gonnet (2016), at pp. 41-42; and Picciotto (2014).
767. OECD TPG, para. 1.65. On risk capacity, see, in particular, Schön (2014), at
p. 17. See also Bullen (2010), at pp. 482-492, with respect to the 1995/2010 OECD TPG
text. For an analysis of the content of the control criterion under the 2017 OECD TPG
in the context of determining IP ownership and allocating profits to R&D financing,
see secs. 22.3.2. and 22.4.3., respectively. See also Schön (2014), at pp. 20-22, where
scepticism is expressed with respect to putting (too much) weight on the performance
of control functions for the purpose of assigning risk to group entities for transfer pri
cing purposes. Schön observes that this OECD approach seems to draw on the Key En-
trepreneurial Risk-Taking (KERT)functions doctrine (see the discussion in sec. 17.4.2.
of this book), originally developed for allocating profits to permanent establishments
of financial enterprises under article 7 of the OECD MTC, and states that the approach
is not as well suited to the context of art. 9 of the OECD MTC (where he argues that
contractual risk allocation should be recognized to a larger extent).
768. See OECD TPG, paras. 9.23 and 9.24.
769. Risk management is defined in OECD TPG, para. 1.61.
771. Id.
215
Allocation Rules
The issue of risk and outsourcing is particularly relevant in the context of

determining the ownership of self-developed manufacturing intangibles,772
where the question is whether the entity that purports to be the owner of the
IP (and thus entitled to the residual profits) controls the outsourced R&D
functions of the group entity that actually creates the IP.
It should, in the author’s view, be brought into question as to whether this

sophisticated doctrine is indeed useful. It will, to some extent, respect
controlled transactions that separate risks (and the connected profits)
from the business activities that generate them.773 The doctrine is there-
fore somewhat suspect in the sense that it may potentially be seen as con-
flicting with the stated purpose behind the 2015 BEPS revision of the in-
tangibles guidance that operating profits from unique intangibles should
be allocated to jurisdictions where the intangible value was created.774
With this doctrine, the OECD has attempted to strike a balance between
the need to respect controlled transactions that may be seen, at least to
some degree, as parallels to third-party outsourcing, while at the same time
preventing structures that enable BEPS. The aim of the guidance is likely
to ensure that structures that are supported by some degree of economic
substance are respected, while those that assign risks to “empty” foreign IP
holding companies and perhaps even cash box companies without employ-
ees that have the necessary technical, R&D or business expertise to make
772. The issue is also relevant in other contexts. For instance, asset ownership alone
will not entail that the risk and opportunities associated with the asset should be al-
located to the owner entity; see OECD TPG, para. 1.85.
773. On a more principal (and partly de lege ferenda) basis, see Schön (2014), at p. 29
(following an extensive analysis), where he reaches the conclusion that contractual risk
allocations should, as a general rule, be recognized for transfer pricing purposes. To
avoid profit shifting, however, Schön (i) does not recognize risk allocations in transac-
tions that purely shift risks among group entities (e.g. intra-group captive insurance
schemes and intra-group debt financing); and (ii) finds that IP income should be taxed
on a commensurate-with-income basis (ex post). The author’s impression is that Schön,
on this point, (i) advocates a more severe look-through of contractual risk allocations
than what is the current OECD TPG solution (but at the same time, the OECD TPG
seem to go further in not recognizing contractual risk allocations based on “control
function” considerations than what would be in line with Schön’s reasoning); and (ii)
argues for an ex post solution akin to the one now adopted in the 2017 OECD TPG (see
the discussion in sec. 16.5. of this book), but at the same time, it seems clear that the
new 2017 OECD TPG “important functions doctrine” goes further than what can be
aligned with Schön’s general line of reasoning.
774. See Schön (2014), at p. 20 for comments on the OECD tendency to increase the
threshold for control in transfer pricing. Note also that the issue of outsourcing and risk
seems to be controversial within the OECD, as there apparently are diverging – and
incompatible – viewpoints on the very same issue depending on the context; see the
discussion in sec. 2.5. of this book and supra n. 217.
216
The aggregation of controlled transactions
decisions on operating risks are not. Only experience will show whether
the doctrine is useful.
6.6.5.5.3. Risks affect pricing, not the other way around
A point emphasized in the 2017 OECD guidance is that pricing arrange-

ments are not determinative of risks.775 The idea is that the manner in
which a group entity is compensated in a controlled transaction may indi-
rectly also be a statement as to which risks it is deemed (by the group) to
incur. Such taxpayer assertions on controlled risks cannot, of course, be
given effect unless the criteria in the OECD guidance on risk are satisfied.
The point made by the OECD is interesting, but should, in all fairness, be
rather obvious, as risk should be influential on the pricing outcome and
not the other way around. A controlled manufacturing entity compen-
sated under the cost-plus method, for instance, should not automatically
be deemed to be shielded from risks pertaining to raw material or other
input price fluctuations or to the utilization of its manufacturing capacity
(even if the pricing could indicate that it does not incur such risks).
6.7. The aggregation of controlled transactions
6.7.1. Introduction
The point of departure under both the US and OECD regimes is that a
specific controlled transaction should be tested against a specific CUT.
Nevertheless, both regimes allow aggregation of both controlled and un-
controlled transactions.
Aggregation of uncontrolled transactions pertains to the issue of using

(under the selected transfer pricing method) profit data from a “pool” of
uncontrolled transactions with varying degrees of comparability to the
controlled transaction. This issue is highly relevant in practice, as more
comparable, disaggregated, transaction-level third party profit data often
will be unavailable. Whether uncontrolled transactions can be aggregated
is a comparability issue and should be addressed in connection with the
analysis of the particular transfer pricing method that triggers the issue.
The author refers to his discussion in section 8.6., where he provides an ex-
tensive analysis of whether aggregated third-party profit data can be used
775. OECD TPG, para. 1.81. On this issue, see Barbera (2003), at p. 71.
217
Allocation Rules
pursuant to the current OECD TPG when applying the TNMM, under
which, this comparability issue is particularly pronounced.
The topic in sections 6.7.2.-6.7.4. will be the aggregation of controlled

transactions. This issue pertains to whether it is appropriate to treat several
different controlled transactions as one for pricing purposes (i.e. deeming
the transactions as a whole and then applying a transfer pricing method to
this whole). Whether such treatment should be applied will depend pri-
marily on which transfer pricing method is suitable for the case and which
third-party comparables are available to benchmark the controlled transac-
tion (profit comparison should refer to the same level of aggregation on the
side of both the controlled and uncontrolled transactions).
The author discusses the US and OECD positions on the aggregation of

controlled transactions in sections 6.7.2. and 6.7.3., respectively. The ag-
gregation of controlled transactions under the OECD TPG was an issue in
the illustrative 2012 Canadian Supreme Court ruling in GlaxoSmithKline
Inc. v. R,776 the ruling of which is discussed in section 6.7.4.
6.7.2. The US regulations
The final 1994 US regulations introduced the rule that the combined ef-
fect of two or more separate controlled transactions could be considered in
determining an arm’s length profit allocation if this would yield the most
reliable result.777 This essentially entailed a requirement to take into ac-
count the entire value chain for the purpose of determining transfer prices.
This approach was necessary in order to ensure proper application of the
profit-based CPM and PSM (introduced in the same regulations).778 For
instance, a CPM allocation will require an aggregated approach, where
the royalty allocable to an IP-owning parent is determined as the residual
profits after a normal market return is allocated to a subsidiary as the tested
party. Thus, the income allocable to the parent is determined indirectly (by
way of testing the transactions of the subsidiary) and not directly through
benchmarking the income of the parent against third-party royalty rates
from CUTs. This, in reality, aggregates the transactions of the tested party
776. 2012 SCC 52 (2012), which affirmed 2010 CAF 201 (F.C.A., 2010), which re-
versed 2008 TCC 324 (T.C.C., 2008).
777. Treas. Regs. § 1.482-1(f)(2)(i). That would normally be the case for related prod-
ucts or services; see Treas. Regs. § 1.6038A–3(c)(7)(vii).
778. The US regulations contain four examples illustrating the 1994 guidance; see
Treas. Regs. § 1.482-1T(a)(i)(E), Examples 1-4.
218
for benchmarking purposes, as it is allocated a net profit as total remunera-

tion for its total contribution of functions, assets and risks.
The US aggregation provisions were significantly expanded in September

2015 to ensure that all intangible value transferred in controlled transactions
is captured for taxation, regardless of how the value transfer has been legally
structured.779 For instance, if a range of different IP is transferred together as a
whole and the value of the package is greater than the sum of the stand-alone
IP values, it will not be appropriate to price the transfer on a transaction-by-
transaction level, as that would not capture all of the value transferred.
The 2015 amendment was driven by two main factors. First, the profit-
based methodology (the CPM and PSM) during the last 2 decades has been
further developed and adapted to the context of valuation, in particular in
the form of the income method and residual profit split model for valu-
ing buy-in transactions in CSAs.780 These methods apply aggregated dis-
counted cash flow (DCF) valuation methodology aimed at also capturing
residual intangible value (e.g. goodwill, going concern value and workforce
in place).781 This methodological development has been combined with an
increased focus on economic substance and value chain analysis. The 2015
aggregated approach to valuation applies generally, also outside the context
of CSA transactions (for which the income method was originally tailored).
Second, experience has amply demonstrated a need for coordinated and ag-
gregated analysis of controlled transactions that are speculatively split into
single transactions by multinationals for treatment under different Internal
Revenue Code (IRC) provisions (typically those of sections 367 and 482)
in an attempt to migrate US-developed intangibles without triggering a tax
charge, or at least a charge less than that what would have been levied had
the complete transaction been assessed as a whole under a single IRC provi-
sion (typically, of section 482 alone).782 The author refers in particular to his
779. T.D. 9738; see Treas. Regs. § 1.482-1T(a)(i)(B), at (A) and (C). See Odintz et al.
(2017), at sec. 2.2.5.2, on this. See also the comments in this book on these new provi-
sions in the context of the relationship between IRC secs. 367 and 482 in sec. 3.5.8.; the
best-method rule in sec. 6.4; and intra-group R&D services (contract R&D arrange-
ments) in sec. 21.4.
780. See the analysis of these methods in secs. 14.2.8.3. and 14.2.8.6., respectively.
781. This is illustrated by the IRS allocation assertion in Veritas Software Corpora-
tion & Subsidiaries v. CIR (133 T.C. No. 14 [U.S. Tax Ct. 2009], IRS nonacquiescence
in AOD-2010-05; see the analysis of the ruling in sec. 14.2.4.), which was based on the
predecessor of the specified 2009 income method.
782. See Treas. Regs. § 1.482-1T(a)(i)(E), Example 9, where the taxpayer transfers
manufacturing intangibles in a sec. 367 transaction, and then subsequently enters into
219
Allocation Rules
discussion in section 3.5.4. While the 2015 provisions on aggregation are

geared towards the relationship between section 367 and 482 transactions,783
they also deal with intangible value transferred in connection with the ren-
dering of intra-group services, targeting contract R&D arrangements.784
The IRS’s aggregation approach was incorporated into IRC section 482
in the 2017 US tax reform, clarifying that the service may, going forward,
base its transfer pricing assessments on aggregated valuations and on the
realistic alternatives principle.785
6.7.3. The OECD TPG
The language contained in the OECD TPG on the aggregation of con-

trolled transactions dates back to the 1995 revision. It permits aggregation
in three scenarios:786
(1) where separate transactions are so closely linked or continuous that
they cannot be evaluated on a separate basis (e.g. licence agreements,
typically priced indirectly through the TNMM).787 The US roundtrip
case law discussed above,788 as well as the Canadian GlaxoSmithKline
Inc. v. R. ruling discussed below,789 are examples of structures involv-
ing closely linked transactions;
a CSA transaction in which marketing intangibles connected to the same product are
contributed as a buy-in. The example finds that the value of the intangibles is greater
in the aggregate due to synergies among the intangibles than if valued as two separate
transactions. It therefore requires that the synergies are taken into account in determin-
ing the arm’s length results for the transactions. See also Treas. Regs. § 1.482-1T(a)(i)
(E), Example 5, on the aggregation of a transfer of ten interrelated patents.
783. For a further analysis of the relationship between secs. 367 and 482 in the context
of CSAs, see the discussion in sec. 3.5.8.
784. See Treas. Regs. § 1.482-1T(a)(i)(E), Example 10; and the discussion in sec. 21.4.
of this book.
785. See sec. 1.7.
786. OECD TPG, paras. 3.9-3.12.
787. Further, as an example of transactions that should be viewed together, the OECD
TPG mention the licensing of manufacturing know-how and the supply of vital compo-
nents to an associated manufacturer. The position of the OECD TPG is that the licens-
ing transaction and the sales transaction “should be evaluated together using the most
appropriate arm’s length method”. Another example is where a transaction is routed
through another group company. The OECD TPG find it more appropriate to consider
the routed transaction chain in its entirety rather than to consider the individual trans-
actions on a separate basis.
220
(2) where it is appropriate to use a portfolio approach. This may, for in-
stance, be where a product is marketed with a low profit or at a loss in
order to create a demand for other products or services;790 and
(3) for so-called “package deals”, e.g. where patent licences, technical and
administrative services and the lease of production facilities are con-
tracted as one integrated transaction.791
The 1995 OECD wording on aggregation of controlled transactions must

now be regarded as being relatively outdated when compared to the updated
2017 guidance contained elsewhere in the OECD TPG (the wording on in-
tangibles, risk, CSAs, etc.) and when taking into account the shift in OECD
preferences with respect to the choice of pricing methodology (from the
CUT method to the profit-based methods, in particular, the PSM). In order
to ensure that the new 2017 guidance can be applied as intended, it is neces-
sary to interpret the older 1995 wording so that it can be harmonized with
the pronouncedly substance-based content of the 2017 wording (on intangi-
bles, in particular). It is the author’s view that the 1995 wording does not re-
strict the aggregation of controlled transactions in any cases in which, after
a thorough functional analysis, it is clear that (i) the economic substance of
the controlled transactions seen together would be better reflected (and thus
be more susceptible to arm’s length pricing) if the transactions were treated
as one; (ii) a transfer pricing methodology can be reliably applied to bench-
mark the aggregated controlled transaction; and (iii) there is third-party
benchmarking profit data available for the selected transfer pricing method.
6.7.4.
GlaxoSmithKline (Canada)
In October 2012, the Supreme Court of Canada announced its first ever
ruling in a transfer pricing case, GlaxoSmithKline Inc. v. R.792 The case
790. OECD TPG, para. 3.10. A portfolio approach is a business strategy in which the
taxpayer bundles certain transactions for the purpose of earning an appropriate return
across the portfolio rather than necessarily on specific single products within the port-
folio.
792. GlaxoSmithKline Inc. v. R., 2008 TCC 324 (T.C.C., 2008), reversed by 2010
CAF 201 (F.C.A., 2010), application/notice of appeal filed in 2010 CarswellNat 4089
(S.C.C., 2010), and leave to appeal allowed by 2011 CarswellNat 682 (S.C.C., 2011),
affirmed by 2012 SCC 52 (S.C.C., 2012). For a discussion of the ruling, see also, e.g.
Schön et al. (2011), at pp. 221-224; Schön (2011b), at p. 64; Pichhadze (2013); Vidal
(2009); Dujsic et al. (2008); and Pichhadze (2015), at sec. 3.1.2.
221
Allocation Rules
shows that a multinational may ensure a specific profit allocation through

a combination of different types of controlled transactions and that it may
be entirely necessary to aggregate them in order to ensure an arm’s length
profit allocation under the OECD TPG.793 Because the case provides a
practical illustration of a typical context in which aggregation of controlled
transactions will be useful under the OECD TPG, as well as of how such
aggregation should be achieved, the author will devote some space to ana-
lysing the case.
6.7.4.2. The factual pattern
GlaxoSmithKline is a UK-headquartered multinational operating in the

pharmaceutical industry.794 The relevant parts of the group structure for
the period under analysis are illustrated in figure 6.1.795
Figure 6.1
Glaxo Holdings PLC
GGL
Glaxochem
Ltd.
Montrose Glaxo
& Far East (Pte) Ltd
Bernard
Castle VIC
Glaxo
Operations
UK Ltd. 1991 Glaxo
Pharmaceuticals
Adechsa Glaxo (Pte) Limited
Glaxo (Switzerland) Canada Glaxo
(Ranitidine
Export Far East 50%
factory)
(Pte) Ltd
GGR UK
1982 Glaxochem Shin
Glaxo 1991 (Pte) Ltd. Nihon 45%
Pharmaceuticals Jitsugoyo
Co Ltd.
UK Ltd.
Glaxo
Kabushiki
Kaishai
793. On this topic, see also Wittendorff (2013).

794. The Glaxo group is among the ten largest pharmaceutical groups in the world.
The case concerns the income years 1990-1993.
795. The figure was produced by the Tax Court of Canada and is appendix III to the
Tax Court ruling in this case (2008 TCC 324). In sec. 6.7.4.3., the author refers to the
Tax Court case as the TC ruling. The TC ruling was written by Judge Gerald J. Rip,
Chief Justice of the Canadian Tax Court.
222
During the income periods at trial, the Glaxo group performed secondary
manufacturing and sales of the ulcer drug Zantac in Canada through the
company GlaxoSmithKline Inc. (Glaxo Canada). Pharmaceutical products
are manufactured in two basic stages, normally referred to as primary and
secondary manufacturing. Primary manufacturing is the making of the ac-
tive pharmaceutical ingredient for a pharmaceutical product. Secondary, or
pharmaceutical, manufacturing refers to the process of placing the active
ingredient into a delivery mechanism (e.g. a tablet, liquid or gel), as well
as packaging. Glaxo Canada was a subsidiary of the UK company Glaxo
Group Ltd. (GGL). GGL owned intangibles required for the manufacture
and sale of Zantac. The essential intangibles in this context were the patent
and trademark required to manufacture and sell Zantac. GGL was a whol-
ly-owned subsidiary of the ultimate parent company of the group Glaxo
Holdings Plc (Glaxo Holdings), also a UK company.
The Zantac value chain was organized as follows. The primary manufac-
turing of the active ingredient used for Zantac, ranitidine, sold in Canada,
was performed by a group entity in Singapore, Glaxo Pharmaceuticals
(Pte) Limited (Singapore Sub). Ranitidine was sold from Singapore Sub to
a group clearing company in Switzerland, Adechsa S.A. (Swiss Sub), and
then resold to Glaxo Canada through a supply agreement between Swiss
Sub and Glaxo Canada. Transfer prices for all stages of the value chain
were determined by the ultimate parent company, Glaxo Holdings. The
necessary intangibles to produce and sell Zantac in Canada were licensed
from GGL to Glaxo Canada through a licence agreement. The strategy
of the Glaxo Group for minimizing its worldwide taxes was to recognize
as much profit as possible in Singapore, then to recognize as much of the
remaining profit as possible in the United Kingdom and to ensure that the
group did not pay tax on the same income twice.796
The price of Zantac in Canada was determined by Glaxo Holdings, based

on an assessment of the quality of the product compared to the best com-
peting product at the time of introduction of Zantac.797 Glaxo Holdings
decided that local distributors for Zantac, such as Glaxo Canada, should
retain a gross margin of 60%,798 and the remaining 40% should be remit-
ted back to Glaxo Group in the United Kingdom in the form of royalties or
transfer prices.799 The transfer price for the active ingredient throughout the
whole value chain was designed in order to realize the 60% Canadian gross
796. 2008 TCC 324, para. 13.

797. Id., para. 45.
798. The term “gross margin” is net sales minus COGS.
799. 2008 TCC 324, para. 47.
223
Allocation Rules
profit retention ratio. An important factor is that Canada levied withhold-

ing tax on both dividend and royalty payments from Canada to the United
Kingdom. The domestic rate was reduced to 10% under the Canada-United
Kingdom treaty.800
The Singapore Sub did not pay any tax in Singapore on the profits it made
from producing and selling ranitidine to the Swiss Sub.801 When the profits
of the Singapore Sub were brought into the United Kingdom through divi-
dend distributions, UK tax was only payable on any excess of the UK tax
rate over the Singapore tax rate.802 The transfer price for the active ingredi-
ent when sold from Singapore Sub to Swiss Sub was therefore set high in
order to realize as much of the profits from the value chain as possible in
Singapore. Singapore Sub earned gross profits of approximately 90% in
Singapore during the income years covered by the case.
In Switzerland, there was an agreement between the Swiss tax authorities

and the Swiss Sub concerning its income taxation. It was agreed that taxes
were to be paid on the basis that the company earned a profit of 4% on the
resale of the active ingredient to Glaxo Canada. If the profits from the sales
exceeded 4%, taxation would be based on actual profits.
Under the licence agreement between Glaxo Canada and GGL, Glaxo Can-
ada paid a royalty of 6% of its net sales of Zantac as consideration for the
provision by GGL of the make-sell rights to Zantac, as well as the right to
use trademarks, receive technical support, use registration materials, ac-
cess new products and improvements, etc. A highly significant point for the
case is that, by virtue of the licence agreement, Glaxo Canada was required
to purchase the active ingredient from a Glaxo group entity and adhere to
Glaxo standards.803
Under the supply agreement, Glaxo Canada purchased ranitidine from

Swiss Sub for generally five times the price that generic ranitidine was
sold for between unrelated parties in Canada during the period at issue.
It was clear that Glaxo ranitidine and generic ranitidine were chemically
800. See Canada-United Kingdom treaty, art. 10, nr. 1 a) for dividends; and art. 12,
nr. 2 for royalties.
801. 2008 TCC 324, para. 47. This was due to a 10-year relief tax holiday that began
in 1982. After this period, the tax rate was 10%. Under the relief programme, the Glaxo
group benefitted from tax sparing between Singapore and the United Kingdom.
802. Apparently, Singapore income was deemed by the UK tax authorities to have
been taxed at the full Singapore tax rate.
803. See TC ruling, para. 86; and Supreme Court ruling (SC ruling), para. 46.
224
equivalent and bioequivalent. Thus, there were no substantial differences

between generic and Glaxo ranitidine.804 In the years covered by the case,
Glaxo Canada made a gross profit on sales on Zantac of approximately
57% – In other words, very close to the targeted 60% gross profit margin
set by Glaxo Holdings.
The Canadian tax authorities increased the income of Glaxo Canada for
the income years at issue on the basis that Glaxo Canada overpaid the
Swiss Sub for the purchase of ranitidine. The increased amount was treated
as deemed dividend distributions to its UK parent company, GGL. The
increased amount was therefore also subject to Canadian withholding tax.
6.7.4.3. The 2008 Tax Court ruling
The question before the Tax Court (TC) was whether the transfer price for
ranitidine between the Swiss Sub and Glaxo Canada was set in accordance
with the domestic transfer pricing provision of the Canadian Income Tax
Act.805 The TC found that the Canadian tax authorities relied on the OECD
Guidelines when assessing transfer pricing matters under domestic law.
The Federal Court of Appeal had also stated that the OECD Commentar-
ies should provide information on the interpretation and application of the
domestic transfer pricing provisions.806
The tax authorities argued before the TC that the arm’s length price for
ranitidine under the domestic arm’s length provision should, as in the reas-
sessment, be set on the basis of the price agreed between the generic com-
panies for the purchase of ranitidine from unrelated suppliers, based on
the CUP method. The unrelated Canadian companies, Apotex and Novop-
harm, had paid to their third-party suppliers of ranitidine a price only one
fifth of the price that Glaxo Canada had paid to the Swiss Sub during the
804. See TC ruling, para. 96.

805. The 1985 version of sec. 69(2) of the Canadian Income Tax Act applicable for
the income years of the case had the following wording: “Where a taxpayer has paid or
agreed to pay to a non-resident person with whom the taxpayer was not dealing at arm’s
length as price, rental, royalty or other payment for or for the use or reproduction of any
property, or as consideration for the carriage of goods or passengers or for other ser-
vices, an amount greater than the amount (in this subsection referred to as ‘the reason-
able amount’) that would have been reasonable in the circumstances if the non-resident
person and the taxpayer had been dealing at arm’s length, the reasonable amount shall,
for the purpose of computing the taxpayer’s income under this Part, be deemed to have
been the amount that was paid or is payable therefor.”
806. TC ruling, para. 59.
225
Allocation Rules
same period. The tax authorities supported their assertion with an analysis
under the cost-plus method.807
Glaxo Canada asserted that the generic ranitidine transactions did not rep-
resent appropriate CUPs for two reasons. First, the actual business circum-
stances of Glaxo Canada were wholly different from those of the unrelated
generic ranitidine purchasing companies, Apotex and Novopharm. The
thrust of this argument was that the pricing under the supply agreement
had to be seen in the context of the licence agreement. Second, the raniti-
dine purchased from the Swiss Sub was not comparable to the ranitidine
purchased by the generic ranitidine purchasing companies Apotex and
Novopharm from their respective unrelated suppliers because Glaxo ran-
itidine was produced under Glaxo good manufacturing practices.
The first question to be assessed by the TC was whether the transfer pricing
of ranitidine under the supply agreement should take the licence agree-
ment into consideration, i.e. whether the controlled transactions should be
aggregated. The tax authorities relied on a domestic Supreme Court case
(Singleton v. R.) for the proposition that the assessment should look at the
transaction at issue and not at the surrounding circumstances.808 Glaxo re-
lied on two other cases in which the courts had found some contracts ancil-
lary and incidental to a genesis contract.809
The TC found that the licence agreement and the supply agreement could
stand alone and that neither was ancillary to the other. Glaxo argued that
the point of the transfer pricing structure was that Glaxo Canada should
retain 60% of the profits, while 40% should be remitted to foreign Glaxo
companies. It further argued that it made no difference whether the 40%
were remitted through transfer pricing, royalties or a combination of the
two.810 This is a key point of the case. The TC disagreed with the view of
Glaxo Canada and stated that the purpose of the case was to find a proper
transfer price for the ranitidine in order to determine the tax liability of
Glaxo Canada.
807. The cost-plus assessment used the cost of manufacturing the ranitidine and add-
ed a suitable profit margin.
808. The 2001 Singleton case, Singleton v. R., 1996 CarswellNat 1816 (T.C.C., 1996),
reversed by 1999 CarswellNat 1009 (Fed. C.A., 1999), leave to appeal allowed by 2000
CarswellNat 653 (S.C.C., 2000) and affirmed by 2001 SCC 61 (S.C.C., 2001).
809. R. v. Koffler Stores Ltd., 1975 CarswellNat 336 (Fed. T.D., 1975), affirmed
by 1976 CarswellNat 200 (Fed. C.A., 1976); and GSW Appliances Ltd. v. R, 1985
CarswellNat 346 (T.C.C., 1985).
226
Royalty payments from Canada were subject to withholding tax. Royalty

payments would go to the intangible-owning UK company, GGL, and be
taxed there. Conversely, the purchase price paid for ranitidine by Glaxo
Canada to the Swiss Sub would not be subject to withholding tax in Can-
ada, and the profits would be taxable in Switzerland, then go to Singapore
and ultimately back to the United Kingdom through dividend payments
from Singapore. The TC was of the view that it could very well be that a
40% remittance of the total profits of Glaxo Canada was reasonable, but
this was not the question in the case.
The question was whether the ranitidine purchase price paid by Glaxo
Canada satisfied the arm’s length requirement. In the view of the TC, the
transfer pricing of Glaxo Canada combined two transactions, namely the
purchase of ranitidine under the supply agreement and the payment for
intangibles under the licence agreement, and thereby ignored the distinct
tax treatment that followed from each.811 The TC supported its view with
the domestic Singleton case,812 as well as the US Tax Court case Bausch &
Lomb, Inc. v. Commissioner.813
Glaxo Canada further argued that the circumstances in which the unrelat-
ed generic ranitidine companies Apotex and Novopharm purchased ran-
itidine were not comparable to the circumstances in which Glaxo Canada
bought ranitidine. Thus, in the view of Glaxo Canada, the unrelated ran-
itidine transactions could not be used as comparable transactions under
the CUP method. Glaxo Canada listed a range of factors, with reference
to the 1995 OECD Guidelines, that supposedly supported its view.814 A
key argument from Glaxo Canada was that it was required to purchase
the ranitidine from the Swiss Sub, pursuant to its licence agreement with
GGL.
The TC found it clear that Glaxo Canada was contractually obligated to

purchase ranitidine from the Swiss Sub, but the question was whether a
person in Canada, dealing at arm’s length, would have accepted the condi-
tions and paid the price that Glaxo Canada did. The evidence before the
TC demonstrated that it was the marketing efforts of Glaxo Canada and
the value of the brand name Zantac that enabled Glaxo Canada to charge a
premium price for Zantac in the Canadian market. There was no evidence
811. TC ruling, para 78.

812. See supra n. 808.
1084 (2nd Cir., 1991). See the discussion of the ruling in sec. 5.2.5.2. of this book.
227
Allocation Rules
that the price or value of ranitidine had any causal relationship with the ex-
ternal market price of Zantac. Further, Glaxo Canada argued that the price
of ranitidine should be high due to the regulatory approval and marketing
assistance it received from GGL and the use of the trademark Zantac to
sell the drug in Canada. The TC found this irrelevant, as the provision of
intangibles should be priced under the licence agreement with GGL and
not under the supply agreement with the Swiss Sub.815
The second main argument from Glaxo Canada was that the ranitidine
transactions between unrelated generic companies and their suppliers were
not comparable to the ranitidine purchases by Glaxo Canada from Swiss
Sub, because the latter ranitidine was produced under the Glaxo groups
good manufacturing practice. After an extensive evidentiary assessment,
the TC rejected the assertion by Glaxo Canada. The TC found that the
Glaxo manufacturing practice did not change the nature of the good. Glaxo
Canada also admitted that the generic ranitidine was chemically and bio-
logically equivalent, as required by the Canadian health authorities. The
TC found that the Glaxo manufacturing practice could have some minor
positive price effect but that it did not affect the comparability with the ra-
nitidine used by the unrelated generic companies. After a thorough transfer
pricing assessment, the TC found the CUP method to be applicable.
Glaxo Canada relied on a specific application of the resale price method.

The resale price method compares gross margins, i.e. sales revenues minus
the COGS.816 As stated by the TC, the resale price method is most reli-
able when it measures the return of only one function, it is isolated to a
particular product and there is a high degree of similarity in functions and
risks between the companies being compared. Because costs not directly
related to goods sold vary extensively between companies – perhaps espe-
cially between companies in different jurisdictions – what will be an arm’s
length gross margin in one country will not necessarily be the same for
another company in another jurisdiction. For example, a company resident
in a country with low marketing expenses will be able to accept a lower
gross margin than a company resident in a jurisdiction that requires large
marketing expenses. Glaxo Canada argued that its gross profit margins
were largely in line with those earned by European distributors. The TC
rejected this argument on the grounds that Glaxo Canada performed many
more functions and assumed more risks than the European licensees. In the

816. COGS excludes all other costs than direct product costs. See the discussion in
sec. 6.2.3.2.
228
eyes of the TC, this fact should have justified a higher gross profit margin
to Glaxo Canada through a lower transfer price for ranitidine.
Further, Glaxo Canada supported its resale price method-based transfer

price for ranitidine with a TNMM analysis, which compared net profits
between companies.817 The TC rejected the analysis by Glaxo Canada on
the basis that the selected companies for comparison were not reasonable
and that there was insufficient evidence of other functions undertaken by
the comparables. The Canadian tax authorities, however, supported their
CUP assessment with the cost-plus method. The cost-plus assessment used
as its point of departure the manufacturing costs for ranitidine, and added
a suitable profit. During the income years at trial, the Singapore Sub had
cost-plus margins between 766% and 1,059% of manufacturing costs. If
the profit margin of the Singapore Sub alternatively had been 25% (as the
Canadian tax authorities viewed as arm’s length) and the profit margin of
the Swiss Sub was 4%, as agreed with the Swiss tax authorities, then the
transfer price for ranitidine to Glaxo Canada would have been approxi-
mately the amount assessed under the CUP method.
The TC concluded that the CUP method was the preferred method for
setting the transfer price for ranitidine and that the generic companies in
Canada were appropriate comparables. Glaxo Canada was therefore, in
computing its taxable income, not allowed to deduct the excess price it
had paid for ranitidine to the Swiss Sub. On this point, the reassessment of
the Canadian tax authorities was therefore upheld (apart from a marginal
adjustment for the value added to the ranitidine through the Glaxo manu-
facturing standard).818
The second question before the TC was whether the treatment by the Ca-
nadian tax authorities of the excess transfer price paid for the ranitidine
should be regarded as deemed dividend payments from Glaxo Canada to
GGL, pursuant to a domestic tax law provision on the allocation of in-
come.819 In the McClurg v. R. case, the purpose of this domestic alloca-
817. See 1995 OECD TPG, para. 2.58.

818. The Tax Court’s CUP analysis was criticized by commentators. See, e.g. Hill
(2008), at p. 269; Rhee (2008), at p. 671; Vincent (2008), at p. 177; Dujsic et al. (2008),
at p. 203; and McCrodan et al. (2009), at sec. 23:1-21.
819. Subsec. 56(2) of the Canadian Income Tax Act, which then contained the follow-
ing wording: “A payment or transfer of property made pursuant to the direction of, or
with the concurrence of, a taxpayer to some other person for the benefit of the taxpayer
or as a benefit that the taxpayer desired to have conferred on the other person (other
than by an assignment of any portion of a retirement pension pursuant to section 65.1
of the Canada Pension Plan or a comparable provision of a provincial pension plan as
229
Allocation Rules
tion provision was characterized as ensuring that payments that other-

wise would have been received by the taxpayer would not be diverted to
third parties as an anti-avoidance technique.820 If the excess amount were
deemed to be a dividend payment to GGL, the amount would be subject to
withholding tax. The TC found that the excess amounts that were paid to
the Swiss Sub were not payments for ranitidine. Further, the Swiss Sub did
not provide any other goods or services to Glaxo Canada, and was there-
fore not entitled to the excess payments.821
The TC viewed GGL as being entitled to the excess payments. Had the
excess amounts not been paid to the Swiss Sub, they would have, at some
point, been distributed to GGL in the form of dividend payments. This fun-
nelling of excess profits to the Swiss Sub was caused by the setting of the
transfer price by Glaxo Holdings and by the concurrence of GGL that the
Swiss Sub could receive funds that GGL was entitled to. The TC therefore
regarded the excess amount as received by GGL as deemed dividend pay-
ments. The reassessment concerning withholding tax on deemed dividend
distributions from Glaxo Canada to GGL was therefore upheld by the TC.822
6.7.4.4. The 2012 Supreme Court ruling
The TC ruling was appealed to the Federal Court of Appeal, which found
that the licence agreement was a circumstance that had to be taken into
account when determining whether the prices paid by Glaxo Canada for
ranitidine were reasonable. This ruling was appealed to the Supreme Court
of Canada (SCC). The issue on appeal before the SCC was the correct in-
terpretation of section 69(2) of the Canadian Income Tax Act, in particular
with regard to which circumstances were to be taken into account when
determining the reasonable arm’s length price.
Similarly to the TC, the SCC found that the OECD TPG were not control-
ling as if they were a Canadian statute and that the test of transactions or
defined in section 3 of that Act or of a prescribed provincial pension plan) shall be in-
cluded in computing the taxpayer’s income to the extent that it would be if the payment
or transfer had been made to the taxpayer.”
820. McClurg v. R., 1984 CarswellNat 369 (T.C.C., 1984), reversed by 1986 Carswell-
Nat 244 (Fed. T.D., 1986), affirmed by 1987 CarswellNat 556 (Fed. C.A., 1987), leave
to appeal allowed by 1988WL876606 (S.C.C., 1988) and affirmed by 1990 CarswellNat
520 (S.C.C., 1990). See the latter ruling, at pp. 1052-1053.
822. Apart from the marginal adjustment for value from the Glaxo manufacturing
process.
230
prices ultimately had to be determined according to the domestic transfer

pricing provision rather than any particular methodology or commentary
set out in the OECD TPG. However, as the domestic transfer pricing provi-
sion did not provide any guidance on concrete transfer pricing methodol-
ogy, the SCC relied on the methods set out in the OECD TPG.823
The SCC interpreted the TC ruling to say that the Singleton case precluded
the TC from considering the licence agreement.824 In its summary of the
TC ruling, the SCC observed that the Canadian tax authorities argued that
earlier SCC cases of Singleton825 and Shell Canada Ltd. v. R.,826 as well as
the OECD TPG, required a transaction-by-transaction approach for the de-
termination of an arm’s length transfer price. The SCC did not agree with
this view. The Court found that both Singleton and Shell were inapplicable
for determination under section 69(2).827
The Court then went on to comment on the arguments from the tax authori-
ties concerning the OECD TPG. The 1995 version of the OECD TPG, rele-
vant for the case at trial, stated that as a point of departure, the arm’s length
principle should be applied on a transaction-by-transaction basis,828 but that
in some situations, where separate transactions are so closely linked or
continuous that they cannot be evaluated adequately on a separate basis,
the transaction-by-transaction approach should give way for a holistic ap-
proach to setting the transfer price.829
The Court found that the OECD TPG required the economically relevant
characteristics between the controlled and uncontrolled transactions to be
sufficiently comparable.830 Only when there are no related transactions or
when related transactions are not relevant to the determination of the rea-
sonableness of the transfer price in issue did the Court find that a transac-
tion-by-transaction approach would be useful.
823. SC ruling, para. 21.

825. See sec. 6.7.4.3.
826. Shell Canada Ltd. v. R., 1997 CarswellNat 401 (T.C.C., 1997), reversed by
1998 CarswellNat 170 (Fed. C.A., 1998), leave to appeal allowed by 1998WL1727462
(S.C.C., 1998), leave to appeal allowed by 1999WL33183768 (S.C.C., 1999) and re-
versed by 1999 CarswellNat 951 (S.C.C., 1999), additional reasons in 1999 Carswell-
Nat 1808 (S.C.C., 1999).
828. See 1995 OECD TPG, para. 1.42.
829. Id.
830. See SC ruling, para. 42; and 1995 OECD TPG, para. 1.15.
231
Allocation Rules
The Court found the rights under the licence agreement between Glaxo
Canada and GGL to be contingent on Glaxo Canada entering into the sup-
ply agreement with the Swiss Sub and paying the ranitidine transfer price
under the agreement as set by Glaxo Group. Further, the Court concluded
that the result of the price paid was to allocate to Glaxo Canada what Glaxo
Group considered suitable compensation for its secondary manufacturing
and marketing activities. The Court recognized that the payments under
the supply agreement for ranitidine included payments for some of the
rights and benefits under the licence agreement.831 The Court further found
that the generic comparables did not reflect the economic and business
reality of Glaxo Canada, nor did they indicate the arm’s length price for
ranitidine. The Court remitted the case back to the TC for re-determination
of the arm’s length price for ranitidine, under the instruction that the trans-
fer price of ranitidine under the supply agreement should be determined in
light of the rights and benefits under the licence agreement. In other words,
the controlled transactions should be aggregated for the purpose of deter-
mining an arm’s length profit allocation.832
6.7.4.5. Observations on the Supreme Court ruling
The Canadian GlaxoSmithKline case illustrates a general point: the trans-

fer prices applied in different types of controlled transactions that all per-
tain to the same value chain will normally be aligned and determined as
a package by the management of the multinational on a centralized global
level with the aim of providing the involved group entities with an arm’s
length total compensation (taking into account all transactions). In the au-
thor’s view, the TC’s ruling represents an analytical and thorough transfer
pricing assessment. In this section, the author presents his comments on
the SCC ruling, with which he partly disagrees. First, however, it must be
stressed that the GlaxoSmithKline case was decided on the basis of domes-
tic Canadian tax law, i.e. the former section 69(2) of the Canadian Income
Tax Act. The author’s comments will be limited to the interpretation and
application of the OECD TPG. Nevertheless, both the TC and SCC rulings
are heavily reliant on the OECD Guidelines in their interpretation of Cana-
dian income tax law. The ruling of the SCC seems, in essence, founded on
its interpretation of paras. 1.15 and 1.42 of the 1995 OECD Guidelines.833

832. The Tax Court was scheduled to hear the case on 12 Jan. 2015, but the matter was
settled between the Canada Revenue Agency and GlaxosmithKline Inc.; see Tax Notes
Intl., 12 Jan. 2015, at p. 124.
833. See SC ruling, paras. 39-43.
232
First, the SCC seems to overstate the significance of the Singleton case
and the Shell case for the assessment by the TC that the transfer price for
ranitidine under the supply agreement should be assessed on a stand-alone
basis. The author’s impression of the TC ruling on this point is that the TC
simply used the Singleton case as a supportive argument for what is largely
an evidentiary assessment of the material content of both the supply and
licence agreements.834
Second, the SCC seems to ask the wrong question. According to the SCC,
the question is of what “an arm’s length purchaser would pay for the prop-
erty and the rights and benefits together where the rights and benefits are
linked to the price for the paid property”.835 The SCC finds that the “result
of the price paid was to allocate to Glaxo Canada what Glaxo Group con-
sidered to be appropriate compensation for its secondary manufacturing
and marketing function”836 and that “whatever price was determined by
Glaxo Group would be subject to s. 69(2) and the requirement that the
transfer pricing transactions be measured against transactions between
parties dealing with each other at arm’s length”.837
It is clear that the only performance rendered by the Swiss Sub under the
supply agreement was the sale of ranitidine.838 Further, there was no doubt
that the Glaxo Canada ranitidine and the generic ranitidine were chemi-
cally and biologically equivalent. It is therefore simply not logical that the
transfer price – for an admittedly generic performance and for that per-
formance only – should be set higher than the established and observable
market price.
Also, it was clear that the excess price paid under the supply agreement
for ranitidine was payment for the benefits bestowed upon Glaxo Canada
834. Under para. 74 of the TC ruling, Judge Rip states “I have made no similar find-
ing of fact. Both the License Agreement and the Supply Agreement can stand alone;
neither is ancillary to the other”. Further, in para. 78, he states: “It may very well be
that a 40 percent total profit to Glaxo Group is reasonable; however, the issue before
me is whether the purchase price of the ranitidine was reasonable. One cannot com-
bine the two transactions and ignore the distinct tax treatments that follow from each.”
The author views these statements as clearly pertaining to an analysis of the material
content – and the legal consequences – of the transfer agreements. Further, the TC also
supported its assessment on this point with the US Tax Court Bausch & Lomb case.
Neither the quoted statements nor Bausch & Lomb are commented by the Supreme
Court of Canada; focus rests with the Singleton case and, to some extent, the Shell case.
838. SC ruling, para. 51, indirectly.
233
Allocation Rules
in the licence agreement with GGL. The significant portion of this value
came from the effect of the trade name Zantac and other intangibles made
available by GGL to Glaxo Canada.839 These intangibles were owned by
GGL, not by the Swiss Sub. Therefore, GGL was entitled to compensation
for the provision of these intangibles.
The view of the SCC, i.e. that the decisive point is whether the total amount
of compensation – across all contracts – to Glaxo Canada is at arm’s length,
seems simplistic in this context. The view would perhaps be defendable in
a purely domestic transfer pricing case, where the income from the differ-
ent contracts were treated equally for tax purposes. In this case, however,
where the different controlled transactions were entered into between dif-
ferent group entities in different jurisdictions, the view must be rejected.
Fundamentally, this is as much of a question of correct allocation, or dis-

aggregation, of income between taxpayers, as it is a transfer pricing issue.
The question is whether a taxpayer, by way of intra-group contractual ar-
rangements, is free to allocate income without regard to the underlying le-
gal relationships. Specifically for this case, this would mean whether GGL
was free to decide – with binding effect for the domestic tax treatment –
that compensation that it was entitled to, as the owner of unique and highly
valuable intangibles (the Zantac trade name, patent and know-how), should
not be paid to the United Kingdom as royalties from Canada, but as an
excessively high transfer price for raw materials to another group company
resident in Switzerland.
In this case, it may be argued that (at least some of) the excess price should
have been reclassified as royalty payments for the provision of intangibles
from GGL to Glaxo Canada. This would be more in line with the legal and
economic reality of the licence agreement and GGL’s ownership of intan-
gibles. Thus, the question should be whether the agreed royalty payments
of 6% under the licence agreement were arm’s length. The facts of the case
may indicate that this was not so. Glaxo Holding decided that 40% of the
gross profits of Glaxo Canada were to be remitted to Glaxo Holding. Thus,
it would seem that the lion’s share of this 40% may have been allocated to
the most tax-beneficial of the controlled transactions in the value chain (in
this case, as an excess transfer price for ranitidine from Glaxo Canada to
Switzerland, from there to Singapore and ultimately to the United King-
dom).
839. SC ruling, para. 52; and TC ruling, para. 91.
234
The Supreme Court touched upon the issue of whether the excess payments
for ranitidine under the supply agreement were linked to the provision of
specific intangibles under the licence agreement,840 but did not conclude
on this issue. It did, however, note that if the purchase price for ranitidine
included compensation for intellectual property rights granted to Glaxo
Canada by GGL, there would have to be consistency between that and
Glaxo Canada’s position with respect to withholding tax.
Further, the SCC held the issue open for reassessment by the TC of wheth-
er compensation for intellectual property rights was justified in this case.
This is a peculiar statement. GGL owned highly valuable and unique intan-
gibles, which it made available to Glaxo Canada. This performance was, of
course, in principle, subject to transfer pricing, and was priced, at 6% of the
net sales of Glaxo Canada. The question is therefore not of whether the pro-
vision of intangibles from GGL in the United Kingdom to Glaxo Canada
should be subject to transfer pricing, but whether the agreed royalty rate
of 6% represented an arm’s length transfer price for the provision by GGL
of rights to the patent, trade name and other intangibles required by Glaxo
Canada in order to perform secondary manufacturing and sales of Zantac
in Canada. The facts indicated that there existed large residual profits in
Glaxo Canada (from the Zantac value chain) equal to the excess payment
for ranitidine and that this excess return was caused by the high price that
Zantac demanded in the Canadian marketplace. The reason for this high
return was largely due to the trade name Zantac, or in other words, IP
owned by GGL. It therefore seems highly questionable whether the agreed
6% royalty represented an arm’s length compensation to GGL from Glaxo
Canada for the use of the patent and trade name for Zantac.
The SCC left considerable leeway for the TC to re-determine the transfer
price for ranitidine under the supply agreement.841 However, for the reasons
above, to the extent that the SCC recommended that the transfer price be
set above the market price for generic ranitidine, the author disagrees. The
TC should not have altered its determination of the transfer price of raniti-
dine as such. Of course, whether the TC would have felt authorized to leave
its previously determined transfer price for ranitidine largely unaltered in
light of the SCC ruling is an open question. Nevertheless, on one decisive
point, the TC should have altered its assessment. The excess amount paid
for the ranitidine should not be classified as dividend distributions, but as
royalty payments from Glaxo Canada to GGL. Reclassification of the ex-

841. See SC ruling, paras. 61-63.
235
Allocation Rules
cess amount from payment for ranitidine to the Swiss Sub to royalties paid
to GGL would bring the tax treatment in line with the legal and economic
realities of the case. GGL owned the IP applied in the Canadian Zantac
value chain and was therefore entitled to compensation.
As it does not seem doubtful that the lion’s share of the excess payments de
facto was due to the high price of Zantac as a result of the marketing power
of the trade name, treatment as royalties will be the natural transfer pricing
classification. Further, the reassessment by the tax authorities, whereby the
entire excess amount was reclassified as deemed dividend distributions,
seems unnecessarily harsh, and also somewhat out of sync with the re-
alities of the case. Royalty payments represent a business cost for Glaxo
Canada and would therefore, contrary to dividend payments, be deduct-
ible for tax purposes in Canada. The only domestic difference between the
treatment as asserted by Glaxo Canada in its original tax return, where the
excess amount was treated in full as payment for ranitidine, and a tax treat-
ment where the excess amount is treated as royalties will be that the latter
payments are subject to withholding tax in Canada. Both payments would,
however, be tax deductible for Glaxo Canada.
In the author’s view, the TC should have, upon re-determination, upheld its
transfer pricing assessment for the payment of ranitidine and classified the
excess amount paid for ranitidine as royalty payments to GGL. This would
have ensured that the point behind aggregating controlled transactions, i.e.
to realize arm’s length compensation among group entities by taking into
account all value transfers, would have been fulfilled, while at the same
time ensuring that the total amount of compensation (as calculated based
on the aggregated approach) would be allocated correctly among the indi-
vidual controlled transactions relating to the value chain. While the solu-
tion drawn up by the SCC managed to allocate in total an arm’s length
compensation to the Canadian group entity, it failed to allocate the com-
pensation correctly among the controlled transactions. The result was that
what should have been deemed to be royalty payments (subject to with-
holding tax and allocable to the United Kingdom) were instead treated as
COGS (not subject to withholding tax and allocable to Switzerland). Thus,
in the end, while the total compensation allocated by the SCC to the aggre-
gated controlled transactions was arm’s length, the portion of this amount
allocated to the individual controlled transactions were not.
236
Chapter 7
Direct Transaction-Based Allocation of Residual

Profits to Unique and Valuable IP: The CUT Method
7.1. Introduction
This chapter is dedicated to a discussion of the transaction-based US and
OECD methodology for directly allocating operating profits for unique
intangible property (IP) used in an intangible value chain, namely the
comparable uncontrolled transaction (CUT) method.842 This method is a
specified pricing method under the US regulations and the OECD Trans-
fer Pricing Guidelines (OECD TPG) 843 and evaluates whether the royalty
charged in a controlled transfer of IP is at arm’s length by reference to
the royalty charged in a CUT. While relatively ideal for the pricing of ge-
neric inputs for which there are efficient markets (e.g. oil, sugar and other
commodities),844 it is generally ill-suited for the purpose of allocating op-
erating profits from intangible value chains based on unique and valuable
intangibles, for which there is no true competition, and therefore no CUTs.
Contrary to the purpose behind it, the CUT method – in the context of
IP-driven value chains – may, in practice, be particularly attractive for tax-
planning purposes, and thus represent an obvious potential for BEPS. Be-
low-arm’s length royalty rates may be sought to be justified through the use
of purported CUTs that, in reality, pertain to IP transfers with economic
characteristics that differ from those of the IP transferred in the controlled
transaction. In combination with ambiguous comparability adjustments,
the outcome of the CUT method may be manipulated to yield the desired
result.
The CUT method may become a “black box” pricing methodology in the
sense that it will only offer a limited degree of transparency with respect
842. For discussions of the comparable uncontrolled transaction (CUT) method, see,
e.g. Wittendorff (2010a), pp. 649-650 and pp. 713-720; Markham (2005), at pp. 94-97;
and Andrus (2007), at p. 643. For a historical overview of the CUT method, see the
1992 International Fiscal Association (IFA) general report in Maisto (1992), at p. 32.
843. Treas. Regs. § 1.482-4(c); and OECD TPG, paras. 2.13-2.20 and 6.146-6.147 (see
also Examples 23 and 26 in the annex to ch. VI of the OECD TPG).
844. See Rocha (2017), at p. 193, where it is stated that most IFA branch jurisdictions
use the CUT method for transactions with commodities (see also the discussion at
pp. 227-229 of the same work).
237
Chapter 7 - Direct Transaction-Based Allocation of Residual Profits to
Unique and Valuable IP: The CUT Method
to the underlying assessments that were determinative for the third-party

royalty rate. It will also normally be difficult to ascertain whether the in-
tangibles transferred in the controlled and uncontrolled transactions in fact
share the same economic characteristics, which will be decisive for ascer-
taining an arm’s length allocation of operating profits among controlled
parties.
For these reasons, the United States severely restricted the application
of the CUT method to allocating profits for IP already in the 1994 final
regulations. The author will analyse the US provision in section 7.2. The
OECD, however, has traditionally embraced the CUT method as its flag-
ship method, but has now distanced itself from its prior position, as will be
elaborated in the analysis in section 7.3.
7.2. The US CUT method
7.2.1. Introduction
To apply the US CUT method for allocating profits for IP, the purported
CUT must pertain to the same intangible as transferred in the controlled
transaction, or to a comparable intangible.845 This will be discussed in sec-
tions 7.2.2. and 7.2.3., respectively. Thus, the topic pertains to the CUT
method comparability requirements that are specific to the context of in-
tangibles. The general US comparability requirements (discussed in sec-
tion 6.6.) form the backdrop for this analysis.
7.2.2. The purported CUT pertains to a transfer of the same

intangible as transferred in the controlled transaction
If the same intangible is transferred in a controlled and uncontrolled trans-

action, the royalty used in the latter transaction may be the most direct and
845. Treas. Regs. § 1.482-4(c). The US regulations reject two types of uncontrolled
transactions as comparables, namely (i) transactions that are not made in the ordinary
course of business (see Treas. Regs. § 1.482-1(d)(4)(iii)(A)(1); and Treas. Regs. § 1.482-
1(d)(4)(iii)(B), Example 1); and (ii) transactions designed to establish an arm’s length
result for the controlled transaction (see Treas. Regs. § 1.482-1(d)(4)(iii)(A)(2); and
Treas. Regs. § 1.482-1(d)(4)(iii)(B), Example 2). See Brauner (2008), at p. 128 on the
US CUT method.
238
The US CUT method
reliable measure of the arm’s length result of the former transaction.846 The
US regulations, however, emphasize that the general comparability factors
must align similarly for both transactions in order for the CUT pricing to be
acceptable as a benchmark. This requirement is illustrated in two examples.
The first example pertains to a US parent that has developed and patented a
new drug and has governmental authorization to make and sell the drug in
the United States and other countries.847 It licenses make-sell rights to the
drug for country X to a local subsidiary and for the neighbouring country
Y to a local unrelated company. The countries are similar in terms of popu-
lation, distribution of income among citizens, as well as the occurrence of
the disease that the drug targets, and they provide similar protection for
intellectual property rights. The costs of manufacturing and marketing the
drug in countries X and Y are expected to be approximately the same.
The terms of the controlled and uncontrolled agreements are materially

identical, including the royalty used. The drug is expected to be sold in
similar quantities and at similar prices in both countries. The conclusion of
the example is that the CUT is a reliable measure of the arm’s length roy-
alty rate.848 The second example uses the same facts as the first example,
but with the twist that the incidence of the disease is much higher in coun-
try Y.849 Thus, the profit potential from exploiting the right to make and sell
the drug is likely to be much higher in country Y than in country X. It is
therefore concluded that the CUT is unlikely to provide a reliable measure
of the arm’s length royalty rate for the controlled agreement.
7.2.3. The purported CUT pertains to a transfer of a different

intangible than that transferred in the controlled
transaction
In order for a purported CUT pertaining to the transfer of a different intan-

gible than the one transferred in the uncontrolled transaction to determine
846. Treas. Regs. § 1.482-4(c)(2)(ii). See also Ainsworth et al. (2012), where a com-
parative study concludes that exact comparables trump all other pricing approaches in
the jurisdictions encompassed by the study.
847. Treas. Regs. § 1.482-4(c)(4), Example 1. The example is reiterated in Treas.
Regs. § 1.482-8, Example 7.
848. See also Roberge (2013), at p. 223 on the use of external CUTs from small markets.
239
the controlled allocation of operating profits, two criteria must be fulfilled

under the US regulations. The IP transferred in the controlled and uncon-
trolled transactions must (i) be used in connection with similar products or
processes within the same industry or market; 850 and (ii) have similar profit
potential.851
The author will focus on the similar profit potential criterion, as this, in
practice, is the single most important requirement under the CUT method
(and has significantly influenced the 2017 OECD TPG).852 The criterion
made its debut in the 1993 temporary regulations853 and was introduced
because the 1992 proposed regulations suggested that the new comparable
profits method (CPM) should be used as a mandatory check on the pric-
ing results yielded by all other pricing methods.854 This proposed “san-
ity check” requirement was scrapped in the 1993 temporary regulations
(which, in and of itself, reduced the relative position of the CPM compared
to the position initially intended for it in the 1992 proposed regulations).
The 1993 preamble observed that the removal was “offset by incorporating
a reference to profit potential in the definition of comparability, reflecting
Congressional concern that royalty rates for “high-profit” intangibles could
“be set on the basis of industry norms for transfers of much less profitable
items”.855
The need to reliably measure profit potential increases as the total amount
of operating profits relative to the investment go up. Reliability is affected
by the extent to which the profit attributable to the IP can be isolated from
the profit attributable to other value chain inputs (e.g. manufacturing and
marketing functions).
The estimation of profit potential is a problematic exercise, even with re-

spect to the controlled transaction (for which detailed information is nor-
mally available). The operating profits from an intangible value chain may
not only need to be split among routine value chain contributions (e.g. con-
tract manufacturing) and unique IP, but it will normally also be necessary
to split the residual profits among two or more unique intangibles (e.g.
850. Treas. Regs. § 1.482-4(c)(2)(iii)(B)(1)(i).

851. Treas. Regs. § 1.482-4(c)(2)(iii)(B)(1)(ii).
852. See the discussions in secs. 6.6.3. and 11.4., with further references.
853. 1993 Temp. Treas. Regs. (58 FR 5263-02) § 1.482-4T(c)(2)(ii)(A)(3).
854. 1992 Prop. Treas. Regs. (57 FR 3571-01) § 1.482-2(d)(4)(i).
855. See 1993 Temp. Treas. Regs. (58 FR 5263-02), Preamble (the discussion of
§ 1.482-4T(c)). See also H.R. Rep. No. 426, 99th Cong., 1st Sess. 424 (1985), at pp. 425-
426.
240
The US CUT method
patents and marketing intangibles that refer to the same product). It may be
unrealistic to accomplish this based on purported CUTs due to a lack of de-
tailed information on the third-party value chains to which the CUTs refer.
The profit potential of the intangibles transferred in the controlled and un-
controlled transactions should, as the main rule, be determined directly.
The author will comment on this in section 7.2.3.2. Nonetheless, a direct
assessment may not always be feasible based on the information available.
The US regulations therefore, in some of these cases, allow the determina-
tion to be based on an indirect assessment. The author will discuss this in
section 7.2.3.3., and will then tie some comments to the determination of
profit potential in other cases in section 7.2.3.4.
7.2.3.2. Direct assessment of profit potential
The US regulations take the position that the profit potential of an intangi-
ble is most reliably measured by directly calculating the net present value
(NPV) of the operating profits allocable to the intangible, typically through
a discounted cash-flow (DCF) analysis.856 A range of factors must be taken
into consideration in this determination, such as the capital investment and
start-up expenses required and the risks to be assumed.857
7.2.3.3. Indirect assessment of profit potential
The US regulations allow the profit potential of certain low-value intan-

gibles to be estimated indirectly when the information necessary to di-
rectly calculate the NPV allocable to the intangible transferred in the un-
controlled transaction is unavailable858 and the “need to reliably measure
profit potential is reduced because the potential profits are relatively small
in terms of total amount and rate of return, [and] comparison of profit
potential may be based upon the factors referred to in paragraph (c)(2)(iii)
(B)(2)”.859
856. Treas. Regs. § 1.482-4(c)(2)(iii)(B)(1)(ii), second sentence.

857. See the discussion of valuation in ch. 13, as well as the more particular issues
connected to the valuation of buy-ins in the context of cost-sharing arrangements
(CSAs) in ch. 14.
858. Treas. Regs. § 1.482-4(c)(2)(iii)(B)(1)(ii), fourth sentence.
859. Treas. Regs. § 1.482-4(c)(2)(iii)(B)(ii). These factors are (i) the terms of the
transfer (whether the IP rights are exclusive, any restrictions on the use or geographic
area in which the rights may be exploited, etc.); (ii) the stage of development of the IP;
(iii) the rights to receive updates; (iv) the uniqueness of the IP (including the degree
and duration of protection); (v) the duration of the licence (including the termination
241
The threshold for bypassing the direct assessment requirement is ambigu-

ously described in the quoted text. It is, however, clear that the authority
is reserved for cases in which the “potential profits are relatively small”.
Based on the wording of the exception, as well as an example contained in
the regulations,860 the author finds that the exception only applies to generic
intangibles that do not account for a substantial part of the operating profits
from the products into which they are used as inputs. Such “normal return”
intangibles will likely be common within certain industries, for instance,
in the computer hardware industry, where competition is fierce and com-
petitors offer relatively similar components, as was the case in Seagate.861
Thus, the exception should be inapplicable if the IP transferred in the con-
trolled transaction is unique and valuable.
There is, of course, not always a distinct line between unique intangibles
that generate considerable residual profits and more generic intangibles
that do not account for a substantial portion of the operating profits from
the value chain. The exception should be limited to cases in which it is
clear that the intangible transferred in the controlled transaction is not a
high-value intangible. The author bases this conclusion on the fact that the
example contained in the regulations lists specific circumstances that justi-
fy an application of the exception (e.g. the existence of competing products
and the fact that it was apparent that the intangible was not responsible for
a significant portion of the profits).862
7.2.3.4. Assessment of profit potential in other cases
The question at issue in this section is how to compare the profit potential
of the IP transferred in the controlled and uncontrolled transactions when
the information available on the purported CUT is insufficient to perform
a direct assessment and the IP transferred in the controlled transaction
is unique, high-value IP that does not qualify for the indirect assessment
or renegotiation rights); (vi) the economic and product liability risks to be assumed by
the transferee; (vii) the existence and extent of any collateral transactions or ongoing
business relationships between the transferee and transferer; and (viii) the functions to
be performed by the transferer and transferee.
result only recommended by AOD-1995-09 (IRS AOD, 1995), and acq. 1995-33 I.R.B.
4 (IRS ACQ, 1995). See the discussion of the case in sec. 5.2.5.5.
862. Another thing entirely is that the example is not wholly logical, in the sense that
it contains so much information on the CUTs that it seems peculiar that it would not be
possible to directly assess their profit potentials.
242
The US CUT method
exception. The author imagines that this is the most relevant question in
practice, as the intangible value chains of multinationals tend to rely on
high-profit IP for which there are no CUTs available.
This issue is not addressed by the US regulations, which offer only the two
options discussed in sections 7.2.3.2.-7.2.3.3. (direct assessment of profit
potential through a DCF analysis or indirectly through a comparison of
terms and conditions). If none of these options are feasible, one could logi-
cally assume that the CUT method simply will be inapplicable. The au-
thor does not favour that interpretation. The CUT method should not be
rejected unless there are no reliable CUTs available. Of course, in order to
conclude on that issue, one must first determine whether the controlled and
uncontrolled transfers have similar profit potential.
In order to resolve this problem, the author suggests that the profit potential
of the IP transferred in the controlled transaction be determined directly
through a DCF valuation. It will normally be possible both for the multi-
national and the tax authorities to do this based on profit data possessed
by the multinational pertaining to its own business and value chains (e.g.
management accounting records). When the NPV has been estimated, the
question becomes how much of this value would be allocated to the trans-
ferer under his best realistic alternative to the licensing transaction.
The realistic alternative will normally be for the licenser to make and sell
the products itself instead of licensing out. An unrelated transferer would
logically not accept an alternative use of his intangible that would put
him worse off than he would be with his best alternative. The NPV result
can be converted into a royalty rate using the same technique as the one
prescribed under the US lump-sum provisions (aimed at constructing the
equivalent royalty rate).863 The result under the best realistic alternative
provides a precise indication of the lowest acceptable royalty rate in the
controlled transaction. If the royalty rate in the purported CUT is lower
than this rate, this will, in the author’s view, be a convincing indication
that the IP transferred in the CUT does not have profit potential similar to
the IP transferred in the controlled transaction. The purported CUT should
therefore be rejected.
This solution falls within an ordinary linguistic understanding of the lan-

guage used in the direct estimation provision, i.e. that “the profit potential
863. See the analysis of the US lump-sum provisions in sec. 16.4.
243
of an intangible is most reliably measured by directly calculating the net

present value of the benefits to be realized”. (Emphasis added) 864
This wording does not require the profit potential of both intangibles to
be directly assessed. The author’s interpretation ensures a direct assess-
ment of the profit potential of the intangible transferred in the controlled
transaction. Economic logic will then dictate that the profit potential of
the intangible transferred in the purported CUT cannot be similar if the
royalty rate in the CUT is below the royalty rate that the controlled trans-
ferer would receive in his best realistic alternative to licensing out the in-
tangible.
Further, the author’s interpretation is supported by the historical link be-

tween the profit potential criterion and the best realistic alternative assess-
ment.
The restrictive effect that the profit potential criterion has on the CUT meth-
od was originally illustrated in the 1993 temporary regulations through a
roundtrip example with a factual pattern akin to that in Bausch.865 The ex-
ample pertained to a US parent that owned the patent to a successful prod-
uct with a market price of USD 6. Even a selling price of USD 2 would have
provided the parent with a reasonable return on its costs, had it produced
the product itself. The parent instead chose to license make-sell rights to
the patent to a foreign subsidiary for a royalty of USD 1, based on a pur-
ported CUT involving the licence of a similar proprietary process. The
parent then purchased finished products from the subsidiary at the market
price of USD 6,866 effectively allocating the entire residual profits from
the value chain to the foreign subsidiary. For the purpose of determining
whether the intangibles had similar profit potential, the example authorized
the US Internal Revenue Service (IRS) to take into account the parent’s
“alternative of producing and selling product X itself as a factor that may
affect the amount P would demand as a royalty for the proprietary process
if dealing with an uncontrolled taxpayer at arm’s length”.867
864. Treas. Regs. § 1.482-4(c)(2)(iii)(B)(1)(ii), second sentence.

865. 1993 Temp. Treas. Regs. (58 FR 5263-02), § 1.482-4T(c)(2)(iv). See also Bausch
& Lomb Inc. v. CIR, 92 T.C. No. 33 (Tax Ct., 1989), affirmed by 933 F.2d 1084 (2nd
Cir., 1991); and the comments on the case in sec. 5.2.5.2.
866. The remaining part of the subsidiary’s production was sold to unrelated parties
for the same price.
867. 1993 Temp. Treas. Regs. (58 FR 5263-02), § 1.482-4T(c)(2)(iv), second para-
graph.
244
The US CUT method
As the parent could have chosen to make and sell the product itself, thus
reaping the whole USD 4 of residual profits, the US parent was “unlikely to
accept a royalty for the proprietary process of less than $4”.868 The intan-
gible transferred in the unrelated transaction with a royalty of USD 1 was
therefore not considered comparable to the intangible transferred in the
controlled transaction, due to having a different profit potential. This result
was in accordance with the IRS’s litigation stance in Bausch.
The example was criticized.869 The OECD was concerned that the relative
position of the CPM would be elevated by the inclusion of the profit po-
tential comparability requirement.870 The OECD’s arguments were that the
taxpayer would have to form an opinion of the profit potential of its own
intangible, which could be speculative, and gather data on the uncontrolled
intangible in order to calculate a profit potential for it.871 The 2nd task force
report stated that it could be “difficult, if not impossible” to obtain the
details of the CUT.
These arguments were, as far as the author can see, circular. The lack of
data on the purported CUT would indicate that the CUT method was not
appropriate in the first place. Indeed, the report seemed to indicate a gener-
ally relaxed attitude towards comparability under the CUT method. It sug-
gested that some of the comparability data required under the temporary
regulations, including the “uniqueness of property, expected economic life
of the property, provisions of third party agreements, and so forth”, would
be difficult or impossible to obtain.872 Even though not directly stated,
the report could be interpreted to express that the CUT method should
be available even in the absence of such information. The Baush-inspired
example was replaced in the final 1994 regulations.873
868. 1993 Temp. Treas. Reg. (58 FR 5263-02), § 1.482-4T(c)(2)(iv). See also Bausch
& Lomb Inc. v. CIR, 92 T.C. No. 33 (Tax Ct., 1989), affirmed by 933 F.2d 1084 (2nd
Cir., 1991); and the comments on the case in sec. 5.2.5.2.
869. See Tax Notes Intl., 1 Mar. 1993, at p. 525.
870. 2nd task force report, paras. 3.8, 3.11 and 3.13.
871. Id., at para. 3.7.
872. Id., at paras. 3.10 and 3.12.
873. The new example pertains to a US parent that has developed a drug to treat mi-
graines (which replaces an existing drug); see Treas. Regs. § 1.482-4(c)(4), Example
4. The question is whether the royalty rate used in an uncontrolled licence agreement
for the existing drug can be used to benchmark the royalty rate for the new drug. The
example rejects the licence agreement for the older drug as a CUT, as the new drug has
superior qualities that will enable it to make larger profits. Where there is no predeces-
sor intangible available that one can use to benchmark the profit potential against, profit
potential should, in the author’s view, be determined on the basis of the realistic alter-
245
In the author’s view, the removal of the Bausch-like example has no bear-
ing on the relevance of the best-realistic-alternatives assessment under the
profit potential criterion. The idea that transfer pricing results should be
compatible with the results that could be achieved through the best options
realistically available is a fundamental principle in US transfer pricing law.
It is expressed as a comparability factor under the general comparability
provisions,874 it lies at the core of what can be accepted as an unspecified
pricing method875 and it is the foundation for the logic behind the income
method of the new cost-sharing regulations.876 It would therefore be wholly
inconsistent if the profit potential of a purported CUT could be deemed
similar to that of the controlled transaction if the royalty rate of the CUT
made the controlled licenser worse off than he would have been with his
best realistic alternative.
The author therefore concludes that the profit potential of unique, high-val-
ue IP transferred in a controlled transaction should not be deemed similar
to the profit potential of IP transferred in a purported CUT if the royalty
rate extracted from the CUT results in a lesser allocation residual profits
to the licenser than he would have achieved with his best realistic alter-
native.877 If the profit potential is not similar pursuant to this assessment,
the CUT should be rejected (and it can thus not be used under the CUT
method).
7.2.4. There are no CUTs available
If there are no uncontrolled transactions available that fulfil the compara-

bility requirements discussed in sections 7.2.2.-7.2.3., the CUT method will
yield an unreliable result. It is therefore not likely that it will be deemed the
best method in such cases.878
natives available to the controlled parties, similarly to the assessment in the original
Bausch-like example of the temporary regulations.
874. Treas. Regs. § 1.482-1(d)(3)(iv)(H).
875. Treas. Regs. § 1.482-4(d).
876. Treas. Regs. § 1.482-7(g)(4). See also Treas. Regs. § 1.482-7(g)(2)(iii). The cost-
sharing regulations are also relevant to controlled agreements outside the context of
CSAs; see Treas. Regs. § 1.482-4(g).
877. Often, the best realistic alternative for the licenser will be to make and sell the
product itself.
878. See Treas. Regs. § 1.482-1(c)(1) and (2)(i), last sentence. See also § 1.482-9(c)
(4), Example 4, where the CUT method is deemed inappropriate when a US parent ren-
ders research and development services to a foreign subsidiary and uses a proprietary
software program (which it owns) to render these services. No uncontrolled parties are
246
The OECD CUT method
7.3. The OECD CUT method
7.3.1. Introduction
The historical OECD position on the applicability of the CUT method for
allocating profits from value chains that make use of unique IP has been
reversed 180 degrees with the 2017 OECD TPG.879 The 1995 consensus
text (which was carried over to the 2010 text without alterations) expressed
a preference for the CUT method and was brief, ambiguous and problem-
atic.880 It contained relaxed guidance on acceptable CUTs,881 which could
be interpreted to convey an acceptance of the use of industry averages and
bids to benchmark controlled royalty provisions.882 The 1995 wording is
now legal history and of little relevance for the interpretation of the new
2017 OECD guidance. While the 2017 text has not made any changes to
the general CUT guidance contained in chapter II of the OECD TPG, it
has introduced significant supplemental guidance on the relevance of the
CUT method in the context of unique IP, emphasizing comparability. The
author will discuss the new guidance on comparability requirements in
section 7.3.2., comparability adjustments in section 7.3.3. and the use of
commercial databases in section 7.3.4. Concluding comments are provided
in section 7.3.5.
identified that perform services identical or with a high degree of similarity to those
performed by the parent, and reliable comparability adjustments cannot be made.
879. The OECD has also changed its view on other important aspects of transfer
pricing. For instance, the OECD no longer upholds its historically critical attitude (as
expressed in the 1993 OECD report commenting on the 1992 proposed US regulations)
towards applying the concept of the realistic options available; see Parekh (2015), at
p. 298.
880. Least problematic was the list of relevant comparability factors; see 1995/2010
OECD TPG, paras. 6.20-6.21.
881. 1995/2010 OECD TPG, para. 6.23. The wording did not seem to require
similar expected benefits from the IP transferred in the controlled and uncontrolled
transactions, as the wording was merely that expected benefits “should be consid-
ered”.
882. See also the discussion of Ciba (Ciba-Geigy Corp. v. CIR, 85 T.C. No. 11 (Tax
Ct., 1985), acquiescence recommended by AOD-1987-21 (IRS AOD, 1987) and acq.
1987 WL 857882 (IRS ACQ, 1987)) in sec. 5.2.3.4., where the Court used royalty offers
and industry average rates to support its position. It should be noted that the Tax Court
ruled on the basis of the 1968 regulations. The author does not find it likely that the
Court would have ruled similarly if the case had been decided under the CUT provi-
sions of the 1994 regulations.
247
7.3.2. Comparability requirements for unique IP

under the CUT method
Unique and valuable intangibles (that generate residual profits) will gener-
ally be exclusive,883 legally protected rights884 within a certain geographical
area.885 These are the basic parameters that must be comparable between
the controlled and uncontrolled transactions. In this section, the author will
focus on the more problematic aspects for which there also must be suf-
ficient comparability.
The IP transferred in the controlled and uncontrolled transactions must be

comparable with respect to useful life, as the value of an intangible is the
present value of the profits allocable to it over its lifetime. Thus, all else
equal, the longer an intangible generates profits, the more valuable it will
be.886 The useful life can be affected by:
− the extent and duration of legal protections and the rate of technologi-
cal change;
− rights to enhancements, revisions and updates; 887
− the purposes for which the intangible is transferred (e.g. the transfer
of make-sell rights or rights to use the IP for further research and
development); 888 and
− the stage of development of the transferred IP (IP relating to estab-
lished products are normally more valuable than IP connected to part-
ly developed or not-yet-commercialized products).889
However, it is likely that the single most important comparability factor is

the expectation of future benefits (as the above factors will be reflected in
the expected NPV of the transferred IP).890 This factor is the OECD’s par-

885. OECD TPG, paras. 6.119-6.120.
886. OECD TPG, paras. 6.121-6.122.
887. OECD TPG, paras. 6.125-6.126.
888. The author discusses this issue in greater detail in the context of valuing buy-in
contributions to CSAs in ch. 14.
889. OECD TPG, paras. 6.123-6.124. On the significance for transfer pricing pur-
poses of the life-cycle stage of the IP, see Dolman (2007), under the section entitled
“Life cycle of IP”.
248
The OECD CUT method
allel to the US profit potential comparability criterion.891 The new OECD

guidance states that:
in the case of a transfer of an intangible … that provides the enterprise with a
unique competitive advantage in the market, purportedly comparable intan-
gibles or transactions should be carefully scrutinised. It is critical to assess
whether potential comparables in fact exhibit similar profit potential.892
If there is a significant discrepancy between the anticipated future ben-

efit from the controlled and uncontrolled intangibles, it will be “difficult
to consider the intangibles as being sufficiently comparable to support a
comparables-based transfer pricing analysis in the absence of reliable com-
parability adjustments”.893
The emphasis on future benefits as a comparability factor is new in the

2017 OECD TPG894 and relatively surprising, as the OECD historically has
disapproved of this as a comparability factor.895
The assessment of future benefits will normally be carried out through a

DCF valuation, in which all factors that may materially affect the future
benefits must be taken into account.896 Relevant factors include risks relat-
ed to the further development of a partially developed intangible and suc-
cessful commercialization, infringement, product liability, etc.897 The new
wording on valuation will be guiding in this assessment. The author refers
to the discussion in chapter 13 for further analysis of valuation techniques
in the context of transfer pricing of unique intangibles.
7.3.3. Comparability adjustments for unique IP

under the CUT method
The 2017 OECD TPG are sceptical towards making comparability adjust-
ments to purported CUTs in the context of unique IP, as such adjustments
891. Treas. Regs. § 1.482-4(c)(2)(iii)(B)(1)(ii). See also the discussion in sec. 7.2.3.1.,
with further references.
893. Id.
894. The notion of profit potential is, however, used in ch. IX of the OECD TPG on
restructurings (see OECD TPG, paras. 9.39-9.47), but then not as a comparability fac-
tor, but as an indicator of whether the stripped-down local entity transferred something
of value to the foreign party in the controlled transaction that requires remuneration.
895. 2nd Task Force Report, Paras. 3.8, 3.11 and 3.13.
896. OECD TPG, paras. 6.153-6.178.
249
can be difficult to carry out reliably.898 Adjustments that are largely relative
to the royalty charged in the purported CUT may indicate that the IP trans-
ferred in the controlled and uncontrolled transactions are not sufficiently
comparable. If so, the position of the OECD is that it may be necessary to
select a different pricing method (in practice, the transactional net margin
method or the profit split method).
7.3.4. Commercial databases
Several commercial actors now offer transfer pricing database services.899

Typical clients include multinationals, law and accounting firms and tax
authorities. The royalty data contained in these databases is extracted from
publicly available licence agreements, most of which are public due to fil-
ing requirements for multinationals that raise capital in the United States,
and may be found on Electronic Data Gathering, Analysis, and Retrieval
(EDGAR), a government database created and maintained by the US Se-
curities and Exchange Commission (SEC).900 A significant portion of the
filed licence agreements have the royalty rate, or other forms of compen-
sation, redacted.901 In addition to this, multinationals often do not license
their unique IP to third parties at all, as there is a need to maintain control
over the entire value chain.902 It should therefore be assumed that the third-
party licence agreements filed with the SEC generally do not pertain to

899. Such companies include Royaltystat (http://www.royaltystat.com), Ktmine (http://
www.ktmine.com) and Royaltyrange (http://www.royaltyrange.com).
900. From 1996, all publicly traded US companies are required to make their filings
on the Electronic Data Gathering, Analysis, and Retrieval (EDGAR), except for filings
made on paper due to a hardship exemption. Third-party filings with respect to these
companies, such as tender offers and schedule 13D, are filed on EDGAR.
901. The background for this practice is the US Securities and Exchange Commis-
sion’s (SEC’s) confidential-treatment-request process. The 1933 US Securities Act Rule
406 (17 CFR 230.406) and the 1934 US Securities Exchange Act of 1934 Rule 24b-2
(17 CFR 240.24b-2) allow the confidential treatment of information contained in a Se-
curities Act or Exchange Act filing, respectively. Other provisions allowing requests for
confidential treatment include 1933 Act Rule 418 (17 CFR 230.418) and 1934 Act Rule
12b-4 (17 CFR 240.12b-4) for certain types of supplemental information requested by
the SEC staff in connection with Securities Act and Exchange Act filings, as well as
Rule 83 of the SEC’s Rules of Practice (17 CFR 200.83). In order to gain confidential
treatment, the relevant information must be exempt from disclosure under the US Free-
dom of Information Act (FOIA) (5 U.S.C.A. § 552). The FOIA contains nine excep-
tions, the most relevant in this context being 5 U.S.C.A. § 552(b)(4) pertaining to “trade
secrets and commercial or financial information obtained from a person and privileged
or confidential”.
902. See the discussion of foreign direct investment in sec. 2.3.
250
The OECD CUT method
unique and valuable IP (that generates residual profits). Thus, the com-
mercial databases that draw upon the EDGAR database (which the author
gathers is most of them) contain incomplete information, which is unsuit-
able for the purpose of benchmarking controlled allocations of operating
profits from value chains that rely on unique IP.
The OECD position on the use of royalty rates extracted from these com-
mercial databases is that it is important to assess whether such publicly
available information is sufficiently detailed to permit an analysis of the
specific IP comparability factors.903 The author agrees with this (for the
reasons mentioned above). Further, tax authorities should question any ap-
plications of the CUT method that rest on purported CUTs extracted from
these commercial databases, as the information will normally not provide
sufficient background to properly carry out the comparability analysis.
7.3.5. Concluding comments
The core problem with the CUT method is that it directly allocates residual
profits by reference to what third parties have agreed in other licensing
transactions. In order for the method to provide a reliable result, it is crucial
that there is an extreme degree of comparability between the IP transferred
in the controlled and uncontrolled transactions. That is an unreasonable
contingency when it comes to unique intangibles. Also, the CUT method is
subject to reduced relevance on a more general level under the 2017 OECD
TPG due to its position relative to the other OECD pricing methods (which
the author will revert to in depth in chapter 11).
251
Chapter 8
Indirect Profit-Based Allocation of Residual

Profits for Unique and Valuable IP:
The CPM (US) and TNMM (OECD)
8.1. Introduction
In this chapter, the author will discuss the US and OECD one-sided profit-
based transfer pricing methodologies as they stand in the current US regu-
lations and the 2017 OECD Transfer Pricing Guidelines (OECD TPG).
The US method is named the comparable profits method (CPM), while its
OECD counterpart is the transactional net margin method (TNMM).904
Both utilize what is essentially the same profit allocation methodology, the
core content of which is to allocate a portion of the net operating profits
from a value chain, equal to a normal market return, to the group entity
that provides only routine value chain inputs.905 The rest of the net operat-
ing profits (residual profits) are thereby indirectly allocated to the other
party to the controlled transaction that contributes unique and valuable
intangible property (IP). The methodology plays a truly significant role in
how residual profits from IP value chains are allocated internationally.906
The author will underline that this chapter should be read in light of the
analysis of the historical development of the contract manufacturer theory
in chapter 5, in particular, the analysis of case law in section 5.2.5., the
design of the basic arm’s length return method (BALRM) in the 1988 US
White Paper in section 5.3.3. and the US and OECD implementation of
profit-based methodology in sections 5.3.4. and 5.4., respectively.
The one-sided profit-based methodology was developed by the US tax

authorities. The basic idea was to deal with the difficult issue of transfer
904. For a historical overview of the methodology, see the 1992 IFA general report
in Maisto (1992), at p. 56. For early and thorough analyses, see Horst (1993); and Cul-
bertson (1995). For an early comparison of the comparable price method (CPM) and
transactional net margin method (TNMM), see Taly (1996); and Rozek et al. (1999).
For a relatively recent and brief discussion of the CPM, see Wittendorff (2010a), at
pp. 651-656 and pp. 735-753. See also Brauner (2008), at p. 130; Hamaekers (2003);
Markham (2005), at. pp. 107-116; and Casley et al. (2003).
905. For a thorough historical analysis of the debate over net income benchmarking,
see Culbertson et al. (2003), at p. 77. See also Maisto (1992), at pp. 42-50.
906. See discussion of the centralized principal model in sec. 2.4.
253
Chapter 8 - Indirect Profit-Based Allocation of Residual Profits for Unique
and Valuable IP: The CPM (US) and TNMM (OECD)
pricing of unique and valuable IP by going through the “back door”. As it

was generally not possible to find comparable uncontrolled transactions
(CUTs) upon which to directly benchmark controlled royalty rates, a far
more accessible approach was developed. This consisted of benchmarking
the profits allocable to the controlled generic value chain inputs (contract
manufacturing, simple distribution functions, etc.) and then asserting that
what was left of the operating profits after such routine inputs had been
remunerated necessarily had to be the residual profits due to the unique and
valuable IP employed in the value chain.
This chapter will focus on three key issues, which the author will analyse
after having presented a lead-in to the methodology in section 8.2. The first
issue is to clarify in which scenarios the one-sided profit-based methodolo-
gy can be applied, i.e. to delineate the scope of application of the methodol-
ogy. This will be analysed in section 8.3. The second issue is to clarify how
the methodology works, i.e. how operating profits can be allocated under
the methodology. This will be analysed in section 8.4. The third issue is
how the comparability standard for the methodology shall be interpreted,
in particular, with respect to the TNMM. The rest of the chapter will be
devoted to this third issue in sections 8.5.-8.9.907
The 2017 OECD TPG regarding intangibles envision a more elevated role
for the profit split method, in practice, likely at the cost of the TNMM. The
author will revert to the relative importance of the TNMM in the context
of the OECD TPG in chapter 11, after having discussed the new OECD
2017 guidance on the profit split method in chapter 9 and the guidance on
the allocation of incremental operating profits from location savings, local
market characteristics and synergies in chapter 10.
8.2. A lead-in to the methodology
The BALRM of the 1988 US White Paper was introduced as a specified

transfer pricing method in the 1994 US regulations in the form of the
CPM.908 The methodology was originally intended as a weapon for the US
907. The author discusses comparability under the CPM in sec. 8.5. and under the
TNMM in sec. 8.6. In sec. 8.7., he discusses whether the TNMM has converged towards
the 1988 White Paper basic arm’s length return method (BALMR). In sec. 8.8., he asks
whether the TNMM comparability standard is a problem in practice, before ending
the chapter with closing comments on comparability under the one-sided, profit-based
methods in sec. 8.9.
908. On this, see also Casley et al. (2003), at p. 162.
254
A lead-in to the methodology
Internal Revenue Service (IRS) against outbound profit-shifting structures

(roundtrip transactions).909 Its introduction was met with criticism.910 The
OECD hesitantly adopted the methodology in the 1995 OECD TPG as
the TNMM, with some modifications. The TNMM is now likely the most
applied transfer pricing method in the world.911 Multinationals began ap-
plying the methodology to their advantage. TNMM-based allocations of
normal market return profits to source jurisdictions where routine func-
tions are carried out are an integral part of the principal model.912 The
other side of such allocations is that the residual profits from local sales are
extracted from taxation at source to a foreign group entity, normally resi-
dent in a low-tax jurisdiction, that holds ownership of the group’s unique
IP (the principal).
In this sense, the one-sided profit-based methodology has gone from being
a tool developed by tax authorities aimed at ensuring that IP ownership
entities in high-tax jurisdictions are allocated residual profits (a purpose
for which it is still used; see, in particular, section 14.2.8.3. on the US in-
come method) to facilitate taxpayer structures in which IP ownership enti-
909. In outbound contexts in which a US parent had a foreign subsidiary that per-
formed routine functions, the method would indirectly allocate the residual profits to
the United States. In inbound contexts in which a US licensee performed routine func-
tions, the method would allocate only a normal return to the United States. Example
12 of the White Paper (Notice 88-123), however, concerning an inbound licensing ar-
rangement in which the foreign parent carried out research and development (R&D),
indicated that the US tax authorities would use a profit split approach in cases in which
the US subsidiary performed functions that were arguably complex. Nevertheless, it
is clear that the final methodology, as adopted in the 1994 regulations, would allocate
only a normal market return to a US entity that contributes only routine inputs to the
value chain.
910. The main arguments against the CPM raised in the 1993 OECD task force report
were that (i) inappropriate comparables could be applied (paras. 3.8, 3.11 and 3.49); (ii)
the CPM could penalize controlled entities that were commercially unsuccessful by
comparing their actual profits to those of more successful third parties (paras. 3.11 and
3.12); (iii) the CPM was reliant on the theory of profit-maximizing companies, which
could be difficult to apply, due to deviating short-term motives, such as market penetra-
tion motives (paras. 3.10 and 3.11); (iv) the use of current profit experience could entail
hindsight (paras. 3.6 and 3.7); (v) the profit split method was preferable to the CPM
(paras. 3.12-3.15); (vi) the CPM would entail compliance burdens for small companies
(para. 3.55 and Executive Summary, section E(f)); and, finally (and, in the author’s
view, much of the reason behind the comprehensive OECD criticism), (vii) the CPM’s
allocation of residual profits was deemed unreasonable (paras. 2.19 and 3.49).
911. Likewise, the CPM seems to be the most applied methodology with respect to
the United States. See Brauner (2010), at p. 2; Chandler et al. (2003); and Hyde et al.
(2005).
912. See the discussion of the centralized principal model in sec. 2.4.
255
ties in low or no-tax jurisdictions are allocated the residual profits.913 The
methodology remains controversial, but the critical audience has changed
from mainly multinationals in the mid-1990s to source-state tax authori-
ties, particularly those of developing countries, due to the modest level of
operating profits allocated to them under these methods (see the analysis
in sections 11.2.-11.3.).914
The CPM and TNMM are not the only methods that are capable of allocat-
ing a normal market return to a group entity that contributes only routine
inputs to the value chain. The resale price method and the cost-plus method
may also be used for this purpose. These latter methods are, however, gross
profit methods, in the sense that they benchmark the gross profits of the
controlled transaction against the gross profits of comparable uncontrolled
transactions (CUTs). In order to do so, they require disaggregated financial
information from third parties so that it is possible to identify the gross
profits. Such information is often not available.915 In contrast, the CPM
and TNMM check net profits, which is data that is normally available.916
While the resale price and cost-plus methods are sensitive to accounting
classifications (only the cost of goods sold (COGS) are included in gross
profits), the CPM and TNMM avoid this problem (net profits include all
operating costs). Thus, even though the resale price and cost-plus methods,
on the one side, and the CPM and TNMM, on the other side, may all be
used to allocate normal market returns to controlled routine input provid-
ers, the latter may be applied when it is not possible to apply the resale
and cost-plus methods in a reliable manner due to information constraints.
This is not to say that that the CPM and the TNMM are not plagued by
913. Therefore also in the context of valuation, as reflected in the US income method
for pricing cost-sharing arrangement (CSA) buy-ins (see the discussion of the investor
model and income method in secs. 14.2.7. and 14.2.8.3., respectively), and, indirectly,
also in the OECD guidance on valuation (see the discussion in ch. 13, in particular, sec.
13.5.) through the use of the realistic alternatives of the controlled parties to guide the
valuation. The TNMM is also popular among tax authorities for the purpose of serving
as the default method for advance pricing agreements (APAs); see Yamakawa (2007),
at p. 6 and p. 19. For an updated US perspective on APAs, see Thomas et al. (2018).
914. Tax authorities in source jurisdictions generally argue that they should be al-
located incremental profits on top of the normal return, due to factors such as cost sav-
ings, local market characteristics or locally developed marketing intangibles. See the
discussion of location savings, market characteristics and synergies in ch. 10. See also
the analysis in secs. 11.2. and 11.3.
915. See Haugen (2005), at p. 224; Casley et al. (2003), at p. 164; Hamaekers (2001),
at p. 35; Cools (1999), at p. 178; and Roberge (2013), at p. 232. See also the analysis in
sec. 6.2.4.
916. Albeit not always on a transaction level, as the author will revert to in detail in
sec. 8.6.
256
The scope of application of the methodology
comparability complications, because they indeed are (see the discussion

in section 8.6.), but not to the extent that is the case for the resale price and
cost-plus methods.
The CPM and TNMM are, in practice (alongside the profit split method),
the leading methods for allocating residual profits. The core methodologies
underlying these profit-based methods are closely related. The allocation of
profits under the profit split method is generally carried out in a two-step
process. First, each controlled party is allocated a normal market return
for its routine value chain contributions. Second, the residual profit is gen-
erally allocated according to the relative values of their unique contribu-
tions. The first step is essentially the same allocation as under the CPM and
TNMM. Thus, the CPM and TNMM are best seen as an application of the
profit split methodology in the specific scenario in which one of the parties
to the controlled transaction (the tested party) does not own any unique IP
and is therefore allocated 0% of the residual profits, while the other party
to the controlled transaction is allocated 100% of those profits.
8.3. The scope of application of the methodology
Because the one-sided profit-based transfer pricing methodology will al-

locate only a normal market return to the tested party, both the US regula-
tions and the OECD TPG limit the scope of application of the methodology
to cases in which the tested party does not possess any unique and valuable
IP.917 The US regulations state that “in most cases the tested party will be
the least complex of the controlled taxpayers and will not own valuable
intangible property or unique assets that distinguish it from potential un-
controlled comparables”.918
The wording of the 1992 proposed and 1993 temporary US regulations

more clearly emphasized that the tested party could not be in possession
917. On this issue, see the Canadian Tax Court ruling in Alberta Printed Circuits Ltd.
v. The Queen (2011 TCC 232). The Canadian tax authorities denied deductions claimed
by a Canadian company for royalty payments made to its subsidiary in Barbados based
on the view that the Barbados subsidiary was a routine entity that did not own any
unique and valuable intangible property (IP) and thus could be allocated profits based
on the TNMM. The Court, however, found that unique and valuable IP was indeed
contributed to the value chain by the Barbados subsidiary, which thus could not serve
as the tested party under the TNMM. For comments on the ruling, see Pankiv (2017), at
p. 87; and MacIsaac et al. (2012). See also Wittendorff (2010a), at p. 739, on the tested
party under the CPM/TNMM.
918. Treas. Regs. § 1.482-5(b)(2)(i).
257
of unique value chain inputs.919 Why this language was watered down in
the 1994 final regulations is unclear. After all, the whole point of the CPM
is to allocate a normal market return to the tested party, thereby indirectly
allocating all residual profits to the other party to the controlled transaction
that owns unique and valuable IP.
In spite of the more relaxed language in the final regulations, the author
finds it clear that the CPM should not be applied where both controlled
parties own unique IP relevant to the examined transaction. The residual
profits should then be split between these parties, making the one-sided
CPM allocation of residual profits irrelevant. The main area of applica-
tion for the CPM is in licence agreements between group entities that
own unique IP, on the one side, and routine contract manufacturers or
low-risk distributors (LRDs), on the other side. The fact that this core
component of the CPM was not diluted in the 1994 final US regulations
is clearly illustrated by an example of its application,920 which is the same
roundtrip transaction example used in the 1992 proposed and 1993 tem-
porary US regulations (with some superficial changes). In the example,
the entire residual profits are allocated to a US parent, while a foreign
manufacturing subsidiary only receives a normal return on its routine
functions.
Parallel to this, the OECD TPG state that the TNMM is applicable where
“one of the parties makes all the unique contributions involved in the
controlled transaction, while the other party does not make any unique
contribution”.921
Thus, the OECD has made the application of its one-sided profit-based
methodology contingent on the tested party not owning any unique and
valuable IP.
919. § 1.482-2(f)(4)(ii) of the 1992 proposed regulations (57 FR 3571-01) stated that
“the tested party is the party to the controlled transaction whose operating income can
be verified using the most reliable data and with the fewest and most accurately quanti-
fiable adjustments. In the case of a transfer of an intangible, the tested party ordinarily
will be the transferee”. § 1.482-5T(b)(1) of the 1993 temporary regulations (58 FR
5263-02) stated that “the tested party ordinarily will be the participant in the controlled
transaction that does not use valuable, non-routine intangibles that it either acquired
from uncontrolled taxpayers and with respect to which it bears significant risks and
possesses the right to significant economic benefits or developed itself”.
920. Treas. Regs. § 1.482-5(e), Example 4.
921. OECD TPG, para. 2.59. See also OECD TPG, paras. 3.18-3.19.
258
How operating profits may be allocated to the tested party under
the methodology
8.4. How operating profits may be allocated to

the tested party under the methodology
8.4.1. Introduction
The determination of an arm’s length result under the CPM and TNMM
is based on the amount of operating profits that the tested party would
have earned from the controlled transaction if its profit level indicator were
equal to that of a comparable unrelated enterprise. The determination of
this hypothetical profit lies at the heart of the methodology and follows
three stages: (i) an appropriate profit level indicator is selected (discussed
in section 8.4.2.); (ii) profit data is extracted from uncontrolled taxpayers
that engage in similar business activities under similar circumstances us-
ing the selected profit level indicator (discussed in section 8.4.3.); and (iii)
the selected profit level indicator data is applied to the financial data of
the tested party to compute the comparable profits (discussed in section
8.4.4.).
8.4.2. Selecting an appropriate profit level indicator
A profit level indicator is a ratio that measures the relationship between

profits and costs incurred or resources employed, e.g. operating profits di-
vided by sales.922 The point of the indicator is to provide a reliable measure
of the operating profitability of independent enterprises that are compara-
ble to the tested party. The US regulations and the OECD TPG do not pose
any limitations on which profit level indicators can be used under the CPM
and the TNMM, respectively.923 As emphasized in the OECD TPG, the
profit level indicator should be tailored to the business activity and value
creation of the tested party.924
922. Profit level indicators should be derived from a sufficient number of years of data
to reasonably measure returns that accrue to uncontrolled comparables. Under the US
regulations, this should, as a minimum, encompass the taxable year under review and
the preceding 2 taxable years; see Treas. Regs. § 1.482-5(b)(4).
923. See Clark (1997), at p. 807, on the choice of profit level indicators, where it is
asserted that the choice may have significant impact on the resulting transfer prices.
924. E.g. sales or operating expenses may be appropriate to use as the denominator
if the tested party is a distribution entity; operating expenses may be appropriate for a
service provider entity; and operating assets may be appropriate if the tested party car-
ries out capital intensive functions (e.g. a contract manufacturer).
259
As the selection of an appropriate profit level indicator will depend on the

specific case, neither the CPM nor the TNMM is a homogeneous method
across applications. When referring to a specific application of the CPM
or TNMM, it should be specified as to which profit level indicator was
chosen in order to facilitate a meaningful discussion, e.g. that a return-
on-capital-employed (ROCE)-based CPM or sales-based TNMM was ap-
plied.
The US regulations systematize profit level indicators into three groups,

namely (i) the ROCE rate; (ii) financial ratios; and (iii) other profit level
indicators. This classification also fits in with the context of the TNMM.
The ROCE rate is the ratio of operating profits to operating assets.925 As

an example, assume that the following data is extracted from unrelated
enterprises. In order to produce an arm’s length ROCE range, one simply
divides the operating profits by the operating assets for each enterprise.
The arm’s length range in this case stretches from a ROCE of 9.3% to
16.8% (see Example 1).
Example 1
Independent Independent Independent Independent Independent

enterprise 1 enterprise 2 enterprise 3 enterprise 4 enterprise 5
Operating assets 15,000 30,000 20,000 50,000 15,000
Operating profit 1,400 5,000 3,000 8,400 1,660
(EBIT)
ROCE (EBIT/CE) 9.3% 16.7% 15.0% 16.8% 11.1%
EBIT = Earnings before interest and taxes; CE = Capital employed
925. See sec. 6.2. for a discussion on operating profits. If the Berry ratio (see infra
n. 927) is used, the relevant numerator will be the gross operating profits. Under the
US CPM provisions, operating assets are defined as the value of all assets used in the
relevant business activity of the tested party, including fixed (e.g. plant and equipment)
and current (e.g. cash, cash equivalents, accounts receivable and inventories) assets;
see Treas. Regs. § 1.482-5(d)(6). Operating assets do not include financial investments,
such as subsidiaries, excess cash and portfolio investments. The OECD TPG use a simi-
lar definition of operating assets (see OECD TPG, para. 2.97). See Yamakawa (2007),
at p. 11, who finds that the return on capital employed (ROCE) is suitable for testing the
presence of barriers to entry.
260
the methodology
The reliability of the ROCE as a profit level indicator increases as oper-

ating assets play a greater role for the creation of operating profits, and
is therefore appropriate for capital-intensive businesses, such as manufac-
turing or shipping. Reliability further depends on the extent to which the
composition of the tested party’s operating assets is similar to that of the
uncontrolled comparable. If there are difficulties in valuing the operating
assets, the reliability of the ROCE will diminish. This may be the case for
businesses that rely on unique IP, e.g. start-ups with modest assets apart
from the goodwill associated with its employees.
Financial ratios measure relationships between gross or net operating prof-

its, on the one side, and sales revenue or costs, on the other side.926 The ratio
of operating profits to sales is commonly used as a profit level indicator.
The ratio of gross profits to operating expenses (the Berry ratio) may also
be relevant, but it is contingent on access to third-party accounting data
that distinguishes between the COGS and operating expenses.927 Such data
will often be unavailable,928 rendering the Berry ratio inapplicable. The
reliability of the Berry ratio will also depend on the extent to which the
composition of the tested party’s operating expenses is similar to that of
the comparable unrelated enterprises.
As an example of the use of financial ratios as the profit level indicator,

assume that the data in Example 2 is extracted from comparable unrelated
enterprises.
926. Treas. Regs. § 1.482-5(b)(4)(ii). Under the OECD TPG, the costs must relate to
the controlled transaction under review; see OECD TPG para. 2.92. See Yamakawa
(2007), at p. 9, for an informed discussion of the use of return on costs as a profit level
indicator. Certain applications of the cost-plus method (using net profit margins) in
practice have, in reality, been disguised (cost-based) applications of the TNMM; see
Haugen (2005), at p. 225.
927. A Berry-ratio coefficient of 1 or more indicates that the enterprise is making
profit above all operating expenses, whereas a coefficient below 1 indicates that the en-
terprise is losing money. The OECD TPG find the Berry ratio suitable where the resale
price or cost-plus methods are inapplicable (see para. 2.101), but not where the operat-
ing expenses of the tested party are materially affected by controlled costs (e.g. head
office charges, rental fees or royalties). See Bhatnagar (2017) for a recent discussion of
the Berry-ratio (including Indian case law), where the author emphasizes that the ratio
is only suitable for benchmarking profit allocation to pure routine group entities (i.e.
limited risk distributors and service providers) when a clear causal link exists between
the operating expenses of the entity (seen as a measure of value added provided by the
entity) and its gross profits. For further comments on the Berry ratio, see, e.g. Witten-
dorff (2010a), at p. 745. For a thorough analysis, see Przysuski et al. (2005).
928. See sec. 6.2.4.
261
Example 2

Capital 15,000 30,000 20,000 50,000 15,000
employed
Sales 10,000 50,000 35,000 65,000 25,000
Operating 1,400 5,000 3,000 8,400 1,660
profit (EBIT)
EBIT/sales 14.0% 10.0% 8.6% 12.9% 6.6%
The ratio of operating profits to sales is selected as the profit level indicator,
as the financial statements of the independent enterprises do not provide
information on gross profits. The arm’s length range in this case stretches
from 6.6% to 14%.
Functional differences generally have a greater effect on financial ratios

than on the ROCE rate.929 For this reason, the US regulations require closer
functional comparability between the tested party and the selected unre-
lated enterprises if a financial ratio is used as the profit level indicator than
what is the case under the ROCE.
Other profit level indicators may be used if they provide reliable measures
of the income than the tested party would have earned if it had dealt with
controlled taxpayers at arm’s length.930
8.4.3. Extracting the profit level indicator data from

comparable independent enterprises
Under both the CPM and TNMM, the profit level indicator data must be
extracted only from comparable unrelated enterprises.931 The greater the
929. Treas. Regs. § 1.482-5(b)(4)(ii).

930. Treas. Regs. § 1.482-5(b)(4)(iii); and OECD TPG, para. 2.76. Under the US reg-
ulations, profit level indicators based solely on internal data may not be used, as they
are not objective measures of profitability derived from uncontrolled taxpayers engaged
in similar business activities under similar circumstances. See Yamakawa (2007), at
p. 10, for a discussion of return on assets, and at p. 11 on return on equity.
931. Treas. Regs. § 1.482-5(a); and OECD TPG, para. 2.58 and paras. 3.27-3.28. The
1993 temporary US regulations also allowed that “other statistical techniques and cri-
teria” could be used for finding the arm’s length range; see Temp. Treas. Regs. (58 FR
5263-02) § 1.482-5T(d)(2)(ii). Arguments for using statistical techniques in transfer
262
the methodology
degree of comparability between the tested party and the selected com-
parable enterprises is, the more reliable the results will be.932 The compa-
rability requirements of the CPM are, in reality, what separates it from its
OECD counterparty.
The 1993 US temporary regulations indirectly expanded the scope of the

CPM by allowing for a relaxed degree of comparability between the tested
party and independent enterprises chosen as comparables. As the CPM
measured the total annual operating profits of the relevant business activity
of the tested party, the comparability requirement for independent enter-
prises was that they “need be only broadly similar, and significant product
diversity and some functional diversity between the controlled and uncon-
trolled transactions is acceptable”.933
Of course, as the OECD wanted to limit the general scope of its CPM im-
plementation to abusive cases, the allowance for broadly similar unrelated
parties as comparables was criticized.934 The 1993 temporary regulations,
however, went further with the introduction of the arm’s length range pro-
vision, a feature of which was to allow the use of profits from independent
enterprises that did not even meet the “broadly similar” comparability
requirement.935 If such “incomparable” comparables were used, only the
interquartile range between the 25th and 75th percentile of the compa-
rable profit interval (CPI) could be used as benchmark data. The inter-
quartile restriction, which had the effect of eliminating the outer borders
of the observed variation in actual third-party return data, was inserted
precisely to compensate for dilution of the CPM comparability require-
ments. The author will elaborate on the CPM comparability requirements
in section 8.5.
The TNMM is burdened with more restrictive comparability requirements

than the CPM. The OECD’s point of departure is that the transactions of
both the tested party and the unrelated enterprises must be segregated so
that the net profit margin of the controlled transaction is benchmarked only
pricing cases had previously been rejected by the courts. See Seagate Technology, Inc. v.
CIR, 102 T.C. No. 9 (Tax Ct., 1994), acquiescence in result only recommended by AOD-
1995-09 (IRS AOD, 1995), and acq., 1995-33 I.R.B. 4 (IRS ACQ, 1995), where the use
of average prices was found wanting.
932. Treas. Regs. § 1.482-5(c)(2)(i).
933. See 1993 Temp. Treas. Regs. (58 FR 5263-02) § 1.482-5T(c)(1).
934. See 2nd Task Force Report, para. 2.21.
935. See 1993 Temp. Treas. Regs. (58 FR 5263-02) § 1.482-5T(d)(2)(ii). See also the
analysis of the US and OECD arm’s length range provisions in sec. 6.5.
263
against the corresponding margins from CUTs. Thus, the OECD demands
transactional comparability. The material content of this important re-
quirement will be thoroughly analysed in section 8.6.
8.4.4. Applying the extracted profit level indicator data

to the tested party
The extracted third-party profit level indicator data is compared to the re-
ported profits of the tested party. A transfer pricing income adjustment
must be performed if the controlled profits lie outside the arm’s length
range. The author will illustrate this. The following data is extracted from
comparable unrelated enterprises:

Capital 15,000 30,000 20,000 50,000 15,000
employed
Sales 10,000 50,000 35,000 65,000 25,000
EBIT 1,400 5,000 3,000 8,400 1,660
ROCE (EBIT/ 9.3% 16.7% 15.0% 16.8% 11.1%
CE)
EBIT/sales 14.0% 10.0% 8.6% 12.9% 6.6%
The data of the tested party is as follows:
Capital employed 45,000

Sales 27,500
EBIT 22,000
ROCE (EBIT/CE) 48.9%
EBIT/sales 80.0%
In this example, it is clear that the tested party has realized an operat-
ing profit significantly above the profits of the comparable independent
enterprises. The tested party realized a ROCE of 48.9%, with the arm’s
length range being 9.3%-16.8%. Further, the tested party realized EBIT/
sales of 80%, while the corresponding arm’s length range is 48.9%. If the
third-party profit level indicators (ROCE and EBIT/sales) are applied to
264
Comparability under the CPM
the financial data of the tested party, the following arm’s length range is
generated:
ROCE (EBIT/CE) 6,300 4,500 3,857 5,815 2,988

EBIT/sales 2,567 4,583 4,125 4,620 3,043
Depending on which profit level indicator provides the most reliable result
in this particular case, the reported operating profits of the tested party
of 22,000 will be adjusted downwards to an appropriate point within the
arm’s length range. If ROCE is chosen, the reported profits will be adjusted
downwards to somewhere within the ROCE arm’s length range of profits of
2,988-6,300. Alternatively, if EBIT/sales is chosen, the reported profit will
be adjusted downwards to a point within the arm’s length range between
2,567 and 4,620.936
8.5. Comparability under the CPM
The analysis of the general comparability requirements under the US regu-

lations in section 6.6. forms the backdrop for the discussion in this section.
The CPM provisions require that the chosen profit level indicator be ex-
tracted from an independent enterprise that engages in “similar business
activities under similar circumstances” as the tested party.937 If the tested
or independent parties carry out other business activities than the activity
connected to the controlled transaction, an allocation of income, costs and
assets between these activities must be performed.938 This allocation can be
done directly, if there is a factual relationship that supports it, or indirectly,
936. The procedure for applying the CPM is illustrated in four examples in the regula-
tions; see Treas. Regs. § 1.482-5(e), Examples 1-4.
937. Treas. Regs. § 1.482-5(a). See also Casley et al. (2003), at p. 165.
938. Treas. Regs. § 1.482-1(f)(2)(iii). The profits of the tested party shall ordinarily
be compared to uncontrolled profits from the same taxable year. It may, however, be ap-
propriate to consider the controlled and uncontrolled profits for 1 or more years before
or after the year under review. If such multiple year data is used, the controlled profits
must still be benchmarked against uncontrolled profits from the same income years.
The 1993 temporary regulations (58 FR 5263-02), § 1.482-1T(b)(2)(iii)(B), allowed
for average controlled profits to be benchmarked against average uncontrolled profits
over the same period, but this option was removed in the final regulations. On carving
out the profits from the controlled transaction from the other transactions of the tested
party, see Yamakawa (2007), at p. 16.
265
through a reasonable allocation formula.939 If direct allocations cannot be

made, the reliability of applying the CPM will be reduced relative to other
methods that require fewer allocations.
The degree of functional comparability required to obtain a reliable result

under the CPM is generally less than that required under the resale price or
cost-plus methods,940 as differences in functions will be reflected in operat-
ing expenses. Taxpayers performing different functions may have different
gross profit margins but nevertheless earn similar levels of operating prof-
its.941 Further, as operating profits are normally less sensitive than gross
profits to product differences, the CPM is not as dependent on product
similarity as the resale price or cost-plus methods.942
Comparability adjustments must be made if there are differences between

the tested party and an independent enterprise that would materially affect
operating profits.943 Direct adjustments of the operating profit of the inde-
pendent party may be needed, for instance, for differences in accounts pay-
able or in the use of stock-based compensation.944 Asset-based adjustments
will indirectly adjust operating profits (e.g. reclassifying from expensing
to capitalizing will increase operating profits) and may require economic
accounting adjustments, such as changing principles for accounting for in-
939. Treas. Regs. § 1.482-5(c)(3)(iii).

940. Treas. Regs. § 1.482-5(c)(2)(ii).
941. The resale price and cost-plus methods will not take differences in functions
performed into account, as the object of investigation under those methods is gross
profits. The CPM has been criticized for providing a less-than-reliable measure of an
arm’s length result, due to the possibility that the method may be affected by differ-
ences in cost structure, management efficiency, etc. between the tested party and the
uncontrolled comparables; see, e.g. Rozek et al. (1999), at p. 1556.
942. Treas. Regs. § 1.482-5(c)(2)(iii).
943. Treas. Regs. § 1.482-5(c)(2)(iv). See Yamakawa (2007), at p. 12 for an informed
discussion of comparability adjustments. See also Wittendorff (2010a), at p. 749 on ac-
counting adjustments.
944. The US regulations contain two comparability adjustment examples. The first
concerns the adjustment of operating assets and profit for differences in accounts re-
ceivable (Treas. Regs. § 1.482-5(e), Example 5). Third-party accounts receivable are
adjusted downwards to achieve the same proportion of accounts receivable to sales as
the tested party, and third-party operating profits are adjusted by deducting imputed in-
terest income on the excess accounts receivable. In the second example, adjustments are
made to operating profits for differences in accounts payable (Treas. Regs. § 1.482-5(e),
Example 6). There are significant differences in the level of accounts payable between
the independent distributors and the US subsidiary in the example, as the latter carries
out all purchases on a cash basis. To adjust for this, the operating profit of the uncon-
trolled distributors is increased to reflect the interest expense imputed to the accounts
payable.
266
Comparability under the TNMM
ventory (e.g. from “last, in first Out” (LIFO) to “first in, first out” (FIFO)),
altering depreciation periods and working capital adjustments.945
8.6. Comparability under the TNMM
8.6.1. Introduction
The analysis of the general OECD TPG comparability requirements in sec-

tion 6.6. forms the backdrop for the discussion in this section.
The TNMM benchmarks the profit margin of the controlled transaction

by reference to the profit margin that the tested party realizes from its own
CUTs (internal comparables), if such comparables are available, and if not
(which normally is the case), by reference to the profit margin earned in
CUTs by an independent enterprise (external comparables).946 The com-
parability considerations discussed in section 8.5. for the CPM are also
relevant under the TNMM.947 For the TNMM, however, the clearly most
significant comparability issue is to what extent third-party profit margins
that stem from a portfolio of unrelated transactions that are both compa-
rable and incomparable to the controlled transaction (so-called “blended
profits”) may be used to benchmark the allocation of operating profits in the
controlled transaction. This comparability problem is among the most sig-
nificant in international transfer pricing. The author is not aware of any le-
gal literature in which this problem has been subject to thorough analysis.948
In principle, this problem applies to all OECD transfer pricing meth-

ods apart from the CUT method. Due to the immense popularity of the
TNMM, however, the problem will, in practice, arise in connection with
the application of this method. As the author will revert to in the analysis
in this section, this problem was the single most controversial aspect of the
OECD’s implementation of the one-sided profit-based pricing methodol-
ogy in the 1995 OECD TPG. It has been the subject of repeated OECD
945. See the discussion on the distinction between accounting and economic adjust-
ments in sec. 8.8., and particularly infra n. 1013.
947. Historically, the comparability standards of the CPM have, however, been per-
ceived as less strict than those of the TNMM. See Culbertson et al. (2003), at p. 89.
The TNMM has nevertheless attracted criticism for having too-relaxed comparability
standards; see, e.g. Fris (2003), at p. 194.
948. E.g. Pankiv (2017), at p. 75 touches upon comparability issues in relation to the
TNMM, but does not provide an analysis.
267
discussions since then. As the OECD TPG are ambiguous regarding this
problem, the author’s task will be to clarify how the OECD consensus text
should be interpreted.
Such clarification is important and necessary, because the OECD compa-

rability standard with respect to this problem will determine the amount
of freedom that taxpayers and tax authorities have to choose third-party
comparables data that are more or less comparable to the data of the tested
party, and thereby also (indirectly) the amount of freedom to determine the
arm’s length range of acceptable normal market return profit margins to be
applied in the controlled transaction. The more freedom there is, the more
the profit margin can be designed to fit within the profit allocation prefer-
ences of the taxpayers or tax authorities applying the TNMM. For instance,
if a multinational wants to extract a maximum amount of residual profits
from a market jurisdiction,949 it will seek to remunerate the market jurisdic-
tion routine distribution entity based on an application of the TNMM that
uses third-party profit data that exhibits a minimum amount of normal mar-
ket returns. The market jurisdiction tax authorities will, on the other side,
seek to identify third-party profit data that exhibits a maximum amount
of normal market returns in order to restrict the multinational’s extraction
of residual profits to a minimum. Thus, the OECD TNMM comparability
standard will determine how far taxpayers and tax authorities can go with
respect to “cherry-picking” third-party comparables in this manner.
8.6.2. The concept of blended profits illustrated

by an example
The author will illustrate the problem of blended profits through an example
(see Example 3). The tested party in this example is a related distribution
entity that distributes golf balls as its sole activity. The tested party realizes
an aggregated net profit margin of 15%. Due to insufficient accounting in-
formation on gross margins, it is determined that a net profit method should
be applied. An unrelated distribution entity that distributes golf balls is
identified. The unrelated entity is comparable to the tested party with re-
spect to functions performed, assets used and risks incurred. In addition to
golf balls, the unrelated enterprise also sells tennis and basketballs.
The only profit information available on the unrelated enterprise is its pub-
lic financial statements. These show that the enterprise earns an aggre
949. See sec. 2.4. on the centralized principal model.
268
gated net profit margin of 16.2%. That margin results from the sale of
golf, tennis and basketballs. It is not possible to deduct from the financial
statements how much of the total net profits stem from the sales of golf
balls and how much of the profit is attributable to the two other groups of
transactions carried out by the independent distributor. Thus, information
on the decomposition of sales, COGS and operating expenses among the
three transaction types is unknown.
The unavailable disaggregated management accounting of the independent

enterprise shows the following breakdown of profits (values are rounded).
Example 3
Golf balls Tennis balls Basketballs Total transactions

transactions transactions transactions (aggregated)
Sales 1,900 880 730 3,510
COGS 1,000 650 620 2,270
Gross profit 900 230 110 1,240
Operating 450 120 100 670
expenses
Net profit 450 110 10 570
Net profit 23.7% 12.5% 1.4% 16.2%
margin
Contribution 54.1% 25.1% 20.8% 100.0%
to total net
margin
Data on the above individual three types of Aggregated
transactions are not available in public financial transaction data is
statements (available only internally to the available in public
management of the unrelated enterprise) financial statements
Let us suppose that the net profit margin from the sale of golf balls is 23.7%,
but that these profits only contribute 54.1% to the total blended profits of
the independent enterprise. The remaining blend comes from the sale of
tennis balls at a net profit margin of 12.5% (which contributes 25.1% to the
total net margin) and from the sale of basketballs at a net profit margin of
1.4% (which contributes 20.8% to the total net profits of the enterprise).
Given that it realistically will not be possible to separate the three groups
of transactions based on the public financial statements of the unrelated
enterprise, the application of a net profit method will have to be based on
269
the aggregated, or blended, net profits of 16.2%. The relatively close prox-
imity of the return position of the tested party of 15% and the blended net
profit of the independent enterprise of 16.2% would likely place the income
of the tested party within the arm’s length range. Thus, no adjustment will
be carried out.
Of course, with further insight into the composition of the blended profits
of the comparable enterprise, one would learn that the true uncontrolled net
profit margin for the distribution of golf balls is 23.7%, representing a gap
of 8.7% from the reported net profits of the tested party. Had the net profits
from the comparable uncontrolled transaction been known, it would likely
have resulted in an adjustment of the taxable income of the tested party.
This example illustrates that the application of net profit methods becomes
problematic when it is not possible, based on the financial information
available on independent enterprises, to separate profits from transactions
that are comparable to the controlled transaction from profits due to trans-
actions that are not comparable as such. This is because the arm’s length
standard of article 9 of the OECD Model Tax Convention (OECD MTC)
requires that the pricing of a controlled transaction be tested against com-
parable transactions only. Had information that would allow the separation
of profits generated from the sale of golf balls from the overall profits of
the third party been available, it would have been unproblematic to use
the profit margin of the comparable golf ball transactions to determine an
arm’s length profit for the controlled transaction of the tested party.
The fundamental issue is the lack of relevant accounting data on uncon-

trolled comparable enterprises. As stated in the 2006 OECD report on
comparability:
[T]hird parties rarely publicly disclose detailed information on their transac-
tions, even on aggregated transactions at the product line level. Indeed, enter-
prises are generally required to disclose financial information only at the com-
pany or even group level, and information on prices or margins per transaction
is often regarded as confidential business information.950
One could argue that this problem may be alleviated if one carefully selects
unrelated comparable entities that only carry out one type of transaction
that is comparable to the controlled transaction. Thus, in Example 3, an
950. Comparability: Public invitation to comment on a Series of draft issues notes

(OECD 2006) [hereafter the 2006 OECD Comparability Report], p. 64., para. 1. See
sec. 8.6.4. for a discussion of the 2006 OECD Comparability Report.
270
unrelated distributor that only sells golf balls should be selected. This ar-
gument is valid – and at the same time, unrealistic – in the vast majority of
cases. As stated in the 2006 OECD report on comparability, “it is in effect
very rare that an independent enterprise limits its activity to a single type
of transaction”.951
One is therefore left with the disadvantageous fact that the unrelated enter-
prises that bear the closest resemblance to the tested party with respect to
functions, assets and risks will normally carry out a significant amount of
different types of transactions. For instance, unrelated distributors that are
selected to measure the profit from the distribution of golf balls may sell
a range of other sporting gear products, often encompassing thousands of
products. If the overall net profits from such enterprises are used to deter-
mine the allocation of income to the tested party, the transfer pricing as-
sessment will be based on a blended profit, which melts the profits from the
sale of golf balls with the profits from a variety of different product groups
with varying profit characteristics. There is no question that applying such
a blended profit margin to allocate income to the tested party represents a
significant comparability problem.
Depending on the degree to which the profit from the uncontrolled compa-
rable transaction is diluted by profits from other transactions carried out by
the independent enterprise, one could, in effect, end up comparing apples
to oranges by using the total net profit margin of the independent enter-
prise to determine the pricing of the controlled transaction. This will have
a clearly negative effect on the reliability of the net profit method being
applied, as the profit determined for the tested party on the basis of the ex-
tracted blended net profit margin will include the profits from transactions
that are not comparable to the controlled transaction.
Normally, it will be clear whether the total net profits of the independent
enterprise consist of profit contributions from transactions other than those
comparable to the controlled transaction. However, the extent to which this
is true will not be readily ascertainable, because public financial state-
ments do not offer the information necessary to disaggregate the total net
profits of the independent enterprise into profits from individual groups of
transactions.
This again poses a dilemma. If the total net profits of the independent en-
terprise are only mildly diluted by the profits from incomparable transac-
951. 2006 OECD Comparability Report, p. 65, para. 5.
271
tions, the negative effects of applying a blended net profit margin may be
negligible. For instance, if Example 3 was altered so that the profit contri-
bution from golf balls accounted for 90% while the profits from tennis and
basketballs each accounted for 5% of the total profits, the total net profits
of the independent enterprise would be approximately 22%.952 This is rela-
tively close to the 23.7% net profit margin generated by the sale of golf
balls on a stand-alone basis – so close, in fact, that use of the blended profit
margin to allocate income to the tested party likely would not significantly
affect the reliability of the results.
Conversely, if the example were altered so that the profit contribution from
sales of golf balls only accounted for 25% while the profits from tennis
and basketballs each accounted for 37.5% of the total profits, the total prof-
its would be approximately 11.1%.953 This is far below the true net profit
margin of the comparable golf ball transactions of 23.7%. If the blended
net profit margin of 11.1% were used to determine the income of the tested
party, uncontrolled transactions that are incomparable to the controlled
transaction would, in fact, have a decisive influence on the allocation of
income under article 9 of the OECD MTC. Such a result is fundamentally
problematic, as the arm’s length principle requires that the income of a re-
lated party be set by reference to the income realized by unrelated parties
through comparable transactions.
This leaves a transfer pricing assessment – particularly under the one-sided

methods – in a precarious position. Depending on the composition of the
total net profits of an independent enterprise, they may be more or less
comparable to the profits of the uncontrolled transaction. Often, there will
be no reliable way of ascertaining the degree to which the total net profits
have been tainted by the profits from incomparable transactions. A signifi-
cant degree of uncertainty surrounding the application of the transfer pric-
ing methods will therefore normally exist if the benchmark profit margin is
extracted from third-party blended profits. The uncertainty, however, goes
both ways. As touched upon above, even if the total transactional portfo-
lio of the comparable enterprise encompasses transactions that are incom-
parable to the controlled transaction, this does not necessarily entail that
the total net profit margin of the enterprise will diverge significantly from
what it would have been had the enterprise only carried out the comparable
transaction, e.g. a distributor that only realizes a normal return across all
of its transactions.
952. 23.7 × 0.9 + 12.5 × 0.05 + 1.4 × 0.05 = 22%

953. 23.7 × 0.25 + 12.5 × 0.375 + 1.4 × 0.375 = 11.1%
272
This will typically be relevant for a range of enterprises that operate under
fierce competition, for instance, a retail computer hardware distributor that
realizes a relatively stable profit margin on the products it distributes. If
Example 3 is changed so that the net profit margins realized from the sale
of golf, tennis and basketballs are 13%, 3.5% and 14%, respectively, the
relative influence from each product group on the total net profit margin of
the distributor will not make much of a difference, as the total net profits in
any case will lie between 13% and 14%.
Again, in practice, it will be difficult to ascertain the degree to which the

total net profit margin of an independent enterprise diverges from the net
profit margin it realizes on the comparable transactions. Decomposing the
total blended profits of an independent enterprise into profits from com-
parable and incomparable transactions would – even if some supporting
information were available – be a discretionary and complex affair.
Even if sales could be tracked to individual groups of transactions, the

COGS and operating expenses would need to be allocated among the trans-
action groups, likely triggering difficult discretionary assessments.954 Real-
istically, it would likely be necessary to have access to the management ac-
counting data of the selected independent enterprise in order to find reliable
information on single product or product group profits. Even such profit
information may be based on discretionary allocations. Further, manage-
ment accounting information is not available in public financial statements.
This is also information that unrelated enterprises cannot be expected to
willingly share with others, for commercial and competition reasons.
8.6.3. The 1995 consensus text on the TNMM with respect to

aggregation of transactions
Prior to the finalization of the 1994 US regulations, commentators feared

that the proposed CPM could be applied to the overall activities of an unre-
lated, complex multinational enterprise, and thus result in the extraction of
an aggregated net profit margin that also encompassed profits from transac-
tions that were irrelevant to the transfer pricing assessment and that con-
tributed differently to the total profits of the enterprise than the contribution
from the comparable transactions. As the level of profits realized from the
different types of transactions could vary, the reliability of basing a transfer
pricing assessment on such blended net profits could be severely impeded.
954. See also 2010 OECD TPG, para. 3.37, fifth sentence.
273
The 1993 OECD task force report presented some views on the CPM of
relevance to the issue of blended profits. First, the report expressed con-
cern that the applicable business classification of the tested party would
be difficult to carry out properly.955 Such classification was of importance
because unrelated enterprises would be selected from within the same
classification. The report expressed concern that this could lead the IRS
to apply sector averages for net profits. The concern was reiterated in the
1994 report, which stated that net profit margins derived from uncontrolled
enterprises that are “only broadly similar” would not necessarily achieve
an arm’s length result.956
At the time, few (if any) countries apart from the United States had im-
plemented a one-sided net profit transfer pricing method in their domestic
legislations. Several OECD countries were sceptical towards this transfer
pricing methodology, and the 1994 OECD discussion draft leading up to
the 1995 revision of the OECD TPG discouraged the use of net profit meth-
ods on a general level.957 It should be noted that the 1994 discussion draft
discussed the implementation of the CPM as such.
Also at that time, the OECD had not yet envisioned restricting the method
to a purely transactional basis. This is evident by the fact that the draft
focuses on comparing the tested party as such to unrelated enterprises.958
In the spring of 1995, the OECD negotiations on the final version of the
1995 OECD TPG were at a standstill. Negotiations had halted on the issue
of whether the guidelines should accept the application of the CPM of the
newly enacted 1994 US regulations.959 Two fractions emerged, one of which
maintained that the CPM was not compatible with the arm’s length principle
of article 9 of the OECD MTC, on the basis that the method could be applied
to the overall profits of complex enterprises, resulting in rough comparisons.
The second fraction took the opposite view, i.e. that the CPM indeed was
compatible with article 9, as the method would be applied to segmented
groups of transactions to extract an uncontrolled net profit margin, which
955. 1993 task force report, para. 3.51.

956. 1994 task force report, para. 2.21.
957. 1994 OECD discussion draft, para. 173.
958. The core concern was that “where differences in the characteristics of the en-
terprises being compared have a material effect on the operating margins being used,
it would not be appropriate to apply the comparable profits method without making
adjustments for such differences”; see 1994 OECD discussion draft, para. 176.
959. See Taly (1996). Michel Taly was head of the tax policy department in the French
Ministry of Finance and Vice Chairman of the OECD Committee on Fiscal Affairs.
274
was considered to be aligned with the arm’s length principle. The views of
the second fraction prevailed. This is evident by the fact that the OECD
distinguished its one-sided net profit method from the CPM of the final
1994 US regulations by emphasizing a transactional character, as signalled
by the chosen name, i.e. the transactional net margin method.960
The final consensus text of the 1995 OECD TPG, as set out in paragraph
3.42, expressed the requirement that the TNMM should be applied only
to the profits of the tested party that were attributable to “particular” con-
trolled transactions.961 As a point of departure, it was deemed inappropriate
to apply the TNMM on a company-wide basis if the tested party engaged
in a variety of different controlled transactions that could not appropriately
be compared on an aggregated basis with those of an unrelated enterprise.
However, the 1995 OECD TPG allowed the aggregation of closely linked
or continuous controlled transactions.962 In scenarios in which this was
allowed, the aggregated transactions would be regarded as one coherent
transaction for pricing purposes, i.e. the aggregated transactions would be
priced as one transaction.
The author will not discuss the aggregation of controlled transactions further.
The reason for this is that it will normally be possible to disaggregate the op-
erating profits of a tested party down to a product or product group level, due
to access to the management accounting data of the tested party. Thus, the
issue of aggregation of the tested party’s transactions is largely a question of
whether it is appropriate to do so. Pursuant to the 1995 OECD TPG, this will
depend on whether the controlled transactions are strongly interrelated (pack-
age deals) or unsuitable for stand-alone pricing (continuous transactions).
The author will instead focus on comparability problems associated with ex-
tracting net profit margins from aggregated and blended third-party profits.
Paragraph 3.42 of the 1995 OECD TPG is of particular relevance for this
discussion. It stated that the following:
[W]hen analysing the transactions between the independent enterprises to the
extent they are needed, profits attributable to transactions that are not similar
to the controlled transactions under examination should be excluded from the
comparison.963 (Emphasis added)
960. See also Lodin (2009), at p. 430.

961. 1995 OECD TPG, para. 3.42.
962. 1995 OECD TPG, para. 1.42. See the similar text in the 2010 OECD TPG, para.
3.9.
963. Id.
275
On the basis of this wording, seen in light of the positions taken by the
OECD on this issue in the preceding task force reports and discussion
drafts, it must be concluded that the 1995 OECD TPG rejected the use of
profit margins extracted from blended profits.
The position of the 1995 OECD TPG, with respect to the extraction of
profit margins, may be summarized in table 8.1 (it is assumed that both
the tested and unrelated enterprise carry out a variety of transactions).964
Table 8.1
1995 OECD TPG

Total transaction portfolio Segregated
(tested party) transactions 964
(tested party)
Total transaction The profit margin The profit margin from
portfolio from company-wide a specific controlled
(unrelated party) controlled transactions is transaction is
benchmarked against a benchmarked against a
company-wide, third-party company-wide, third-party
profit margin (not allowed) profit margin (not allowed)
Segregated The profit margin The profit margin from
transactions from company-wide a specific controlled
benchmarked against a benchmarked against
company-wide, third-party the profit margin from a
profit margin (not allowed) specific CUT (the only
application of the TNMM
allowed under the 1995
OECD TPG)
964. The 1995 OECD TPG allowed aggregation of the tested party’s transactions
where the transactions formed a continuum or were closely linked, such as long-term
contracts for the supply of commodities, IP licence agreements and the pricing of close-
ly linked products (product lines); see 1995 OECD TPG, para. 1.42. See also OECD
TPG, para. 3.9, which contains an almost identical text. The aggregated controlled
transactions would then be regarded as one coherent transaction and would be priced as
such. The author disregards this aggregation option for the purposes of table 8.1, as the
issue here is to illustrate comparability problems associated with testing the net profit
of different transaction types that are incomparable. This is, to some extent at least,
distinguishable from aggregating transaction types that either are strongly interrelated
(package deals) or transactions that are not suitable for stand-alone pricing (continuous
transactions).
276
8.6.4. The 2006 comparability report
The segregation requirement of the 1995 OECD TPG proved to be prob-

lematic in practice. As part of its ongoing monitoring of the implemen-
tation of the 1995 OECD TPG, Working Party 6 identified that further
development was necessary with respect to the guidance on comparability
issues encountered when applying the OECD pricing methods, as well as
on the application of the transactional profit methods. The first issue was
addressed in the 2006 comparability report (the 2006 report).
As a way of background for the 2006 report, transfer pricing practice sub-
sequent to the 1995 OECD TPG applied net profit margins extracted from
blended third-party profits to a significant degree. The 2006 report found
that:
[…] practitioners … often try to overcome the difficulty of comparing the
gross or net profit margin arising from their particular controlled transac-
tions with the gross or net profit margins earned on aggregated transactions
performed by independent enterprises performing comparable activities.965
(Emphasis added)
The argument underlying these practices was typically that, in the absence
of better alternatives, net profit margins from blended profits should be al-
lowed, as long as the third-party enterprise from which the profit margin
was extracted was functionally comparable to the tested party.
Most of the business commentators who responded to the 2006 compara-

bility questionnaire issued in connection with the 2006 report stressed that
it was difficult to find third-party data on a transactional level. This was
simply because the information available in public financial statements was
not presented on a transactional level. Further, the 2006 report stated that
public financial statements provided no insight into the pricing decisions
or cost allocations pertaining to particular products that would explain the
profit margins reflected in the financial statements. Some commentators
stated that “there may not always be a relevant difference between third
party transactional data and third party company-level data … which …
should be acceptable unless transactional data are available and provide a
more reliable measure”.966
965. 2006 OECD comparability report, para. 7.

966. Id., at para. 8.
277
Practitioners tended to consider that the transactional focus of the OECD

TPG was sufficiently respected by ensuring that the comparisons were
made with third parties that exclusively or mainly undertook the type of
transaction under analysis.967
The point of departure in the 2006 report was that aggregation of third-
party transactions generally reduced the reliability of a comparability anal-
ysis. A transactional analysis would generally be preferred when transac-
tional data were available.968 The problem was that “in practice such data
may not be available”,969 and therefore:
OECD countries agreed that there was a need to recognize that in practice
there are circumstances in which aggregated comparable data should be used,
but only where its use provided the most reliable available evidence to inform
the arm’s length nature of transfers between associated enterprises.970
Thus, the stance of the OECD with respect to the segregation of third-
party transactions was softened in the 2006 report relative to the hard-
line approach taken by the 1995 OECD TPG. Working Party 6 was of the
opinion that a transactional analysis should be retained as “the starting
point” for the comparability analysis. It was not found desirable that the
OECD TPG should allow the use of highly aggregated data on third-party
profits as comparables,971 in particular, because such an approach could be
interpreted as an endorsement of using “default” comparables in typical
scenarios. The 2006 report emphasized that the transactional focus of the
OECD transfer pricing methods did not mean that transfer pricing should
be analysed at “the level of each individual sale”.972 Further, third-party
transactions could be aggregated along the lines allowed by the OECD
TPG for aggregating the transactions of the tested party.973
For instance, if the tested party was an LRD, it could be that, in the market
concerned, there were only a few unrelated enterprises with publicly avail-
able financial statements that undertook low-risk distribution functions.
These LRDs could then be chosen as comparables regardless of whether
967. Id., at para. 7.

968. Id., at para. 9.
969. Id.
970. Id.
971. Id., at para. 10.
972. Id., at para. 14.
973. See supra n. 964. Similarly, the 2006 report suggested that third-party transac-
tions could be aggregated where they “form an aggregate of comparable transactions
that are subject to the same economic factors and decision making process”; see 2006
OECD comparability report, para. 13.
278
their transactions were comparable to the controlled transaction. This was

viewed as posing a risk of transfer pricing assessments becoming more
formulaic.
The 2006 report recognized that it would normally be possible to segregate

the transactions of the tested party. Management accounting information
from the tested party would allow a breakdown of individual transactions.
The relevant question was whether aggregation of the individual transac-
tions of the tested party was appropriate.
The problem faced with respect to third-party transactions was the reverse.
Here, the issue was the lack of accounting data sufficiently detailed to seg-
regate the third-party transactions. The 2006 report recognized that the
relevant question was whether aggregated third-party accounting data at
all could be segregated. Normally, this would not be the case.
The report found that aggregated third-party profit data could be useful
when the third party engaged in a single type of transaction that was com-
parable to the controlled transaction.974 The problem was there when the
third party also carried out other transactions that had different economic
profiles than the controlled transaction and these other transactions had a
material impact on the total profits of the third party.975 In other words, the
relevant problem was the use of net profit margins extracted from blended
profits.
The 2006 report found that aggregated third-party profits could be used
where the transactions of the third party were subject to the “same or simi-
lar economic factors” and therefore could “be regarded as one transaction”
for comparability purposes.976 This is illustrated in an example in which
the controlled transaction is a related party’s importation of a particular
type of blender from a foreign group entity for local resale to third par-
ties.977 The resale price method is chosen as the transfer pricing method.978
An unrelated retailer is identified, but this retailer sells dozens of different

blenders from a range of suppliers. The example states that it is not neces-
sary to examine each individual blender transaction from each individual
supplier. One should instead exercise judgement and determine whether
974. 2006 OECD comparability report, Aggregation of Transactions, para. 17.

975. Id.
976. Id.
978. Thus, gross profit is the relevant measure of profitability.
279
the independent party’s purchase and sale of the blenders are subject to the
same or similar economic conditions and therefore can be regarded as one
transaction for comparability purposes.
An important point is that the example does not elaborate on what it means
by “similar economic conditions”. The author interprets the wording as
expressing whether it reasonably can be assumed that the gross profits from
resale of the different blenders are similar.
The author will add two points to this. First, the “similar economic condi-
tions” exception, even though well-intended, seems rather useless. The rea-
son for this is that, in practice, it will not be possible to determine whether
the different third-party transactions indeed are subject to similar eco-
nomic conditions. Product-level profit information is simply not publicly
available. Second, the exception could therefore (unintentionally) become
a safe harbour for taxpayers. It will be difficult in practice to disprove a
taxpayer assertion that the different transactions are subject to similar eco-
nomic conditions; the actual data that will verify or reject the assertion is
unavailable.
A twist to the example is provided, where the unrelated distributor now

sells, in addition to blenders, toasters.979 Thus, the profits of the unrelated
distributors are blended. No CUT exists, and the financial statements of
the unrelated distributor do not provide information on gross profits. The
TNMM is therefore selected as the appropriate pricing method. The ques-
tion is whether a net profit margin extracted from the blended profits of the
unrelated distributor can be used under the TNMM.
The report states that this will “not necessarily” be the case.980 The blended
net profit margin can only be used if it can reasonably be assumed either
that (i) the five comparability factors are satisfied, in particular, that the
incomparable toaster transactions have “similar functional analysis and
have similar economic characteristics” as the sale of blenders; or (ii) the
toaster transactions “represent a negligible portion of the total” profits of
the unrelated distributor.
The author’s interpretation of the first criterion is that the net profits gener-
ated by the incomparable third-party transactions must be similar to the
net profits generated by the comparable transactions. Otherwise, the profits

980. Id.
280
from the incomparable transactions would cause the total net profits of
the enterprise to deviate from the profits of the comparable transactions.
Where to draw the line with respect to which profits are “similar” is not ad-
dressed in the 2006 report. With respect to the second alternative criterion,
the author’s interpretation is that blended net profit margins may be used if
the net profit margins of the incomparable transactions have only margin-
ally affected the total net profits of the unrelated distributor.
It will, in practice, be difficult or impossible to verify whether the two alter-

native criteria are fulfilled. The 2006 report also admits this.981 Thus, there
is an obvious risk that this will provide a safe harbour for taxpayers, as it
will be very challenging to disprove a taxpayer assertion that the blended
net profit margin applied under the TNMM is valid, either because the
incomparable transactions that are included in the blended profits have a
net profit margin similar to the comparable transactions or that the incom-
parable transactions only have a negligible impact on the blended profits.
In summary, it is the author’s view that the criteria proposed by the 2006
report for determining whether aggregated third-party accounting infor-
mation can be used as comparables under the OECD pricing methods were
theoretically well-founded. The criteria sought to ensure comparability be-
tween the controlled and uncontrolled transactions. Nevertheless, the main
goal of the 2006 report was to address the practical problems encountered
in applying the pricing methods. In the author’s view, the report failed to
find a useful solution to this. The wording of the 2006 report on this issue is
also decidedly soft, in that the aggregated company-level data “might not”
or will “not necessarily” provide satisfactory comparable data, indicating
that the criteria were not absolute in any case.
The 2006 report did, however, draw a bright line against comparing the
aggregated net profits of the tested party and an unrelated party solely
because they operate in the same industry sector. Such a comparison was
found unacceptable because it would not take account of the differences
between the tested party and the independent party with respect to intangi-
bles, risks and organizational models.
Also, such a comparison would lie close to using industry averages as com-
parables. The 2006 report found it inconsistent with the arm’s length prin-
ciple to make company-wide comparisons of all distributors of electronic
goods without giving proper regard to whether the transactions performed
981. Id., at para. 18: “[It] might be very difficult to demonstrate in practice.”
281
by these companies satisfied the comparability factors as well as the ag-

gregation criteria. Thus, the 2006 report rejected typical database-search-
based applications of the TNMM in which no comparability adjustments
would be made to the results from the search.
The positions of the 2006 report can be summarized in table 8.2.982 983
Table 8.2
2006 comparability discussion draft

Total transaction portfolio Segregated transactions982
(tested party) (tested party)
Total The profit margin from The profit margin from a
transaction company-wide controlled specific controlled transaction
portfolio transactions is benchmarked is benchmarked against a
(unrelated against a company-wide, company-wide, third-party
party) third-party profit margin (not profit margin (allowed if (i)
allowed) the transaction portfolio of
the third party (including
both comparable and
incomparable transactions)
have similar profit margins;
or (ii) the profits from the
incomparable transactions
only have a negligible
impact on the total profits of
the third party) 982
Segregated The profit margin from The profit margin from
transactions company-wide controlled a specific controlled
(unrelated transactions is benchmarked transaction is benchmarked
party) against a company-wide, against the profit margin from
third-party profit margin (not a specific CUT (allowed)
allowed)
8.6.5. The 2008 discussion draft
The work of the OECD on the application of the transactional profit meth-
ods resulted in a 2008 discussion draft (the 2008 report).984 It also ad-
982. See supra n. 964.

983. See 2006 OECD comparability report, paras. 13, 15, 17 and 18.
984. Transactional profit methods, Discussion draft for public comment (OECD
2008) [hereinafter 2008 OECD discussion draft].
282
dressed problems pertaining to the use of aggregated third-party account-

ing data as comparables under the pricing methods.
The 2008 report recognized that, due to the lack of publicly available and
disaggregated third-party accounting information, the segmentation of the
tested party’s profits would be less problematic than the segmentation of
a third party’s profits, as long as the tested party’s accounting system was
able to track profitability on an appropriately segmented basis.985 Similarly
to the 2006 report, it admitted that there could be cases where aggregated
third-party transactions could provide valid comparables for a particular,
non-aggregated controlled transaction.
A new tone was evident in the 2008 report. It drew the conclusion that, due
to the lack of public third-party accounting data that could allow profit mar-
gins to be determined for specific transactions or groups of transactions,
there needed to be sufficient comparability between the “economically sig-
nificant functions of the tested party and of the third party comparables”.986
(Emphasis added)
The functions performed by the third party in its total operations had to
be closely aligned to the functions performed by the tested party with re-
spect to the controlled transaction in order to allow a net profit margin
extracted from the blended third-party profits to be used to determine an
arm’s length outcome.
The conclusion of the 2008 report was that when it was impossible in
practice to achieve the transactional level set out as the ideal in the OECD
TPG, it would still be important to achieve the highest level of compara-
bility possible through making suitable adjustments based on the avail-
able evidence.987 Strong concerns had been expressed by several countries
in relation to the segregation of transactions, with respect to both the
tested party and third parties. The draft therefore concluded that para-
graph 3.42 of the 1995 OECD TPG should be amended in accordance
with the views expressed in the 2006 and 2008 reports.988 This conclusion
was ambiguous, as the proposed solutions in the 2006 and 2008 reports
were not necessarily reconcilable. The author will revert to this issue in
section 8.6.6.
985. 2008 OECD discussion draft, para. 110.

986. Id., at para. 115.
987. Id., at para. 117.
988. Id., at para. 120.
283
8.6.6. The final 2010 OECD TPG on the use of aggregated

third-party profits as comparables
The final 2010 OECD TPG express three different norms for determining
whether it is allowed to use aggregated third-party profits as compara-
bles under the OECD pricing methods. At least as a point of departure,
these norms are not immediately reconcilable. The fragmented and seem-
ingly conflicting character of the OECD TPG on this issue must be seen
in light of the differing opinions of the OECD member countries. Ap-
parently, there are difficulties in reaching a clear consensus with respect
to the use of aggregated third-party profits as comparables. In sections
8.6.6.2.-8.6.6.6., the author will endeavour to interpret these three norms
and reconcile his interpretations in order to provide a coherent norm for
the use of aggregated third-party profits as comparables under article 9 of
the OECD MTC.
8.6.6.2. The first norm: Aggregated third-party profits may be used

as comparables as long as they are the result of “similar”
third-party transactions
The 1995 consensus text in paragraph 3.42 of the 1995 OECD TPG on the
use of aggregated third-party accounting data as comparables expressed
that “when analysing the transactions between the independent enterprises
to the extent they are needed, profits attributable to transactions that are
not similar to the controlled transactions under examination should be ex-
cluded from the comparison”.
In the discussion in section 8.6.3., the author concluded that this wording
rejected the use of profit margins extracted from blended profits.989
In September 2009, the OECD published proposed revisions to chapters

I-III of the OECD TPG.990 The revision kept the text of paragraph 3.42 of
the 1995 OECD TPG intact. The author finds this rather surprising, given
the conclusions of the 2006 and 2008 reports. The original 1995 wording
on the use of aggregated third-party accounting data is now incorporated
989. See the discussion of the 1995 OECD TPG text in sec. 8.6.3.
990. Proposed revision of chapters 1-3 of the transfer pricing guidelines (OECD
2009) [hereinafter 2009 OECD proposed revisions].
284
into paragraph 2.79 of the 2010 OECD TPG. Paragraph 2.79 is placed in
chapter II of the OECD TPG on transfer pricing methods as part of the
TNMM guidance.
Thus, pursuant to paragraph 2.79, the use of aggregated third-party profit

data as comparables should be rejected if the total profits are influenced
by the profits from transactions that are not comparable to the controlled
transaction.
8.6.6.3. The second norm: Aggregated third-party profits may be

used as comparables as long as they are not the result of
“materially different” third-party transactions
The 2009 revision draft added a new last sentence to paragraph 3.42 of the
1995 OECD TPG. The sentence read: “[S]ee paragraphs 3.9-3.12 on the
evaluation of a taxpayer’s separate and combined transactions and para-
graph 3.36 on the use of non-transactional third party data”.991
This sentence is now incorporated into paragraph 2.79 of the 2010 OECD
TPG. The 2009 proposed paragraph 3.36 addressed whether aggregated
third-party accounting information could provide reasonably reliable com-
parables.992 The 2009 report draft text for paragraph 3.36 was incorporated
into paragraph 3.37 of the 2010 OECD TPG without alterations. Paragraph
3.37 is placed in chapter III of the OECD TPG on comparability analyses
as part of the guidance on comparable uncontrolled transactions.993
991. 2009 OECD proposed revisions, para. 2.143.

992. Thus, the proposed text did not address the more impractical issue of whether
the tested party’s aggregated company-wide transactions (in cases where it carried out
more than one type of transaction) could be reliably compared to the results of aggre-
gated third-party transactions. This was presumably because, in most cases, it will be
feasible to segregate the tested party’s transactions, due to the insight of the tax authori-
ties into the management accounting documents of the taxpayer.
993. The OECD TPG also state that company-wide, third-party data may provide bet-
ter comparables than segmented third-party data in certain circumstances, for instance,
where the activities underlying the company-wide, third-party profit data correspond
to the set of controlled transactions. This is difficult to accept. The whole point of seg-
regating third-party transactions will be to separate third-party transactions that are
comparable to the controlled transaction from those that are not. If such segregation is
carried out, resulting in the identification of third-party comparables, it seems difficult
to accept that third-party, company-wide profit data nevertheless should be preferred,
regardless of whether the purported degree of functional similarity is largest with re-
spect to the company-wide or segregated third-party transactions.
285
With respect to the use of aggregated third-party accounting information

as comparables, the 2009 text took the stance that this would depend in
particular on: “whether the third party performs a range of materially dif-
ferent transactions”.994
It is not entirely clear what is meant by “materially different transactions”.

The wording itself may indicate that if unrelated enterprises carry out sig-
nificantly different types of transactions, it will not be appropriate to use
the aggregated profit data as comparables. This interpretation entails that
the requirement is relaxed as compared to the 1995 text, which demanded
the exclusion of third-party transactions as long as they were not “similar”
to the controlled transaction.995
Seen in light of the discussion in the 2006 report, the author finds it rea-
sonably clear that the guidance seeks to restrict the use of blended profit
margins that are the result of third-party transactions that contribute sig-
nificantly different net margins to the total profits. Of course, the decisive
question is how large the deviation must be between the net profit margins
contributed by the different third-party transactions before they are con-
sidered “materially different”. That question is left unanswered in the 2009
text.
In practice, the answer will also be irrelevant. The reason for this is that
even if a specific threshold were set, it would not be possible to determine
whether it was reached, given the limitations of publicly available account-
ing data.996
Thus, the author’s conclusion is that the wording of paragraph 3.37 of the
OECD TPG rejects the use of blended third-party profits as comparables,

995. See 2010 OECD TPG, para. 2.79, which reiterates the original 1995 OECD TPG
(para. 3.42) text on this point.
996. Further, the guidance does not clearly state that aggregated third-party profit
data cannot be used as comparables, even if the data stems from third-party transac-
tions that are materially different to the controlled transaction. On the other side, it is
not easy to see that there could be any other purpose to the language than to restrict the
use of aggregated third-party profit data as comparables. The language goes on to state
the rather obvious fact that where segmented third-party data is available, it can provide
better comparables than company-wide, non-segmented, third-party profit data, due to
its more transactional focus. The guidance does, however, recognize that segmented
data may trigger issues in relation to the allocation of business expenses among the
segmented groups of transactions.
286
to the extent that the third-party transactions contribute “materially differ-

ent” net profit margins to the total blended profits.
8.6.6.4. The third norm: Aggregated third-party profits may be

used as comparables if the total functions performed by
the third party are closely aligned with the functions
performed by the tested party with respect to
the controlled transaction
The 2009 report also introduced a new paragraph 2.144, which was in-
corporated into the 2010 OECD TPG in paragraph 2.103 without material
changes. Paragraph 2.103 addresses the use of aggregated third-party ac-
counting information as comparables. Paragraph 2.103 is placed in chapter
II of the OECD TPG on transfer pricing methods, as part of the TNMM
guidance.
Paragraph 2.103 specifically states that comparability issues connected to

the use of aggregated third-party accounting data are not particular to the
TNMM. Rather, the problem in practice is more acute under the TNMM as
compared to the gross profit methods, due to the heavy reliance on external
comparables under the TNMM.997 The paragraph recognizes that there is
often insufficient accounting data publicly available to allow for third-par-
ty net profit margins to be extracted on a transactional level. In light of this:
the functions performed by the third party in its total operations must be
closely aligned to those functions performed by the tested party with respect
to its controlled transactions in order to allow the former to be used to deter-
mine an arm’s length outcome for the latter.998 (Emphasis added)
This wording is materially the same as that of the 2008 report.999 It is dif-
ficult to interpret it in any other way than approving the use of aggregated
third-party accounting data on net profits as a comparable in a controlled
transaction when the total functions performed by the third party are close-
ly similar to the functions performed by the tested party with respect to its
controlled transaction. This regardless of whether, as well as the extent to
which, the aggregated third-party profits are blended.
997. This explanation is not convincing. The gross profit methods place the same
amount of reliance on external comparables as the TNMM. The author therefore as-
sumes that the added actuality of the problem under the TNMM is simply due to the
fact that the TNMM is the most commonly applied transfer pricing method.
999. See 2008 OECD discussion draft, para. 115.
287
This interpretation is supported by the realistic acknowledgement in para-

graph 2.103, where it is stated that in the case that it is impossible in prac-
tice to achieve the transactional level set out as the ideal in the OECD TPG,
it is still important “to try” to find the most reliable comparables through
making adjustments based on the evidence that is available.1000 In compari-
son, the proposed 2009 revision text plainly stated that it was “important
to achieve the highest level of comparability”.1001
The final 2010 OECD TPG text therefore strikes a more liberal tone than
the 2009 proposed text. This again lends support to the interpretation that
the OECD TPG allow the use of non-segregated third-party profit data.
In the author’s view, the consensus text in paragraph 2.103 of the OECD
TPG is fundamentally important, and something of an earthquake in the
landscape of the OECD TPG. It indeed allows the use of non-transactional
third-party profit data as comparables under article 9 of the OECD MTC.
Also significant is the fact that the guidance should not be regarded as
limited to the TNMM, but should apply also to the other pricing methods.
The rule entails, for instance, that the total net profit margins extracted
from third-party manufacturers and distributors that produce or distribute
a range of different products may, where it is not possible to segregate the
data, be used as comparables against a specific controlled transaction car-
ried out by the tested party, given that there is a strong functional similarity
between the tested party and the selected third-party enterprises.
No guidance is provided as to what functions should be regarded as simi-

lar. Presumably, this will often be self-evident. In complex, unrelated en-
terprises that perform a range of different functions, however, it will not
necessarily be easy to determine whether the quantitative and qualitative
aspects of such functions are similar to the functions performed by the
tested party with respect to the controlled functions. The relevant thresh-
old under paragraph 2.103 is that the total third-party functions must be
“closely aligned” with the controlled functions.
The language itself could indicate that the relevant third-party functions
must almost be identical to the ones performed by the tested party with re-
spect to the controlled transaction. The author objects to this interpretation.
In his view, an overly detailed assessment on this point may contribute to
continued ambiguity in the OECD TPG on the important issue of the use of
1000. OECD TPG, para. 2.103.

288
third-party account data as comparables. Instead, the language should be

interpreted in light of the overall purpose of the TNMM, which is to allo-
cate a normal market return to the tested party. The single most important
aspect will therefore be to ensure that the third party only performs routine
functions, that it does not own any unique and valuable intangibles and that
it only incurs normal levels of risk. If the third party is closely similar to
the tested party on these parameters, the use of its blended profits to extract
a profit margin to be used under the OECD pricing methods should be ac-
cepted pursuant to paragraph 2.103.
Further, as long as the selected third-party enterprise is similar to the

tested party with respect to functions, assets and risks, the negative ef-
fects on comparability of using different blended profits should be limited.
The tested party will often be a contract manufacturer or an LRD. Simi-
lar third-party enterprises will typically only earn a normal return, due to
their generic functions, routine assets and limited risks.
Thus, it may, to some extent, be argued that even if the profit margins on
the different products or services sold vary, the variance may be limited,
as all transactions may be assumed to only yield a normal return. This
supports the argument that the “closely aligned” guidance should be in-
terpreted to mean that the third party must be similar to the tested party
with respect to functions, assets and risks, as opposed to an overly-detailed
focus on the performance of particular functions.
8.6.6.5. Harmonizing the three norms through interpretation
In summary, the following stances are taken in the 2010 OECD TPG with
respect to the use of aggregated third-party profits as comparables:
− the first norm rejects such comparables if some of the third-party
transactions are “not similar to the controlled transactions” (paragraph
2.79);
− the second norm rejects such comparables if the third party performs
a “range of materially different transactions” (paragraph 3.37); and
− the third norm rejects such comparables if it cannot be concluded that
the “functions performed by the third party in its total operations [are]
closely aligned with the functions performed by the tested party with
respect to its controlled transactions” (paragraph 2.103).
Clearly, there seems to be a conflict between the three norms. This ambi-
guity must be seen in light of the fact that there are differing views among
289
the OECD member countries on whether profit margins extracted from

blended profits may be used under the OECD pricing methods. It was
therefore likely difficult to achieve a clear consensus.
On the one hand, some countries likely acknowledged the dire need to
adapt the OECD TPG to what is clearly common transfer pricing prac-
tice internationally, as well as to recognize that, in the vast majority of
cases, there simply will not be sufficiently detailed accounting data avail-
able to segregate the transactions of third parties. This view is reflected
in paragraph 2.103. Other countries likely took the more principal (and
rather theoretical) view that using company-wide data is inconsistent with
the traditional transactional focus of the OECD TPG and that comparables
that do not succumb to these requirements should be rejected, as seemingly
expressed in paragraph 3.37.
It is of importance to determine a coherent interpretation of the OECD

TPG with respect to the acceptability of using aggregated third-party prof-
its. Such profits will, presumably, in almost every case, be blended, in the
sense that the total profits consist of profits from third-party transactions
that are both comparable and incomparable to the controlled transaction.
First of all, paragraph 2.79 (“not similar”) is a remnant from the 1995 con-
sensus text, which was directly carried over to the 2010 OECD TPG. This
is surprising, given the considerable development both in global transfer
pricing practice since 1995 and the OECD movement on this issue, as re-
flected in the 2006 and 2008 reports and in paragraphs 3.37 and 2.103 of
the OECD TPG. It was not commented as to why this text was carried over
to the 2010 OECD TPG.
Further, the text is difficult to reconcile with paragraph 3.37. Both norms
pertain to the composition of the third-party transactions. Paragraph 2.79
draws the line at “not similar”, and paragraph 3.37 at “materially differ-
ent”. Thus, these norms address the same object, but the paragraph 3.37
norm is more relaxed than its predecessor norm in paragraph 2.79. The
author does not find these two norms to be reconcilable. His conclusion is
that the predecessor norm in paragraph 2.79 should be rejected on the basis
of the lex posterior principle. Simply put, the “materially different” norm
expressed in paragraph 3.37 of the 2010 OECD TPG would not have any
effect if the 1995 “not similar” norm in paragraph 2.79 were to apply. Such
a result could not possibly have been the intention when the 2010 OECD
TPG text was agreed upon. In this section, the author will therefore disre-
gard the “not similar” norm in paragraph 2.79.
290
The question is then whether the use of aggregated third-party net profit
data is allowed when that data reflects materially different transactions
(blended profits), but the total functions carried out by the third party are
closely similar to the functions carried out by the tested party with respect
to the controlled transaction, i.e. whether blended profit comparables are
allowed in situations in which the guidance both in paragraphs 3.37 and
2.103 is triggered. The author refers to his interpretation of paragraphs 3.37
and 2.103 in sections 8.6.6.3. and 8.6.6.4., respectively. The question now is
how the paragraphs relate to each other.
First, for the purpose of deducting a coherent rule for the use of aggregated
third-party data as comparables, it is, in the author’s view, important to rec-
ognize that the guidance contained in paragraph 3.37 is ambiguous, while
paragraph 2.103 is relatively clear. It does not seem meaningful to give
paragraph 3.37 priority over paragraph 2.103, as that would entail that one
replaces clear guidance with ambiguous guidance.
Second, it is difficult to disregard the fact that the guidance in paragraph

3.37 will not be possible to apply in real-world transfer pricing cases. The
extent to which third-party transactions are materially different will be un-
known, due to insufficiently detailed publicly available accounting data. In
contrast, it may be relatively easy to ascertain whether the functions, assets
and risks of the independent enterprise are similar to those of the tested
party. In the author’s view, the “materially different transactions” guidance
in paragraph 3.37 is not useful, while the “closely aligned” functions guid-
ance in paragraph 2.103 is. This is a clear argument for giving paragraph
2.103 priority over paragraph 3.37.
Third, it cannot with certainty be concluded that the guidance in paragraph

3.37 on materially different transactions is wholly incompatible with the
guidance in paragraph 2.103 on similar functions due to the ambiguous
language of the former. Paragraph 3.37 makes the use of aggregated third-
party profit data contingent on whether the data can provide “reliable”
comparables. Paragraph 3.37 finds that this will depend on whether the
third party performs a range of “materially different transactions”. Thus,
the problem is seen as a reliability issue. Of course, comparability is not
black and white; the extent of reliability will vary depending on the com-
parability factors.
A plausible interpretation of paragraph 3.37 therefore seems to be that ag-

gregated third-party data may be used, even if such data encompasses ma-
terially different transactions. However, as the degree of materially differ-
291
ent transactions increases (i.e. as profits become more blended), the degree
of reliability associated with using the aggregated data as a comparable
will decrease correspondingly. If this interpretation of paragraph 3.37 is
adopted, one might end up in a situation where the degree of comparability,
and thus reliability, associated with aggregated profit data as a comparable
is questionable, but in principle, acceptable under paragraph 3.37.
Fourth, the question is whether there are any other transfer pricing alterna-
tives available that will yield a more comparable and reliable result than
basing the pricing on a profit margin extracted from blended profits. The
author does not find this likely, as the problem of blended third-party net
profits pertains to all OECD transfer pricing methods, apart from the CUT
method.
Should it be, however, that a superior third-party comparable for which

it is possible to ascertain profits on a transactional basis does exist, the
best-method rule will dictate that a method is chosen that selects the best
comparable available.1002 A more realistic scenario seems to be the one
described in paragraph 2.103 that the only third-party data available is ag-
gregated net profit information. In this situation, there simply will be no
alternative comparables that are more reliable. This is important, as the
paragraph 3.37 argument of reduced reliability becomes less important if
there are no better alternative comparables upon which to base the transfer
pricing assessment. This point also supports giving paragraph 2.103 prior-
ity over paragraph 3.37.
8.6.6.6. Conclusion
In light of the discussions in sections 8.6.6.2.-8.6.6.5., the author adopts

the following interpretation of the OECD TPG with respect to the use of
aggregated third-party profit data as comparables.
As the point of departure, the OECD TPG allow the use of aggregated
third-party profit data, even if those data encompass transactions that are
materially different. If the profit data are blended, however, the reliability
of the transfer pricing analysis will be reduced as compared to the reliabil-
ity that would have been the result had the third-party profit data only en-
compassed segregated transactions that were comparable to the controlled
transaction.
1002. See OECD TPG, para. 2.2.
292
The decisive threshold for whether blended third-party profit data may
be used as comparables is the paragraph 2.103 instruction that the func-
tions performed by the third party in its total operations must be closely
aligned with the functions performed by the tested party with respect to its
controlled transaction. Under this interpretation, there will be coherence
between the guidance in paragraph 3.37 on “materially different transac-
tions” and the guidance in paragraph 2.103 on functions “closely aligned”.
Because the guidance in paragraph 2.103 pursuant to the above interpreta-

tion will be decisive with respect to whether aggregated third-party profit
data may be used as comparables under the OECD pricing methods, it is,
in the author’s view, important that the sparse guidance of the paragraph is
interpreted in a purposeful manner. In particular, the single most important
aspect will be to ensure that the third party only performs routine func-
tions, that it does not own any unique and valuable intangibles and that it
only incurs normal levels of risk. If the unrelated enterprise satisfies these
criteria, its blended profits should be allowed to be used as comparables.
The stance of the current OECD TPG with respect to aggregation of trans-
actions may be summarized in table 8.3 (assuming that both the tested and
unrelated enterprises carry out a variety of transactions).1003
Table 8.3
2010 OECD TPG

Total transaction portfolio Segregated
(tested party) transactions1003
(tested party)
Total transaction The profit margin The profit margin from
portfolio from company-wide a specific controlled
benchmarked against a benchmarked against a
company-wide, third-party company-wide, third-party
profit margin (not allowed) profit margin (allowed)
Segregated The profit margin The profit margin from
transactions from company-wide a specific controlled
benchmarked against a benchmarked against
company-wide, third-party the profit margin from a
profit margin (not allowed) specific CUT (allowed)
1003. See supra n. 964.
293
8.6.7. Applying the 2010 OECD TPG rule to the golf ball

example
The author will illustrate the implications of the interpretation of the

OECD TPG in section 8.6.6.6. by reverting to the golf ball example (see
Example 3 in section 8.6.2.).
The tested party realizes a 15% net margin on its controlled golf ball sales.
The independent enterprise realizes the following margins: the net profit
margin from the sale of golf balls is 23.7%, which contributes 54.1% to the
total blended profits of the independent enterprise. The remaining blend
comes from the sale of tennis balls at a net profit margin of 12.5%, which
contributes 25.1% to the total net margin, and from the sale of basketballs
at a net profit margin of 1.4%, which contributes 20.8% to the total net
profits of the enterprise.
Thus, the aggregated third-party net profits in this example are blended
and consist of contributions from clearly materially different transactions.
The comparable transaction yields net profits of 23.7%, while the two in-
comparable transactions yield 12.5% and 1.4%. The impact of the incom-
parable transactions on the blended profits is significant, as the difference
between the net profit of the comparable transaction of 23.7% and the over-
all blended profits of 16.2% is 7.5%. Pursuant to the author’s interpretation
of the OECD TPG in section 8.6.6.6., neither the fact that the aggregated
third-party profits are tainted by the incomparable transactions nor the de-
gree to which they are are relevant for determining the admissibility of
using the profits as a comparable under the OECD pricing methods.
This is an important recognition, because the actual result of using this net
profit margin will be erroneous. The true arm’s length net profit margin
in this case is 23.7%. Nevertheless, the end result of applying the TNMM
using the blended net profit margin is that the return position of the tested
party of 15% is measured against the blended “arm’s length” profit margin
of 16.2%. This result is regrettable, as it does not yield a true arm’s length
allocation of income.
Of course, in the real world, it will not be possible to ascertain the degree
to which the result of using a blended net profit margin deviates from the
true arm’s length net profit margin. However, one can logically be reason-
ably sure that a blended profit margin will not reflect the precise and true
arm’s length net profit margins.
294
Has the TNMM converged towards the 1988 White Paper BALRM?
It is important to see the reduced reliability of using blended profit margins

in the bigger picture. The use of such margins is relevant mainly under the
one-sided OECD pricing methods. With respect to the transfer pricing of
intangibles, the TNMM is, in practice, the only method for which the issue
is normally relevant. The purpose of the one-sided methods in general, and
the TNMM in particular, is to allocate a normal market return to the tested
party. As long as a proper functional analysis is carried out and it is con-
cluded with a high degree of certainty that the independent enterprise does
not perform non-routine functions, own unique and valuable intangibles
or incur extraordinary risks, the transactions of the independent enterprise
will likely yield a normal return. Thus, for the purpose of ensuring a proper
allocation of residual profits among jurisdictions, it does not seem to be a
detrimental problem that there is some uncertainty connected to ascertain-
ing an arm’s length profit margin that provides a normal return. The simple
reason for this is that a normal return is, in itself, limited to an average or
typical market return. This will narrow the potential interval for variance
significantly.
Ultimately, transfer pricing is not a precise science. Even though the use
of net profit margins extracted from third-party blended profits undoubt-
edly will entail a great deal of uncertainty with respect to their predictive
value as indicators of arm’s length results, there cannot, in the author’s
view, be any doubt that such margins will, regardless, allocate a normal
return to the tested party. In practice, that normal return will be impre-
cise. In the bigger picture, however, the most important result is that the
tested party is not allocated any intangible super profits. The interpreta-
tion of the OECD TPG in section 8.6.6.6., at least, seems to ensure this
objective.
8.7. Has the TNMM converged towards the 1988

White Paper BALRM?
The question is whether the material content of the TNMM, as understood
in the above interpretation (see section 8.6.6.6.) and in current transfer
pricing practice, has converged towards the content of the BALRM and
CPM. If so, that would be ironic, in light of the significant historical resist-
ance from factions within the OECD against the US approach.
The BALRM is rooted in microeconomic theory, which professed that, in

competitive markets, where the factors of production are homogeneous and
mobile across business sectors, competition would cause above-normal re-
295
turns to become zero in the long run.1004 A fundamental conclusion was

that there would be equality between the revenues of an enterprise and
the sum of the market returns that could be earned by all of the factors of
production employed by the enterprise.1005 The White Paper drew the con-
clusion that this relationship could be used to determine the allocation of
income among related parties by identifying and valuing the factors of pro-
duction owned by a related entity and assigning the market return that each
factor could have earned on its best alternative use in the marketplace.1006
The equality between revenue and production factor market returns could
not be assumed to be sustained in two specific and practical scenarios,
namely in (i) monopoly situations; and (ii) the presence of unique IP.1007
The existence of super profits could not be ruled out in these scenarios, due
to lack of competition.
The first step under the BALRM was to identify functions that utilized
generic factors of production (as opposed to unique IP).1008 The second step
was to assign income to these functions by way of extracting normal mar-
ket return profit data from third-party enterprises that were functionally
similar to, and owned similar assets as, the tested party, thereby indirectly
also allocating the residual profits from unique and valuable IP to the other
party to the controlled transaction.1009
Transfer pricing practice, as well as paragraph 2.103 of the 2010 OECD

TPG, has sought to limit the negative consequences of using blended profit
1004. White Paper (Notice 88-123), p. 83. See also Lipsey et al. (1981), at pp. 229-231;
and Henderson et al. (1980), at pp. 107-110. This did not entail that a related party
that operated in a competitive market using standard and mobile factors of production
would earn a taxable income of zero, but rather that the entity would earn just a normal
rate of return and not residual profits.
1005. White Paper (Notice 88-123), p. 84.
1006. Id.
1007. Id., at p. 85.
1008. Id., at p. 96.
1009. Under the BALRM, assets were divided into cash working capital and other op
erating assets. Working capital was assigned its actual return, while the return on op-
erating assets was identified or estimated. The BALRM was presented in the White
Paper (Notice 88-123) primarily as a method that measured returns on operating assets,
but the model was also open to using other financial ratios; see White Paper, p. 97. The
Berry ratio (see sec. 8.4.2. and supra n. 927), for instance, could be useful in measuring
the returns on service activities and in other situations in which operating assets are
difficult to measure consistently. Even though the White Paper did not go into details
on how the BALRM would identify the relevant market returns, there would not be any
alternatives to using transaction-based measures of return (e.g. the rate of return on
operating assets or other financial indicators).
296
Is reduced transactional comparability under the TNMM
a significant problem?
margins under the TNMM by requiring functional comparability between

the tested party and the selected uncontrolled enterprises.1010 Thus, the jus-
tification for asserting that the selected uncontrolled enterprises are com-
parable to the tested party will not, in practice, be similarities between the
controlled and uncontrolled transactions themselves, but rather between
the functions performed, assets used and risks incurred by the tested and
independent enterprises.
On this basis, it must be concluded that the TNMM, as interpreted in sec-

tion 8.6.6.6. and applied in practice, displays fundamental similarities to
the BALRM of the 1988 White Paper. There is little basis left on which to
contrast the TNMM against the BALRM and the CPM.
8.8. Is reduced transactional comparability under

the TNMM a significant problem?
A typical application of the TNMM will entail uncertainty. It is unlikely
that a blended profit margin extracted from independent enterprises (that
are comparable to the tested party with respect to functions, assets and
risks) will be fully comparable to the profit margin of the controlled trans-
action. The question here is thus whether this poses a significant problem
in reality.
First, the author presumes that the use of blended net profit margins will
normally allocate an income to the tested party that is roughly equal to an
arm’s length income. In other words, the result of applying such margins
will likely not significantly distort the allocation of income between related
parties relative to the “true” arm’s length allocation of income that would
have been the result had there been full transactional comparability be-
tween the extracted third-party profit margin and the controlled profit mar-
gin. The reason for this is that the TNMM can only be applied where the
tested party is a “simple” entity that contributes only routine value chain
inputs. Such entities will, due to competition from third parties, only yield
normal market returns on all of their transactions. The extracted blended
profit margin will therefore likely only vary within a limited normal re-
turn range. Measurement errors will also be mitigated by the arm’s length
1010. In addition to functional comparability, unrelated comparable enterprises are,

in practice, typically also selected on the basis of similarity in the areas of business,
assets, size, employees, etc.
297
range (which was the reason why it was introduced concurrently with the
CPM).1011
Second, one should surrender the ideal of a perfectly comparable third-par-

ty transaction.1012 As discussed above, the reason why blended profits are
used under the TNMM is the simple fact that public third-party financial
statements do not provide enough disaggregated information to separate
gross or net profit margins on a product (or product group) level. Even if
the necessary management accounting information to carry out such seg-
regation were publicly available, it would be difficult to ensure full com-
parability between the tested and uncontrolled transactions because the
gross and net profit margins for single products (or product groups) are not
“objective” sizes. Such margins depend on a range of highly discretionary
management assessments for which it may be difficult to carry out reli-
able comparability adjustments (e.g. allocation of variable and fixed costs
among different products).1013
1011. See the analysis of the “arm’s length range” concept in sec. 6.5. It should be
noted that the OECD TPG hesitate to accept that the arm’s length range may mitigate
the potential level of inaccuracy produced by the TNMM, as it may not “account for
situations where a taxpayer’s profits are increased or reduced by a factor unique to that
taxpayer” (see OECD TPG, para. 2.73). The stated rationale behind this position is that
the arm’s length range may then not include points representing the profits of independ-
ent enterprises that are affected in a similar manner by a unique factor. The author
finds this unclear. First, if the tested party is in possession of unique functions or assets
(e.g. uniquely skilled research personnel or unique IP), it is clear that the TNMM (or
any other one-sided transfer pricing method, for that matter) should not be applied to
allocate profits to the tested party in the first place. Second, the arm’s length range is a
statistical tool, designed specifically to alleviate problems associated with information
on the transactions of unrelated enterprises being insufficiently complete to ensure that
there are no material differences between them and the controlled transactions of the
tested party. It is difficult, in this light, to see why the arm’s length range should not
offer the same comfort in this situation.
1012. See also Luckhaupt et al. (2011), at p. 116, where it is emphasized that there is a
large degree of discretionary assessment associated with applying the OECD transfer
pricing methods.
1013. Two types of comparability adjustments are, in practice, relevant with respect
to third-party profit data, i.e. (i) accounting adjustments; and (ii) “economic” adjust-
ments. Accounting adjustments do not affect the profit-and-loss statement, and thus the
net profit, of the unrelated enterprise of which the profit data is being adjusted, while
economic adjustments do. Accounting adjustments are performed in order to ensure
that the tested party and the unrelated enterprise has accounted for income and ex-
pense items in a similar manner. Typical accounting adjustments include reclassifying
expenses among costs of goods sold and operating expenses. Accounting adjustments
are irrelevant under net profit methods. Economic adjustments include adjusting the
principles for accounting for inventory (LIFO/FIFO), reclassifying from expensing to
capitalizing costs, working capital adjustments, adjustments to assets (e.g. deprecia-
tion/amortization periods), adjustments for foreign exchange items, capacity utilization
298
Closing comments on comparability under the one-sided, profit-based
methodology
Third, transfer pricing imprecision is not limited to problems associated

with the use of blended profit margins under the one-sided, profit-based
pricing methodology. For instance, the use of royalty rates from purported
CUTs involving similar IP (inexact comparables) under the CUT method
may – even if a range of comparability adjustments are carried out – de-
viate significantly from the “true” arm’s length allocation of income. A
measurement error under the CUT method, when applied in order to al-
locate residual profits from unique and valuable IP, will generally result in
significantly larger allocation errors (in absolute amounts) than measure-
ment errors from applying a blended profit margin under the one-sided
methods used to allocate normal market returns. The same can be said for
the profit split method, which allocates residual profits based on a broad
and discretionary assessment of relative values (see the analysis in chapter
9). For instance, the question under the CUT method may be whether the
royalty should be 20-30% of sales, while the issue under the one-sided
methods may be whether the tested party should be entitled to a normal
market return for its routine functions of, for instance, 5-10% on costs.
The author therefore concludes that the partial lack of transactional com-
parability from using blended profit margins under the TNMM, although
not ideal, should not be regarded as a detrimental comparability problem
in transfer pricing.
8.9. Closing comments on comparability under

the o ne-sided, profit-based methodology
As illustrated in the discussions in this chapter, attaining a high degree of
comparability under the CPM and the TNMM is not an easy task. Practice
has shown that even for group entities that contribute solely routine inputs
to the value chain, it is generally difficult to find truly comparable third-
party profit data upon which to base benchmarking of controlled profit
allocations under the CPM and TNMM. This may be due to several fac-
tors, but one particularly pronounced problem is that multinationals often
segregate and disperse their functions among group entities in a highly
fragmented and specialized manner, while third-party comparables candi-
adjustments and stock option adjustments. Such adjustments alter the accounting profits
of the unrelated enterprise and are relevant both under gross and net profit methods.
Due to the lack of detailed accounting information (see secs. 6.2.4. and 6.2.5.) and the
complexity associated with making accounting and economic adjustments, such com-
parability adjustments are often not carried out in practice.
299
dates normally carry out a more holistic set of business functions. Further,
it may also be that there are no true independent third-party comparable
enterprises available. It may be that the closest that one comes to finding a
comparable unrelated routine input provider is an enterprise that belongs
to a third-party group. That is a problem if the transactions of that entity
mainly consist of transactions with other entities within its group (as will
often be the case). The author gathers that, in practice, it is often necessary
to resort to such comparables for a lack of better alternatives.1014 The com-
parability problems associated with the one-sided, profit-based methodol-
ogy are among the reasons why the OECD now has extended the scope of
application of the profit split method in the 2017 OECD TPG (see the dis-
cussion in section 9.2.3.) and generally favours the profit split method over
the other pricing methods (see the discussion in section 11.4.). The problem
is that this methodological development triggers new and additional trans-
fer pricing issues, which will be discussed in chapter 9.
1014. For instance, operating profit margins extracted from group entities within ver-
tically integrated contract manufacturing multinationals, e.g. Flextronics (see http://
www.flextronics.com), Celestica (see http://www.celestica.com) or Jabil (see http://
www.jabil.com), may, in practice, be used to benchmark the profits of the tested party.
300
Chapter 9
Direct Profit-Based Allocation of Residual Profits to

Unique and Valuable IP: The Profit Split Method
9.1. Introduction
In this chapter, the author will discuss the US and OECD two-sided, prof-
it-based transfer pricing (profit split) methodologies as they stand in the
current US regulations and the 2017 OECD Transfer Pricing Guidelines
(OECD TPG).1015 The core content of the PSM (PSM) is that it allocates
the net operating profits from a value chain among the group entities that
contribute value chain inputs in proportion to the relative value of those
contributions. The method is particularly relevant in the context of intangi-
ble value chains, as it may be used to allocate residual profits from unique
and valuable intangible property (IP) without the need for comparable un-
controlled transactions (CUTs), which generally will be unavailable.
The author would like to emphasize that this chapter should be read in light
of the analysis of the historical development of profit-based transfer pricing
methodology in chapter 5, in particular the analysis of Eli Lilly in section
5.2.4.3. and of the US and OECD implementation of the profit-based meth-
odology in sections 5.3. and 5.4., respectively.
From being developed through US case law as a reaction to – and a way of

dealing with – the lack of comparables upon which to base application of
1015. The following discussion with respect to the OECD profit split method (PSM)
is based on the OECD BEPS Action 10: Revised guidance on profit splits discussion
paper (OECD 2017), as the final consensus text has not yet been released at the time of
writing. It is anticipated that the final text will not deviate on significant points from
the discussion paper. For a historical overview of the methodology, see the 1992 IFA
general report in Maisto (1992), at pp. 51-53. For a recent comparative overview of the
application of profit splits in the context of value chains in different (IFA branch) juris-
dictions, see Rocha (2017), at p. 229. See also Chand et al. (2014) for a recent discussion
of the PSM under the OECD Transfer Pricing Guidelines (OECD TPG), taking into
account the new BEPS country-by-country documentation requirements. For recent
commentary on the OECD PSM guidance, see Robillard (2017a). For a comparison
between the PSM and formulary apportionment, see Kroppen et al. (2011), where a
positive outlook on the PSM is presented due to its ability to allocate, in an arm’s
length manner, profits without the need for direct third-party comparables. For further
recent discussions on the PSM, see, e.g. Milewska et al. (2010); Kadet (2015); Robillard
(2017b); and Feinschreiber et al. (2016).
301
Chapter 9 - Direct Profit-Based Allocation of Residual Profits to Unique and
Valuable IP: The Profit Split Method
the traditional, transaction-based transfer pricing methods, the PSM has,

over time, grown into the role of being one of the most practically im-
portant transfer pricing methods internationally. The 2017 OECD TPG on
intangibles envision a significantly more central role for the methodology
than the OECD originally intended when it introduced the method into the
OECD TPG in 1995. The author will revert to the relative importance of
the method in the context of the OECD TPG in chapter 11, after discuss-
ing the new OECD guidance on the allocation of incremental operating
profits from location savings, local market characteristics and synergies in
chapter 10.
This chapter will focus on two key issues. The first issue is the clarification of
in which scenarios the PSM can be applied, i.e. the delineation of the scope
of application for the methodology. This will be analysed in section 9.2. The
second issue is the clarification of how the method works, i.e. how operating
profits can be allocated under the methodology. This will be analysed in sec-
tion 9.3. The author will then comment on how the PSM can be applied in
valuation scenarios in section 9.4. Lastly, in section 9.5., the author discusses
how the OECD PSM is limited to information known or reasonably foresee-
able at the time at which the controlled transaction is entered into.
9.2. The scope of application of the methodology
9.2.1. Introduction
Having a critical approach to the scope of application for the PSM is im-
portant because the methodology relies greatly on discretionary assess-
ments and was originally designed only to be applied where no CUT-based
methods could be used instead. If the scope of application becomes very
broad, this could trigger the risk that both taxpayers and tax authorities re-
sort to the method even though there may be other transfer pricing methods
available that could offer more reliable and precise profit allocation results.
9.2.2. The US PSM
In the United States, the PSM was first introduced into the Internal Reve-
nue Code (IRC) section 482 (temporary) regulations in 1993.1016 The meth-
1016. 58 FR 5263-02; see the notice of proposed rulemaking in 58 FR 5310-01, at

§ 1.482-6T. See Brauner (2008), at p. 132, on the US PSM.
302
od played a surprisingly modest role in those regulations, given the fact that
the methodology had been applied frequently in case law over decades and
that it was aligned with the new commensurate-with-income approach.1017
Application of the method was limited to cases in which “each controlled
taxpayer owns a valuable, non-routine intangible”.1018 This limitation was
scrapped in the final version of the PSM in the 1994 US regulations, which
still applies today. Subsequently, there were some alterations to the PSM
provision in connection with the introduction of the 2009 intra-group ser-
vices regulations,1019 most notably, the inclusion of a definition of non-rou-
tine contributions.1020
The US PSM can, in principle, be applied even if one of the controlled

parties does not contribute valuable and unique IP to the value chain.1021
The best-method rule, however, will normally restrict application of the
method to cases in which both parties indeed contribute such value chain
inputs. Where one of the parties contributes only routine inputs, the data
and assumptions pertaining to this party will likely be significantly more
reliable than for the party contributing unique IP, typically favouring the
comparable profits method (CPM).1022 Further, the strict requirements for
1017. This “backseat” position was, in the author’s view, likely due to the fact that
(i) the US, in outbound transfers of US-developed unique intangibles, where research
and development (R&D) expenses normally had been deducted concurrently through-
out the development phase against US income, simply did not deem a split of residual
profits with the transferee jurisdiction to be reasonable; and (ii) the US preferred the
comparable price method CPM) as its flagship method in order to prevent BEPS.
1018. See 58 FR 5310-01, § 1.482-6T(b)(1). This triggered discussions of what quali-
fied as non-routine intangibles. Langbein (2005), at p. 1071, seems to attribute the re-
strictions of the 1993 temporary regulations on the use of the PSM to the political
commitment of the US to the arm’s length principle (as opposed to formulary appor-
tionment).
1019. 74 FR 38830-01. The general US PSM in Treas. Regs. § 1.482-6 applies also
to controlled service transactions, along with some specific guidance in Treas. Regs.
§ 1.482-9.
1020. See Treas. Regs. § 1.482-6(c)(3)(i)(B), where the US regulations define a non-
routine contribution negatively (simply as a contribution that is not accounted for as
a routine contribution). This definition was introduced following a debate that began
already in 1993, when the notice of proposed rulemaking (58 FR 5310-01) outlining
the proposed PSM requested comments on a possible definition. See the analysis in sec.
3.4.3. on non-unique value chain contributions.
1021. This, however, is not the case for the PSM for intra-group services. See Treas.
Regs. § 1.482-9(g)(1), which states that the method cannot be used where only one
controlled taxpayer makes significant non-routine contributions to the controlled trans-
action.
1022. See the discussion on the scope of application for the CPM in sec. 8.3. See also
the preamble to Treas. Regs. § 1.482-6 in the final 1994 regulations (59 FR 34971-01).
See also Treas. Regs. § 1.482-6(c)(2)(ii)(D).
303
using the comparable profit split allocation pattern (see the discussion in
section 9.3.2.) make the residual profit split allocation the practical applica-
tion of the PSM. The latter allocation pattern presupposes that both parties
own valuable and unique IP that contributes to the residual profits. Thus, in
spite of not being a formal requirement, there is little doubt that the practi-
cal area of application for the general US PSM is to controlled transactions
in which both parties contribute unique and valuable IP to the value chain.
As the split of residual profits under the PSM will ordinarily not be based
on third-party data, the US Internal Revenue Service (IRS) asserts that the
PSM should generally be considered a method of last resort.1023 The author
takes issue with this. The PSM will normally be the only method appli-
cable when both parties to a controlled transaction contribute non-routine
inputs, as the CUT method will generally be unavailable due to the lack of
CUTs.1024 In these situations, the PSM should be deemed the most appro-
priate method, not a last resort. Further, the author finds the basic premise
behind the IRS’s position, i.e. that the results of pricing methods that apply
internal data in all cases will be inferior to those that apply external data,
to be questionable.1025 Meaningful and reliable external benchmarks for
splitting residual profits will be unavailable in the vast majority of cases.
Also, the PSM partly relies on external data for the allocation of income, as
the first step of allocating normal market returns to routine functions uses a
comparable profits method (CPM) approach. Further, splitting the residual
profits on the basis of a concrete assessment of causality has proven to yield
reliable results in practice, e.g. in Eli Lilly.1026
9.2.3. The OECD PSM
The 1993 introduction of the PSM into the US regulations was the motiva-
tion behind the OECD’s adaptation of the PSM (along with the transac-
tional net margin method (TNMM)) in the 1995 OECD TPG. The OECD’s
implementation was sceptical towards the profit-based methods in general
1023. See the preamble to the 1994 regulations (59 FR 34971-01).

1024. Treas. Regs. § 1.482-1(c)(1).
1025. The US Internal Revenue Service’s (IRS’s) position may possibly be due to an
excessive focus on typical outbound scenarios, where unique US-developed intangibles
employed in the value chain are owned by a US entity and the foreign entity is a mere
routine input provider (i.e. “standard” CPM scenarios). This may, in practice, very well
be the most relevant scenario, but it should not be forgotten that there necessarily also
will be cases in which the foreign entity does provide unique inputs to the intangible
value chain. In such cases, there will likely be no alternative to the PSM.
1026. See the analysis of the case in sec. 5.2.4.3.
304
and limited their application in ways that the US regulations did not (e.g.
with the transactional approach and ex ante approach).1027 The historical
point of departure, as reflected in paragraph 2.109 of the 1995/2010 OECD
TPG, was that the PSM should not be used when one party to the con-
trolled transaction only contributed routine functions, such as contract
manufacturing or low-risk distribution. While this point of departure was
not discarded in the new 2017 consensus text, it has certainly been relaxed
through a broadening of the scope of application of the PSM.
The 2017 OECD TPG allow the application of the PSM in three main sce-
narios. First, the OECD PSM can be applied if both parties to the controlled
transaction contribute unique and valuable IP to the value chain.1028 This is
still the core area of application for the method. Thus, it will normally not
be appropriate to apply the PSM when one party to the controlled transac-
tion performs only routine functions.1029 Sharing profits in such cases will
be unlikely to reflect arm’s length profit allocation.1030 This first area of
application of the OECD method is parallel to the main area of application
of the US PSM.
Second, the OECD PSM can be applied to allocate profits from highly in-
tegrated operations where a one-sided method would not be appropriate.1031
While the 1995/2010 OECD TPG did mention that the method could offer
a solution for highly integrated operations,1032 this area of application of
the methodology was quite notably elaborated on in the 2017 text, indicat-
1027. The PSM, like the transactional net margin method (TNMM), may only be ap-
plied on a transactional basis. In other words, the OECD TPG restrict the extent to
which both the controlled and uncontrolled transactions may be aggregated. See the
discussions of the aggregation of controlled and uncontrolled transactions in sec. 6.7.
and sec. 8.6., respectively. With respect to controlled transactions, the OECD trans-
actional requirement largely mirrors the “relevant business activity” delimitation of
the US PSM; see Treas. Regs. § 1.482-6(a), which states that the combined operating
profits must be derived from the most narrowly identifiable business activity of the con-
trolled taxpayers for which data is available that includes the controlled transactions.
The aggregation problem is not nearly as pronounced under the PSM as under the one-
sided methods, as the allocation of residual profits under the PSM will normally not be
based on third-party profit data.
1028. BEPS Action 10, Revised guidance on profit splits, para. 6. The draft text con-
tains a definition of unique and valuable value chain contributions in para. 16. This
definition is materially similar to that contained in para. 6.17. The author refers to the
analysis in sec. 3.4.3.
1029. The TNMM will normally be applied in these scenarios; see the discussion in
sec. 8.3.
1030. Id., at para. 14.
1031. Id., at para. 7.
1032. 1995/2010 OECD TPG, para. 2.109.
305
ing the intention of the OECD that the method should be applied going
forward, also outside of cases in which both parties contribute unique and
valuable IP to the value chain. The logic underlying this “expansion” of
the scope of application of the PSM is not immediately apparent.1033 The
business operations and value chains of multinationals will normally be in-
tegrated. The economic gain that can be achieved through such integration
is one of the core reasons why some businesses choose to become multi-
nationals in the first place.1034 The question then becomes how to separate
“normal” multinational enterprise (MNE) integration, which will not in
itself indicate that the PSM could be applied, from “high” MNE integra-
tion, which – now according to the 2017 OECD TPG – will indicate that
the PSM can be applied. The 1995/2010 version of the OECD TPG did not
define the concept of highly integrated operations, but the 2017 version
does, as follows:
[O]ne party to the transaction [that] performs functions, uses assets and as-
sumes risks is interlinked with, and cannot reliably be evaluated in isolation
from, the way in which another party to the transaction performs functions,
uses assets and assumes risks.1035
This wording indicates that the “highly integrated expansion” will be rel-
evant when it is not possible to apply a one-sided pricing method to re-
munerate one of the parties to the controlled transaction. That will be the
case when the relevant party contributes unique and valuable IP to the con-
trolled transaction. Outside of such cases, however, it should almost always
be possible to use a one-sided pricing method, taking into account that the
comparability requirements generally are relaxed.1036
Even still, there will undoubtedly be situations in which it will be problem-

atic to find reliable comparables. Multinationals tend to separate functions
into different legal entities, with single entities devoted to logistics, ware-
housing, marketing, etc., making it difficult to find similar unrelated enter-
prises. Third-party enterprises that are candidates for being used as com-
parables under one-sided methods will typically carry out a range of other
activities and transactions in addition to those comparable to the functions
and transactions of the tested party. In such cases, however, it should nor-
mally be preferable to resort to comparability adjustments than to apply
1033. See also, in this direction, Robillard (2017a), where it is argued that the applica-
tion of the PSM may depart from third-party profit allocations in integrated business
models, such as franchising structures.
1034. See the discussions in secs. 2.2. and 2.3.
1035. Id., at para. 19.
1036. See the analysis in sec. 8.6. with respect to the popular TNMM.
306
the PSM. The reason for this is that if the group entity to be remunerated
is a routine provider, it should only make a normal market return (which
is likely what the third-party comparables also are making). Thus, even if
significant comparability adjustments must be made, these will likely not
affect the end profit allocation result much. The routine entity will only be
allocated a normal market return. Application of the PSM, however, could
open up even more discretionary and imprecise profit allocations in such
scenarios. There can even be a risk that the routine entity is allocated above
normal market returns if the PSM is applied. It is therefore not clear that
the “highly integrated expansion” can serve a useful purpose.
The ambiguity underlying the “highly integrated expansion” is reflected on

several points in the 2017 OECD TPG. For instance, the text argues that the
expansion will be relevant if both parties to the controlled transaction con-
tribute significant assets to a long-term arrangement.1037 If the contributed
assets are unique and valuable IP, the PSM would be applicable, even under
its “traditional” scope of application, without any need for the “highly inte-
grated expansion”. If, however, the significant assets are not unique but are
generic, their contribution should normally be susceptible to pricing based
on either the CUT method or the one-sided methods. There should then not
be any need for the expansion.
Nevertheless, it should not, in principle, be ruled out that the expansion

could be relevant in some narrow cases outside of scenarios in which both
parties contribute unique and valuable IP. This could, for instance, be the
case when both parties contribute relatively unique and significantly valu-
able assets to a long-term cooperation when there are no realistic options
available for replacing the contributed assets with equivalent value chain
input factors. An example could be a business cooperation within the oil
sector, where one group company contributes unique IP in the form of seis-
mic technology (software, know-how, etc.) and the other group company
contributes a highly valuable and one-of-a-kind specialized vessel. A profit
split could be justified here under the “highly integrated operations” expan-
sion. The profit split allocation to the group entity that does not contribute
unique IP (i.e. the entity that contributes the vessel) should then be subject
to a “sanity check” under a one-sided method (cost-plus or TNMM) in or-
der to have some assurance that it is not allocated too much profit.
The 2017 OECD TPG further argue that the highly integrated expansion
can be relevant when the contributions are highly interrelated or interde-
1037. BEPS Action 10, Revised guidance on profit splits, para. 21.
307
pendent.1038 This is illustrated by reference to a situation in which the con-

trolled parties contribute unique and valuable IP to the value chain and
where combinations of IP are transferred intra-group. The 2017 text does
not contain any examples of a situation in which a highly integrated struc-
ture by itself (i.e. without the presence of contributions of unique and valu-
able IP from both parties) justifies the application of the PSM.1039 It seems
peculiar that the OECD has chosen to expand the scope of the “highly
integrated” application of the PSM without coming up with a single practi-
cal example of how the doctrine can be applied.
It is the author’s impression that the “highly integrated expansion” is so

ambiguous that it can be misused, for instance, by asserting that the con-
tributions from what is, in reality, a routine input entity in a low-tax juris-
diction shall be allocated a portion of the residual profits from unique IP
(that it does not own) under the assertion that its value chain contribution
is “highly integrated”. If it is possible to remunerate the relevant group en-
tity based on a one-sided pricing method, tax authorities should not accept
allocations based on the “highly integrated expansion” of the PSM. Such
allocations will likely be more imprecise than profit allocations based on
one-sided pricing methodologies and may yield what are, in reality, non-
arm’s length results (e.g. if a routine entity is allocated residual profits).
Comparability problems associated with applying a one-sided pricing
method should not justify the application of the PSM if, in principle, it is
clear that a one-sided pricing structure would be more appropriate, taking
into account the nature of the controlled parties’ value chain contributions.
Third, the OECD PSM can be applied if both parties to the controlled
transaction share the assumption of one or more economically significant
risks pertaining to the controlled transaction1040 – including when the eco-
nomically significant risks are separately assumed by each controlled party
– if the risks are so closely interrelated that the playing out of the risks can-
not be isolated from each other.1041
This “economically significant risks” expansion is ambiguous.1042 Outside

of transfer pricing of financial services (where risk is key for allocating
1038. Id., at para. 22.

1039. The 22 June 2017 OECD discussion draft on the amendment of the PSM guid-
ance even requested input on examples of scenarios in which “highly integrated” by
itself would lead to application of the PSM.
1040. Id., at paras. 6 and 25.
1041. Id., at para. 26.
1042. See also, in this direction, the critical comments in Robillard (2017a).
308
profits),1043 it is unclear whether the expansion is meaningful on a stand-

alone basis. This point is illustrated by the examples contained in the 2017
OECD TPG. One example professes that the PSM should be applied to
allocate profits among two group entities that each assume economically
significant risks when the risks are closely interrelated and interdepend-
ent.1044 Ironically, the PSM would be applicable in the example regardless
of the risks, as both group entities contribute unique and valuable IP to the
value chain. Thus, the expansion has no true independent significance in
the example. Another example states that the PSM should be applicable to
allocating profits among two car manufacturing group entities that trade
car components among themselves, as their manufacturing risks are highly
interdependent.1045 Also here, both group entities contribute unique and
valuable IP to the value chain. The method would thus be applicable even
if the risks are not highly interdependent.
In summary, while the above expansions may increase the availability of

the methodology, there is obvious danger in allowing profit splits in these
ambiguously defined scenarios. A profit split will – regardless of the cho-
sen allocation key – always have some degree of inherent imprecision.1046
Such discretionary profit allocation may be relatively unproblematic when
it is clear that both parties do indeed contribute unique and valuable IP to
the value chain. Even if the exact extent to which the parties contribute
value is unclear, it will, in such scenarios, be near certain that both parties
do contribute significantly. If, however, discretionary profit splits are also
allowed among two group entities that contribute significantly different
values (e.g. one contributes unique and valuable IP while the other does
not), there is a risk that a profit split will allocate too little income to one of
the parties and too much to the other.
The revised “highly integrated” and the new “economically significant

risks” expansions should therefore be applied in only very narrow cases.
The line should be drawn where it is possible to remunerate one of the
group entities using a one-sided method. This goes to the core of the arm’s
length principle. A one-sided method will only allocate a normal market
return to a group entity that contributes routine inputs to the value chain
because that is what similar third parties earn. Likewise, super profits can
1043. See id., Example 7, in which the controlled parties share the assumption of fund
management risks.
1044. Id., at Example 3.
1045. Id., at Example 10.
1046. See also, e.g. Roberge (2013), at p. 234.
309
only be earned through unique and valuable assets, typically IP. An appli-
cation of the PSM should never result in profit allocations that contradict
these fundamental economic axioms. Thus, if the PSM is applied in the
“highly integrated” or “economically significant risks” scenarios, the allo-
cation should be subject to a sanity check under a one-sided method. Such
a dual methodology approach will, to some extent at least, mitigate the
potential dangers of the 2017 OECD expansion of the PSM.1047
This also comes to another point. The 2017 expansion of the OECD PSM
contributes to a dilution of the boundary between the scope of application
of the PSM and the one-sided methods. There may now be overlapping
areas of application of the two classes of transfer pricing methodologies.
This overlap underlines the need to focus on the profit allocation result that
is yielded by the applied methodology and to substantiate that result by way
of reference to more than one transfer pricing method. For instance, the
result yielded through applying the PSM (based on “highly integrated” or
“economically significant risks” scenarios) could be supported through an
application of the cost-plus method or the TNMM to one of the group enti-
ties participating in the controlled transaction, as well as a check against
the “realistic alternatives” pricing principle.
The 2017 OECD TPG expansion of the PSM should be seen in light of the
generally dominant position that the OECD envisions that this methodol-
ogy shall have in relation to the other transfer pricing methods (a topic
that is discussed in section 11.4.), as well as the profit split-like“important
functions doctrine” that the OECD adopted in the 2017 OECD TPG for de-
termining IP ownership (which is discussed in section 22.3.2.). The OECD
has completely rejected its historical preference for the CUT method (with
respect to the transfer pricing of IP) and legal ownership (for determin-
ing ownership of IP developed intra-group) as expressed in the 1995/2010
OECD TPG in favour of a holistic profit split approach that spans across
both the transfer pricing and IP ownership provisions of the 2017 OECD
TPG. The potential downside of this comprehensive OECD preference for
the PSM is that almost every profit allocation problem can now arguably be
dealt with using the highly discretionary PSM, resulting in imprecise profit
allocations that are difficult to verify the accuracy of and that may lead to
irreconcilable profit allocations among the involved jurisdictions, and thus
ultimately also to double taxation.
1047. On the use of more than one transfer pricing method (under the previous genera-
tion of the OECD TPG), see, e.g. Ahmadov (2011), at p. 191.
310
How operating profits may be split under the methodology
9.3. How operating profits may be split under

the methodology
9.3.1. Introduction
The PSM allocates operating profits from a particular value chain among
the group entities that contribute inputs to that value chain in proportion
to the relative values of those inputs.1048 An illustrative example of this is
Eli Lilly,1049 where the core issue was the allocation of the residual profits
earned by the Puerto Rican subsidiary from sales of the Darvon pill among
the manufacturing IP owned by the Puerto Rican subsidiary and the mar-
keting IP owned by the US parent company. The question that will be dis-
cussed in sections 9.3.2.-9.3.3. is that of how profits can be split under the
method, i.e. what kinds of assessments can be used to estimate the relative
values of the value chain contributions to allocate profits to the involved
group entities. This is a question of which allocation patterns are allowed
under the methodology.

the US regulations
The US regulations allow two specified alternatives for allocating prof-

its under the PSM, namely1050 (i) the comparative profit split; and (ii) the
1048. See Robillard (2017b) for critical comments on the OECD guidance on the de-
lineation of the profits to be split. Robillard argues that the delineation may trigger dif-
ficult accounting documentation (must often resort to internal management accounting
to document product-level profits) and classification issues (e.g. discretionary alloca-
tion of costs among different product lines).
1049. See the analysis in sec. 5.2.4.3.
1050. The 1993 temporary US regulations contained two additional profit split pat-
terns, which were scrapped in the final 1994 regulations. The first of these was the
“capital employed allocation” (see 58 FR 5310-01, § 1.482-6T(c)(3)), which divided
the residual profits according to the average capital employed by each controlled party
in relation to the total capital employed in the intangible value chain. The preamble to
the final 1994 regulations found that, with one exception, it had not been possible to de-
scribe a case in which it would be possible to conclude with certainty that two or more
controlled taxpayers faced equal levels of risk (see 59 FR 34971-01). The exception
was that of joint venture agreements in which the parties, ex ante, agree to share costs
and benefits proportionately. Such scenarios are addressed by the cost-sharing arrange-
ment regulations. The capital employed allocation pattern was therefore omitted in the
final 1994 US regulations. See also Wittendorff (2010a), at pp. 759-766, on allocation
patterns under the PSM. The second additional profit split pattern was the “unspecified
profit split” (see 58 FR 5310-01, § 1.482-6T(c)(5)).
311
residual profit split.1051 In addition, profit splits as an unspecified pricing

method can be used.
The comparable profit split alternative benchmarks the controlled split of

operating profits against splits adopted by unrelated enterprises engaged in
similar transactions under similar circumstances.1052 Thus, this application
of the method splits the profits in the same fraction as is done in a CUT.
The more unique and valuable the contributions from the parties to the
controlled transaction are, the more difficult it will be to find a CUT, and
therefore, it will be more unlikely that this allocation pattern can be used
in practice.1053
In cases in which it is at all possible to identify a CUT, it will be highly

likely that comparability adjustments must be made to the selected CUT.1054
It will be practical to adjust for differences in the financial accounting data
of the controlled and uncontrolled taxpayers that could materially affect
the division of operating profits.1055 It will first be necessary to ensure that
the accounting data of both parties are properly segregated “between the
relevant business activity and the participants’ other business activities”
so that the third-party profits used to benchmark the controlled profits are
truly comparable (i.e. they stem from a comparable value chain).1056 Fur-
ther, as the PSM is a net profit method, differences in accounting classifica-
tions with respect to the profits in the controlled and CUT will not affect
measurement (e.g. classification of expenses as costs of goods sold (COGS)
or operating expenses). Economic accounting differences, however, are
clearly relevant (e.g. whether an item is capitalized or expensed).1057 The
accounting treatment of, for instance, research and development (R&D)
expenses and the depreciation and amortization of assets must be checked
for consistency.
1051. See Treas. Regs. § 1.482-6(c)(2) and (3), respectively. For critical reflections on
the restrictions imposed on allocation patterns under the US PSM, see Horst (1993).
1052. Treas. Regs. § 1.482-6(c)(2)(i). The comparable profit split allocation pattern
was first introduced in the 1988 White Paper (see Example 10) and then in the 1993
temporary regulations (see 58 FR 5310-01, § 1.482-6T(c)(4)).
1053. See Treas. Regs. § 1.482-6(2)(ii)(D), which asserts that reliability “may be en-
hanced by the fact that all parties to the controlled transaction are evaluated under the
comparable profit split”. This ambiguous wording may refer to the general point that the
PSM is a two-sided method. If so, the argument must be rejected, as a one-sided method
clearly may be better suited for pricing than a two-sided method, given the facts and
circumstances of a particular case.
1054. See Treas. Regs. § 1.482-5(c)(2).
1055. See Treas. Regs. § 1.482-1(d)(2).
1056. Treas. Regs. § 1.482-6(2)(ii)(C)(1).
1057. See supra n. 1013 with respect to accounting adjustments.
312
The comparable profit split alternative is unavailable if the return on capital

employed (ROCE) earned by the selected uncontrolled enterprises varies
significantly from that of the controlled taxpayers.1058 Such scenarios will
indicate that the profit potential of the unique IP owned by the controlled
and uncontrolled taxpayers differs significantly. It must then be assumed
that the CUTs from which the third-party profit split is extrapolated are un-
reliable. The strict criteria set out in the US regulations for the application
of the comparable profit split will make the method unlikely to be applied
much in practice.1059
The residual profit split allocation pattern was first introduced in the 1993
temporary US regulations.1060 The pattern bears some resemblance to the
methodology applied in Eli Lilly1061 and in the White Paper’s basic arm’s
length return method (BALMR) with profit split,1062 as it splits the op-
erating profits from a value chain between the controlled taxpayers that
contribute value chain inputs through a two-step process. First, a normal
market return is allocated to each controlled party as compensation for
its routine contributions.1063 Thereafter, the residual profits are split based
on the “relative value” of the controlled parties’ non-routine value chain
contributions.
A non-routine contribution will, in practice, be unique IP, the relative value

of which may be measured through1064 (i) external market benchmarks (i.e.
CUTs); (ii) capitalized R&D costs; or (iii) the amount of current R&D
expenditures in recent years.1065 External benchmarks will be generally be
1058. Return on capital employed (ROCE) is operating profits as a percentage of as-

sets. The regulations do not elaborate on what is meant by a significant departure. This
will necessarily create uncertainty, but likely no major practical problems, as the com-
parable profit split pattern will rarely be applicable.
1059. As recognized already in the White Paper (Notice 88-123), at p. 88, it is unlikely
that a comparable uncontrolled price will be available if the multinational does not also
license the intangible transferred in the controlled transaction to unrelated parties (e.g.
in certain territories).
1060. See 58 FR 5310-01, § 1.482-6T(c)(2).
1061. 84 T.C. No. 65 (Tax Ct., 1985), affirmed in part, reversed in part by 856 F.2d 855
(7th Cir., 1988).
1062. See Notice 88-123 and its Example 11.
1063. The normal return should be determined under a suitable pricing method.
In practice, a one-sided method, likely the CPM, will be applied; see Treas. Regs.
§§ 1.482-6(c)(3)(i)(A), 1.482-3, 1.482-4, 1.482-5 and 1.482-9.
1064. Treas. Regs. § 1.482-6(c)(3)(i)(B)(2).
1065. This last measure will only be relevant where R&D costs are relatively constant
over time and the useful lives of the IP contributed by all parties is approximately the
same.
313
unavailable in practice.1066 One is therefore left with estimating the relative

values of the contributed IP based on capitalized or current R&D costs.
The author finds this restriction somewhat peculiar for two reasons. First,
there tends to be little correlation between R&D costs and the value of
successful IP.1067 Intangible development costs are therefore generally un-
suitable as proxies for the value of unique intangibles. The use of such data
for benchmarking purposes may yield unreliable transfer pricing alloca-
tions. The author shares the view of Professor Langbein that the specified
US PSM “in effect … is a method of fractional apportionment that used
a single allocation factor – intangible property development costs – to ac-
complish the allocation of combined profit”.1068
Second, the determination of IP development costs is dependent on a range

of management and financial accounting assessments, for instance, on how
indirect costs shall be allocated between the relevant business activity and
the other activities of the controlled taxpayer, as well as assumptions re-
garding the useful life of the IP. Such assessments may be manipulated in
order to provide a certain outcome.1069 Why the regulations in light of this
have made a profit split based on R&D costs the default US solution is a
good question. Presumably, this is due to the fact that R&D costs are verifi-
able “objective” amounts.1070
1066. The 1993 proposed PSM provisions (§ 1.482-6T in 74 FR 38830-01) recognized

that external benchmarks for the relative value of unique intangibles employed in a
controlled transaction would normally be unavailable. See also the 1988 White Paper
(Notice 88-123), at p. 101.
1067. The 1988 White Paper (Notice 88-123), the preamble to the 1994 regulations (59
FR 34971-01), as well as the final PSM provisions themselves in Treas. Regs. § 1.482-6
recognize that IP development costs generally will have no bearing on the true value of
the developed IP.
1068. Langbein (2005), at p. 1071.
1069. Treas. Regs. § 1.482-6(c)(3)(ii)(C)(3).
1070. The author also finds it unfortunate that the US regulations chose to illustrate the
PSM solely through an example in which capitalized R&D costs are used as proxies
for relative values; see Treas. Regs. § 1.482-6(c)(3)(iii). The example pertains to a US
parent that has developed a patented material (Nulon), which it licenses to its European
subsidiary. To determine the royalty rate to be paid by the subsidiary, the example
uses the ratio of capitalized R&D and marketing expenditures divided by sales. The
fundamental problem with the example is that R&D costs generally cannot serve as
an estimate of the value of the developed IP (simply because the value of the IP rarely
will have any causal link to the development costs). Thus, there is no way of knowing
whether the royalty rate professed by the example also allocates business profits from
the locally owned marketing IP to the parent, even though the parent should have no
stake in that profit. It can also be noted that the 2009 final services regulations (74 FR
38830-01) provide two profit-split examples; see Treas. Regs. § 1.482-9(g)(2), Exam-
ples 1 and 2. Both are rather voluminous and ultimately provide little guidance.
314
The residual allocation pattern of the specified US PSM should be seen as

inherently unreliable due to its dependence on R&D costs. The methodol-
ogy should generally be rejected under the best-method rule.
Instead, an unspecified application of a residual profit split should be used.

The proposed profit split provisions included in the temporary 1993 US
regulations allowed “other methods of approximating the relative values”,
i.e. an unspecified profit split.1071 To apply this pattern, it had to be shown,
to the satisfaction of the IRS, that the selected measure provided an eco-
nomically valid basis for the allocation of the residual profits. This profit
allocation pattern was omitted as a specified allocation pattern in the 1994
final regulations, due to the general availability of using unspecified pric-
ing methods.1072 Thus, an unspecified application of the residual profit split
will, in fact, constitute an unspecified transfer pricing method under the
US regulations.
This unspecified method should split the residual profits pursuant to a criti-
cal facts-and-circumstances-based assessment of the causal relationship
between unique IP contributed to the value chain and the residual profits
akin to the assessments applied in pre-1994 regulation case law, with Eli
Lilly being the prime example. The assessments rested on a thorough func-
tional analysis of the intangible value chain and had a tendency towards an
equal split of the residual profits.1073
The major innovation in Eli Lilly and the 1988 White Paper, and later the
CPM, was to decompose the value chain inputs into routine and non-rou-
tine contributions. Allocating normal returns to routine contributions may
1071. 58 FR 5310-01.
1072. See Treas. Regs. § 1.482-4(d).
1073. In PPG Industries Inc. v. CIR (55 T.C. 928 [Tax Ct., 1970]), the split was 55/45;
in Eli Lilly & Co. v. US (372 F.2d 990 [Cl.Ct, 1967]), the split was 68/32; in Eli Lilly
and Company and Subsidiaries v. CIR (84 T.C. No. 65 [Tax Ct., 1985], affirmed in part,
reversed in part by 856 F.2d 855 [7th Cir., 1988]), the split was 45/55; in G.D. Searle &
Co. v. CIR (88 T.C. 252 [Tax Ct., 1987]), the split was 25% of the foreign subsidiary’s
net sales, allocated to the United States (the split of the residual profits is unknown, but
presumably, it was rather equal); in Bausch & Lomb Inc. v. CIR (92 T.C. No. 33 [Tax
Ct., 1989], affirmed by 933 F.2d 1084 [2nd Cir., 1991]), the split was 50/50; the precise
split in E.I. Du Pont de Nemours and Co. v. US (1978 WL 3449 [Cl.Ct., 1978], adopted
by 221 Ct.Cl. 333 [Ct.Cl., 1979], certiorari denied by 445 U.S. 962 [S.Ct., 1980], and
judgment entered by 226 Ct.Cl. 720 [Ct.Cl., 1980]) is unknown, but a considerable por-
tion of the residual profits were allocated to the United States; and the split in Hospital
Corporation of America v. CIR (81 T.C. No. 31 [Tax Ct., 1983], nonacquiescence rec-
ommended by AOD- 1987-22 [IRS AOD, 1987], and Nonacq. 1987 WL 857897 [IRS
ACQ, 1987]) was 75/25, in favour of the United States.
315
be relatively straightforward once these contributions have been identified.

The crux lies in the allocation of residual profits to two or more items of
unique IP that are all necessary components in the value chain. It is the
author’s assertion that this particular problem will normally not be possible
to solve with any objective accuracy. One should have realistic expecta-
tions of the precision of profit splits. Metaphorically put, both hot water
and ground coffee beans are necessary ingredients to make a good cup of
coffee, and one would be of little value without the other for the purpose of
yielding the intended result. The ideal of an accurate allocation of residual
profits that align with the actual relative value of the involved IP based on
causality should realistically be relinquished.1074
Weighing the relative values then becomes a matter of judgement. In the

author’s view, there should be two guiding stars for this assessment. First,
unique IP should only be allocated residual profits that it actually has con-
tributed to. In other words, causality must be proven, even if the extent
of causality is unknown. Second, if all items of IP employed in the value
chain are necessary components for the generation of residual profits, there
should be a presumption of an equal profit split.1075 This will limit the ex-
tent of potential allocation errors. The author finds this reasoning to be
largely in line with the following description of the BALRM with the profit
split addition in the 1988 White Paper:
In splitting this residual amount between the related parties, it is not neces-
sary to place a specific value on each party’s intangible assets, only a rela-
tive value. Of course, it is easier to state this principle than to describe in
detail how it is to be applied in practice. In many cases, there will be little
or no unrelated party information that will be useful in determining how the
split would be determined in an arm’s length setting. Furthermore, the costs
of developing intangibles, even if known, may bear no relationship to value,
especially in the case of legally protected intangibles, and generally should
1074. It may often be possible to estimate the value of IP underlying successful

products (with established track records, current cash flows and projections of future
growth) with reasonable reliability. The Darvon pill in Eli Lilly and the Zantac pill in
GlaxoSmithKline Holdings (Americas) are examples of such products. Even though
such IP can be valued in and of itself, it will be difficult to allocate the value chain
profits among the IP value chain contributions.
1075. Alternatively, it may be that one of the intangibles employed in the value chain
clearly is of lesser importance for the generation of residual profits. For instance, sales
of patented pharmaceutical products are often driven by the extent to which physicians
prescribe the pharmaceuticals to their patients. It is reasonable to assume that phy-
sicians are mainly motivated by the qualitative aspects of the compounds, including
side-effects and required dosages, and less by the trademarks they are sold under. In
such cases, it may be reasonable to attribute the lion’s share of the residual profits to the
manufacturing intangible.
316
not be used to assign relative values to the parties’ intangible assets. Splitting
the intangible income in such cases will largely be a matter of judgment.1076
(Emphasis added)
Such a facts-and-circumstances-based assessment stands in stark contrast

to the arbitrary allocation of residual profits yielded by the default R&D
cost solution under the specified US PSM. The dangers of not basing the
allocation of residual profits on underlying causal relationships has also
been amply demonstrated in practice, where multinationals have used the
“Nulon” example (contained in the US profit split method regulations) to
actively argue for unrealistically low buy-in payments under the previous
generation of US cost sharing regulations.1077

the OECD TPG
Like the US regulations, the OECD TPG also describe two main allocation
patterns for the PSM, i.e. contribution analysis and residual analysis.1078
The main content of these allocation patterns is essentially the same as
under the US regulations. The contribution analysis splits the profits based
on what independent enterprises would have adopted in comparable trans-
actions, while the residual analysis first allocates a normal market return to
routine value chain contributions before the residual profits are split among
the unique IP value chain contributions.
Profits may be split under both allocation patterns based on the division
extrapolated from CUTs, such as from joint venture arrangements, phar-
maceutical collaborations or co-marketing or co-promotion agreements.1079
If there are no reliable CUTs upon which to base the profit split (which
will normally be the case), the assumption will be that independent parties
would have split the relevant profits in proportion to the value of their re-
spective contributions to the generation of profits in the controlled transac-
tion.1080 The split may then be based on an assessment of the relative value
1076. See White Paper (Notice 88-123), at p. 101.

1077. See Treas. Regs. § 1.482-6(c)(3)(iii); and supra n. 1070.
1078. BEPS Action 10, Revised guidance on profit splits, paras. 35 and 37, respectively.
1079. Id., at para. 52.
1080. See the critical comments in Robillard (2017a) and Robillard (2017b), where it
is argued that this application of the PSM resembles formulary apportionment, tak-
ing into account that external benchmarking data is not applied to allocate the profits.
While the author sees the logic in this, he does not agree. It must be assumed that
unrelated parties would have allocated profits according to their respective bargaining
317
contributions as measured by the contributed functions, assets and risks

through profit-splitting factors.1081
Contrary to the US regulations, the OECD TPG do not set out any real
restrictions on which profit-splitting factors may be used to allocate profits,
as long as the factors reflect the key contributions to value in relation to the
controlled transaction.1082 Asset-based profit-splitting factors (tangibles,
intangibles, capital employed, etc.) may be used when there is a strong
correlation between the contribution of the relevant assets and the creation
of value in the controlled transaction.1083 The valuation of such assets for
the purpose of determining relative values will be highly discretionary.1084
Market values should, in the author’s view, generally be used, as historical
cost values (e.g. based on capitalized R&D expenses) are generally not
indicative of real value. Further, cost-based profit-splitting factors can be
used when it is possible to identify a strong correlation between relative
expenses incurred and relative value contributed to the controlled transac-
tion.1085 Other profit-splitting factors are also allowed, such as incremental
sales, employee compensation and hours worked.1086 Such factors will gen-
erally not be appropriate for allocating residual profits, as they will lack
influence on the relative values of unique value chain contributions.
In summary, the 2017 OECD TPG contain a broad description of allowed

profit-splitting factors, which, in essence, does not restrict how the PSM
can be applied; “all bets are off”, so to speak.1087 In the author’s view, the
core area of concern for the OECD should have been to provide guidance
on how to ensure that the profit-splitting factors selected are truly corre-
lated with the creation of value in the controlled transaction. After all, the
powers and realistic alternatives, as indicated by the relative worth of their value chain
contributions.
1081. BEPS Action 10, Revised guidance on profit splits, para. 53. This is, in reality,
the same as a split according to the relative bargaining powers of the respective parties.
On use of the concept of bargaining power in transfer pricing, see Blessing (2010b),
where it is argued that application of the concept should be confined to splitting residual
profits from IP value chains under the PSM. Blessing also ties the concept to the realis-
tic alternatives of the controlled parties; see Blessing (2010b), at p. 170. See also infra
n. 1125.
1082. BEPS Action 10, Revised guidance on profit splits, at para. 54.
1083. Id., at para. 64.
1084. Id., at para. 60.
1085. Id., at para. 66.
1086. Id., at para. 57.
1087. This may trigger a risk of diverging profit split assessments from taxpayers and
jurisdictions. Poor harmonization among jurisdictions in the application of profit splits
has already been highlighted by IFA branches; see Rocha (2017), at p. 231.
318
whole point is to ensure that the profit split reflects the relative value of the
controlled value chain contributions. To achieve this, it is crucial that the
profit-splitting factors chosen are capable of dividing profits pursuant to
underlying causality of value drivers. The focus of the current OECD TPG
on describing possible profit-splitting factors is likely due to the fact that
the PSM has been perceived to be subjective, supposedly caused by alloca-
tion keys that can be difficult to verify from objective evidence. Due to this,
there has likely been a drive towards presenting profit-splitting factors as
being objective.
The author is sceptical of this “objective” allocation key approach. Even

if the keys themselves are objective, in the sense that the number of em-
ployees, costs, etc. are easily verifiable, the relationship between the keys
and residual profits are not. Quite on the contrary, “objective” allocation
keys will normally be poor indicators of the relative value of unique IP
employed in a value chain. If “objective” allocation keys are used to dis-
tribute residual profits without any sanity check with respect to whether the
chosen allocation key actually is capable of reflecting underlying causality,
this will, in effect, be formulary apportionment.1088
The author finds it important to bear in mind that the PSM does not – and
cannot – allocate income with objective precision.1089 In his view, the most
reliable allocation of residual profits under the PSM will be based on a
concrete, facts-and-circumstances assessment of causality, not formulaic
allocation keys. When the precise influence (causality) of each item of IP
on the residual profits is impossible to ascertain objectively, an equal profit
split may realistically be the most sensible solution. In the author’s opinion,
no formulaic allocation key may replace the effectiveness of a critical func-
tional analysis and thorough assessment of the evidence in each particular
case. The elaboration on “objective” allocation keys contained in the 2017
OECD TPG can be seen as a step in the wrong direction.
An example of the unfortunate consequences that basing the split of re-

sidual profits under the PSM on “objective” formulaic allocation keys may
have is the experience of the US tax authorities under the previous cost-
sharing regulations, where taxpayers argued in favour of low buy-in valu-
ations based on rapid depreciation of capitalized intangible development
costs, resulting in the migration of valuable, US-developed intangibles at
1088. For discussions on theoretical proposals for profit splits based on formulary ap-
portionment, see, e.g. Avi-Yonah (2010); and Avi-Yonah et al. (2009).
1089. In this direction, see also Schön (2010a), at p. 248.
319
low valuations.1090 In the author’s view, there is a clear risk of BEPS if “ob-
jective” allocation keys are further embraced by the OECD. Instead, the
OECD TPG should more clearly emphasize that the split of residual profits
over the entire duration of the controlled agreement should reflect the rela-
tive values of the non-routine contributions of the controlled parties to the
intangible value chain. Only then will the allocation of intangible operat-
ing profits be aligned with the creation of intangible value.
Nevertheless, the use of “objective” allocation keys to split operating prof-

its may be more reliable when the PSM is applied to allocate operating
profits from value chains that rely solely on routine contributions, such as
the sale of generic products or services, as an alternative to the one-sided
methods in cases in which there are reliability concerns surrounding the
available comparables (e.g. in the “economically significant risks” applica-
tions discussed in section 9.2.). It may be reasonable to assume that the
potential causal link between such “objective” keys and routine market
returns is stronger than the case would be for a causal link between such
“objective” keys and residual profits.
9.4. The PSM in valuation scenarios
The PSM will, in practice, typically be used to determine an annual roy-

alty rate for intra-group transfers of limited rights to unique IP (licensing
scenario), i.e. to allocate profits among controlled parties on a yearly basis.
The method may, however, also be used for the purpose of determining the
value of full rights to IP that is transferred intra-group (valuation scenario).
The 2017 OECD TPG contain some guidance on this latter application of
the PSM.1091 While the US regulations lack a direct counterparty to this
general OECD profit split valuation guidance, the US cost-sharing regula-
tions allow the application of the PSM for the purpose of determining the
value of a buy-in payment (which is valuation in a specific context). The
author will tie some comments to this, with a focus on the OECD guidance.
1090. See the discussion of the 2007 US CIP valuation approach in sec. 14.2.3.
1091. The 2013 OECD Revised discussion draft on transfer pricing aspects of intangi-
bles (2013 RDD) introduced guidance on the application of the PSM in some particular
valuation contexts (including some examples); see paras. 166-170. The wording was
brought over without changes to para. 6.146 of the 2014 OECD Guidance on Transfer
Pricing Aspects of Intangibles, OECD/G20 Base Erosion and Profit Shifting Project,
Action 8: 2014 Deliverable (2014D), but was not finalized (as opposed to the guidance
for a range of other issues), indicating that there was controversy surrounding para.
6.146. The 2015 text contains the same language as that in the 2014D.
320
The PSM in valuation scenarios
Firstly, the 2017 OECD TPG assert that the PSM may be applied to price
a sale of full rights to IP.1092 This means that the method will be applied to
allocate among the controlled parties the value estimate that results from
a discounted cash-flow (DCF)-based valuation.1093 The author is not con-
vinced by the logic of the OECD TPG on this point. If the purpose is to
price the transfer of ownership of a specific item of IP from one controlled
entity to another, why should there be a basis to apply a profit split? Surely,
the transferer should be entitled to the entire value of an intangible that it
owns. In the author’s view, the method should be inapplicable in this case,
because the transferee will not have contributed value that should entitle it
to a split of the profits. He refers to the discussion of the US income method
to price the contribution of a pre-existing intangible to a cost-sharing ar-
rangement (CSA) in section 14.2.8.3., which is significantly similar to the
issue at hand in most respects.
Further, the 2017 OECD TPG suggest that the PSM, through a DCF valu-
ation, can be applied to pricing transfers of partially developed intangi-
bles.1094 The relative value of the IP-development contributions rendered
prior to the transfer (as embodied in the transferred intangible) and after
the transfer may then be used to determine the split. The value of partially
developed unique IP comes from its potential for coming to fruition, i.e.
reaching completion and being applied in a value chain to earn profits. A
valuation must therefore, in the author’s view, rest on two main estimates.
First, the future profits allocable to the completed IP must be determined,

likely using a DCF valuation. This is the “normal” assessment under the
PSM, albeit in a valuation context. The assessment will refer to the value
chain for a particular future product. The valuation must be based on the
best estimates of revenue, costs and growth in future periods, discounted
using a risk-adjusted rate. The net present value (NPV) of the operating
profits from the sales of the product must – as always under the PSM – be
allocated between the routine and non-routine contributions to the value
chain, with normal market returns and residual profits, respectively. If the
non-routine contributions also include other intangibles than the completed
IP, the valuation must include a split of the NPV of the residual profits
between these intangibles, based on the normal assessment of the relative
values of the non-routine contributions.
1092. OECD TPG, para. 6.149.

1093. The new guidance on valuation must be taken into account; see OECD TPG,
paras. 6.153-6.180. See also the discussion of the new 2015 guidance on valuation in ch.
13.
1094. OECD TPG, para. 6.150.
321
Second, it must be determined how much of the value allocable to the com-
pleted IP is due to the partially developed intangible. The question here
will, in practice, be to determine whether the transferer, subsequent to the
transfer of the partially developed intangible, will continue to perform the
important development functions.1095 If he does, the entire value allocated
to the completed IP should be used as the price for the partially devel-
oped intangible. For instance, it may be that the transferer has carried out
the important functions pertaining to R&D up until the transfer, and will
continue to do so subsequent to the transfer, while the transferee funds the
remaining development.
In this case, the transferer should be entitled to the entire residual profits
after a risk-adjusted return has been allocated to the intangibles’ develop-
ment funding.1096 It will normally be contrary to the alternatives realisti-
cally available to the transferer entity to not allocate all residual profits
to it. In reality, this is akin to applying the TNMM in the context of a
valuation. However, if the transferee does contribute important develop-
ment functions that contribute to the completion of the partially developed
intangible, the transferee should be allocated a portion of the NPV of the
completed IP that corresponds to the relative value of his development
contributions. This will reduce the value allocable to the transferer, and
thereby the arm’s length transfer price of the partially developed intangi-
ble. This solution would be akin to applying the PSM in the context of a
valuation.
The recent US cost-sharing regulations include provisions on the valua-

tion of pre-existing intangibles that are contributed to a CSA.1097 These
rules assume that the entire value of the pre-existing intangibles should
be allocated to the transferer, based on the view that the first-generation
intangible is the platform and the key value driver for subsequent ver-
sions of the intangible developed under the CSA. The regulations allow
a profit split only if the transferee contributes other unique inputs to the
cost-sharing activity.1098 Outside of CSAs, it cannot, in the author’s view,
always be presumed that the transferee should not be entitled to any
residual profits from a partially developed intangible. This issue must
be assessed concretely, and the question will be whether the transferee
1096. See sec. 22.4. for an analysis of the remuneration of IP development funding
under the 2017 OECD TPG.
1097. Treas. Regs. § 1.482-7.
1098. Treas. Regs. § 1.482-7(g)(7).
322
The OECD PSM is limited to information known or reasonably
foreseeable at the outset
contributes any important R&D functions to the remaining development

process.
Lastly, the 2017 OECD TPG state that the PSM can be applied in order to
price the transfer of limited rights to a fully developed intangible.1099 In
this scenario, the method is used to evaluate the relative values of the non-
routine contributions to the intangible value chain in order to determine the
split of the residual profits. The question is what portion of the combined
operating profits is allocable to the transfer of the limited rights to this
particular IP. In other words, this is the traditional licensing scenario seen
from the point of view of one of the parties contributing unique intangi-
bles to the value chain. The author refers to the discussion of the endorsed
OECD profit split allocation patterns in section 9.3.3.
9.5. The OECD PSM is limited to information known or

reasonably foreseeable at the outset
The 1995/2010 OECD TPG restricted application of the PSM to antici-
pated profits,1100 i.e. to profits that were “known or reasonably foreseeable”
at the time at which the controlled transaction was entered into.1101 The US
PSM is not limited in the same way, due to the US commensurate-with-
income approach, which ensures that the allocation of operating profits
under the method is aligned – in every income period – with the actual IP
value chain contributions of the controlled parties. The 1995/2010 OECD
limitation was relevant in two contexts: (i) when the initial taxpayer pric-
ing was based on the PSM; and (ii) when other pricing methods were ap-
plied.1102
With respect to context (i), the 1995/2010 OECD TPG warned that the
taxpayer “could not have known what the actual profit experience of the
business activity would be at the time that the conditions of the controlled
transaction were established”.1103
1099. OECD TPG, para. 6.152.

1100. 1995/2010 OECD TPG, paras. 2.128 and 2.130.
1101. This should be seen in light of the general stance that the 1995/2010 OECD TPG
took on the issue of periodic adjustments to controlled intangibles transfers with fixed
pricing structures. See the analysis of periodic adjustments in ch. 16.
1102. This structure stems from the 1994 OECD discussion draft (see paras. 139 and
141).
1103. OECD TPG, para. 2.128. This wording is, in essence, the same as in the 1994
OECD discussion draft, para. 139.
323
This logic is, in the author’s view, flawed. Expected profits are relevant, but
not decisive, for agreeing on a split. The point under the PSM is that the
controlled parties beforehand agree to split the operating profits, whatever
they may prove to be, based on an assessment of the relative value of the
functions, assets and risks that they are going to contribute to the intangible
value chain. Thus, the focus will be on the relative value of the inputs that
each controlled party brings to the value chain, not on the absolute amount
of monetary outputs. Logically, it should not matter whether the actual
operating profits prove to be significantly smaller or larger than projected
at the time at which the controlled transaction was entered into. Contra-
ry to the 1995/2010 OECD TPG restrictions on periodic adjustments of
controlled transactions that use fixed pricing structures, controlled PSM-
pricing provisions should be seen as dynamic, as the relative values of the
non-routine contributions may fluctuate during the course of the controlled
agreement.
With respect to context (ii), where the PSM is used to review an initial
pricing based on a method other than the PSM (e.g. the CUT method), the
1995/2010 OECD TPG limited application of the PSM to taking into ac-
count the “information known or reasonably foreseeable by the associated
enterprises at the time the transactions were entered into, in order to avoid
the use of hindsight”.1104
Here, the problem was whether the initial, and typically fixed, pricing
should be shielded from a subsequent PSM-based reassessment if the ac-
tual profits were outside what reasonably could have been expected at the
time at which the controlled transaction was entered into. This issue would
rarely be triggered in practice because the initial allocation of residual prof-
its would then have been founded on a CUT-based method, which would be
a rare occurrence.1105
Much has happened in international transfer pricing jurisprudence since

the “known or reasonably foreseeable” PSM delimitation was introduced
into the 1995 OECD TPG. The allocation of operating profits based on data
of actual profits is now commonplace. The widespread use of the TNMM
and taxpayer year-end adjustments is a testament to this.1106 The 2017
1104. OECD TPG, para. 2.130. See also paras. 2.11 and 3.74; and the discussion in the
1994 OECD discussion draft, para. 141.
1105. Such a reassessment would, in effect, constitute a periodic adjustment. The au-
thor refers to the discussion on periodic adjustments in ch. 16.
1106. See the analysis of year-end adjustments in ch. 15.
324
The OECD PSM is limited to information known or reasonably
foreseeable at the outset
OECD TPG have kept the “known or reasonably foreseeable” limitation,

but have watered it down considerably.1107 Controlled profit splits of actual
profits are now allowed when the controlled parties share the assumption
of the same economically significant risks or separately assume closely
related economically significant risks.1108 Outside of such scenarios, how-
ever, controlled profit splits are limited to anticipated profits.1109
Regardless of whether the agreed profit split refers to anticipated or actual

profits, the basis upon which the profits are split (i.e. the selection of profit
splitting factors, the calculation of profits, etc.) must generally be deter-
mined using information known or reasonably foreseeable at the time at
which the controlled transaction was entered into, unless major unforeseen
developments have occurred that would have resulted in a renegotiation of
the agreement had it occurred between independent parties. In order for
this restriction to bear any meaning, it must be interpreted as “locking”
the basis of the profit split at the time at which the controlled transaction
is entered into.
An arm’s length profit split should not, in the author’s view, be locked at the
outset. It should reflect the actual controlled value contributions on an an-
nual basis. It will often be the case that the relative value of the controlled
parties’ non-routine contributions to the intangible value chain will change
over the years. For instance, with a patent being the most valuable contri-
bution initially, the marketing IP becomes more valuable in later phases. If
the split is fixed at the outset and does not take into account such changes,
the split should not, in the author’s view, be regarded as set pursuant to the
relative values of the parties’ contributions, as causality will be lacking in
subsequent periods.
This is a problem. For instance, if the split was locked at 60/40 at the out-
set based on estimates of the values that would be contributed to the value
chain by the controlled parties while it later turns out that the split should
have been 50/50 if it was to reflect the actual values contributed, the locked
royalty will overcompensate one group entity and undercompensate the
other group entity based on the actual relative values. Such results should
not be accepted, as they do not reflect profit allocations that are aligned
with value creation. Profit splits that are locked at the outset represent ob-
vious risks of BEPS, as it will be difficult for tax authorities to prove that
1107. BEPS Action 10, Revised guidance on profit splits, para. 46.
1108. Id., at para. 44.
1109. Id., at para. 45.
325
the basis for the split was not founded on sound information at the time at
which the controlled transaction was entered into.
The OECD should, in the author’s view, discard the current remnants of
the “known or reasonably foreseeable” limitation and adopt an approach
akin to the US commensurate-with-income standard to ensure that the only
profit splits that are accepted are those that, on an annual basis, reflect the
actual relative values of the controlled parties’ value chain contributions.
326
Chapter 10
Location Savings, Local Market

Characteristics and Synergies
10.1. Introduction
In this chapter, the author will analyse the 2017 OECD Transfer Pricing
Guidelines (OECD TPG) on the allocation of incremental operating profits
due to location savings, other local market features and synergies.1110 These
profit allocation issues have not previously attracted much attention in the
OECD TPG.1111 The 2017 guidance should be seen as a response from the
OECD to the growing discontent among some jurisdictions, particularly
high-growth developing economies (e.g. China and India), with respect to
the modest income assigned to them by multinationals under controlled
profit allocations based on the one-sided pricing methods.1112 These juris-
dictions feel entitled to something more than just a normal market return
on routine functions.1113 It was therefore important for the OECD that these
issues were addressed as part of the larger BEPS Project.1114 The debate on
where incremental operating profits due to local market features should be
taxed is best seen as part of the larger discussion on the alignment of oper-
ating profits with value creation. The topic is not whether such incremental
profits should be taxed, but where (more precisely, whether the source state
should be entitled to tax these incremental operating profits).1115
Regarding terminology, both “location savings” and “other local mar-

ket features” fall under the umbrella term “location-specific advantages”
(LSAs).1116 In this chapter, however, the author will refer to location savings
separately in order to follow the structure of the 2017 OECD TPG. Further,
in line with established terminology, he will refer to incremental operating
1110. OECD TPG, paras. 1.139-1.173.

1111. Some brief and general comments were made in the 2010 OECD TPG; see paras
1.55 and 1.57.
1112. See the analyses in sec. 2.4. on the principal model and in ch. 8 on the one-sided
profit-based methods.
1113. For comments on the UN transfer pricing manual, see, e.g. Wittendorff (2012a).
For comments on the India chapter in the manual, see PwC (2012).
1114. The BRICS countries (Brazil, Russia, India, China and South Africa) partici-
pated in the deliberations of Working Party 6.
1115. On this topic, see Francescucci (2004a), at p. 72.
1116. On this topic, see also Schön et al. (2011), at pp. 82-83; and Wilkie (2014a).
327
Chapter 10 - Location Savings, Local Market Characteristics and Synergies
profits from location savings and LSAs as cost savings and location rents,
respectively.
10.2. A lead-in to the topic: The incremental operating

profits at stake
As discussed in the previous chapters, the OECD transfer pricing methods
rest on the fundamental distinction between routine and non-routine value
chain contributions. The methods fundamentally assume that the residual
profits that are left after a normal market return has been attributed to rou-
tine inputs are due solely to the non-routine inputs employed in the value
chain (normally, unique intangible property (IP)). The methods are, in this
sense, rudimentary tools. They see only two things, namely (i) routine in-
puts entitled to a normal market return; and (ii) non-routine inputs entitled
to the residual profits.1117 While the economic logic behind this paradigm
is undoubtedly sound, critical voices are questioning the view on causality
inherent to – as well as the equity of – the assumption that residual profits
are entirely due to unique IP.
This problem goes to the core of why multinationals exist. These enter-
prises reap a range of benefits that solely domestic enterprises do not. The
critical voices argue that residual profits in fact include some incremental
operating profits that are caused by cost savings, other local market char-
acteristics and synergies, and that these profits should be taxable at the
source (as opposed to being allocated as royalties to a foreign IP-owning
group entity). This argument is fuelled by the perception that there may not
always be a bright line between unique IP and LSAs.1118
The author would like to illustrate the issue at hand through an example.
Imagine that a multinational based in Norway owns valuable patents and
trademarks for an established luxury yacht brand with an outstanding repu-
tation. It decides to enter the Chinese market and distribute and sell yachts
there. China has a large and growing market for such luxury products,
consumers with significant purchasing power and a preference for Western
products. Also, the degree of local competition for this specific type of
boat is limited, and the infrastructure for doing business in China is ideal.
These local market features will likely interact with the unique IP that is
1117. See also Francescucci (2004a), at p. 73, on this point.

1118. The author will revert to this issue in the discussion of the allocation of profits
from LSAs in sec. 10.7.
328
A lead-in to the topic: The incremental operating profits at stake
applied in the yacht value chain, in the sense that, under the specific market
conditions in China, the IP is more valuable in this market than in some
other markets.
Let us assume that the group distribution and sales entity in China realizes
the following operating profits in 2017:
Sales 10,000
Cost of goods sold (COGS) 2,200
Gross profit 7,800
Operating expenses 3,300
Operating profit 4,500
The activities performed locally in China are routine assembly, distribution

and sales functions. Let us say that the transactional net margin method
(TNMM) is applied to remunerate these functions, and the arm’s length
range displays a median of a 6% operating margin. This will entail that
the Chinese distribution entity should be allocated an operating profit of
600 as taxable income. The residual profits of 3,900 are allocated to the
Norwegian parent as royalty payments for the unique IP licensed to the
Chinese distribution entity. Similar results are likely not entirely uncom-
mon in practice under the widespread application of the one-sided methods
by multinationals.
Let us for now assume that, in theory, it is possible to break down the Chi-
nese operating profits of 4,500 as follows:
Operating profits due to routine functions 600

Operating profits due to unique intangibles 3,300
Operating profits due to cost savings 200
Operating profits due to other LSAs 400
As in the outset scenario, it will be straightforward to determine that the

normal market return for routine functions of 600 should be allocated to
China and that the residual profits due to the unique intangibles of 3,300
should be allocated to Norway. What is more challenging is to determine
which jurisdiction should be allocated the right to tax the 200 in cost sav-
ings and the 400 in location rent. That question – or more specifically, the
329
OECD’s position on it – will be the theme of discussion in the rest of this

chapter.
Before the author embarks on that analysis, however, he would like to

comment on some basic positions pertaining to how incremental operating
profits from location savings and LSAs arguably should be allocated.
10.3. Which jurisdiction should be entitled to tax

incremental operating profits: The basic arguments
The transfer pricing rules generally turn on causality. For instance, the
one-sided methods allocate normal market returns on routine value chain
contributions because such inputs create normal returns. Further, the
profit split method (PSM) splits residual profits among each unique item
of IP employed in the value chain based on the assumption that each
item of IP contributes to the creation of the residual profits (in propor-
tion to the relative IP values). The same logic applied to cost savings and
location rents results in the argument that incremental operating profits
should be allocated to the jurisdiction where they materialize, i.e. the
source state.
The typical counter-argument to this causality position is that incremental

operating profits should be allocated based on the bargaining power of the
group entities participating in the controlled transaction. This argument is
closely linked to the underlying logic of the current OECD transfer pricing
methodology, which attributes the greatest bargaining position to group
entities that contribute non-routine value chain inputs. Thus, group enti-
ties that perform only routine functions (e.g. contract manufacturing and
distribution) will have a lesser bargaining position than group entities that
own unique IP. This reasoning will result in the extraction of cost savings
and location rents from source state taxation by foreign group entities that
own the unique intangibles.
Further, a multinational could object to the allocation of incremental op-

erating profits to source states’ routine value chain contribution entities
by arguing that it could alternatively have outsourced the relevant routine
functions to a third-party service provider in the same market for a price
that would pass along the incremental operating profits to it as a buyer and
only allocate a normal market return to the provider as compensation for
its generic functions. This is equal to claiming that comparable third-party
enterprises in the same market are unable to retain cost savings and loca-
330
What are location savings?
tion rents due to price pressure from competitors. This argument goes to
the heart of the arm’s length principle. If the third-party routine function
providers were truly comparable to the tested party, it would be contrary to
the arm’s length principle to allocate more income to the controlled party
than to an unrelated party that carries out comparable transactions under
similar circumstances. The outsourcing argument is forceful, but only
valid if it indeed is a realistic alternative for the multinational in question
to outsource the relevant activity to third parties. That may not always be
the case. Factors such as highly integrated value chains, the protection of
sensitive business information from potential competitors and quality con-
trol may keep it from outsourcing its business activity to third parties. A
multinational will generally want to protect its valuable IP.1119
Notions of equity may also be used as arguments for allocating incremen-

tal operating profits, from the perspective of both source states and mul-
tinationals. Perceptions of equity are subjective. The author’s view is that
cost savings and location rents should be allocated to the source state if
the following two factors are present: (i) it indeed is possible to distinguish
incremental operating profits caused by location savings and LSAs from
profits due to the unique IP employed in the value chain; and (ii) the busi-
ness activity of the local entity could not realistically have been outsourced
to third parties. This may be the case if the local entity contributes unique
inputs to the value chain (e.g. goodwill or know-how) and therefore is enti-
tled to a cut of the residual profits. It may also be the case if the local entity
is a mere routine input provider, but the multinational, for business reasons
(secrecy, quality control, etc.), is unable or unwilling to outsource the rel-
evant functions to third parties.
10.4. What are location savings?
The 2017 OECD TPG refer to location savings as “cost savings attributable
to operating in a particular market”.1120 The notion of location savings is
also discussed in the business restructuring guidance, where it is stated
that “location savings can be derived by an MNE [multinational enterprise]
group that relocates some of its activities to a place where costs … are
lower than in the location where the activities were initially performed”.1121
1119. See sec. 2.3. for comments on business rationales underlying intangible property
(IP) value chains.
1120. OECD TPG, para. 1.139.
1121. OECD TPG, para. 9.126.
331
Thus, location savings fundamentally make up a relative concept that is

only meaningful if the costs of performing an activity in one market are
compared to the costs associated with a realistic alternative of performing
the same activities in another market. This is, as the author will revert to
below, a problematic point of departure. Arm’s length pricing is based on
benchmarking controlled profit allocations against the profits of compara-
ble third parties, which implies that the controlled and third parties must
carry out their business in the same market.
In the context of a business restructuring, it will be easy to identify the

alternative market. This will be the high-cost country in which the activity
was performed prior to the relocation. Location savings are the specific
cost savings of performing the business activity in the country to which
the activity was relocated, compared to the costs that were incurred in the
country where the activity was performed prior to the restructuring.
In other contexts, however, it does not seem obvious as to how cost savings
should be determined. For instance, a multinational carries out distribu-
tion, marketing and sales in many jurisdictions. The salaries paid to the
local employees, as well as other operating expenditures, such as housing
costs and social security payments, may be lower in China and India as
compared to, for instance, France or Norway. If one were to determine the
location savings that accrue to the Chinese entity, which country should
one choose as the cost benchmark? Obviously, there is no true alternative
country available, in the sense that distribution, marketing and sales in
China can only be carried out there. The 2017 OECD TPG do not offer any
indication of how to solve this.
One possible solution could be to use comparative average cost data from
the multinational’s operations in other jurisdictions (average hourly wage,
rent per square feet, etc.). Alternatively, one could use general statistical
data available on wages, rent, etc. in different jurisdictions. This data could
then be compared to the costs of the relevant local entity in order to deter-
mine whether any location savings have accrued to it.
These approaches would result in hypothetical cost savings from not per-
forming the activities in other markets, as opposed to cost savings com-
pared to similar unrelated entities operating in the same market. Unrelated
entities in other markets are not comparable to the tested party, as they op-
erate under different market conditions. The approach would be contrary to
the arm’s length principle, which requires parity in the treatment between
group entities and comparable third-party enterprises. By allocating such
332
The allocation of cost savings
hypothetical cost savings to a local entity, the local jurisdiction would eat
into operating profits that should be taxed in the jurisdiction where the
other party to the controlled transaction is resident.
Thus, if the Chinese entity in the example in section 10.2. is a routine con-
tract manufacturer, its profits should be compared to similar unrelated con-
tract manufacturers in China. If the profits of the Chinese entity lie within
the arm’s length range of the operating margins realized by the comparable
third-party contract manufacturers, no additional income should be attrib-
uted to it. The author will not go further into this here, as the allocation of
cost savings will be discussed in section 10.5. However, because the notion
of cost savings is a relative concept, it was necessary to provide these com-
ments in order to introduce the concept.
10.5. The allocation of cost savings
10.5.1. Introduction
The 2017 OECD TPG on the allocation of cost savings draw upon the
solutions carved out in the 2009 business restructuring guidance.1122 The
OECD position is that retained cost savings should be allocated as inde-
pendent enterprises operating under similar circumstances would have
done.1123 Two scenarios are discussed in the OECD TPG, namely the allo-
cation of cost savings in the presence and the absence of local comparables,
which are analysed in sections 10.5.2.-10.5.3.
10.5.2. Local comparables are available
The OECD’s position is that location savings most reliably can be allocated
based on profit data extracted from “comparable entities and transactions
in the local market”.1124 If such local comparable third-party enterprises
can be identified, no comparability adjustments shall be performed. This
position will, in practice, likely result in full extraction of incremental op-
erating profits due to cost savings from the source jurisdiction to a foreign
group entity. The result is contrary to the causality argument that value
should be taxed where it is created. It may also be of questionable equity.
1122. 2010 OECD TPG, paras. 9.148-9.153; see also OECD TPG, para. 1.140.
1123. For the US position on location savings, see the discussion in sec. 6.6.5.4.
1124. OECD TPG, para. 1.142.
333
The result is, however, consistent with the view that group entities that pos-
sess little in the way of bargaining power should only reap a normal rate of
return on their routine value chain contributions.1125
The result is also consistent with a multinational’s realistic alternative of

outsourcing routine functions to third parties for a normal market fee that
passes along the cost savings to it. This argument is, as mentioned, not rel-
evant if the multinational is unable or unwilling to outsource local func-
tions.1126 If so, the use of local comparables to benchmark the profits of the
tested party will break away from the fundamental principle that transfer
pricing should be aligned with the alternatives realistically available to the
controlled parties. This is not a trivial problem. The question is whether the
realistic alternatives metanorm, or the specific rule that local comparables
may be used if they are available, should prevail. The wording of the specific
rule seems to convey a conscious position of the OECD on this issue. The
author’s understanding of paragraph 1.142 of the OECD TPG is that it forces
the use of local comparables, even if it concretely would not be realistic for
the multinational to outsource the functions performed by the tested party.
The specific rule should thus trump the realistic alternatives principle based
on a lex specialis view.
The use of local comparables to remunerate group entities in markets with

pronounced location savings (and other LSAs) may cause disagreements
between a multinational and the foreign jurisdiction, on the one side, and
the source jurisdiction, on the other. The multinational and the foreign ju-
risdiction will likely argue that they have complied with the OECD TPG
as long as they have remunerated the local group entity under a one-sided
method using local comparables. The source state could argue that the ap-
plication of one-sided pricing methodology using local comparables will
not be reliable unless comparability adjustments are carried out (specifical-
1125. For an informed discussion of the use of the bargaining power concept in transfer
pricing, see Blessing (2010b), at p. 168. Blessing warns, at p. 173, that the concept is
not suitable for application outside the context of allocating residual profits from IP
value chains under the profit split method. His text was written before the 2017 OECD
guidance on the allocation of incremental profits from location savings, local market
characteristics and synergies was issued. Based on Blessing’s reasoning, however, the
author finds it likely that his warning would not apply to the use of the bargaining
power concept for allocating incremental profits from location savings and the like, as
this indeed lies close to the allocation of residual profits under the profit split method.
For further information on bargaining power, see Parekh (2015), at p. 305.
1126. See the comments on the reasons underlying foreign direct investment (FDI) in
sec. 2.3.
334
ly, that the operating profits of the local comparables should be increased to
reflect hypothetical cost savings).
The author will illustrate this through an example. Let us assume that a
multinational in the consumer electronics industry has a low-risk distribu-
tion entity in China that performs routine distribution, marketing and sales
functions. The entity licenses valuable marketing IP from a foreign group
entity. Using Chinese third-party distributors of electronic devices as com-
parables, the arm’s length range under the TNMM shows that the low-risk
distributor (LRD) should earn an operating margin between 6-10%, with
a median of 8%.
Chinese LRD TNMM

(with cost savings) (local comparables)
Sales 10,000
COGS 3,000
Gross profit 7,000
Operating expenses 2,000
Operating profit 5,000 800
Operating margin 50% 8%
This will result in the allocation of an operating profit of 800 to the Chi-
nese LRD. The residual profits of 4,200 are, in their entirety, allocated to
the foreign licenser entity as royalties for the licensed marketing IP. It is,
however, clear that the Chinese LRD benefits from hypothetical location
savings in the form of low labour and housing costs. Compared to the aver-
age costs of these items in the other jurisdictions in which the multinational
operates, the Chinese LRD realizes cost savings of 20%. If one adjusts for
these location savings, the following operating profit data is achieved:
Chinese LRD Chinese LRD Difference

(with cost savings) (without cost savings)
Sales 10,000 10,000
COGS 3,000 3,600 600
Gross profit 7,000 5,800
Operating expenses 2,000 2,400 400
Operating profit 5,000 4,000 1,000
Operating margin 50% 40%
335
The hypothetical cost savings are 1,000. If these are added to the normal
market return of 800 that the LRD is supposed to earn pursuant to the
TNMM, the total income taxable in China would be 1,800, with the resid-
ual profits of 3,200 allocated to the foreign licenser entity. This theoretical
example of a source state reassessment illustrates two things.
First, a comparability adjustment was carried out for the local third-party
comparables in order to take into account hypothetical location savings
that were not reflected by said comparables at the outset. Through this ad-
justment, the source state was able to identify that 1,000 of the LRD’s op-
erating profits were due to hypothetical location savings. As discussed, the
OECD’s position is that such adjustments should not be performed.1127 This
position, if a one-sided method is applied to allocate income to a tested
party, will ensure that the entire local operating profits, apart from a nor-
mal market return on the routine functions carried out by the local LRD,
are allocated to the other party to the controlled transaction, similar to the
outset scenario in the example above.
Second, the example allocated the entirety of the hypothetical location

savings to the tested party. This may arguably be justified by the cau-
sality argument: because the hypothetical cost savings were created in
the source state, in the sense that the multinational would incur rela-
tively higher costs had it alternatively performed the same functions in
a different, high-cost, jurisdiction, the hypothetical incremental profits
should be taxed in the source state. To this, a multinational could object
on two grounds. First, it would be contrary to the arm’s length principle
to allocate hypothetical cost savings to the source state. The Chinese
group entity would be made worse off than comparable unrelated con-
tract manufacturers also operating in China. Second, the multinational
alternatively could have outsourced the routine functions performed by
its local entity to unrelated third parties in the same jurisdiction for a
normal market fee. The MNE would then be able to extract the location
savings.
The author finds that the economic logic underlying the OECD’s position
is convincing. Transfer pricing rules must ensure that group entities are
subject to the same treatment as comparable unrelated enterprises. How-
ever, the author is not entirely convinced that this result necessarily will
1127. OECD TPG, para. 1.142, last sentence.
336
provide an equitable allocation of operating profits among jurisdictions in

all cases.
At least to this author’s knowledge, one of the first transfer pricing cases
that properly dealt with the issue of location savings was Eli Lilly.1128 In
that case, the location savings that were derived from manufacturing the
Darvon pill in Puerto Rico were entirely allocated to the local contract
manufacturing subsidiary. At that time, the US Tax Court did not even
question whether this was the correct allocation choice. After all, it was
clear that the location savings were sourced from Puerto Rico and that they
were distinct from profits both from the manufacturing patents owned by
the Puerto Rican subsidiary and the marketing IP owned by the US parent.
The author finds this worthy of contemplation.1129
It seems, however, that multinationals over time have designed their trans-
fer pricing structures so that profits from location savings are extracted
from local market jurisdictions, typically in connection with the transfer
pricing of IP, labelling incremental profits as residual profits. This practice
is, of course, fully in line with the accepted OECD transfer pricing meth-
ods, in particular the TNMM. As a commentary on these practices, the
2017 OECD TPG express that “the transfer of intangibles … may make it
possible for one party to the transaction to gain the benefit of local market
advantages … in a manner that would not have been possible in the absence
of the … transfer of the intangibles”.1130
The tax authorities in developing countries, such as China and India, have
clearly stated that they do not find it equitable that the entire profits from
cost savings (or other LSAs) are extracted from source taxation.1131 Chi-
nese tax authorities have expressed that they find that the extraordinarily
1128. See the analysis of the case in sec. 5.2.4.3.

1129. It should, however, be noted that in Eli Lilly and Company and Subsidiaries
v. CIR (84 T.C. No. 65 [Tax Ct., 1985], affirmed in part, reversed in part by 856 F.2d
855 [7th Cir., 1988]), the relevant manufacturing functions were migrated from the
United States to Puerto Rico, giving the factual pattern a distinct business restructuring
flavour. The allocation of incremental operating profits from cost savings in Eli Lilly
should therefore not be taken as a strong argument in favour of source state taxation
outside the context of business restructurings.
1130. OECD TPG, para. 1.148.
1131. It was apparently important for the United States to reject the proposal from
China in the run-up to the 2015 revision of the OECD TPG providing that incremental
profits from cost savings and local market characteristics should be allocable to the lo-
cal group entity. See Brauner (2017), at sec. 2.3.1.
337
high profits displayed by local group entities are “rightly earned” by the
Chinese taxpayer entity1132 and that they will tax the entire location savings
from contract R&D carried out by Chinese group entities.1133
Comparatively, the Indian tax authorities favour the PSM for allocating
cost savings (and location rents).1134 They view the competitiveness of the
Indian market, the availability of substitutes and cost structure as key ele-
ments in determining the bargaining power of the controlled parties. This
is, as far as the author can see, effectively the same as deeming these fac-
tors unique value chain contributions. The Indian tax authorities will thus
likely not accept full allocation of cost savings to a foreign group entity,
even if the local group entity pursuant to a traditional functional analysis
focused solely on functions, assets and risks is sufficiently remunerated
under a one-sided pricing method.
While the Chinese and Indian tax authorities differ in their methodological
approaches, they both agree that local comparables cannot be used to de-
termine the benefit from cost savings.1135 The Indian tax authorities argue
that benchmarking the tested party based on local comparables would not
be consistent with the arm’s length principle, as any arm’s length trans-
action fundamentally will be reliant on both parties benefiting from the
transaction. Thus, cost savings (and location rents) should be split between
the controlled parties in a manner that leaves neither party with none or all
of the incremental profits,1136 seemingly in contrast to the resolute view of
the Chinese tax authorities.
1132. 2017 United Nations Practical Manual on Transfer Pricing for Developing Coun-
tries (UN PMTP 2017), para. D.2.4.4.7.
1133. UN PMTP 2017, para D.2.4.4.9, provides an example of a contract R&D scenario
in which this position is applied. Taxation of the location savings is carried out by way
of rejecting the costs of the local Chinese subsidiary as the base upon which the profits
are calculated, instead using the average cost base for the multinational’s R&D centres
in developed countries. In the example, the cost base of the Chinese entity is 100, while
the average cost base of R&D centres in developed countries is 150. The relevant full
cost mark-up is 8%. If applied to the cost base of the Chinese entity, it would yield a
taxable income of 100 × 0.08 = 8. The income is, however, calculated on the basis of the
higher cost base of the R&D centres in developed countries of 150, yielding an income
for the Chinese entity of 150 × 0.08 = 12, thereby taxing the entirety of the location
savings in China.
1134. UN PMTP 2017, para. D.3.7.3.
1135. UN PMTP 2017, paras. D.2.4.3 and D.3.7.4 for China and India, respectively.
1136. Compare UN PMTP 2017, paras. D.2.4.4.11 and D.2.4.4.12 for China with para.
D.3.7.3 for India.
338
In conclusion, the current OECD stance with respect to intra-group alloca-

tion of cost savings in the presence of local comparables is clear. The al-
location must be based on local comparables, and no comparability adjust-
ments shall be made to the data.1137 The author suspects that this position
will result in continued use by multinationals of the TNMM to remunerate
their local routine entities. It appears that the objections from develop-
ing countries have not been accepted by the OECD. It remains to be seen
whether these jurisdictions will adhere to this rule.1138
10.5.3. Local comparables are unavailable
There will often be local comparables available on which to base bench-

marking of local routine entities under a one-sided method. This will, in
part, be due to the relaxed comparability requirements under the TNMM.1139
Nevertheless, local comparables may be unavailable in some cases, for in-
stance, if the tested party performs highly specialized functions within a
small niche market. The 2017 OECD TPG state that the allocation of cost
savings in such cases shall be based on the fact that “the conditions that
would be agreed between independent parties would normally depend on
the functions, assets and risks of each party and on their respective bar-
gaining powers”.1140
This position is illustrated through an example in which a multinational

relocates the routine manufacturing of branded clothes to a low-cost coun-
1137. OECD TPG, para. 1.142. This will result in the extraction of incremental profits
due to cost savings from the source state, which is the same result as under the 1995
OECD TPG; see Francescucci (2004a), at p. 73.
1138. For insightful and critical comments on the OECD position, see also Brauner
(2016), at p. 104, as well as at p. 109, where it is stated that “ the final report does not
pick up on the locational savings point which rejects the demand of developing coun-
tries such as China and India for profit allocations to market economies. This decision
demonstrates the failure of the developing countries in BEPS, the importance of agenda
setting, where changes require consensus and are otherwise rejected to the benefit of
those enjoying the status quo … and the success of the OECD in the deferral of the dis-
cussion of the demands of source countries for more taxation with its pseudo economic
agenda of arm’s length tweaking”.
1139. See the conclusions on the interpretation of the transactional net margin meth-
od’s (TNMM’s) comparability requirements in sec. 8.6.6.6.
1140. OECD TPG, para. 1.143, see also OECD TPG, para. 9.149. See the critical com-
ments on this solution in Brauner (2016), at p. 104: “At what point of difference would it
be better to consider a transaction as not having appropriate comparables and deserving
of special measures (beyond arm’s length)? These questions were left unresolved.”
339
try.1141 With respect to whether local cost savings should be allocated to the
low-cost country manufacturing entity, the OECD TPG state:
[I]n such a situation, a contract manufacturer at arm’s length would gener-
ally be attributed very little, if any, part of the location savings. Doing other-
wise would put the associated manufacturer in a situation different from the
situation of an independent manufacturer, and would be contrary to the arm’s
length principle.1142
This position is based on the premise that the multinational “has the op-
tion realistically available to it to use either the affiliate … or a third party
manufacturer”.1143 The logic behind this is that the OECD assumes that a
third-party manufacturing entity would pass along all of its cost savings
to the multinational. While that may be true, it cannot automatically be
assumed that it would be a realistic option for a multinational to outsource
part of its value chain to third parties, as there may be a need to, for in-
stance, protect business secrets.1144 Nevertheless, this “bargaining power”
rule that the OECD has now established to allocate profits from cost sav-
ings will generally strip local routine group entities from all profits from
cost savings, as such entities normally will be deemed to have inferior bar-
gaining power as compared to group entities that own unique and valuable
IP.1145
Even though the above-quoted wording is relatively clear on the point that
the tested (routine) party shall normally not retain any incremental profits
from cost savings, it must be equally clear that the OECD’s position only
applies to group entities that do not contribute any unique value chain in-
puts. Thus, if a local manufacturing or distribution entity owns valuable
know-how or local marketing IP, it may have a relatively strong bargaining
position and should likely be allocated a portion of the profits from cost
savings.
The OECD’s position is that “in appropriate circumstances … e.g. if there

are significant unique contributions such as intangibles used by both …
the use of a transactional profit split may be considered”. Thus, the OECD
approves of the use of the profit split method here to allocate profits (in-
1141. OECD TPG, paras. 9.150-9.151.

1142. Id.
1143. OECD TPG, para. 9.151.
1145. See Blessing, supra n. 1125.
340
What are other LSAs?
cluding incremental profits from cost savings).1146 The 2017 OECD TPG
include an example of when such a profit split approach would be appropri-
ate, in which a multinational outsources specialized engineering functions
to a subsidiary in a country where wage costs are significantly lower and
the subsidiary is in possession of technical know-how.1147 Cost savings are
allocated along with the residual profits in proportion to the relative values
of the contributed IP, based on the logic that the relative IP values are in-
dicative of the bargaining positions of the parties, and therefore that they
should determine also the allocation pattern for location savings.
10.6. What are other LSAs?
The following discussion is not limited to the business restructuring con-

text, but is geared towards the more general situation in which a multina-
tional performs its activity in a range of jurisdictions that offer varying
degrees of LSAs.
According to the 2017 OECD TPG, “other local market features” include
(i) purchasing power and product preferences of local buyers; (ii) whether
the local market is growing or contracting; (iii) the degree of competi-
tion; (iv) local infrastructure; (v) the availability of skilled workers; (vi) the
proximity to markets; and (vii) other “similar” factors that may affect local
operating profits.1148 These factors affect distribution, marketing and sales.
Thus, most LSAs pertain to the marketing and sales part of a value chain.
Further, LSAs will normally be more pronounced in developing than in
developed countries, typically due to market growth and increasing con-
sumer bases, a preference for Western products, the availability of skilled
workers, etc. LSAs attract both horizontal and vertical foreign direct in-
vestment.1149 For instance, Western multinationals in the consumer elec-
tronic business may invest in China to take advantage of highly-educated,
low-cost labour and a sophisticated network of suppliers and operate as-
sembly plants in China in order to take advantage of being in close proxim-
ity to the market.1150
1146. Incremental profits would be split also under the 1995 OECD TPG if the profit
split method were applied; see Francescucci (2004a), at p. 73.
1147. OECD TPG, paras. 9.152-9.153.
1148. OECD TPG, para. 1.144.
1149. See the discussions on FDI in secs. 2.2. and 2.3.
1150. See UN PMTP 2017, para. D.2.4.4.1.
341
The fundamental assumption is that LSAs contribute to increased oper-

ating profits for the local group entity, either through increased sales or
decreased COGS or operating expenses, relative to what the profits would
have been without the LSAs. Nevertheless, LSAs are elusive. First, while
hypothetical cost savings can be estimated with reasonable reliability
based on third-party benchmarks of cost levels in other markets, the same
cannot be said for most LSAs. For instance, how is one to determine the
amount of increased sales in China due to specific local consumer pref-
erences? Second, and likely most important, some LSAs may be closely
connected to the concept of unique IP.1151 It may, for instance, be difficult
to determine when an LSA (e.g. local consumer preferences) becomes so
particular that it must be deemed unique IP (e.g. goodwill) owned by the
local group entity. In such cases, it may be challenging to separate the loca-
tion rent from the particular LSA from the residual profits generated by the
unique IP employed in the value chain.
Further, in markets where there are significant LSAs present, there will
likely be strong interactions between the unique IP employed in the value
chain and the LSAs. Trademarks and goodwill related to specific products
will typically interact with LSAs, such as consumer preferences and pur-
chasing power, adding increased local value to the IP. It might therefore
be useful to think of LSAs as drivers that can add more value to unique IP
owned by a multinational. The more valuable the unique IP is (e.g. the Ap-
ple logo) and the more pronounced the LSAs are (e.g. Chinese consumers’
preference for Western products), the stronger this interaction will likely be.
An interesting question is whether the added value resulting from this in-
teraction should be viewed as a non-routine value chain contribution from
the local group entity. If so, it will be clear that the local entity shall be
entitled, as the owner of the local IP, to a portion of the residual profits
from local sales, i.e. a profit split approach shall be applied. If not, the local
entity will only be entitled to a normal market return on its routine contri-
butions under a one-sided approach.
The point of departure is, of course, that LSAs are not IP. If the interactions
between the unique IP owned by a multinational and LSAs are strong and
particular enough, however, the existence of local marketing IP should be
recognized (e.g. know-how or goodwill).
1151. On this issue, see also Wilkie (2014a), at p. 356.
342
The allocation of location rents
10.7. The allocation of location rents
As for location savings, the existence and extent of location rents must be de-
termined, as well as whether they are passed along to suppliers or customers.
If location rents that are not passed along are found to exist, the question will
be of how these shall be allocated among the group entities contributing to
the controlled transaction under the 2017 OECD TPG. This will be discussed
in the following sections, with respect to scenarios in which local compara-
bles are present and absent in sections 10.7.2. and 10.7.3., respectively.
10.7.2. Local comparables are available
The position taken in the 2017 OECD TPG is that location rents should be
allocated among group entities based on:
[…] comparable uncontrolled transactions in that geographic market between
independent enterprises performing similar functions, assuming similar risks
… such transactions are carried out under the same market conditions as the
controlled transaction, and, accordingly… specific adjustments for features of
the local market should not be required.1152
It follows from this that incremental operating profits due to location rents
may be allocated based on unadjusted local comparables. This entails that
profit data from functionally comparable local third-party enterprises can
be used under the TNMM to allocate a normal market return to the tested
party.1153 This OECD’s position will likely extract all or most of the loca-
tion rents from source taxation.1154
10.7.3. Local comparables are unavailable
If profit data from comparable local third-party enterprises upon which to

base the allocation of incremental operating profits due to location rents
1152. OECD TPG, para. 1.145.

1153. See also Wittendorff (2016), at sec. III.D. See the analysis of the TNMM compa-
rability requirements for the use of aggregated third-party profit data in sec. 8.6. of this
book.
1154. This is the same result as under the 1995 OECD TPG; see Francescucci (2004a),
at p. 73. See also the comments from Brauner (2016), supra n. 1138.
343
cannot be identified, the position taken in the 2017 OECD TPG is that it
will be necessary to consider the following:1155
− whether a market advantage or disadvantage exists;
− the amount of any increase or decrease in operating profits, vis-à-vis
identified comparables from other markets, that are attributable to the
local market’s advantages or disadvantages;
− the degree to which benefits or burdens of local market features are
passed on to independent customers or suppliers, as well as where they
are not fully passed on; and
− the manner in which independent enterprises operating under similar
circumstances would allocate such benefits or burdens among them-
selves.
Foreign comparables may be used under the 2017 OECD TPG to ascer-
tain the amount of location rent. This will result in a hypothetical location
rent, in the sense that the tested party is compared to third-party enterprises
operating in other markets. The author imagines that it will likely be chal-
lenging to ascertain whether the incremental operating profits of the foreign
comparables are in fact due to foreign LSAs or to unique IP. Once the hy-
pothetical location rent has been identified, the question is of how it shall
be allocated. As for the allocation of cost savings, this will depend on the
bargaining power of the group entities involved in the controlled transac-
tion.1156 If the local party is a routine input provider, the location rent should
be allocated in full to the party that contributes non-routine inputs to the
value chain. If the local entity contributes unique IP, it is entitled to a portion
of the residual profits (including location rents) that is commensurate with
the relative value of its non-routine contribution to the value chain. Thus, a
profit split approach is adopted by the OECD in this latter scenario.1157
10.8. Synergies
The 2017 OECD TPG state:

MNE groups … may benefit from interactions or synergies amongst group
members that would not generally be available to similarly situated independ-
ent enterprises. Such group synergies can arise … as a result of combined pur-
1155. OECD TPG, para. 1.146. See the critical comments on this OECD position in
Brauner (2016), supra n. 1140.
1156. OECD TPG, para. 1.146. See also OECD TPG, para. 9.153; and Blessing, supra
n. 1125.
1157. Incremental profits would be split also under the 1995 OECD TPG if the profit
split method were applied; see Francescucci (2004a), at p. 73.
344
Synergies
chasing power or economies of scale, combined and integrated computer and

communication systems, integrated management, elimination of duplication,
increased borrowing capacity, and numerous similar factors.1158
Positive synergies may result in incremental operating profits.1159 While the

previous generation of the OECD TPG did not specifically address the is-
sue of how incremental operating profits due to group synergies shall be
allocated among group entities, the 2017 OECD TPG deal with the issue.
The point of departure taken in the OECD TPG is that no separate alloca-
tion is necessary for incremental operating profits due to incidental benefits
that a group entity realizes solely as the result of being part of a larger MNE
group.1160 This will be the situation when a subsidiary, due to the high credit
rating of its parent, is able to borrow from unrelated lenders at a lower rate
than it could have attained on a stand-alone basis and then borrows at the
same rate from a related lender.1161 The interest rate charged on the internal
loan is at arm’s length because it is the same rate charged in the uncon-
trolled transaction. No adjustment is required with respect to the synergy
that enabled the subsidiary to borrow at the low rate from the unrelated
lender, as it arose from the subsidiary’s group membership alone and not
from any deliberate, concerted action by the MNE. As there is no require-
ment to separately allocate incremental operating profits from this category
of synergies, the profits will blend into the residual profits from unique IP.
Separate allocation is, however, required for incremental operating profits

due to synergies from deliberate, concerted group actions,1162 for instance,
benefits in the form of group discounts due to centralized group procure-
ment1163 or if the parent in the example above had provided a guarantee to
1158. OECD TPG, para. 1.157. On the allocation of profits from synergies, see, in par-
ticular, the thorough discussion in Peng (2016). For a recent comparative overview of the
treatment of synergies in different (IFA branch) jurisdictions, see Rocha (2017), at p. 218.
1159. See Kane (2014) for an analysis of whether synergy value should be character-
ized as an intangible, with the profits allocable by way of a formula. See also Koomen
(2015b), at p. 240 on synergies, as well as Brauner (2015), at p. 78.
1160. OECD TPG, para. 1.158 (see also para. 7.13). This is also the result under the
US regulations (which deal with synergies as comparability factors), pursuant to which
passive association benefits are not compensable; see Treas. Regs. § 1.482-9(l)(3)(v).
See Odintz et al. (2017), at sec. 2.2.4.2 on this point. On the topic of passive association,
see, in particular, Blessing (2010b), at p. 155.
1161. OECD TPG, paras. 1.164-1.166, Example 1.
1162. OECD TPG, para. 1.161. This is also the result under the US regulations, which
require compensation for purposeful group activities; see Treas. Regs. § 1.482-9(l)(1).
See Odintz et al. (2017), at sec. 2.2.4.2 on this issue.
1163. OECD TPG, para. 1.160. For a very thorough (and interesting) discussion of
procurement and allocation of profits, see Peng (2016), at pp. 385-392.
345
the third-party lender that put the subsidiary in the position to borrow at the
lower rate.1164 The incremental profits from this category of synergies shall
be allocated among group members “in proportion to their contribution
to the creation of the synergy”.1165 For instance, incremental profits due to
large-scale procurement strategies will typically be allocated pursuant to
each group entity’s purchase volume. In other cases, such as in the guar-
antee example, it will be more appropriate to price the specific synergistic
benefit separately. Either way, incremental operating profits from deliber-
ate, concerted group action synergies must be specifically allocated among
group entities.
The 2017 OECD TPG do not offer any justification as to why they distin-
guish between incremental operating profits that arise from incidental ben-
efits, on the one side, and from synergies from deliberate, concerted group
actions, on the other side.1166 The author imagines that it may be difficult to
reliably identify some of the more ambiguous synergy benefits (incidental)
and that it therefore would not be meaningful to require specific alloca-
tion of such benefits. However, in the example in which the subsidiary is
able to borrow at a lower rate solely due to its status as a group member,
the benefit is easily identifiable. The author therefore struggles to see that
there is a logical reason for why this benefit should not be attributed to the
subsidiary.
1164. OECD TPG, para. 1.167. On the transfer pricing of guarantees, see, in particular,
the informed discussion in Blessing (2010b), at p. 162, which includes comments on
General Electric Capital Canada Inc. v. The Queen (2010 DTC 1007, affirmed by the
Federal Court of Appeal [2011 DTC 5011]), where the disallowance by the Canadian
tax authorities of a tax deduction for a 1% guarantee fee taken by the Canadian subsidi-
ary of a US parent was rejected on the basis that the fee was found to be consistent with
what would have been paid to an arm’s length guarantor. On this case, see also Horst
(2011); Ward (2010); Roin (2011), at p. 216; and Koomen (2015b), at p. 240. For further
information on guarantees, see also, e.g. Pankiv (2017), at p. 58; Ledure et al. (2010);
Riedy et al. (2010); and Boidman (2011).
1165. OECDTPG, para. 1.162. See Peng (2016), at p. 380 and p. 385, where he un-
derlines the positive aspects of applying a profit split approach to allocate the synergy
profits.
1166. See also critical comments on this in Brauner (2016), at p. 105, as well as p. 109,
where it is stated that “[t]he problem is that synergies cannot really be contributed to
– they occur because of the affiliation of the parties involved. It is nonsense, therefore,
to seriously attempt to allocate them to particular parties”. On the distinction, see, in
particular, Peng (2016), at p. 382.
346
Chapter 11
Transfer Pricing of Intangibles in

the Post-BEPS Era under the OECD TPG
11.1. Introduction
In this chapter, the author will briefly analyse how the 2017 OECD Trans-
fer Pricing Guidelines (OECD TPG) envision the application of the cur-
rently accepted pricing methodology in the context of transfer pricing of
intangible property (IP). The relevant methods for allocating operating
profits from IP value chains are, in practice, the methods that were the sub-
ject of analysis in the previous chapters, namely (i) the comparable uncon-
trolled transaction (CUT) method; (ii) the transactional net margin method
(TNMM); and (iii) the profit split method (PSM). Historically, there is no
question that the OECD has preferred the CUT method. The recent BEPS
revision of the intangibles chapter of the OECD TPG reveals an altered at-
titude of the OECD. The CUT method is no longer in vogue, but the PSM
is. This PSM endorsement can be seen as a reaction to the shortcomings of
both the CUT method (i.e. a lack of reliable comparables for unique IP) and
the TNMM (i.e. that it allocates too little profits to source jurisdictions).1167
This new methodological preference from the OECD triggers an interest-
ing question: Does it entail that the one-sided allocation of normal market
return profits to source jurisdictions under the TNMM must be modified or
brought to a stop? The author will discuss this in sections 11.3. and 11.4.,
respectively, after making some contextualizing remarks in section 11.2.
11.2. A transfer pricing paradigm under pressure
Perhaps the most basic trait of the transfer pricing rules is that they are de-
signed to allocate residual profits among jurisdictions in a way that mirrors
where the intangible value is created. An implication of this is that residual
profits shall not be allocated to jurisdictions that do not partake in IP devel-
opment. This paradigm has drastic consequences for the international al-
location of taxing rights over the operating profits of multinationals. There
1167. With respect to the transactional net margin method (TNMM), in this direction,
see also Linares (2005), at p. 37. See also Vidal (2002), at p. 414.
347
Chapter 11 - Transfer Pricing of Intangibles in the Post-BEPS Era under
the OECD TPG
may be vast differences in the amount of taxable income allocated to coun-

tries that are entitled to residual profits and those that are not.
The author will illustrate this through an example. Imagine that a multi-
national, through its Norwegian parent, performs and funds research and
development (R&D) and holds ownership of all resulting IP. The multina-
tional is present in 40 other countries, where it carries out routine distri-
bution and marketing. Because the Norwegian entity alone provides the
unique value chain inputs, it may extract all foreign operating profits that
are left after TNMM-based normal market returns are allocated to the
local routine input entities. Seen from the perspective of the Norwegian
group entity, the transfer pricing rules function as “extraction tools” for the
residual profits.1168
This form of centralized allocation of operating profits is controversial. A

source jurisdiction in which a multinational carries out routine functions
may observe that the local group entity has many employees, owns valuable
tangible assets and carries out transactions that result in significant operat-
ing profits locally. Yet, very little of this profit is left for taxation at source.
In other words, there may be a perceived mismatch between the level of
local economic activity and local taxation. This appeared to be the case
in the well-publicized Amazon.com and Starbucks hearings in the United
Kingdom in 2012.1169 Such a perceived mismatch may trigger a sense of
inequity, particularly in jurisdictions that offer significant location-specific
advantages (LSAs) that multinationals benefit from, which may be further
exacerbated if there is an impression that the extracted residual profits are
not subject to much taxation in the jurisdiction where the foreign owner of
the unique IP employed in the value chain is resident.
This criticism is a protest against the core content of the current transfer
pricing paradigm. Critics seem to be under the impression that the inter-
national community would be better off with transfer pricing rules that al-
locate operating profits commensurately with the overall economic activity
carried out in source states, for instance, as indicated by the number of local
employees, tangible assets, revenues, etc. (so-called “formulary apportion-
ment”). The current rules are not designed for this purpose; rather, they
seek to align the allocation of operating profits among the different entities
within a multinational as third parties acting under similar circumstances
1168. See also sec. 2.4. for an overview of the centralized principal model.
1169. See, for instance, http://www.bbc.com/news/business-20559791 (accessed 1 Sept.
2015).
348
A transfer pricing paradigm under pressure
would have done. Specifically, the current rules distinguish between rou-
tine and non-routine inputs to IP value chains, where the former inputs are
only entitled to a normal market return, ensuring that residual profits are
allocated only to the jurisdictions where intangible value is created.
The formulary apportionment alternative professed by the critics would

spread the residual profits across all jurisdictions in which the multina-
tional does business. This would overcompensate most jurisdictions, in the
sense that they would be allocated more operating profits than the routine
value chain inputs provided by group entities resident in them are worth.
The jurisdictions where intangible value is created would be undercom-
pensated by not being allocated the entirety of the residual profits. In other
words, they would not be assigned operating profits that are commensurate
with the worth of the unique value chain inputs provided by the group enti-
ties resident there. Such a system would provide a form of equality, in the
sense that taxing rights over the worldwide profits of a multinational would
be divided equally among jurisdictions based on some predetermined al-
location key (employees, assets, etc.).
This alternative system would not, however, be rooted in economic reason-

ing, as it assumes that all value chain inputs, routine and non-routine, are
worth the same. It would be unable to emulate the profit allocation results
of transactions among third-party enterprises with genuinely conflicting
interests, the hallmark of which is that competition will drive the prices
for routine value chain contributions down to a level where the providers
only reap a normal market return. This alternative system would therefore
fail to provide parity in the tax treatment of unrelated and related parties.
Even though the current system both allows and invites the use of one-
sided pricing methods to remunerate related routine input providers, there
is the current perception that multinationals are applying the TNMM too
aggressively and that the returns allocated to related routine value chain
input providers in source states are unreasonably sparse. This criticism is
rooted in the perspective of source states. The OECD has sought to address
these concerns through the combination of the new rules on allocation of
profits from location savings, LSAs and synergies (analysed in chapter 10),
as well as the new guidance on the relative role of the transfer pricing
methods, which will be analysed in this chapter.
Before the author begins his analysis, however, he would like to point out
that the perspective of the jurisdictions in which IP value is created must
not be forgotten. These jurisdictions will likely not share the views of the
349
the OECD TPG
source states. The argument will be that since the unique IP employed in
the value chain – which, after all, is responsible for generating residual
profits – is created there, the taxing rights to these profits should also be
assigned there. The argument is supported by economic logic, which dic-
tates that non-routine contributions should be allocated residual profits. In
order to accomplish this allocation, it is necessary to apply the one-sided
methods (in practice, the TNMM) to allocate a normal market return only
to jurisdictions in which related routine value chain input contributors are
resident so that no residual profits “leak” into these jurisdictions.
Thus, there is tension between the interests of jurisdictions where the

unique value chain inputs are developed and the jurisdictions in which they
are exploited and their value materializes. The 2017 OECD TPG on IP
ownership are designed to protect the first group of jurisdictions through
the use of the “important functions doctrine” for assigning entitlement to
residual profits.1170 At the same time, the new guidance on location savings,
LSAs and synergies, as well as the new language on the use of transfer
pricing methods in the context of IP value chains, seeks to take into ac-
count the perspective of the source states. This is not an easy balancing
act for the OECD. The fundamental question is whether it is possible to
allocate something more than a normal market return for the routine value
chain inputs emanating from source states without diluting or fracturing
the core economic logic upon which the arm’s length transfer pricing sys-
tem is based.
11.3. Shall something more now be allocated to source

jurisdictions?
The TNMM is the king of the transfer pricing methods. It is likely the most
applied method globally. Nevertheless, the 2017 OECD TPG display a lack
of endorsement for the methodology in the context of IP value chains. This
is done indirectly in two ways.
First, a generally sceptical attitude towards the use of one-sided methods is

expressed as follows:1171 “[O]ne sided methods, including the resale price
method and the TNMM, are generally not reliable methods for directly
valuing intangibles.”1172 This point is reiterated in paragraph 6.58 of the
1171. See OECD TPG, paras. 6.129, 6.131, 6.138, 6.141, 6.146 and 6.209.
1172. OECD TPG, para. 6.141.
350
Shall something more now be allocated to source jurisdictions?
OECD TPG, which states that “the reliability of a one-sided transfer pric-
ing method will be substantially reduced if the party or parties perform-
ing significant portions of the important functions are treated as the tested
party or parties”.1173
Given that it is outside the scope of application of the TNMM to allocate

income to a tested party that contributes non-routine inputs to the intan-
gible value chain, these statements convey a readily apparent truth, but
nevertheless tone down the relevance of the one-sided methodology for IP
pricing purposes.
Second, the fact that the 2017 OECD TPG contain significant supplemental
language on the CUT method and the PSM and introduce entirely new and
thorough guidance on valuation techniques while adding nothing new for
the TNMM further signals a reduced position of the TNMM.1174 Also, the
new language on the PSM in the intangibles chapter of the OECD TPG
states that “it should not be assumed that all of the residual profit after
functional returns would necessarily be allocated to the licensor/transferor
in a profit split analysis related to a licensing agreement”.1175 This can be
read as a warning against applying the TNMM to remunerate a group en-
tity that only provides routine inputs to the controlled transaction.
In light of this generally negative attitude towards the TNMM, it can be

asked whether the new 2017 OECD TPG entail a requirement to allocate
something more to routine group entities in scenarios in which the TNMM
nevertheless is applied.
The clear point of departure is that no changes are made to the general
guidance contained in the OECD TPG on the TNMM.1176 However, it is
now stated:
In some circumstances such mechanisms can be utilised to indirectly value
intangibles by determining values for some functions using those methods
and deriving a residual value for intangibles. However, the principles of para-
1173. The wording of the 2013 Revised discussion draft on transfer pricing aspects of
intangibles (2013 RDD) is slightly different; see 2013 RDD, para. 81, which states that
“the reliability of a one-sided transfer pricing method will be substantially reduced if
the party performing the important functions is treated as the tested party”. (Emphasis
added) The alteration is clarifying, but has no material impact on the message deliv-
ered.
1174. See OECD TPG, paras. 6.146-6.180.
1175. See OECD TPG, para. 6.152, last sentence.
1176. OECD TPG, paras. 2.56-2.107.
351
the OECD TPG
graph 6.133 are important when following such approaches and care should
be exercised to ensure that all functions, risks, assets and other factors con-
tributing to the generation of income are properly identified and evaluated.1177
(Emphasis added)
The question is whether the quoted wording entails that the TNMM can
no longer be applied as described in chapter II of the OECD TPG, or more
specifically, whether an additional return must be allocated to source states
on top of the normal TNMM market return for local routine value chain
inputs. If so, the emphasized portion of the statement should be regarded
as nothing less than an earthquake in international transfer pricing juris-
prudence.
The quoted statement is ambiguous. First, the distinction between directly

and indirectly valuing intangibles is not entirely clear. Both the TNMM
and the PSM fundamentally rely on first allocating a normal return to rou-
tine functions before allocating residual profits. In principle, the difference
between the profit-based methods is, as mentioned, that under the TNMM,
the residual profits are, in their entirety, allocated to one of the parties
to the controlled transaction, while the PSM divides the residual profits
among them. Thus, both methods “indirectly value intangibles”. In this
regard, the TNMM is no different from the PSM.
Second, the wording seems to limit the application of the TNMM to “some
circumstances”, without clarifying what these are. Given that the TNMM
already at the outset is a narrow method, confined to being applied only
where the tested party is a pure routine input provider, it is not imme-
diately intuitive whether the wording seeks to further narrow its field of
application. The author’s impression is that it does indeed intend to do so.
This interpretation is supported by the following sentence, which states
that if the TNMM is applied, all functions, risks, assets and “other fac-
tors” contributing to the generation of profits must be properly identified
and evaluated. This seems to indicate that a normal market return will no
longer be sufficient, as routine group entities shall also be rewarded for
“other factors”.
Third, the delimitation is not technically clear. Paragraph 6.141 refers to

paragraph 6.133, which, in broad language, lists a variety of factors to take
into consideration (and presumably remunerate), such as “risks borne, spe-
1177. OECD TPG, para. 6.141.
352
Shall something more now be allocated to source jurisdictions?
cific market characteristics, location, business strategies, and [multination-

al enterprise] group synergies among others”.
The position taken by the 2017 OECD TPG on location savings and LSAs
is to allow the use of local comparables to benchmark the profits that are
attributable to local routine function providers.1178 This will normally result
in the extraction of cost savings and location rents from source jurisdic-
tions.1179 Thus, even though paragraphs 6.133 and 6.141 seem to indicate
that the TNMM shall allocate an “additional” return to source states, the
acceptance of local comparables effectively puts an end to that ambition in
most practical scenarios.
What if there are no local comparables available? The bargaining power

of the controlled parties will then be key for allocating income.1180 If the
local entity is a routine function provider while the foreign party to the
transaction contributes unique IP, it is likely that cost savings and loca-
tion rents will be extracted from the source. Thus, the result is similar to
a standard application of the TNMM.1181 The new guidance on synergies
will normally entail a separate pricing, which will not affect the general
application of the TNMM.1182
Further, the reference in paragraph 6.133 to “risks borne”, “business strate-

gies” and “others” introduces a general uncertainty as to which specific
factors shall now attract remuneration pursuant to the 2017 OECD TPG, as
well as under which methodology income shall be allocated to them and
whether the remuneration shall be some form of a normal market return or
a portion of the residual profits from unique IP.
With respect to the factors of “others” and “business strategies”, an answer

to this will necessarily first require a clarification of what is referred to by
1178. See OECD TPG, para. 1.142 for location savings (see also the discussion in sec.
10.5.2.); and OECD TPG, para. 1.145 for location-specific advantages (LSAs) (see also
the discussion in sec. 10.7.2.).
1179. This was also the result under the 1995 OECD TPG; see the discussion in Fran
cescucci (2004a), at p. 72. See also Roberge (2013), at p. 222.
1180. On the concept of bargaining power in transfer pricing, see Blessing (2010b). See
also infra n. 1125.
1181. Alternatively, local comparables may be unavailable if the local group entity
owns unique intangibles. The TNMM will then be inapplicable (this method only al-
locates a normal market return to controlled routine input providers), necessitating an
application of the profit split method (PSM). The cost savings and location rents will be
split among the controlled parties, along with the residual profits.
1182. See the discussion on the allocation of profits from synergies in sec. 10.7.3.
353
the OECD TPG
these terms. “Risks borne”, in the context of unique IP, presumably refers
to financial risk incurred through R&D funding, but this is far from clear.
Financial risk should be rewarded a financial return commensurate with the
risks taken, as determined on a separate basis. Such funding should not at-
tract a general profit split.1183 “Business strategies” should not, in the author’s
view, in and of themselves be subject to compensation. Of course, strategies
will result in transactions, functions, assets and risks that are compensable,
but that is a different matter entirely, not to mention that it will likely be
impossible to reliably price “strategies” on a separate basis in practice.
In conclusion, it is the author’s view that paragraphs 6.133 and 6.141 will
not limit the application of the TNMM. It will still be possible to apply
the method to allocate a normal return only to local routine functions and
thereby extract residual profits from the source.
The author will also add that he sees no sound basis for restricting applica-
tion of the TNMM. The methodology should, in the author’s view, apply
with full strength in the narrow circumstances for which it was designed,
i.e. when the tested party renders only routine value chain contributions. If
access to the TNMM is cut off in such ordinary scenarios, multinationals
would likely be better off by outsourcing routine functions to third parties
willing to perform them for a normal market return.1184 In other words, the
allocation of an “additional” return to group entities that render routine
value chain inputs will result in less favourable taxation of related parties
than for unrelated parties. Transfer pricing rules must seek to attain parity
in the taxation of controlled and uncontrolled transactions.
Another thing is that tax authorities should be vigilant with respect to

whether multinationals indeed do apply the TNMM in the manner that
it was intended. With this, the author refers to the possibility that local IP
should be recognized to a larger extent than what currently is the case. It is
likely that a range of local “routine input providers” have developed good-
will, know-how or outright local marketing IP. If that is the case, the local
group entity cannot be treated as the tested party under the TNMM. The
method will be inapplicable in these scenarios. That may leave no other
alternative than to apply the PSM. That will attract not only residual profits
1183. See the analysis of research and development funding remuneration under the
OECD TPG in sec. 22.4.
1184. There may, however, be costs connected to such outsourcing; see the discussion
on the choice between third-party outsourcing and foreign direct investment in sec. 2.3.
354
The relative roles of the CUT method, TNMM and PSM for the transfer
pricing of intangibles
from the local marketing IP to the source jurisdiction, but also a cut of the
location savings and rent.1185
11.4. The relative roles of the CUT method, TNMM

and PSM for the transfer pricing of intangibles
The 2017 OECD TPG state that “the transfer pricing methods most likely
to prove useful in matters involving transfers of one or more intangibles are
the CUP method and the transactional profit split method”.1186
The danger with the CUT method is that it does not directly take into ac-
count the value chain contributions from the controlled parties. Instead,
it refers to how third parties have priced a purportedly comparable IP
transfer. If the transferred intangibles are not fully comparable, the CUT
method may result in profit allocations that conflict with economic logic,
e.g. that residual profits are allocated to an IP holding company without
substance or to a group entity that performs only routine functions. In light
of the new OECD emphasis on similar profit potential as a comparability
requirement for CUTs involving unique IP,1187 as well as the scepticism to-
wards comparability adjustments to such transactions,1188 the CUT method
should, in practice, rarely be relevant for allocating operating profits from
IP value chains.1189
That leaves the TNMM and the PSM. In contrast to the CUT method,
where the main concern is the lack of reliability, the meagre OECD enthu-
siasm for the TNMM is due to the fact that it allocates a bare minimum of
income to source states. In contrast, the PSM will provide more balanced
results, allocating a portion of the residual profits from unique IP, as well
1185. The incremental profits from location savings and rent will be allocated along
with the residual profits pursuant to the bargaining powers of the controlled parties, as
determined by the relative values of their unique value chain contributions.
1186. OECD TPG, para. 6.145.
1188. OECD TPG, para. 6.129.
1189. The author finds reason to add that the CUT method is also subject to reduced
relevance on a more general level. The TNMM and the PSM will normally be easier to
apply in the context of unique intangibles and will provide enough flexibility to render
the CUT method largely redundant, both from the perspective of multinationals and tax
authorities. The CUT method was more relevant in earlier transfer pricing practices,
when there were fewer and less developed specified pricing methods available.
355
the OECD TPG
as location savings and rent to source states.1190 Thus, an application of the

PSM will likely be favoured by source states, but less so by multinationals
and jurisdictions in which intangible value is created.
There is no question that the 2017 OECD TPG prefer the PSM for the
transfer pricing of intangibles.1191 As analysed in chapter 9, the 2017 OECD
TPG have extended the scope of application for the PSM so that it now
partly overlaps the scope of application of the TNMM. Further, the core
profit allocation methodology underlying the PSM has gained increased
significance in the OECD TPG through the new IP ownership provisions.
The OECD guidance on IP ownership states that “in identifying arm’s
length prices for transactions among associated enterprises, the contribu-
tions of members of the group related to the creation of intangible value
should be considered and appropriately rewarded”.1192
The quoted statement signals a holistic approach to the allocation of oper-

ating profits from IP value chains, encompassing both the rules on transfer
pricing and IP ownership. The difference between profit allocation assess-
ments carried out under the OECD IP ownership rules and the PSM lies
mainly the context in which the assessments are performed. The assess-
ment of IP ownership pertains to the phase prior to the completion of the IP
development. This will typically be before the intangible starts to generate
cash flows. After this time, the IP may be employed in a value chain, along
with other non-routine and routine contributions, and the PSM may be ap-
plied to allocate the residual profits.
The important question is whether this comprehensive OECD preference

towards the PSM has resulted in any concrete requirements to use the
method in scenarios in which this previously was not the case. Given that
the only alternative to the PSM in practice will be the TNMM, this ques-
tion can largely be rephrased as an issue of whether the new guidance lim-
its the application of the TNMM. The author’s conclusion in section 11.3.
was that this is likely not the case.
1190. OECD TPG, para. 6.2. It should be noted that OECD TPG, para. 6.132 also em-
phasizes the importance of considering the economic consequences of the transaction
when selecting a transfer pricing method.
1191. See Schön et al. (2011), at pp. 267-290, for a discussion of the future role of the
PSM in transfer pricing (based on the 2010 OECD TPG) and a comparison with formu-
lary apportionment.
356
The relative roles of the CUT method, TNMM and PSM for the transfer
pricing of intangibles
Thus, while the 2017 OECD TPG clearly encourage the application of the
PSM, the author finds that there is little basis for claiming that there are
any new requirements to do so. As the scope of application of the TNMM
seems largely untouched, the principal point of departure remains that the
TNMM can be applied when one party to the controlled transaction only
contributes routine inputs. Going forward, the TNMM can therefore, as
before, be used to extract residual profits from source jurisdictions by way
of allocating only a modest normal market return to the routine group enti-
ties resident there. In this sense, the 2017 OECD TPG have brought little
in the way of change.
357
Chapter 12
Allocation of Residual Profits for Unique and

Valuable IP Based on Unspecified Pricing Methods
12.1. Introductory comments
Unspecified pricing methods are normally only used when the specified
methods (i.e. the comparable uncontrolled transaction (CUT) method,
resale price method, cost-plus method, transactional net margin method
(TNMM) and profit split method (PSM)) are perceived as insufficient to al-
locate income from the controlled intangible property (IP) transaction.1193
Historically, unspecified methods have served as important drivers for the
development of new specified methods. The TNMM, for instance, start-
ed out as an unspecified method under US law. The US Internal Revenue
Service (IRS) developed the so-called “contract manufacturer” theory,
which it used as its primary pricing argument in the roundtrip cases liti-
gated in the 1970s-1980s, then as a “fourth method” under the 1968 US
regulations.1194 According to this theory, a foreign group entity that has
only contributed routine inputs to the value chain should not be allocated
any residual profits. This theorem later resulted in the basic arm’s length
return method (BALRM) of the 1988 White Paper and the comparable
profits method (CPM) of the final 1994 regulations, which then led to the
international adaptation of the TNMM in the 1995 OECD Transfer Pricing
Guidelines (OECD TPG). Another example, and a modern-day parallel
to the contract manufacturer theory, is the IRS’s assertion for the pricing
of US-developed IP under the pre-2009 cost-sharing regulations, pursuant
to which foreign group participants in a cost-sharing arrangement (CSA)
that only contribute IP development funding should not be allocated re-
sidual profits.1195 The theory has now been codified as the specified income
method of the current cost-sharing regulations.1196
1193. On the use of unspecified methods, see, e.g., Wittendorff (2010a), at pp. 658-
662.
1194. The author refers to the discussions in sec. 5.2.5.
1195. The 2007 Coordinated Issue Paper (CIP) outlines the unspecified income meth-
od; see the analysis in sec. 14.2.3. See also the discussion of Veritas Software Corpora-
tion & Subsidiaries v. CIR (133 T.C. No. 14 [U.S.Tax Ct. 2009], IRS nonacquiescence
in AOD-2010-05) in sec. 14.2.4.
1196. Treas. Regs. § 1.482-7(g)(4).
359
Chapter 12 - Allocation of Residual Profits for Unique and Valuable IP
Based on Unspecified Pricing Methods
This illustrates that there is an interaction between the current specified

transfer pricing methods, on the one side, and the development of unspeci-
fied methods, on the other. First, the more developed the specified transfer
pricing methods are, the less need there will be to resort to unspecified
methods. The point of the above IRS assertions was to allocate residual
profits in an arm’s length manner in the context of new value chain struc-
tures that gave rise to challenges that could not properly be addressed
through the then-available specified methods. Thus, unspecified methods
may serve as “contingency valves” that can be used to allocate residual
profits from controlled structures that are not susceptible to pricing under
the current specified methods.1197
Second, there is an unclear line between dynamic interpretation of the cur-

rent specified methods and the use of unspecified methods. It will depend
on local jurisprudence and legal traditions and to what degree courts are
willing to interpret and apply specified methods in a manner that they ar-
guably were not originally designed for. In the case of the United States,
the author’s impression is that the courts tend to follow the wording of
the regulations closely and express an unwillingness to resort to dynamic
interpretations based on the fundamental principles embodied in the rel-
evant specified pricing method. An illustrative example of this is the US
Tax Court’s rejection of the unspecified income method assertion of the
IRS in Veritas. In the author’s view, the assertion could largely be seen as
a tailored application of the existing specified CPM (due to the material
similarities between the CPM and the income method as professed by the
IRS). The IRS did not, however, argue along this line, so the issue was not
addressed by the Court.1198
Given the sophistication of the specified methods available for the pricing
of controlled IP transfers in the current US regulations and the 2017 OECD
TPG, unspecified methods will generally rarely be relevant. To the extent
that they are relevant, they will likely mainly serve as tools for tax authori-
ties (e.g. the contract manufacturer and income method theories discussed
above).1199 The one example of a taxpayer application of an unspecified
1197. It should also be acknowledged that the five specified pricing methods are flex-
ible. A single specified pricing method may accommodate different pricing approaches.
For example, an array of allocation parameters are allowed under the transactional net
margin method (TNMM) and the OECD profit split method (PSM) (see the discussions
in secs. 8.4. and 9.3.3., respectively).
1198. See sec. 14.2.4. for further discussion.
1199. This is because (i) the specified methods are so wide that they can normally
accommodate taxpayer pricing assertions; and (ii) there may be increased penalty
risks for taxpayers connected to the use of unspecified methods under US law; see IRC
360
Unspecified methods under the US regulations
method that the author is familiar with pertains to US multinationals that,

in their pricing of buy-ins under the pre-2009 cost-sharing regulations,
used capitalized and depreciated research and development (R&D) costs
to allocate residual profits under the PSM.1200
In sections 12.2. and 12.3., respectively, the author will discuss the extent
to which tax authorities and taxpayers may apply unspecified transfer pric-
ing methods under the US regulations and the 2017 OECD TPG to allocate
residual profits in controlled IP transfers.
12.2. Unspecified methods under the US regulations
The US regulations allow the application of an unspecified pricing method

to allocate profits in controlled IP transfers.1201 The method must then “take
into account the general principle that uncontrolled taxpayers evaluate the
terms of a transaction by considering the realistic alternatives to that trans-
action, and only enter into a particular transaction if none of the alterna-
tives is preferable to it”.1202 Further, “an unspecified method should provide
information on the prices or profits that the controlled taxpayer could have
realized by choosing a realistic alternative to the controlled transaction”.1203
Thus, the realistic alternatives of the controlled parties is the decisive pri
cing principle that all unspecified methods applied to allocate residual
profits from controlled IP transfers under US law must adhere to.1204 The
regulations contain an example that illustrates this restriction.1205 It per-
tains to a US parent that makes and sells a proprietary adhesive product in
the United States. The parent licenses make-sell rights to the manufactur-
ing IP on which the product is based for the European market to a local
sec. 6662(e), 26 U.S.C.A. § 6662, supplemented by the 2011 regulations in 26 C.F.R.

§ 1.6662-6. For an instructive and rather detailed analysis of transfer pricing penalties
under US law, see Levey et al. (2007), at pp. 209-224, with further references.
1200. See sec. 14.2.3.
1201. Treas. Regs. § 1.482-4(a). If used, the methodology must comply with the gen-
eral provisions of the regulations, including the best-method rule, the comparability
requirements and the provisions for the arm’s length range; see Treas. Regs. § 1.482-1.
The controlled allocation of IP income will be subject to periodic adjustments; see
Treas. Regs. § 1.482-4(f)(2). The author analyses the US periodic adjustment provision
in ch. 16.
1202. Treas. Regs. § 1.482-4(d)(1).
1203. Id.
1204. In this direction, see also Odintz et al. (2017), at sec. 2.2.5.1.
1205. Treas. Regs. § 1.482-4(d)(2).
361
subsidiary, which in turn sells the product to related and unrelated parties
for the market price of 550 per tonne. The controlled royalty rate is set to
100 per tonne. In determining whether the rate is at arm’s length, the real-
istic alternative for the parent to producing and selling the adhesive product
itself for the European market is considered. Reasonably reliable estimates
indicate that if the parent directly supplies the product to the European
market, a selling price of 300 per tonne would cover its costs and provide it
with a reasonable profit for its functions, risks and investment of capital. In
other words, by licensing the manufacturing IP to the subsidiary, the parent
forgoes a profit of 250 per tonne compared to the income that would accrue
to it with its best realistic alternative. The example therefore concludes that
the controlled royalty of 100 is not at arm’s length.
This illustrates that the realistic alternatives requirement may serve as an

effective barrier against controlled non-arm’s length allocations of oper-
ating profits. The focal point of the realistic alternatives assessment will
normally be the allocation of profits to the transferer. The reason for this
is simple: the transferer is the owner of the unique IP and should therefore,
as the point of departure, be entitled to all residual profits from its own in-
tangible. The realistic alternatives principle, as well as economic logic (and
even common sense), dictate that there must be convincing reasons as to
why the transferer should relinquish any of his residual profits to the trans-
feree.1206 Such reasons will not exist according to the realistic alternatives
principle if the controlled allocation of operating profits places the trans-
ferer worse off than he would have been with his best realistic alternative.
The realistic alternatives principle is of fundamental importance in trans-

fer pricing.1207 It underpins all transfer pricing methods and is the basic ma-
terial link between the specified and unspecified methods. The CPM and
the income method, for instance, may be seen as concrete manifestations
1206. This is the same basic reasoning that formed the foundation for the buy-in pri
cing approach of the current US regulations. See Brauner (2016), at p. 114, where he
describes the approach of the 2005 proposed cost-sharing regulations as follows: “The
thought was that the party with existing intangibles and research capability would not
permit a third party to invest in future development of crown jewel intangibles and
receive profits beyond normal financial profits.” See also infra n. 1986.
1207. The significance of the realistic alternatives principle is pronounced in the con-
text of valuation. The author refers to the discussions of the 2017 OECD valuation
guidance in sec. 13.5. and of the income method of the US cost-sharing regulations (see
Treas. Regs. § 1.482-7(g)(4)) in sec. 14.2.8.3. See also Bullen (2010), at pp. 544-562,
on the role of the realistic alternatives principle in the context of non-recognition under
the 1995/2010 OECD TPG, as well as for reflections pertaining to use of the principle
in the context of transfer pricing at pp. 736-737.
362
Unspecified methods under the OECD TPG
of the principle in the sense that the tested party is only allocated a normal
return. This result aligns with the best realistic alternatives of a transferer
(non-tested party) that transfers unique IP to a transferee (tested party) that
only contributes routine inputs. The principle is also significant in deter-
mining whether there is sufficient comparability between the controlled
transaction and purported CUTs, in particular, with respect to the issue of
whether the IP transferred in the controlled and uncontrolled transactions
have similar profit potential.1208
Further, the US regulations contain a rule that coordinates the CSA provi-
sions with the general transfer pricing provisions applicable to intra-group
IP transfers.1209 The point of the provision is to ensure that the principles,
methods and comparability requirements set out in the CSA regulations
are applied analogically to the pricing of IP transfers that do not qualify as
CSAs (i.e. controlled IP development efforts where the participating enti-
ties share costs and risks under an arrangement that does not qualify as
a CSA). Nonetheless, as the United States’ unspecified-method rule and
CSA income method (the specified go-to method for buy-in pricing in the
context of CSAs) are both based on the “realistic alternatives” principle,
the primary consequence of this coordination rule will likely be to merely
provide guidance on how to apply the highly-developed CSA pricing meth-
ods analogically, not to dictate pricing results that diverge from those that
follow from the CSA pricing methods.
12.3. Unspecified methods under the OECD TPG
As indicated in section 12.1., the development and use of unspecified meth-

ods is often linked to domestic law, e.g. as reactions to perceived insuf-
ficiencies of the specified pricing methods available or as a way to ad-
dress transfer pricing problems that may be rather specific to a particular
jurisdiction. Methodology akin to the unspecified IRS-professed income
method in Veritas, for example, may not be relevant for jurisdictions where
little R&D is carried out or where the use of CSAs is uncommon.
1208. For comments on the similar profit potential comparability criterion under the
US regulations, see sec. 7.2.3. on the US CUT method, sec. 14.2.3. on the valuation
approach of the 2007 CIP (LMSB-04-0907-62) and sec. 14.2.4. on Veritas Software
Corporation & Subsidiaries v. CIR (133 T.C. No. 14 [U.S.Tax Ct. 2009], IRS nonac-
quiescence in AOD-2010-05). With regard to this criterion under the OECD TPG, see
secs. 7.3. and 11.4. on the OECD CUT method.
1209. Treas. Regs. § 1.482-4(g).
363
International tax law (here, articles 9 and 7 of the OECD Model Tax Con-
vention) will be decisive for the question of whether the controlled alloca-
tion of income pursuant to an unspecified method can be upheld with effect
for the distribution of operating profits among the contracting jurisdictions.
The 2017 OECD TPG allow the use of unspecified pricing methods to
allocate income in controlled IP transactions.1210 The use of such meth-
ods should be supported by an explanation as to why the specified OECD
pricing methods were regarded as less appropriate and the reason why the
selected unspecified method was regarded as providing a better solution.1211
The application of an unspecified pricing method must comply with the

general OECD pricing requirements. The functional analysis must take
into account “all factors that contribute to value creation”.1212 If the pricing
is not based on a CUT, the applied methodology must take into account
factors such as the functions, assets and risks of the controlled parties,
as well as the options realistically available to each of the parties to the
transaction.1213
With respect to IP valuation techniques, the 2017 OECD TPG state that
“depending on the facts and circumstances, valuation techniques may be
used … as a part of one of the five OECD transfer pricing methods de-
scribed in Chapter II, or as a tool that can be usefully applied in identifying
an arm’s length price”.1214
It is not entirely clear whether the use of valuation as a stand-alone tool

should be regarded as a specified or unspecified method under the OECD
TPG. A discounted cash-flow analysis may not necessarily constitute a
1210. See OECD TPG, para. 2.9 and para. 6.136, last sentence.
1211. The wording of OECD TPG, para. 2.9 on unspecified methods (the current
wording is the 2010 consensus text) was to be revised as part of the 2015 BEPS pack-
age to take account of the new guidance on valuation techniques; see 2014D, at p. 24.
It was also to be considered whether the current wording should be modified to reflect
other guidance on transfer pricing methods and special measures that were adopted in
connection with the 2015 work on BEPS. For some reason (likely pressure to complete
the BEPS deliverables on time), para. 2.9 was not revised as part of the 2015 BEPS
package.
1212. OECD TPG, para. 6.133.
1213. OECD TPG, para. 6.139. Other relevant factors include the competitive advan-
tages conferred by the intangibles, especially the relative profitability of products and
services related to the IP, and the expected future economic benefits from the transac-
tion. On applying the options-realistically-available principle in the absence of compa-
rables, see Parekh (2015), at p. 307.
1214. OECD TPG, para. 6.153.
364
Unspecified methods under the OECD TPG
stand-alone transfer pricing method, but could rather be characterized as

an overlay for the specified pricing methods. Regardless, the methodology
must comply with the OECD guidance on pricing methods in chapters I-III
of the OECD TPG, the realistic-alternatives pricing principle1215 and the
guidance on controlled risk allocations.1216
Further, the OECD TPG allow the use of unspecified cost-based pricing
approaches for intra-group IP transfers in limited circumstances.1217 Such
methodology will likely only be reasonable to apply when the aim is to
allocate operating profits from routine IP that does not generate residual
profits. This may include IP developed for internal use, such as in-house
software systems. The 2017 OECD TPG warn, however, that cost-based
pricing approaches may raise serious comparability issues if applied for
pricing IP that relates to products that are sold in the marketplace and that
they are deemed inappropriate for pricing IP when there is a first-mover
advantage,1218 where the IP carries legal protections or exclusivity charac-
teristics1219 or for partially developed IP.1220
Lastly, the 2017 OECD TPG reject the use of rules of thumb to allocate
residual profits in intra-group IP transfers.1221 Rules of thumb have likely
been applied to some extent in practice. For instance, licensing rates may
have been regarded as “ok” if the licenser was left with a royalty of one
third of the licensee’s profits. Such practices, however, have no legal basis
in the OECD TPG.

1216. OECD TPG, paras. 1.56-1.106. See also the comments on the 2017 OECD TPG
on risk in sec. 6.6.5.5.
1217. OECD TPG, para. 6.143.
1218. A future product that is meant to be similar to an already available product may
not be as valuable as the latter. The current product may have additional value due to
being the first product to reach the marketplace. Estimated future replacement costs
(with respect to the IP connected to the future product) will therefore not be a good
proxy for the value of IP transferred currently (the IP connected to the available prod-
uct).
1219. In this case, the IP will exclude potential competitors from the marketplace,
likely resulting in residual profits. Such additional value will not be reflected in replace-
ment costs.
1220. Research and development costs will not reflect the income potential connected
to such IP.
1221. OECD TPG, para. 6.144.
365
Chapter 13
Allocation of Residual Profits to Unique

and Valuable IP through Valuation
13.1. Introduction
In this chapter, the author will analyse how valuation techniques may
be applied to estimate an arm’s length transfer price for a unique and
valuable intangible under the 2017 OECD Transfer Pricing Guidelines
(OECD TPG).1222 The focus will be on intangible property (IP), typi-
cally a patent or a trademark, that is used in an existing value chain for
a specific successful product or service (e.g. for sales of product X in ju-
risdiction Y). The valuation of such intangibles is a significant problem
in practice and is also the central point of attention in the new OECD
guidance.
IP value can be ascertained through a two-step process. Step (1) entails

estimating the future operating profits from sales of the relevant product
or services (in the relevant jurisdictions) and then discounting the future
profits to present value. This necessitates estimates of the size and tim-
ing of sales, costs of goods sold and operating expenses allocable to the
product for each year of its life, as well as the determination of a suitable
risk-adjusted discount rate. Step (2) entails that the present value must
then be allocated, pursuant to the transfer pricing methods, among the
different value chain inputs (which include the IP). This is where the
real transfer pricing assessment comes into play. Thus, step (1) finds the
total operating profits, and step (2) allocates them among the controlled
parties.
1222. On transfer pricing valuation of intangible property (IP), see, in particular,

Brauner (2008). For a discussion on the now historical 2010 OECD TPG on the valua-
tion of intangibles, see, e.g. Wittendorff (2010a), at pp. 673-674; Wittendorff (2010b);
and Wittendorff (2010c). See also Levey et al. (2008a); Levey et al. (2008b); Levey et
al. (2008c); Verlinden et al. (2010); Hughes et al. (2010); Oestreicher (2011); and Laine
(2012). See Boos (2003), at chs. 3 and 4, for a discussion on the valuation of intangibles
from an economics perspective. See also Anson et al. (2005); and Lagarden (2014), at
p. 338, on IP valuation. See also the comprehensive (but rather general) European Com-
mission Study on the Application of Economic Valuation Techniques for Determining
Transfer Prices of Cross Border Transactions between Members of Multinational En-
terprise Groups in the EU (2016) and comments on the study in Treidler et al. (2017).
367
Chapter 13 - Allocation of Residual Profits to Unique and Valuable IP
through Valuation
The same two-step approach can, in principle, be used to value IP that is

only partially developed. There will, however, be more uncertainty con-
nected to the valuation parameters, as it will remain to be seen whether
the product or services that the IP relates to actually will become suc-
cessful.
This chapter will not address the valuation of IP that is not specifically con-
nected to a value chain. Such IP may, for instance, be technical know-how
of a general character or high-level marketing IP (e.g. a firm logo or name).
While such IP certainly may contribute to the profits of an enterprise, it
may be difficult to establish a concrete cause-and-effect relationship with
individual products and services, and thus to directly estimate the profits
it has contributed to on a value chain level. The most fruitful approach in
such cases will likely be to simply value the group entity which owns the
IP on a going-concern basis (i.e. an ordinary business valuation) and then
allocate an appropriate portion of the total asset value of the enterprise to
the IP, akin to a purchase price allocation analysis. It may be challenging
to separate the value of the high-level IP from residual intangible value
(goodwill, going concern value, etc.). Such ordinary business valuations
will, in principle, follow the same overall methodology discussed in this
chapter, albeit on a more aggregated level (which includes profits from all
value chains that the relevant entity contributes to).
Further, in some contexts, several items of IP may be transferred together

as a package, for instance, when know-how, patents and marketing IP
connected to a product are transferred intra-group in the same agree-
ment. In these cases, it may be argued based on the current OECD TPG
that the package transfer should not be valued as simply the sum of the
stand-alone IP values, but rather deemed a going-concern business trans-
fer, given that the latter approach would better reflect an arm’s length
value for the package. As such transfers typically are carried out through
cost-sharing arrangements (CSAs), the author refers to the discussions in
chapter 14.1223
The author discusses the valuation techniques endorsed by the 2017 OECD
TPG in section 13.2., valuation parameters in discounted cash flow (DCF)-
based valuation in section 13.3., allocation of the valuation amount among
the controlled value chain contributions in section 13.4. and the options
realistically available as a restriction on the possible allocation outcomes
1223. See, in particular, sec. 14.2.3., with further references.
368
The valuation techniques accepted under the OECD TPG
in section 13.5. Lastly, in section 13.6., the author asks whether there is a
legal basis for applying a discount to the transfer price.
13.2. The valuation techniques accepted under

the OECD TPG
There are, in general, three categories of valuation approaches.1224 The first
category is intrinsic valuation, where the value of an asset normally is de-
termined using DCF analysis as the present value of the expected future
cash flows that the asset will generate over its economic life.1225 DCF valu-
ation has not previously been embraced by the OECD TPG as a method
for valuing intangibles. Under the 2010 OECD TPG, the expected benefits
from an intangible were relevant mainly as a comparability factor for deter-
mining whether a purported comparable uncontrolled transaction (CUT)
could be used to benchmark the controlled royalty.1226
The second category of valuation techniques is relative valuation, where

the value of an asset is estimated based on how the market prices similar
assets.1227 The traditional transfer pricing methods (the comparable uncon-
trolled price (CUT), resale price and cost-plus methods), as well as the
transactional net margin method (TNMM) and, to some extent, the re-
sidual profit split method1228 are fundamentally relative pricing methods
due to their reliance on comparables (but these methods have a year-to-year
focus, as opposed to valuations that encompass all income years in which
the intangible generates value).
1224. The following discussion draws, to a large extent, on the classification of subject
matter and terminology applied in Damodaran (2006) and Damodaran (2012).
1225. A discounted cash flow (DCF) valuation requires the following inputs: (i) an
estimate of the size and timing of cash flows generated by the intangible over its life;
and (ii) an estimate of the discount rate to apply to the cash flows to convert them into
present values. These inputs may be challenging to estimate and may be influenced by
biases of the person carrying out the valuation.
1226. 2010 OECD TPG, para. 6.20. The 2010 OECD TPG also mentioned that DCF
analysis could be applied as part of a transfer pricing analysis under the profit split
method (PSM) in para. 2.123.
1227. Relative valuation approaches are generally most appropriate when a large num-
ber of assets comparable to the one being valued exist, the comparable assets are fre-
quently priced in a market through a large number of transactions and some common
variable exists that may be used to standardize the price (e.g. market price to book
value).
1228. The normal return allocated under the (first step of the) PSM will be determined
with reference to comparables. See OECD TPG, para. 2.121; and the discussion in ch. 9
of this book.
369
through Valuation
The third category is that of option-based valuation techniques, in which

the value of an asset is estimated based on its relationship with an un-
derlying asset.1229 No guidance is provided in the OECD TPG on such
techniques. In the terminology of the OECD TPG, however, “valuation
techniques” also encompass techniques other than the DCF method.1230 It
should therefore be clear that the OECD TPG, in principle, do not reject the
use of option-based valuation.1231
Thus, it may be observed that the OECD TPG, in principle, allow all three
classes of valuation techniques. The technique employed must, however,
be used consistently with the arm’s length principle,1232 in particular, the
realistically-available-options premise. Valuations of intangibles based on
(development) costs will generally not be in accordance with the arm’s
length principle.1233 The main focus of the new guidance is on the DCF
valuation technique.1234 The comments in this chapter will be limited to
this methodology.
1229. Option pricing may be relevant for investments that would be difficult to price
under DCF or relative valuation models, such as research and development (R&D)
investments or transfers of unproven intangibles at an early stage of development. A
range of different option-pricing models is available, the Black & Scholes option-pri
cing model being perhaps the best-known model; see Bodie et al. (2005), at pp. 758-
766, with further references. Normally, in valuation, risk decreases the value of an
asset. For options, however, risk or variation increases the value of the option, because
the downside of an option is limited to its cost price. The inputs required for option
valuation are difficult to estimate and may be uncertain.
1230. OECD TPG, para. 6.153.
1231. “Real options” are contingent business decisions or the ability of an enterprise to
respond to uncertainty (e.g. the alternative of an enterprise of entering into an unprofit-
able project if that project makes it possible to develop a new and profitable project,
the option to postpone an investment or the option to abandon a project that proves un-
profitable). The value of real options may be priced using financial theory. During the
public consultation hearing on the 2014 Public discussion draft BEPS Action 10: Dis-
cussion Draft on the use of profit splits in the context of global value chains (PSMDD)
held at the OECD in Paris on 19-20 March 2015, the lack of guidance on real-option
pricing in the context of unique intangibles was criticized by business commentators.
The core of the criticism was that a typical DCF valuation will value an R&D project
based on the presumption that the project has a fixed life without taking into account
the value of the real options that are available to the group entity performing the R&D.
1232. OECD TPG, para. 6.154.
1233. OECD TPG, paras. 6.142 and 6.156. See Brauner (2008), at p. 109, for thorough
and critical reflections on the cost approach to IP valuation. See also Lagarden (2014),
at p. 339, where costs as a basis for valuation are rejected.
1234. OECD TPG, para. 6.153.
370
The valuation parameters in DCF-based valuation
13.3. The valuation parameters in DCF-based valuation
The purpose of a DCF analysis in the context of this chapter is to estimate

the present value of future operating profits from an intangible value chain.
The analysis is dependent on estimates of a range of parameters (future op-
erating profits, useful life, growth rates, terminal value and discount rate).
The author will tie some comments to the OECD guidance on these param-
eters in sections 13.3.2.-13.3.4.1235
13.3.2. The estimation of future operating profits
Arguably, the most important parameter in a DCF valuation is the estimate

of future operating profits generated by the product that makes use of the
intangible in question.1236 All else equal, the larger the estimated future
operating profits, the larger the amount yielded by the valuation will be.
Taxpayers will tend to argue for relatively low future operating profits, and
the opposite goes for tax authorities.
Whether the profit projections later turn out to be correct or not (i.e. wheth-
er they correspond with the profits actually realized) will depend on devel-
opments in the marketplace that are unknown at the time at which the valu-
ation is undertaken. The projections are therefore speculative in nature.
Their underlying assumptions must be carefully examined, as there will
1235. In line with this, the OECD finds that the reliability of a valuation may be influ-
enced by the purpose for which it was produced; see OECD TPG, para. 6.155. Valua-
tions performed for non-transfer pricing purposes are generally deemed to be more reli-
able than those prepared exclusively for transfer pricing purposes. Further, valuations
of intangibles reflected in a purchase price allocation (PPA) performed for financial
accounting purposes are not decisive for transfer pricing purposes. The most impor-
tant global accounting standard governing PPAs is International Financial Reporting
Standard (IFRS) 3 (Business combinations). In the United States, a PPA will normally
be carried out pursuant to the Financial Accounting Standards Board’s (FASBs) Finan-
cial Accounting Standard (FAS) 141 (Business Combinations) and FAS 142 (Goodwill
and Other Intangible Assets). See also Chandler et al. (2010).
1236. Thus, the estimates of financial projections will pertain to accrual-based meas-
ures of income as opposed to cash flows. The new guidance suggests that accrual-based
measures may not properly reflect the timing of cash flows, which can generate a differ-
ence in the outcome between an income and cash flow-based approach; see 2015TPR,
at n. 19. It is, however, also recognized that the use of income projections may yield a
more reliable result than cash-flow projections.
371
through Valuation
normally be strong biases both on the side of the taxpayer and of the tax
authorities for skewing the estimates in the desired direction.1237
If the product that makes use of the transferred intangible has a proven
track record, its historical sales and costs may serve as useful indicators
of future performance.1238 Projections for products that are still in develop-
ment or that have yet to be commercialized are inherently less reliable. The
further into the future the intangible in question is expected to yield profits,
the less reliable the projections of income and expenses are likely to be.1239
A relevant question is whether intangible development costs should be in-

cluded in the projections of operating profits. If the transferred intangible
is continually developed and has an indefinite useful life, the OECD finds
it appropriate to include future development costs.1240 For intangibles that
are fully developed and will not serve as a platform for the development
of other intangibles, however, no development costs should be included.1241
13.3.3. The estimation of useful life, growth rates and

terminal value
A DCF analysis estimates the operating profits from an intangible value

chain over the useful life of the products to which the relevant intangibles
refer.1242 The useful life estimate is one of the critical assumptions sup-
porting the valuation model. All else equal, the longer the useful life, the
larger the amount yielded by the valuation will be. Taxpayers will tend
to argue for relatively short useful lives, and the other way around for tax
authorities. This was the case in both Veritas and Amazon.com, where the
taxpayer argued for a short useful life of the transferred intangibles, while
1237. Valuation methods normally leave substantial room for discretionary judge-
ment for the person carrying out the valuation. In a transfer pricing context, this is a
problem. Multinationals and tax authorities will normally be biased towards low and
high valuations, respectively. Also, there is significant information asymmetry between
multinationals and tax authorities. The R&D personnel of a multinational will be in a
vastly superior position relative to the tax authorities when it comes to judging whether
a partially developed medicine has the potential to be successfully developed and com-
mercialized. A tax authority may struggle to critically assess the information provided
by the multinational due to a lack of R&D knowledge and specialized business exper-
tise.
1238. OECD TPG, para. 6.166.
1239. OECD TPG, para. 6.165.
1240. OECD TPG, para. 6.167.
1241. Id.
1242. OECD TPG, para. 6.174.
372
the US Internal Revenue Service (IRS) applied a going-concern perpetuity

approach.1243 The useful life may be affected by the nature and duration of
the legal protections afforded to the intangible, the rate of technological
change in the industry and other factors affecting competition. The author
discusses useful-life assessments more in depth in the analysis of CSAs in
chapter 14.1244
Further, a key element in many DCF valuations is the projected growth

rate.1245 Growth is normally estimated as a yearly percentage increase in
revenues. All else equal, the larger the growth rate, the larger the amount
yielded by the valuation will be. Taxpayers will tend to argue for relatively
modest growth rates, and the opposite goes for tax authorities (as the case
was in both Veritas and Amazon.com).
Enterprises that reinvest portions of their earnings into the business and
generate high returns on these investments should grow at high rates. As
the enterprise grows, it will gradually become more difficult to maintain
a high growth rate. Eventually, the company will not be able to grow at a
rate larger than the growth rate of the economy of which it is a part. This
growth rate, the “stable growth rate”, may, in theory, be sustained in per-
petuity. This premise makes it possible to value all operating profits that
are generated in the stable growth phase as a terminal value for a going
concern, which is the present value at a future point in time of all future
operating profits from the value chain.
A terminal value may be useful when it is clear that the intangible value
chain in question is likely to generate profits beyond the period for which
reasonable financial projections exist.1246 Some intangibles may yield prof-
its in years after the legal protections have expired or the products to which
1243. Veritas Software Corporation & Subsidiaries v. CIR (133 T.C. No. 14 [U.S.Tax
Ct. 2009], IRS nonacquiescence in AOD-2010-05); and Amazon.com v. CIR (148 TC
No 8). See the analysis in secs. 14.2.4. and 14.2.5., respectively.
1244. On useful life, see also the discussion of the 2007 Coordinated Issue Paper (CIP)
(LMSB-04-0907-62) valuation approach in sec. 14.2.3.; the comments on Veritas in
sec. 14.2.4., the comments on Amazon.com in sec. 14.2.5., the discussion of the in-
come method of the US cost-sharing regulations (Treas. Regs. § 1.482-7(g)(4)) in sec.
14.2.8.3.; and the discussion of the general 2017 OECD periodic adjustment authority
in sec. 16.5. See also the discussion of comparability factors under the US regulations
in sec. 6.6.5.
1245. OECD TPG, para. 6.169.
1246. OECD TPG, para. 6.177. Estimating operating profits beyond this period is not
practical and involves a range of estimation risks, reducing the credibility of the esti-
mates.
373
through Valuation
they relate have ceased to exist,1247 e.g. when one generation of intangibles
forms the basis for the research and development (R&D) of future genera-
tions. The OECD finds that it may be appropriate that some portion of the
continuing profits from projected new products should be attributed to oth-
erwise expired intangibles when such follow-up effects exist.1248
The new guidance warns, however, that while some intangibles may have
an indeterminate useful life at the time of valuation, this does not imply
that residual profits will be allocable for such intangibles in perpetuity.1249
It is unusual for revenues derived from a particular product group to grow
at a steady rate over a long period.1250 Thus, tax authorities should exercise
caution in accepting simple models containing linear growth rates that are
not justified on the basis of either experience with similar products and
markets or a reasonable evaluation of likely future market conditions. It
will presumably be difficult to find reliable growth rates for value chains
driven by unique intangibles. It is normally unrealistic to acquire access
to disaggregated profit data from unrelated enterprises for specific compa-
rable products.1251 The internal profit data of a multinational from similar
products may be relevant.
There is reason to be sceptical of the use of terminal values in DCF valu-

ations. The terminal value is generally dependent on pronouncedly sub-
jective assessments, and is thus susceptible to manipulation. The terminal
value may represent a significant portion of the total estimated operating
profits. For instance, in an example illustrating the application of the in-
come method with a terminal value calculation, the total value of the intan-
gible (the buy-in payment) was set to 1,361,1252 of which the terminal value
accounted for approximately 62%. The author also refers to the analysis of
Veritas in section 14.2.4. and of Amazon.com in section 14.2.5., where the
IRS valuation assertions were based on the presumption of perpetual use-
ful lives for the transferred intangibles, combined with high growth rates,
resulting in significant terminal values. If the valuation includes a terminal
value, it is crucial that multinationals and tax authorities clearly outline the
assumptions used in the valuation and provide detailed explanations of the
rationale behind the assumptions.
1247. OECD TPG, para. 6.176.

1248. Id.
1249. Id.
1250. OECD TPG, para. 6.169.
1251. See the discussion of publicly available third-party financial accounting profit
data in sec. 6.2.5.
1252. Treas. Regs. § 1.482-7(g)(4)(viii), Example 7.
374
13.3.4. The estimation of discount rates
The discount rate used in a DCF valuation represents the investors’ cost
of capital.1253 All else equal, the larger the discount rate, the smaller the
present value amount yielded by the valuation will be. Taxpayers will tend
to argue for relatively high discount rates, and the opposite goes for tax
authorities (as also illustrated in Veritas and Amazon.com).
The OECD position is that no single measure for a discount rate is deemed
appropriate in all instances for transfer pricing purposes.1254 When the pur-
pose of the valuation is to value the total capital of a business, the weighted
average cost of capital (WACC) is used.1255 In this context, however, where
the purpose is to value an intangible, the WACC will generally not be a
suitable cost of capital, as the discount rate should be consistent with the
riskiness of the profits being discounted. The guidance emphasizes that
intangibles, particularly those still in development, may be among the most
risky components of a multinational’s business. Risk will, however, gener-
ally be moderate for IP connected to products and services that are suc-
cessfully established in the marketplace.
Risks should be taken into account either in the estimates of future operat-
ing profits or in the discount rate, but not in both.1256 Thus, if the estimated
operating profits have been reduced to take into account potential risk, the
discount rate should be lower than it would have been without such a re-
duction, and vice versa. Both multinationals and tax administrations must
document and explain the assumptions behind the discount rate.1257
1253. It is the minimum return that the investor requires in order to invest in the asset.
The discount rate consists of two elements. The first element is compensation for the
time value of money. Had the investor not invested the funds in the asset in question,
he could alternatively have invested in other income-generating assets. The second ele-
ment is compensation for the risk of receiving cash flows that are smaller or that mate-
rialize later than estimated.
1254. OECD TPG, para. 6.171.
1255. The weighted average cost of capital (WACC) is the cost of equity weighted by
the ratio of equity to total capital, plus the cost of debt after taxes weighted by the ratio
of debt to total capital. There is a variety of competing financial models available for
setting the cost of equity for an enterprise. The available models are usually variants of
the capital asset pricing model (CAPM) (see Bodie el al. (2005), at p. 281, with further
references). The CAPM states that the expected return on equity for an investor is equal
to a risk-free rate plus an additional interest rate for non-diversifiable risk. The cost
of debt is the rate at which the enterprise may currently borrow. The cost of debt will
reflect the default risk of the enterprise and the level of interest rates in the market.
1256. OECD TPG, para. 6.173.
1257. OECD TPG, para. 6.170.
375
through Valuation
13.3.5. What are the consequences of using unreliable

valuation parameters?
The 2017 OECD TPG on valuation are linked to the 2017 wording on peri-
odic adjustments for hard-to-value intangibles.1258 If there is any reason to
believe that the “projections behind the valuation” of an intangible are “un-
reliable or speculative”, the guidance on hard-to-value intangibles is appli-
cable.1259 The author observes that the quoted wording seems to reserve the
link to the periodic adjustment guidance solely for the financial projections
used in the valuation. Nevertheless, he finds no sensible reason as to why
financial projections in this context should not be read to also include the
terminal value used in the valuation, as well as the parameters underlying
it (useful life and growth rate) and the discount rate. The author concludes
that if any of the parameters used in the valuation are deemed unreliable or
speculative, the 2017 guidance on periodic adjustments must apply.
13.4. Allocation of the valuation amount among

the controlled value chain contributions
The outcome of the valuation is an estimate of the present value of the
future operating profits from the relevant intangible value chain (i.e. sales
of the product that makes use of the unique intangible in question). The
problem then turns into how this value shall be allocated among the con-
trolled value chain inputs.1260 This is the real transfer pricing assessment in
the context of valuations.
The new valuation guidance does not specifically address this issue, as
the allocation, in principle, relies on an application of the relevant transfer
1258. For an analysis of how the valuation may be adjusted if the hard-to-value intan-
gibles guidance on periodic adjustments is triggered, see the discussion in sec. 16.5.
1259. OECD TPG, para. 6.168.
1260. The OECD TPG leave substantial freedom for the multinational to define the
form of payment for the transfer of an intangible (see OECD TPG, paras. 6.179-6.180).
Payment may take the form of a single lump-sum payment, periodic payments or pay-
ments contingent on future events (e.g. sales thresholds). The payment form may affect
the level of risk incurred by a controlled party. For example, the contingent payment
form will generally involve greater risk for the transferer than a lump-sum payment
at the time of the transfer of the IP. The form of payment must be consistent with the
economic substance of the controlled transaction. See the discussion of controlled risk
allocations in sec. 6.6.5.5. Payments should reflect the relevant time value of money and
risk features of the chosen form of payment (see OECD TPG, para. 6.180).
376
Allocation of the valuation amount among the controlled value chain
contributions
pricing methodology.1261 The transfer pricing methods and the new guid-
ance on location savings, local market characteristics and synergies1262
will split the valuation amount so that (i) routine value chain contributions
(manufacturing, distribution, etc.) are allocated a normal market return;
(ii) location savings and rent are allocated incremental profits; and (iii)
residual profits are allocated among the non-routine value chain contribu-
tions (unique IP).
A complicating factor is that the structure of the value chain contributions

may change over the valuation horizon (for instance, relatively more mar-
keting efforts during the first years and less in later years).1263 In principle,
it should be estimated how much of each value chain contribution is used
in each income period covered by the valuation, but this may be difficult to
achieve in practice. It will likely, in general, be necessary to presume that
the current structure will be applicable throughout the valuation horizon.
If the product that makes use of the transferred intangible only relies on this
particular intangible while all other value chain inputs are routine inputs,
the entirety of the residual profits from the value chain will be allocable for
this intangible. Thus, in these cases, the total net present value of the prof-
its left after the routine inputs have been allocated a normal market return
pursuant to a one-sided method and incremental profits have been allocated
to location savings and rent will be the value of the intangible. Normally,
however, the residual profits will stem from more than one intangible. In
these cases, the residual profits must be disaggregated and split among the
intangibles so that the transfer price only contains the profits connected to
the intangible that is actually transferred in the controlled transaction.1264
In the absence of CUTs, the profit split method (PSM) must be applied,
the core of which is to allocate the residual profits pursuant to the relative
value of the unique intangibles employed in the value chain. The author
refers to the previous discussions of the PSM.1265 A complicating factor in
this context will be that the relative values of the unique intangibles must
be estimated for each of the future income periods encompassed by the
valuation. This may be problematic, as the relative values may change over
1261. The author refers to the analysis of the transfer pricing methods in chs. 7-9 and
11.
1263. This will be relevant in particular for long-term transfer pricing structures. See,
for instance, the discussion of French in sec. 5.2.3.3.
1264. Such analysis will be demanding. On this, see also Brauner (2008), at p. 116.
1265. The US and OECD PSMs are discussed in ch. 9.
377
through Valuation
time. This was the case in French, where the value of the patent gradually
became less important for the generation of profits than the continually
developing manufacturing know-how.
An additional layer of complexity will be added to the profit allocation if

the transferred intangible is not fully developed. There will be a legal basis
under the 2017 OECD guidance on intangible ownership to also allocate re-
sidual profits to the transferee if it contributes important R&D functions.1266
The allocation of residual profits between the transferer and the transferee
must then depend on the extent to which the transferred intangible was devel-
oped at the time of the transfer. The intangible is only capable of generating
profits once completed, and in these cases, the transferee may complete the
R&D. If there is a genuine contribution of value through R&D performed
by the transferee, the residual profits from the fully developed intangible
should be split in proportion to the intangible value added through the par-
ties’ R&D contributions.1267 It is crucial that this assessment is performed
with a critical view, as transfers of partially developed intangibles are often
carried out to shift intangible profits away from high-tax jurisdictions.1268
Under the 2017 OECD guidance (as opposed to the previous stance ex-
pressed in the 2009 OECD TPG text), the transferee should not be allo-
cated residual profits if the transferer is a group R&D centre that develops
the initial stages of the intangible and then transfers it (typically to a for-
eign “cash box” entity transferee) and continues to perform R&D after the
transfer on a contract research basis (or under similar arrangements) while
the funding of the R&D efforts is provided by the foreign transferee. The
transferee should nevertheless be allocated an arm’s length return on its
funding investment.1269
1266. OECD TPG, para. 6.56. See the analysis of the “important functions doctrine” in
sec. 22.3.2.
1267. See the discussion of the use of profit splits for transfers of partially developed
intangibles (OECD TPG, para. 6.150) in sec. 9.4.
1268. See the analysis of the 2017 OECD guidance on hard-to-value intangibles in
sec. 16.5. Under US law, problems associated with the transfer of partially developed
intangibles have proven problematic in the context of CSAs, in which the transferer
contributes a pre-existing intangible to a CSA while the other CSA participants contrib-
ute intangible development costs. The author analyses the US CSA regulations (Treas.
Regs. § 1.482-7) in ch. 14. The methods developed for valuing the contribution of pre-
existing intangibles (buy-in pricing) also apply for IP transfers that fall outside the
scope of Treas. Regs. § 1.482-7, including arrangements for the sharing of costs and
risks of developing IP in arrangements other than CSAs. See the discussion of unspeci-
fied methods in ch. 12.
1269. OECD TPG, para. 6.62. See the analysis of profit allocation for R&D funding
contributions in sec. 22.4.
378
The options realistically available as a restriction on the possible
allocation outcomes
13.5. The options realistically available as a restriction

on the possible allocation outcomes
The 2017 guidance envisions that the “realistic options available” principle
shall play a key role in the allocation of operating profits for the unique in-
tangible being priced, primarily as a restriction on which allocation results
are acceptable under the arm’s length principle.
The premise for the following discussion is that a DCF valuation has been
performed, resulting in an estimate of the present value of the future oper-
ating profits from the intangible value chain. Further, this value, pursuant
to the transfer pricing methodology, has been allocated among the value
chain contributions, including residual profits to the intangibles employed
in the value chain. The question is whether the realistic-options-available
principle can be used as a “sanity check” with respect to whether the profits
allocated for the relevant intangible are at arm’s length, i.e. whether the
principle restricts the possible allocation outcomes.
The realistic-options-available principle is a fundamental pricing assump-

tion underlying the OECD transfer pricing methodology. It dictates that
the transferer and transferee in a controlled transaction must be presumed
to act in accordance with rational economic reasoning. The core of the
principle is to demand that controlled resources are utilized in an optimal
manner. If controlled allocations of residual profits that do not adhere to
this basic logic were to be accepted with effect for the distribution of op-
erating profits among jurisdictions, the results yielded by international tax
law would depart from basic economic reasoning and third-party market
behaviour. Intangible income would then not be allocated according to in-
tangible value creation.
The principle is powerful when applied in a valuation context because it

may be relatively straightforward to calculate the profits associated with
the transferer’s alternative of not transferring the intangible, but rather car-
rying on its business activities as before the controlled transaction took
place. The profits from this “status quo” alternative will normally represent
a lower limit for an acceptable transfer price. Correspondingly, the best
realistic option available for the transferee may be to not enter into the
controlled transaction. Normally, this will lead to the conclusion that the
transferee must receive at least a normal market return on its investment in
order to not be placed worse off by entering into the controlled transaction.
379
through Valuation
Even though the realistic options available must be assessed from the per-
spectives of both the transferer and the transferee,1270 the main focus in a
transfer pricing valuation will generally be on the allocation of profits to
the transferer (which, as opposed to the transferee, typically is resident in
a high-tax jurisdiction and has often developed the IP). The reason for this
is that the transferer is the owner of the intangible, and thus entitled to all
residual profits from it. Had it not transferred the intangible, but instead
utilized it to make and sell the derivative products itself, it would have been
in a position to reap the entire residual profits alone. A transfer of the in-
tangible should not, according to the realistic-options-available principle,
make it worse off. The principle demands that there must be a convincing
reason as to why the transferer should relinquish the residual profits that it
is entitled to. That reason must be an arm’s length compensation.1271
The 2017 OECD TPG contain an illustration of the significant role played
by the realistic-options-available principle in the context of a valuation.1272
It pertains to a parent that has developed patents and trademarks related to
product F, which it has manufactured and sold to local distribution subsidi-
aries throughout the world. The parent establishes a subsidiary in a foreign
jurisdiction where the tax rate and production costs are significantly lower
than in its jurisdiction, and transfers its production to the subsidiary and
sells the product F patents and trademarks to the subsidiary for a lump
sum. The question is how an arm’s length price should be set for the trans-
ferred intangibles.
The example considers three alternative perspectives for determining the

lower and upper boundaries of an acceptable transfer price:
(1) The first scenario is seen from the perspective of the transferring par-
ent. The present value of the residual profits from the intangibles is
estimated based on the operating profits derived by the parent from the
manufacturing and sale of product F in the “status quo” scenario where
it owns the intangibles and makes and sells the products itself. This is
the value that the intangibles would have for the parent had it instead
1270. OECD TPG, para. 6.139.

1271. See also, in this direction, Brauner (2010), at p. 17, with respect to the similar
issue in the context of buy-ins under the US CSA regulations.
1272. OECD TPG, annex to ch. 6, Example 29 (Pervichnyi). The 2013 Revised discus-
sion draft on transfer pricing aspects of intangibles (2013 RDD) version of the example
was materially the same as that of the 2014D and the final 2015 text, but in addition,
contained detailed calculations explaining the numbers arrived at.
380
The options realistically available as a restriction on the possible
allocation outcomes
chosen not to enter into the controlled transaction. The present value
of the residual profits in this scenario is 601.
(2) The second scenario is relatively the same, but with the twist that the
parent contracts the subsidiary to perform manufacturing in order to
reap the benefits of lower local production costs. The present value of
the residual profits for the parent in this scenario is 853.
(3) The third scenario is seen from the perspective of the subsidiary, where
it owns the intangibles and makes and sells product F. Without taking
into account the investment of the subsidiary required to purchase the
intangibles, the present value of the residual profits generated by the
intangibles is 1,097.
As the second scenario is deemed to be a realistic option for the parent, that
option must be taken into account for the purpose of determining the trans-
fer price. Otherwise, the transfer price cannot be deemed to be at arm’s
length. The new guidance finds that “there is no reason” for the parent to
sell the intangibles at a price that would yield an after-tax cash flow with a
present value of less than 853.1273 Thus, the present value of the best realis-
tic option for the parent is the lowest acceptable transfer price.
Further, the guidance finds that the subsidiary would not be expected to
pay a price for the transferred intangibles that would leave it with an af-
ter-tax return lower than which could be achieved by not engaging in the
controlled transaction.1274 The upper limit for the price of the intangibles
is therefore 1,097. If the subsidiary pays more than that, it would not only
miss out on a return, but it would incur a loss.
In general, as long as the net present value of entering into the controlled
transaction is calculated using the discount rate of the buyer, any transfer
price that is less than the net present value of his best realistic alternative
will place it better off by entering into the controlled transaction. This must
be the fundamental criterion when applying the realistic-options-available
principle from the side of the buyer, as it then will be ensured an invest-
ment with a positive net present value.
The author finds the emphasis of the 2017 OECD guidance on the role
of the realistic-options-available principle in the context of valuation both
1273. OECD TPG, annex to ch. 6, para. 108.

1274. OECD TPG, annex to cha. 6, Example 29, para. 109.
381
through Valuation
necessary and potent. It is necessary in the sense that it would be deprived

of economic logic to force a sales price upon the seller that is lower than
which he could have achieved with a realistic alternative. It is potent in the
sense that the net present value of the seller’s best alternative will be per-
suasive for the transfer price determination, and will thus form an efficient
minimum arm’s length amount.
13.6. Is there a legal basis for applying a discount

to the transfer price?
The valuation of a unique intangible will likely entail a great deal of un-
certainty, regardless of where the transferred intangible is in its life cycle
at the time of transfer.1275 Further, the value will be theoretical, in the sense
that it will not necessarily be a value that the intangible actually could be
sold for to an unrelated party. For instance, a partially developed medicine
would likely be difficult to sell to a buyer who does not possess the multi-
national’s ongoing R&D capabilities. It is conceivable that such factors in
practice could open up for price negotiations. The question is whether there
is a legal basis to apply a “discount” to the theoretical value yielded by the
transfer pricing valuation akin to a discount that a potential third-party
buyer might demand in order to purchase the intangible.
In the absence of a CUT stating otherwise, there can be no legal basis for
a discount. The reason for this is simple. The one-sided methods and the
PSM allocate operating profits among the controlled parties based on their
contributions of routine and non-routine inputs to the value chain. Any
uncertainty in the valuation will be exhaustively taken into account either
in the estimation of the future operating profits from the intangible value
chain or in the determination of the applicable discount rate used to convert
the estimated future profits to present values.
1275. On the significance for transfer pricing purposes of the life cycle stage of the IP,
see Dolman (2007), under the section “Life cycle of IP”.
382
Chapter 14
Allocation of Residual Profits to Unique and

Valuable IP through Cost-Sharing Structures
14.1. Introduction
This chapter is dedicated to a discussion of how operating profits are al-
located through cost-sharing arrangement (CSA) “buy-in” pricing under
the US regulations and the OECD Transfer Pricing Guidelines (OECD
TPG).1276
Controlled CSAs are transactions in which group entities pool resources

to co-develop intangibles.1277 Each CSA participant is deemed to “own”
specified rights to the intangibles developed under the CSA.1278 No royal-
ties are therefore paid among the CSA participants with respect to their
subsequent exploitation of the cost-shared intangibles.1279 Each participant
1276. For informed discussions on cost sharing, see the writings of Brauner, in par-
ticular, Brauner (2010) and Brauner (2016). For a recent comparative study of cost-
sharing arrangements (CSAs), see Okten (2013). For a broad analysis (not limited to
buy-ins) and criticism of the 1995-generation US cost-sharing regulations, see Bensha-
lom (2007). Brauner (2017), at sec. 2.3.1.2, as well as Brauner (2016), at pp. 98, 102 and
112, asserts that the US cost-sharing regulations were likely the single most important
trigger behind the OECD BEPS Project. There is, in the author’s view, probably much
truth in this due to the publicity afforded to US multinationals that based their tax plan-
ning on cost sharing (Apple, Google, etc.) and the growing political will to tighten tax
rules. In this general direction, see also Avi-Yonah (2017), at p. 3; Musselli et al. (2017),
at pp. 333 and 335; and Zuurbier (2016), at p. 12.
1277. The OECD’s terminology corresponding to the US terminology of CSAs is “cost
contribution arrangements” (CCAs). For the sake of simplicity, the author will use the
term CSA regardless of whether the legal context is US or OECD transfer pricing juris-
prudence.
1278. The author will, in this chapter, refer to intangibles developed under a CSA as
“cost-shared intangibles”.
1279. The fact that CSAs are at all recognized as valid legal vehicles in current in-
ternational transfer pricing has been criticized due to the obvious potential for misuse
(migration of IP ownership from high-tax jurisdictions), based on the simple obser-
vation that group entities that contribute valuable research and development (R&D)
efforts to the CSA will normally be resident in high-tax jurisdictions, while the other
participating group entities often are “cash box” entities in low-tax jurisdictions. If
the ongoing R&D efforts are not based on pre-existing intangible property (IP) (thus
with no buy-in), the foreign entities will end up with (co-)ownership of the developed
IP based on cash contributions only. See, e.g. Brauner (2010), at p. 18; and Bensha-
lom (2007), at pp. 651 and 676. In Brauner (2016), at p. 112, recognition of CSAs is
described as representing “extraordinary benefit endowed to the [multinational enter-
383
through Cost-Sharing Structures
is, due to its “ownership”, entitled to the residual profits from the exploita-
tion of its rights.
CSA participants generally contribute three main types of development

inputs, namely (i) pre-existing intangibles; (ii) ongoing research and de-
velopment (R&D); and (iii) financing of R&D.1280 It is, in particular, the
contribution of pre-existing (often established high-profit) intangibles that
triggers profit allocation concerns.1281 These inputs represent the accumu-
lated value of prior R&D efforts and are generally the platform upon which
the further R&D that is to be carried out through the CSA will build. Con-
versely, ongoing R&D activities and financing represent concurrent contri-
butions that normally do not require compensation at the outset of a CSA.
The buy-in refers to the arm’s length charge that the other participants in
the CSA, which gain access to the contributed pre-existing intangible (first-
generation intangible property, or IP) and subsequently receive ownership
stakes in new versions of it developed under the CSA (second-generation
IP), must pay the contributing participant in return for the contribution.
Seen from the perspective of the R&D jurisdiction in which the CSA par-
ticipant that developed the pre-existing intangible is resident, the contribu-
tion of the intangible to a CSA, in reality, represents a migration of it. The
R&D jurisdiction will generally be allocated only a modest fraction of the
future residual profits from subsequent versions of the migrated intangi-
ble developed through the CSA.1282 The buy-in pricing represents the last
prises (MNEs)]”. See also Avi-Yonah (2017), at p. 4, where it is argued for repeal of
the cost-sharing regulations. The traditional pro-argument for continued recognition
of CSAs is reduced compliance burdens combined with a safe harbour for taxpayers;
see Benshalom (2007), at p. 658. See also Odintz et al. (2017), under sec. 2.2.6, where
the US CSA rules are described as focusing on the cost contributions of the controlled
parties, as opposed to the 2017 OECD IP ownership provisions that focus on value
contributions, i.e. the United States (through the CSA regulations) allows IP ownership
to be determined by way of costs, while the OECD provisions are only willing to as-
sign ownership to internally developed manufacturing IP to group entities that perform
development, enhancement, maintenance, protection and exploitation (DEMPE) func-
tions (see the discussion of the OECD’s “important functions doctrine” in sec. 22.3.2.).
1280. The bulk of the cash contributions will normally come from a foreign cash-box
entity located in a jurisdiction with a regime that imposes low or no tax on intangible
income. Conversely, pre-existing IP will normally come from a group entity located in
a high-tax jurisdiction. See also the analysis of the parallel problem under the US and
OECD IP ownership rules with respect to internally developed manufacturing intangi-
bles in secs. 21.2. and 22.2., respectively.
1281. On this point, see Brauner (2010), at p. 7.
384
Buy-in pricing under the US regulations
chance for the R&D jurisdiction to tax the full value of the pre-existing
intangible that was developed within its borders.
The buy-in is the most significant tax issue connected to CSAs.1283 Buy-in-
related reallocations have garnered significant controversy in recent years,
as illustrated by Veritas Software Corp. v. CIR and Amazon.com Inc. v.
CIR,1284 in which the US Internal Revenue Service (IRS) carried out sig-
nificant reassessments of the taxpayers’ buy-in values.
The author discusses buy-in pricing under the US regulations and articles 9
and 7 of the OECD Model Tax Convention (OECD MTC) in sections 14.2.,
14.3. and 17.5., respectively. The relationship between the US and OECD
buy-in pricing is discussed in section 14.4.
14.2. Buy-in pricing under the US regulations
The author begins by tying some comments to the development of the US

cost-sharing regulations in section 14.2.2. He then discusses the valuation
approach taken by the IRS in a 2007 Coordinated Issue Paper in section
14.2.3., before tying some comments to the US Tax Court rulings in Veritas
and Amazon.com in sections 14.2.4. and 14.2.5., respectively. Concluding
comments on the rulings are provided in section 14.2.6. The key concepts
and buy-in methodology of the current CSA regulations are discussed in
sections 14.2.7. and 14.2.8., respectively.
14.2.2. The development of the US cost-sharing regulations
The development of the US buy-in provisions largely mirrors that of the

general transfer pricing rules discussed in section 14.1.1285 In particular,
1283. The buy-in issue was also key under the previous-generation (1995) cost-sharing
regulations. For an interesting discussion, see Culbertson et al. (2003), at p. 102.
1284. Veritas Software Corporation & Subsidiaries v. CIR, 133 T.C. No. 14 (U.S. Tax
Ct. 2009), IRS nonacquiescence in AOD-2010-05. Veritas is analysed in sec. 14.2.4.
Amazon.com Inc. v. CIR, 148 TC No. 8 (U.S. Tax Ct. 2017). Amazon.com is analysed
in sec. 14.2.5.
1285. For a thorough and analytical discussion of the historical development of the US
cost-sharing regulations, see Brauner (2010), at p. 5; and Brauner (2016), at pp. 111-120.
385
the absence of CUTs provided breeding ground for the development of

unspecified pricing methods.
The 1968 regulations contained a brief provision for the sharing of costs
and risks,1286 which essentially required the terms of a CSA to be compara-
ble to those that would have been adopted by unrelated parties under simi-
lar conditions in order to be considered arm’s length.1287 The 1968 regula-
tions contained no buy-in requirement.
When the commensurate-with-income standard was introduced into the

US Internal Revenue Code (IRC) section 482 almost 2 decades later, the
legislature indicated how the 1968 CSA provision should be revised.1288
The position was that CSAs would be respected to the extent that their prof-
it allocation reasonably reflected the actual economic activity undertaken
by the parties to the CSA.1289 Further, if one party actually contributed
funds towards R&D at a significantly earlier point in time than the other or
was effectively putting its funds at risk to a greater extent than the other, an
appropriate return to that party would be required to reflect its investment.
The 1988 White Paper addressed the problem of multinationals selecting

specific high-profit unique intangibles for migration via CSAs (“cherry-
picking”).1290 A buy-in requirement was suggested, pursuant to which the
contribution of a pre-existing intangible to a CSA should command an
arm’s length compensation from the other participants. The White Paper
contained several interesting positions, some of which would later be re-
1286. 1968 Treas. Regs. (33 Fed. Reg. 5848) § 1.482-2(d)(4); see also the 1979 OECD
report on Transfer Pricing and Multinational Enterprises, para. 109. The final regula-
tions were preceded by the 1966 proposed regulations (31 Fed. Reg. 10394) § 1.482-
2(d)(4), which contained extensive rules for cost sharing. These were scrapped in fa-
vour of the high-level requirements contained in the 1968 regulations.
1287. Further, after 1982, US Internal Revenue Code (IRC) sec. 936(h) required that
the intangible income of a domestic company that qualified for the possessions tax
credit had to be included in the income of the US shareholder. One available way to
reduce the addition to US income was to choose a cost-sharing option.
1288. H.R. Conf. Rep. No. 841, 99th Cong., 2nd Sess. II-638 (1986), at p. 638.
1289. Also, the entity that contributed cash to the CSA, typically a foreign entity,
should bear its portion of all R&D costs, encompassing both unsuccessful and success-
ful products within a product area, at all stages of the R&D. Further, the allocation of
costs should be proportionate to the profit before deduction for R&D costs.
1290. This “cherry-picking” practice entailed that the profits from massively success-
ful intangibles were shifted abroad, combined with US deductions for R&D costs (in-
cluding also costs for unsuccessful R&D projects). To counter this, the White Paper
(Notice 88-123), at p. 127, suggested that CSAs had to encompass broad product areas
(based on the 3-digit SIC code). On cherry-picking, see also Benshalom (2007), at
p. 662; and Avi-Yonah (2017), at p. 4.
386
vised. One of which was that, apart from make-sell rights,1291 all rights to
pre-existing intangibles contributed to a CSA – most importantly, the right
to use the intangible for the purpose of further research – should command
a buy-in amount equal to the full market value.1292
The 1992 proposed regulations required an arm’s length compensation for

the transfer of a pre-existing intangible to a CSA.1293 No detailed guidance
was provided for the determination of the buy-in amount. Contrary to the
intention of the White Paper, the 1992 provisions contained no delimitation
for make-sell rights to pre-existing intangibles. The final 1995 cost-sharing
regulations largely carried over the 1992 proposed buy-in provisions.1294
1291. The White Paper did not intend for make-sell rights to be encompassed by the
buy-in requirement. Such rights should command ordinary royalty payments from the
licensee.
1292. The buy-in should also compensate participants for contributing basic R&D that
was not connected to any particular products, as well as the going concern value con-
nected to participant’s research facilities and capabilities that could be utilized in the
CSA. The White Paper was open to declining basis buy-in payments, since “intangi-
bles generally have greater value in the earlier stages of their life cycle” (White Paper,
p. 122). The White Paper did not require the buy-in compensation to be paid as a lump
sum. Also, periodic payments over the average expected life of contributed intangibles
were deemed acceptable.
1293. 1992 Prop. Treas. Regs. (57 FR 3571-01) § 1.482-2(g)(4)(iv). See Brauner (2016),
at p. 113, where he notes that there is a similarity between the 1992 US proposed re-
quirement that CSA participants had to use the CSA-developed IP in active conduct of
trade or business (the requirement was scrapped in the 1996 revision of the CSA regula-
tions) and the 2017 OECD TPG rule that pure cash-box entities cannot end up with the
residual profits from unique IP.
1294. 1995 Treas. Regs. § 1.482-7 (60 FR 65553-01). Some alterations were made to
the 1992 examples illustrating the buy-in provisions, none of which pertained to the
determination of the buy-in amount. In 2003, in the aftermath of Xilinx Inc. and Sub-
sidiaries v. CIR (125 T.C. No. 4 [Tax Ct., 2005], affirmed by 598 F.3d 1191 [9th Cir.,
2010], recommendation regarding acquiescence AOD-2010-03 [IRS AOD, 2010], and
acq. in result, 2010-33 I.R.B. 240 [IRS ACQ, 2010]), guidance requiring the inclusion
of stock-based compensation as intangibles development costs was introduced into the
cost-sharing regulations. For comments on Xilinx, see Brauner (2010), at p. 10; White
(2016), at p. 200; the 2010 report from the Joint Committee on Taxation (JCX-37-10),
at pp. 32-33; Blessing (2010b), at p. 174 (including a general discussion on the account-
ing of equity-based compensation in CSAs); Schön et al. (2011), at pp. 204-206; Horst
(2009); Lang et al. (2011), at p. 233; Levey et al. (2010), at pp. 156-158; Avi-Yonah
(2009a); Benshalom (2007), at p. 699 (see also p. 672); and Dix (2010). In Xilinx, a
US parent entered into a CSA with a wholly owned Irish subsidiary. Both entities per-
formed R&D, manufacturing and sales. All new technology developed under the CSA
would be jointly owned, and the intangibles development costs would be shared pursu-
ant to the reasonably anticipated benefit (RAB) shares of the participants. At issue was
the treatment of costs connected to employee benefits in the form of stock options and
employee stock purchase plans for stock in the parent, granted to the R&D staff of the
parent and subsidiary. The parent deducted the costs of employee stock benefits as busi-
ness expenses over the income years at issue. There was a reimbursement agreement in
387
Experience with the 1995 regulations demonstrated a need for additional

buy-in guidance.1295 Multinationals applied a range of different delimita-
tions (e.g. short intangible life spans) in their CSAs to justify low buy-in
valuations, resulting in offshore migration of significant rights to highly
valuable US-developed intangibles, for what was presumably only a frac-
tion of their true worth.1296
The 2005 proposed regulations were introduced to tackle these challeng-

es.1297 Similar to the observations made in the White Paper almost 2 decades
earlier,1298 no evidence was found that agreements comparable to typical
controlled CSAs actually existed among third parties.1299 Purported CUTs
(e.g. agreements on cost-plus “generic” research) presented to the IRS by
taxpayers were “not viewed as analogous to cost sharing agreements”,1300
as they normally involved materially different divisions of benefits, costs
and risks than those applied in controlled CSAs.
This absence of CUTs was the reason why the 2005 proposed regulations
adopted the basic buy-in pricing premise that the allocation of residual
profits should mirror the result that would have been realized if unrelated
parties under similar circumstances had engaged in a transaction similar
to the CSA (investor model). Inherent to this premise was the assumption
that a controlled party would act as an unrelated, rational investor would.
place between the parent and the subsidiary so that costs connected to employee stock
benefits for the Irish R&D staff would be levied on the subsidiary. The employee stock
benefit costs were not included as shared costs under the CSA. The US Internal Rev-
enue Service (IRS) took issue with this treatment. The taxpayer ultimately prevailed.
Xilinx is interesting on several levels, but its relevance for the purposes of this book is
limited. It does not touch upon the buy-in issue, and the interpretation of the court in
Xilinx is (in the author’s view, thankfully) no longer of relevance for CSAs entered into
subsequent to the 2003 amendment of the CSA regulations.
1295. See the preamble to the 2005 Prop. Treas. Regs. (70 FR 51116-01) § 1.482-7. See
also Dolan (1991); and Koomen (2015a), at p. 147.
1296. See Ballentine (2016), where typical buy-in pricing under the 1995 US CSA
regulations is described as “wildly inconsistent with the arm’s length standard”.
1297. On these proposed regulations, see Kirschenbaum et al. (2005).
1298. See White Paper, ch. 12, sec. A.
1299. This, of course, triggers the question as to why CSAs, with their obvious poten-
tial for misuse, were allowed in the first place. The prevailing view is that the accept-
ance of CSAs in transfer pricing arose organically, without being backed up by any
form of real study or justification, likely based on a naive belief that such agreements
were common among third parties and that all participants in a CSA would bring con-
tributions of somewhat equal value to the table. For good reflections on this important
issue, see Brauner (2010), at pp. 7, 15 and 17.
1300. See the preamble to the 2005 Prop. Treas. Regs. (70 FR 51116-01) § 1.482-7.
388
Unrelated parties would not grant valuable, pre-existing intangibles to a

CSA unless they received an appropriate award for doing so.
14.2.3. The 2007 Coordinated Issue Paper
In 2007, the IRS finalized a Coordinated Issue Paper (CIP) on buy-in

adjustments,1301 aimed at providing guidance for IRS auditors and appeals
settlements.1302 The CIP was withdrawn in 2012, but its material content
remains relevant for audit cases for income years prior to those governed
by the current 2011 cost-sharing regulations.1303 More importantly, the CIP
provides valuable context for the interpretation of the current cost-sharing
regulations (which has further developed the CIP buy-in doctrine). The
author therefore finds a discussion of the CIP to be justified.
The CIP provides guidance on the valuation of pre-existing IP that is made

available to a CSA. The valuation will also incorporate the pricing of rights
that are not transferred for the purpose of research, on an aggregate ba-
sis.1304 The CIP valuation approach was an answer to speculative taxpayer
applications of the specified pricing methods of the 1995 regulations to
value buy-ins.1305
A popular taxpayer approach was to use uncontrolled make-sell licences of

the pre-existing intangible as a basis for applying the CUT method to value
1301. LMSB-04-0907-62. For comments on the IRS’s 2007 Coordinated Issue Paper
(CIP), see Femia et al. (2008).
1302. The majority of US buy-in disputes were resolved at the examination or appeals
stages. At the time at which the CIP was finalized, only one buy-in case was pending
before the US Tax Court, Veritas, which was a major priority for the IRS.
1303. The current regulations apply to transactions entered into on or after 5 January
2009.
1304. Typically, the US parent also contemporaneously licenses make-sell rights in
the current generation of the pre-existing intangible to the foreign subsidiary (that is
party to the CSA). The CIP finds that licensed make-sell rights do not constitute buy-in
intangibles, as they are not “made available for purposes of research”; see Treas. Regs.
§ 1.482-7(g)(2). Thus, standing alone, such rights do not give rise to an obligation to
make a buy-in payment (the arm’s length payment for such licences should be deter-
mined under sec. 1.482-4). Nevertheless, an important aspect of the CIP approach is
that the buy-in intangible and the make-sell licence is valued on an aggregate basis.
See also the comments on the US best-method rule in sec. 6.4. and the aggregation of
controlled transactions in sec. 6.7.2.
1305. The specified methods were incapable of allocating profits to intangibles de-
velopment funding. See the analysis of R&D funding remuneration under the current
OECD TPG in sec. 22.4.
389
the buy-in.1306 The implicit taxpayer assertion was that uncontrolled make-
sell licences for a pre-existing intangible were comparable to making the
same intangible available to a CSA for the purpose of further R&D.1307
Royalties extracted from such CUTs would normally be adjusted down-
wards throughout the useful life period to account for obsolescence of the
licensed intangible.
The result was that the lion’s share of the ownership of, and residual profits
from, future versions of the pre-existing intangible could be migrated for a
price not larger than the present value of the royalties from standard make-
sell rights to the first-generation intangible. Thus, none of the residual prof-
its from the future intangibles developed under the CSA would be allocated
to the United States (where the pre-existing intangible, and thus the basis
for the subsequent research, had been created). The CIP rejected these ap-
plications of the CUT method.1308 The IRS made significant efforts in court
to defend its CIP doctrine against taxpayers’ CUT method applications for
buy-in pricing in the high-profile Veritas (2009) and Amazon.com (2017)
cases, but did not succeed.1309
Another taxpayer strategy was to use the residual profit split method
(RPSM) to determine the buy-in amount. The taxpayer application of
the RPSM placed the risks that are assumed by the foreign subsidiary
in making cost-sharing payments with respect to subsequent-generation
intangibles on par with the risks assumed by the US parent in prior years
1306. The useful life was often determined on the basis of the current-version product
life. This tended to limit the period during which buy-in compensation was to be paid
by the foreign subsidiary, often to a period of only 2-4 years following the formation of
the CSA.
1307. The result was that the buy-in payments would consist only of rapidly declin-
ing royalties over the period in which the current make-sell rights become obsolete.
The value of all residual rights to the contributed pre-existing intangible, including the
rights to use the pre-existing intangible as a basis for further research, would then fall
outside the scope of the buy-in.
1308. The CIP claims that make-sell rights generally will not be comparable to R&D
rights to the pre-existing intangible, in particular with respect to the similar profit po-
tential requirement (see the discussions in secs. 7.2.2. and 7.2.3.). The author finds it
fairly obvious that this must be correct. Make-sell rights simply allow the licensee to
reap some profits from the manufacturing and sale of a current-generation product over
its lifetime. R&D rights in the same intangible entail that the CSA participants receive
co-ownership rights to future versions of the intangible. R&D rights are therefore, in
reality, a vehicle to migrate ownership of an intangible.
1309. The rulings are analysed in secs. 14.2.4. and 14.2.5., respectively, with conclud-
ing comments in sec. 14.2.6.
390
when it developed the first-generation, pre-existing intangible.1310 The

residual profits from the cost-shared intangibles were split in propor-
tion to the foreign subsidiary’s capitalized and amortized share of ongo-
ing R&D costs, as compared to the sum of the US parent’s capitalized
and amortized past development costs for the pre-existing intangible
plus its capitalized and amortized share of ongoing R&D costs.1311 In
other words, if the foreign subsidiary funded 80% of the intangible’s
development costs, it would be allocated 80% of the residual profits.
Multinationals claimed that this was consistent with the profit split
method (PSM) provisions of the regulations.1312 The CIP rejected this
approach.1313
Further, certain valuation methods were used by taxpayers to attribute sub-

stantial positive net present value to the projected R&D expenses of for-
eign “cash box” subsidiaries that only provided R&D funding. The logic
behind this methodology was that because the current version of the buy-in
intangible was expected to become obsolete, over time, the cost-shared in-
tangibles resulting from the R&D should account for an increasing share
1310. The CIP finds that there is unlikely to be a parity of risks in the funding of past
and future R&D.
1311. The buy-in payments would decline (“ramp down”) significantly over time, as
the US parent’s past development costs are fully amortized (the useful life would often
be set to a short period of 2-4 years, based on the useful life of the make-sell rights to
the pre-existing intangible). After this period, the residual profit from the cost-shared
intangibles would be allocated between the parent and the subsidiary pursuant to their
RAB shares.
1312. Treas. Regs. § 1.482-6(c)(3).
1313. The example in Treas. Regs. § 1.482-6(c)(3) pertains to the determination of an
appropriate make-sell royalty for a licence for a fully-developed intangible when both
parties contribute pre-existing intangibles to the operations of the foreign subsidiary.
While the regulations prefer a split of the foreign subsidiary’s residual profits on the
basis of the relative value of the intangible property contributed by each controlled
taxpayer, they do allow the use of capitalized and amortized costs of development as
an allocation key. The CIP, however, argues that the use of cost as an allocation key
is unreliable, as the foreign subsidiary in a CSA context bears only prospective risk,
while the US parent brings both its past risks as well as its commitment to bear a share
of the prospective development risks, and these respective risks are not comparable.
The author agrees with the IRS’s conclusion. However, the principal reason as to why
the residual profit split method (RPSM) should be inapplicable in these cases is the
general requirement that the method may only be applied when both controlled parties
provide unique inputs. The residual profits should be allocated in full to the party that
contributes unique inputs, while the funding entity should be allocated a risk-adjusted
“normal” return on its financial contribution. The CPM would have been a candidate
for application, had it not been for the difficulty of obtaining reliable data on the return
on intangibles development funding contributions. See sec. 22.4. on the allocation of
profits to R&D funding under the OECD TPG; and sec. 22.4.7. with respect to the US
CSA regulations.
391
of the residual profits from future operations. Thus, for instance, if the
US parent and the subsidiary funded 20% and 80%, respectively, of the
ongoing R&D costs, the parent and subsidiary would be allocated 20%
and 80% of the residual profits, respectively. The CIP found that such ap-
proaches violated established principles of transfer pricing – most notably,
that controlled pricing should respect the pricing boundaries dictated by
the best realistic alternatives available to the controlled parties – and there-
fore should be rejected.1314
The CIP’s answer to these base-eroding taxpayer buy-in valuations was to

introduce the “unspecified income method”, which:
[…] determines the value of the buy-in intangible (along with any licensed
make-sell rights) as the present discounted value of the stream of projected
operating profits of the [controlled foreign company (CFC)] (or affiliated
CFCs), after reduction by routine returns and projected cost sharing payments
for the CFC’s share of the R&D projected under the CSA.
All intangibles that, at the time of the buy-in, were reasonably anticipated
to contribute to the R&D under the CSA were deemed to be buy-in intan-
gibles, for which an arm’s length charge had to be paid.1315
In other words, the buy-in amount was estimated by first valuing the en-
tire foreign subsidiary. From this value, deductions were made for normal
market returns allocable to the subsidiary, as well as the subsidiary’s share
of the projected cost-sharing payments under the CSA. This allocated the
entire residual profits to the US parent.1316
This valuation approach was the reverse of the standard “transaction-by-

transaction” transfer pricing methodology, as its objective was to provide
an aggregated valuation.1317 The method simply projected and discounted
1314. The CIP asserts that at arm’s length, the US parent would seek to retain that
potential future premium for its own benefit (aligned with its best realistic alternative),
unless the subsidiary provided sufficient compensation to make its net present value
zero (the subsidiary then receives an allocation of income equal to its discount rate).
See also the discussion of the best-realistic-alternatives principle in the context of valu-
ation under the OECD TPG in sec. 13.5.
1315. The CIP argued that the synergy value of the research team responsible for the
buy-in intangible had to be included in the buy-in, regardless of whether the research
workforce as such was deemed to be an intangible.
1316. The income method applies the discounted cash-flow (DCF) business valuation
approach. See the analysis of valuation under the OECD TPG in ch. 13.
1317. The aggregated valuation approach of the CIP would become influential for the
further development of the US IRC sec. 482 regulations; see the discussion in sec. 6.7.2.
392
the subsidiary’s total future profits without taking into account which spe-
cific assets, functions and risks were responsible for the profit.
The author finds no obvious reason why it should be appropriate to include

in the buy-in valuation, as the CIP income method did, in the value of
rights other than the right to use the contributed pre-existing intangible for
the purpose of further research, as such other rights may not contribute to
the R&D under the CSA (e.g. make-sell rights to the current-generation
intangible).1318 On this point, the CIP approach was too aggressive. The
author suspects that such rights were included because the method would
be difficult to apply otherwise.
The justification offered by the CIP for its aggregated valuation approach
was, however, not this practical consideration, but a more principled argu-
ment: if the US parent combined a transfer of the rights to use the pre-ex-
isting intangible for the purpose of further R&D with a make-sell licence
for the same intangible, the total transfer would be so comprehensive that
it should be deemed as being akin to a sale of the current-generation in-
tangible. This entailed that the valuation could be carried out on a going-
concern basis, taking into account future income streams for perpetuity, as
opposed to a limited number of years.
This methodology purportedly had a number of advantages over the speci-

fied methods available at the time. First, it provided remuneration for R&D
funding, in the form of an expected return on cost-sharing payments. This
was not possible to accomplish through the existing specified methods,
such as the comparable profits method (CPM). Second, as it took into ac-
count the projected operating profits of only the foreign subsidiary, it auto-
matically excluded the effect of any operating intangibles that solely ben-
efitted the US parent (i.e. intangibles that were not transferred), such as US
goodwill and going-concern value. This distinguished the income method
from the market capitalization method that was also introduced into the
1318. Where transactions between the US parent and the foreign (cash-box) subsidi-
ary are limited to one value chain, an aggregated valuation approach seems useful.
If, however, the foreign subsidiary is an established entity with functions, assets and
risks besides those relevant to the examined CSA, a reliable application of the income
method will be contingent on a successful separation of the profits connected to the
CSA intangibles (including any licensed make-sell rights and marketing intangibles)
from the other business profits of the subsidiary so that the buy-in valuation does not
encompass profits from intangibles other than those relevant to the CSA. This was an
issue in Amazon.com; see the discussion in sec. 14.2.5.
393
CIP.1319 Third, the method could, simultaneously with the determination of

the buy-in amount, also set an arm’s length charge for other pre-existing
operating intangibles owned by the US parent that were used by the foreign
subsidiary in its operations (aggregated valuation).
The main weakness of the unspecified CIP income method was, in the
author’s view, its ambiguous approach to the allocation of operating profits
to the R&D funding provided by the foreign subsidiary.1320 The method fo-
cused on the allocation of income to the US parent and on the best realistic
options available to it. The CIP should have more thoroughly considered
the controlled transaction from the point of view of the foreign subsidiary,
in particular, that the subsidiary’s best realistic option for entering into the
CSA would be to invest its R&D funding in alternative projects with simi-
lar risk-return profiles and should thus be allocated a risk-adjusted return
through the calculation of the buy-in amount.
14.2.4. Case law: Veritas (2009)
The 2009 US Tax Court ruling in Veritas Software Corporation v. CIR

dealt with the valuation of a buy-in payment.1321 The ruling is significant,
as it illustrates several key issues pertaining to buy-in valuations. The IRS
1319. The market capitalization method was introduced as an unspecified method to

evaluate an initial buy-in. The total value of all intangibles owned by the US parent was
determined as the difference between market capitalization of the US parent (market
value of equity plus liabilities), minus the value of tangible property. The initial buy-in
amount under the market capitalization method was set as the foreign subsidiary’s pro
rata share of the buy-in intangible amount, where the pro rata share was the RAB share
of the subsidiary.
1320. The CIP could arguably be interpreted so that routine returns shall be allocated
only to the functions of the subsidiary, while its cost-sharing payments shall be refund-
ed at their nominal amounts. The subsidiary would, of course, have a better realistic
option available relative to entering into a CSA if it were to receive no return on its CSA
funding contribution. The author refers to the discussion of the arm’s length remunera-
tion of intangible development funding in the context of the OECD TPG in sec. 22.4.,
which should also be relevant here (the transfer pricing considerations are the same).
1321. Veritas Software Corporation & Subsidiaries v. CIR, 133 T.C. No. 14 (U.S. Tax
Ct. 2009), IRS nonacquiescence in AOD-2010-05. Veritas was presided by Tax Court
Judge Foley, who also ruled in Xilinx Inc. and Subsidiaries v. CIR, 125 T.C. No. 4
(Tax Ct., 2005), affirmed by 598 F.3d 1191 (9th Cir., 2010), recommendation regarding
acquiescence AOD-2010-03 (IRS AOD, 2010), and acq. in result, 2010-33 I.R.B. 240
(IRS ACQ, 2010). For discussions of Veritas, see Brauner (2010), at p. 13; JCX-37-10,
at pp. 33-34; Schön et al. (2011), at pp. 206-210; Wittendorff (2010a), at pp. 644-645,
651 and 660; Levey et al. (2010), at pp. 152-153; Hughes et al. (2010); Lin et al. (2010);
and Odintz et al. (2017), at sec. 2.2.5. See also Brauner (2016), at p. 117 on the ruling.
394
argued that the CIP valuation approach should be applied under the 1995
cost-sharing regulations. As the specified income method of the present
CSA regulations represents a further development of the 2007 CIP ap-
proach, many aspects of the case are relevant to the interpretation of the
current regulations. Veritas is also relatively unique, as it is one of only two
transfer pricing cases (that the author is aware of) that provides a detailed
buy-in valuation analysis, with the other case being Amazon.com.1322
The case pertained to a US parent that, in 1999, entered into a CSA with a
newly established Irish subsidiary for the purpose of developing and man-
ufacturing software products. The CSA consisted of a package of three
agreements, in which the parent (i) assigned all existing sales agreements
with European-based group entities to the subsidiary; (ii) entered into an
agreement with the subsidiary for the sharing of R&D costs, for which the
parties agreed to pool their respective resources and R&D efforts related to
software products and manufacturing processes and to share the R&D risks
going forward; and (iii) entered into a technology licence agreement with
the subsidiary in which make-sell rights to the covered products in Europe,
the Middle East, Africa and the Asia-Pacific region were granted. In ex-
change for these rights, the subsidiary agreed to pay royalties to the parent.
The subsidiary was a marketing and distribution entity, largely performing

routine functions. No R&D was carried out in Ireland. The parent had de-
veloped a range of pre-existing software that it contributed to the CSA. The
products were commercially established and well known. At the time at
which the CSA was entered into, the parent sold its products either direct-
ly to customers or through third-party original equipment manufacturers
(OEMs), distributors and resellers, including Sun, HP and Dell Products.
The calculation of royalties under the OEM licence agreements was based
on list price, revenues or profits, and ranged from 10% to 40% for bundled
products and 5% to 48% for unbundled products, depending on the profit
potential and sales volume of the products. The market for the Veritas soft-
ware was intensely competitive. The current-version software products lost
value quickly, as new functions were duplicated by competitors. At the
time at which the CSA was entered into, the Veritas current-version prod-
ucts, on average, had a useful life of 4 years.
The core of the IRS’s argument was that the parent’s transfer of pre-exist-
ing intangibles was “akin to a sale” and that the transferred assets collec-
1322. 148 TC No. 8; see the discussion in sec. 14.2.5.
395
tively possessed synergies that provided the whole with greater value than
each asset standing alone, and that the controlled transactions therefore
should be aggregated for valuation purposes.1323 The IRS argued that this
approach was permissible under the general aggregation provision in sec-
tion 1.482-1(f)(2)(i)(A) of the US IRC section 482 regulations, according
to which transactions could be aggregated if they produced the “most reli-
able means of determining the arm’s length consideration for the controlled
transactions”.
The taxpayer asserted that this valuation approach was motivated by the
fact that, if valued separately, the transferred intangibles should be deemed
to have finite useful lives, but if valued aggregately as a business transfer,
a perpetuity or going-concern approach could be applied, taking into ac-
count intangibles that were subsequently developed under the CSA rather
than pre-existing at the time at which the agreement was entered into. A
valuation premised on perpetual lives for the underlying intangibles would
yield a higher buy-in price than a valuation premised on finite useful lives,
as the valuation then would encompass estimated operating profits from
more future income years.
The basic discounted cash-flow (DCF) methodology employed by both the

taxpayer and the IRS was fundamentally the same. The differences be-
tween their pricing approaches pertained to the component elements in the
valuations. The taxpayer based its estimates of income on royalty rates ex-
tracted from uncontrolled make-sell licence agreements with OEMs, while
the IRS used income projections. Further, the taxpayer assumed finite in-
come streams limited to the duration of the make-sell licences, while the
IRS assumed perpetual income.1324 The taxpayer valuation therefore did
not contain a terminal value, and no growth rate was used.1325 Further, the
taxpayer applied a higher discount rate than the IRS.1326
1323. Under the now-current regulations, the buy-in pricing would have been subject
to the specified income method; see Treas. Regs. § 1.482-7(g)(4), discussed in sec.
14.2.8.3. See also the comments on the US best-method rule in sec. 6.4.; and on the
aggregation of controlled transactions in sec. 6.7.2.
1324. The longer the useful lives of the intangibles being valued, the larger the output
of the valuation will be. Thus, a taxpayer will normally be inclined to argue for short
useful lives, and the opposite goes for the tax authorities.
1325. The higher the growth rate used in a valuation, the larger the value of the valua-
tion will be (the growth rate will reduce the discount rate used to calculate the terminal
value, thus increasing the terminal value and thereby the total valuation amount). Thus,
a taxpayer will normally be inclined to use a low growth rate, and the opposite goes for
the tax authorities.
1326. The lower the discount rate used in a valuation, the larger the value of the valua-
tion will be. Thus, a taxpayer will normally be inclined to use a large discount rate, and
396
The fundamental problem with the ruling, as the author sees it, is that it
fails to carry out a proper transfer pricing analysis of the transferred rights
to use the existing software for the purpose of further research. The tax-
payer hand-picked uncontrolled make-sell agreements between the parent
and third-party OEMs to determine a 20% appropriate starting royalty rate
for the buy-in payment with a useful life of between 2 to 4 years, which was
ramped down over the buy-in period. The IRS argued that the agreements
were incomparable to the controlled transaction, but this was rejected by
the Court.1327
The author struggles to follow the reasoning of the Court, as the parent, on
top of the transfer of make-sell rights, also transferred rights to use the pre-
existing intangibles for the purpose of further research. As a consequence,
the Irish subsidiary would become the owner of the rights to exploit fu-
ture software developed under the CSA in its markets. Ownership rights
to future intangibles are clearly distinct from make-sell rights to existing
software; the income from current-version, pre-existing intangibles will
be generated over the finite useful life period of the software. The income
from future software developed under the CSA will continue to be gener-
ated as long as new programs are developed. The income streams from
these two types of rights in the pre-existing software, and thus the associ-
ated profit potential, therefore differ significantly.
The parent, by entering into the CSA, in reality relinquished the owner-
ship rights to future software developed under the CSA with respect to the
subsidiary’s markets. It is the loss of income potential resulting from this
relinquishment for which the buy-in should compensate the parent. The in-
come from the make-sell rights to the existing software would, in any case,
be taxable to the parent, as it was the owner of those rights.
Thus, what the Court ended up doing in Veritas was pricing the make-sell
rights that the parent transferred to the subsidiary, not the rights to use the
the opposite goes for the tax authorities.

1327. The author takes issue with this. Under the comparable uncontrolled transaction
(CUT) method, there must be sufficient comparability between the significant eco-
nomic activities under the uncontrolled and controlled transactions; see Treas. Regs.
§ 1.482-4(c)(2)(iii)(A). A clearly significant activity under the CSA was the perfor-
mance of R&D. Third-party licences were purely make-sell agreements, under which
no R&D was to be performed. The uncontrolled and controlled transactions were there-
fore not comparable on this point. The justification of the court for disregarding this
fact, i.e. that its interpretation of sec. 482-7(g)(2) did not include income from future
software, is not convincing. In reality, the court failed to respect the comparability
requirements of the CUT method.
397
existing software for the purpose of further R&D. Even if the 1995 cost-
sharing regulations contained only sparse guidance on the issue of buy-in
valuation (relative to the current CSA regulations), it must be recognized
that the background law on transfer pricing was quite well developed. The
Court focused on the words “pre-existing intangible property” in the US
Treasury Regulations (Treas. Regs.) § 1.482-7(g)(2) and interpreted this to
mean that the buy-in valuation should only take into consideration income
from the existing versions of the software, not income from software sub-
sequently developed through the CSA. What the Court in fact did here was
to infer a stance on the transfer pricing of the buy-in from the first sentence
of section 482-7(g)(2), an issue that the provision did not address. In the
author’s opinion, the focus of the Court should instead have rested on the
second sentence of section 482-7(g)(2): “[T]he buy-in payment by each
such other controlled participant is the arm’s length charge for the use of
the intangible under the rules of §§ 1.482-1 and 1.482-4 through 1.482-6,
multiplied by the controlled participant’s share of reasonably anticipated
benefits.”
Had the Court instead focused on the transfer pricing principles expressed
in the 1994 regulations – as the 1995 buy-in provision specifically instruct-
ed – the outcome of the case should have been different.1328 Most impor-
tantly, the Court should have taken the alternatives realistically available
to the parent and the Irish subsidiary into consideration. This approach
would have had a solid legal basis in the 1994 regulations, as the realis-
tic alternatives principle was expressed in several provisions therein.1329
The fact that the principle was applicable also when applying the transfer
pricing methods, regardless of whether the relevant methodology was a
specified method or not, was clearly expressed in the preamble to the 1994
regulations:
[A]n unspecified method should provide information on the prices or profits
that the controlled taxpayer could have realized by choosing a realistic alter-
native to the controlled transaction. This guidance has been included because
it is a principle that is consistent with all methods that apply the arm’s length
standard. (Emphasis added)
1328. See Treas. Regs. (59 FR 34971-01), sec. 482-7(g)(2).

1329. E.g. in Treas. Regs. (59 FR 34971-01) sec. 1.482-1(d)(3)(iv)(H) as a comparabil-
ity consideration with respect to economic conditions; in sec. 1.482-3(b)(2)(ii)(B)(8) as
a comparability consideration with respect to the transfer pricing of tangibles; as well
as a principle that must be complied with when using an unspecified method to price
tangibles and intangibles in secs. 1.482-3(e)(1) and 1.482-4(d)(1), respectively.
398
Had the Court applied the realistic alternatives principle, it should have
recognized that if the parent had not entered into the CSA, but instead
developed new versions of the software itself, it would have been en-
titled to the residual profits from the future software from all regions
of the world. By entering into the CSA, it effectively relinquished the
residual profits from future versions of the software from the markets of
the subsidiary (which included all worldwide markets apart from the US
markets).
The buy-in payment should compensate the parent for this value transfer
and put it at least as well off as it would have been had it instead developed
the future software itself and licensed out make-sell rights.
From the point of view of the subsidiary, the best alternative to entering
into the CSA would have been to let the parent develop the future software.
It then would not have incurred any of the financial risks associated with
the intangible development, as it would have no obligation to pay any share
of the R&D costs associated with the CSA software development. It would
simply have licensed the fully developed intangible at an arm’s length roy-
alty rate.
After this analysis of the options realistically available, the Court should
have selected an appropriate transfer pricing method. A functional analysis
of the parent and the subsidiary should have resulted in the acknowledge-
ment that the parent provided all unique inputs to the intangible develop-
ment. All R&D was carried out in the United States by the parent, based
on its unique, self-developed intangibles, using its existing R&D teams.
The Irish subsidiary only contributed R&D funding. As only one of the
parties to the controlled transaction furnished unique inputs, a one-sided
methodology akin to the CPM would be the natural choice for determin-
ing the buy-in payment. The CPM would set the buy-in amount so that all
residual profits from future versions of the software developed under the
CSA would be allocated to the parent, while the subsidiary would receive
a normal return on its routine contributions.
Nevertheless, the CPM could not have been applied directly, as there would
be no comparable data available that reflected an arm’s length return on the
subsidiary’s R&D funding.1330 An unspecified method that compensated
1330. See also the analysis of the remuneration of R&D funding under the OECD TPG
in sec. 22.4. See also Odintz et al. (2017), at sec. 2.2.5.1, where it is argued that remu-
neration should be provided for R&D funding under applications of the unspecified
income method.
399
the subsidiary for its funding should therefore have been used. The main
trait of the CPM, i.e. that all residual profits are allocated to the party
that provides unique inputs, however, should have been carried over to this
unspecified method.1331 Thus, the subsidiary should have been allocated
an arm’s length remuneration for its functions and intangible development
funding only, as opposed to the residual profits effectively allocated to it
by the Tax Court. Remuneration of the subsidiary’s funding contributions
would have triggered challenging issues. Provided that the CSA R&D ac-
tivity pertained to further development of established and successful soft-
ware products, the remuneration should be based on the assumption that
the intangible development costs were subject to relatively modest risk lev-
els.1332
In conclusion, it may be observed that the parent was compensated for its
loss of ownership rights to the future intangibles developed under the CSA
(with respect to the subsidiary’s markets) by discounting an arm’s length
make-sell royalty for the income generated by the current-version, pre-
existing software. This did not compensate the parent for its actual loss of
value. The parent instead received a value that it would have been entitled
to regardless of its participation in the CSA (as the pre-existing software
was developed prior to the formation of the CSA). It is therefore difficult
to accept the ruling as an appropriate application of the transfer pricing
methods and the best-options-realistically-available principle. The author
does not find the result in Veritas to be compatible with the IRC section 482
arm’s length standard.1333
1331. As the goal would be to compensate the parent for its loss of income potential
from the future software, the useful life of the software should be assumed to be per-
petual if there are no clear indications that the CSA development will cease at some
specific future point, or at least significantly longer than the useful life of the current-
version pre-existing software. The discount rate applied by the taxpayer should likely
be accepted. The growth rate applied by the IRS should be rejected; the rate used in
the terminal value should be a “steady state” growth rate. All residual profits from the
subsidiary’s exploitation of the marketing intangibles should be allocated to the parent,
given that the subsidiary does not incur marketing expenses that exceed an arm’s length
level.
1332. See the analysis of the remuneration of R&D funding under the OECD TPG in
sec. 22.4.
1333. Brauner (2010) is also critical of the ruling. Brauner (2010), at p. 14; and Brauner
(2016), at p. 118, however, seems to read the ruling as having implications not only for
the interpretation of the previous-generation (1995) US cost-sharing regulations (the
interpretation of which was at issue in the case), but also for the current, new-generation
regulations. The author’s impression is that Brauner reads too much into the ruling. The
author’s view is that the ruling has no direct relevance for the interpretation of the cur-
rent regulations.
400
14.2.5. Case law: Amazon.com (2017)
The primary issue in the 2017 Tax Court ruling in Amazon.com v. CIR
was the determination of an arm’s length buy-in payment for an outbound
transfer of US-developed intangibles.1334
The factual pattern of the case was, in essence, as follows:1335 A US en-

tity in the Amazon group had developed, through its own R&D, a range
of valuable intangibles pertaining to the group’s global, web-based retail
business, including (i) software technology; (ii) marketing intangibles
(Amazon domain names, trademarks, trade names, etc.); and (iii) custom-
er information. In 2005, this US entity entered into a CSA with a Luxem-
bourg group entity and contributed its self-developed intangibles to the
CSA for the purpose of further developing the IP on a joint basis with
the Luxembourg entity. All R&D efforts going forward would, as before,
be carried out by the US entity, but the Luxembourg entity would now
co-fund the R&D costs. The US and Luxembourg group entities would
become co-owners of any new intangibles developed under the CSA and
thus be entitled to the residual profits from these in proportion to their
respective ownership shares. Prior to the formation of the CSA, the US
entity was entitled to all residual profits from the worldwide exploitation
of its IP. By entering into the CSA, it surrendered to the Luxembourg en-
tity its right to residual profits from exploitation of the current and future
IP in the European market. The core issue of the case was to determine
how the US entity should be compensated for its contribution of IP items
(i)-(ii) to the CSA.
As in Veritas, the case pertained to income periods governed by the previ-

ous-generation (1995) cost-sharing regulations, which did not contain the
income method as a specified method for pricing buy-ins. The available
methodologies for pricing contributions of pre-existing IP to a CSA were
those of the CUT method, CPM, PSM and unspecified methods.1336
1334. 148 TC No. 8. At issue was also the allocation under Treas. Regs. § 1.482-7(d)(2)
of certain cost items to the intangibles development cost pool (which the cost-sharing
payments were based on), in light of the invalidation of the regulation in the 2005 Tax
Court ruling in Altera v. CIR (145 TC 91). The author will limit his comments to the
buy-in issue. See also, e.g. Odintz et al. (2017), at sec. 2.2.5; and Avi-Yonah (2017) on
the Amazon.com ruling.
1335. The factual pattern of the case is comprehensive (the ruling spans 207 pages).
The author has somewhat simplified it for the purpose of this discussion.
1336. Treas. Regs. § 1.482-1(c).
401
The taxpayer valued the transfer of IP groups (i)-(iii) separately and then
summarized the individual values, thereby not triggering any residual in-
tangible value (goodwill, going concern value and workforce in place) that
could have arisen if the transfer had been valued as a whole. The taxpayer’s
argument was that the CSA contribution consisted of transfers of several
single assets (each with a 7-year useful life), not of an ongoing business.1337
The result was a value of USD 254.5 million.
Faithful to the CIP approach,1338 the IRS’s reassessment valued the transfer
of IP items (i)-(iii) as a “package” (i.e. a business transfer), based on the
rationale that the buy-in pricing should be consistent with the best realistic
alternatives available to the parties. The alternative for the transferring US
entity would be to continue its full IP ownership position, assume all costs
and risks of further IP development and license out the European rights to
the resulting IP to the Luxembourg entity, thereby retaining entitlement to
all residual profits from worldwide exploitation of future versions of its IP.
According to the IRS, the buy-in price should be set so that the US trans-
ferer would not be put worse off than if it instead had chosen this alterna-
tive. This is simply because the US entity would never have transferred the
IP to a third party for a lower price. The result was a value of USD 3.466
million.
The Court rejected the IRS’s approach for several reasons. First, even
though it did not seem to take issue with the fundamental assertion that
no third party would have entered into a CSA and transferred its IP if the
price was below what the third party could have achieved with a realistic
alternative, the Court simply observed that such “realistic alternatives pric-
ing” would make the option of entering into a CSA, which the regulations
endorsed, meaningless.1339 The reason is because the price of the buy-in
would then be set as if there were no CSA. The Court found that “realistic
alternatives pricing” would be equal to a restructuring of the actual trans-
action, which the regulations only allowed if it lacked economic substance
(which it did not).1340
This was, in the author’s view, an important misstep by the Court. IRC
section 482 requires that intra-group profit allocation must mirror arm’s
1337. The taxpayer valuation itself was DCF-based, with the future estimation of in-
come based on royalties from purportedly comparable third-party agreements under
the CUT method.
1338. LMSB-04-0907-62. See also the discussion in sec. 14.2.3.
1339. See the ruling in 148 TC No. 8, p. 83.
1340. Treas. Regs. § 1.482-1(f)(2)(ii)(A).
402
length, third-party dealings. The Court recognized this principle, but dis-
regarded it. Applying transfer pricing to a controlled transaction to adjust
the controlled price so that it corresponds to what a third party would have
charged is a world apart from restructuring the actual transaction. The
pricing assertion of the IRS was, in principle, nothing more than an ordi-
nary transfer pricing adjustment. Thus, the logic of the Tax Court seems to
imply that all transfer pricing reassessments (i.e. adjustments to the con-
trolled price using transfer pricing methodology) would be equal to a re-
structuring of the actual transaction. Such a view is, of course, nonsensical.
The IRS pricing did respect the actual controlled transaction. Even with
the IRS pricing, the Luxembourg entity would still be obliged to contribute
intangible development costs (IDCs), and for this, would become co-owner
of future IP developed under the CSA.
The point of the IRS was that the price required to be paid by the Luxem-
bourg entity for the future benefits it would receive under the CSA should
be equal to the present value of the benefits. It is true that with such a high
price, the CSA alternative would not have been as attractive for the Lux-
embourg entity. It would not stand to gain or lose anything; the price would
simply be equal to the worth of what it got in return. For the US entity,
however, this high price would be the only rationale for entering into the
CSA. It would simply not make any sense to enter into the CSA at a lower
price, as that would mean that the US entity would give up value for noth-
ing in return. No third party would do so.
It is the author’s view that the Tax Court failed to adhere to the fundamen-
tal core of section 482 when it dismissed the IRS’s “realistic alternatives”
pricing assertion. Even if the cost-sharing regulations applicable to the in-
come years at issue did not include the income method as a specified trans-
fer pricing method for buy-in pricing, the Court could have – as in Veritas
–found legal basis for the IRS’s assertion by way of applying a one-sided
method (e.g. the CPM with some adjustments) to the Luxembourg entity,
thereby leaving the residual profits to be allocated to the US entity through
the buy-in. Such one-sided methodology would likely not have been en-
tirely precise in this case, but would nevertheless have aligned the result
with sound transfer pricing logic and respected the fundamental purpose
of IRC section 482.
Further, the Court supported its rejection of the IRS’s pricing approach on
the basis that it did not provide the most reliable means of establishing an
arm’s length charge for the buy-in transaction. This was founded on two
403
main points.1341 First, the IRS’s approach mixed together IP that was com-
pensable under the applicable previous-generation cost-sharing regulations
(pre-existing IP contributed to the CSA) and IP that was not (IP subse-
quently developed under the CSA).1342
It is true that the IRS’s pricing approach treated the contribution of pre-
existing IP to the CSA as a business transfer that effectively transferred all
future income streams from the European part of the Amazon business to
the Luxembourg entity. The IRS simply valued the business of the Luxem-
bourg entity, and from this, deducted the present value of certain expen-
ditures (mainly IDCs). In effect, this treated the pre-existing intangibles
that were contributed to the CSA as if they had indeterminate useful lives,
thereby also encompassing subsequently developed intangibles.
There is no doubt that the question under the applicable cost-sharing regu-
lations was to determine the arm’s length buy-in value for the pre-existing
IP that was contributed by the US entity to the CSA.1343 However, that
does not imply that the pricing had to be artificially restricted to a sepa-
rate group-by-group valuation when the US IP items (i)-(iii), in fact, were
transferred together to the Luxembourg entity as a package of assets. The
effect of the CSA arrangement as a whole was to transfer the right to future
residual profits from European exploitation of the pre-existing and future
IP created – using the pre-existing IP as the research platform – under the
CSA to the Luxembourg entity. The software, marketing IP and customer
information were designed to be used as an integrated whole. The value of
each IP category would be highly contingent on being used together with
the other IP categories transferred. Also, proper compensation should in-
clude the research value of the pre-existing IP. That value could necessarily
only be reflected as a portion of the value of the successor IP. It seems that
the focus of the Court –as in Veritas – was merely on the exploitation value
of the pre-existing IP (which, of course, was limited) at the expense of a
more thorough and critical examination of the research value inherent to it.
The Court’s tunnel-vision-based application of the cost-sharing regulations
ignored that the controlled pricing should align with the pricing boundaries
dictated by the realistic alternatives of the controlled parties.1344
1341. Ruling in 148 TC No. 8, p. 82.

1342. See Treas. Regs. § 1.482-7(a)(2) and (g)(2).
1343. Id.
1344. The author uses the term “tunnel vision” here much in the same way as Brauner
(2016), at p. 99 applies the term “literal arm’s length” to indicate rather formalistic
transaction-by-transaction pricing assessments based on limited studies of comparabil-
ity.
404
Second, the Court rejected the IRS’s pricing assertion on the basis that it
also entailed compensation for residual intangible value that did not satisfy
the pre-2018 version of the US IP definition referred to in IRC section 482.
This objection from the Court does, in the author’s view, carry substantial
weight. US transfer pricing is indeed restricted by the IRC section 936
definition.1345 While the current-generation cost-sharing regulations bypass
this restriction through the platform contribution concept,1346 there was no
such bypassing in the previous-generation (1995) regulations applicable
to the income years at issue in this case. The IRS’s reassessment valua-
tion violated the pre-2018 version of the section 936 definition’s restriction.
The Tax Court, however, did not properly assess the classification of the
IRS’s value among specified IP (which falls under the pre-2018 section 936
definition) and residual intangible value (which falls outside the pre-2018
section 936 definition). It is therefore unclear as to what extent the IRS’s
valuation would have had to have been reduced in order to comply with the
pre-2018 section 936 definition.
Further, the Court rejected the underlying premise of the IRS’s valuation
– in essence, an implication of applying realistic alternatives pricing – that
the residual profits from the European exploitation of IP developed under
the CSA should be allocated to the group entity that provided the most
valuable intangible development contributions to the CSA. This was the US
entity, as it provided the R&D. The IRS allocated relatively generous com-
pensation to the co-funding contributions of the Luxembourg entity of 18%
of the IDCs.1347 The Court, however, found this unacceptable, as it deemed
the Luxembourg entity to own the European rights to the intangibles, and
therefore held that it should be entitled to the entire residual profits from
these intangibles. The Court’s justification on this point was that the appli-
cable cost-sharing regulations did not distinguish between actual technol-
ogy development and CSA contributions in the form of cash. That observa-
tion is, in isolation, correct. However, in order for cost-sharing to work as
intended, there must be an arm’s length buy-in payment when a party to
the agreement contributes pre-existing IP. The determination of that buy-in
must take into account the realistic alternatives of the controlled parties,
pursuant to which it undoubtedly will be relevant to assess whether a third
party that performs promising and valuable R&D would be willing to give
up future fruits from that R&D simply for being reimbursed for its R&D
costs, which, of course, would be unlikely.
1345. See the analysis of the US IP definition in sec. 3.2.

1346. See the discussion of platform contributions in sec. 14.2.7.5.
1347. Ruling in 148 TC No. 8, p. 86.
405
The Court’s rejection of the CIP-based aggregated pricing approach alone

in reality determined, in broad strokes, the outcome of the case, as the im-
plication of this stance was that IP categories (i)-(iii) then had to be valued
separately. That meant that no residual intangible value due to the integra-
tion between the transferred IP (goodwill, going concern value, etc.) could
be allocated to the US entity through the buy-in. The author will comment
on some of the main points pertaining to the Court’s buy-in valuation of the
separate IP categories (i)-(iii).
IP category (i): Website technology (software)
As the IRS’s CIP-pricing approach had been rejected, the parties agreed
that the CUT method should be used to price the buy-in on the basis of roy-
alty rates extracted from transactions between Amazon and independent
retailers. There was disagreement on the applicable royalty rate, the useful
life of the transferred IP, the revenue base on which to apply the royalty
rate, as well as the appropriate discount rate.
The most interesting aspect of the ruling on this point is the Court’s as-
sessment of the useful life of the transferred IP.1348 Contrary to the IRS’s
“infinite life” assertion, the Court found that the transferred IP on average
had a useful life of only 7 years. Further, the Court found that the concepts
and code of the existing software became obsolete relatively quickly and
that there was no indication that new software developed under the CSA
relied on old code or prior ideas to any significant extent. The Court did,
however, recognize that the transferred IP had residual value that was not
captured by the 7-year premise, and therefore added a “tail period” of
an additional 3.5 years with reduced royalty payments to compensate for
this residual value. The Court seemed to have been in some doubt with
respect to whether this tail actually captured the residual value appropri-
ately, but ultimately found it to be sufficient, as the IRS had not offered
any alternative means of calculating an appropriate tail period. The author
is sceptical as to whether the Court, through this approach, managed to
appropriately allocate income to the US entity in order to compensate it
for the research value inherent to the transferred software technology. It
also seems fairly obvious that the Luxembourg entity, through the CSA,
acquired access to the know-how, experience and workforce in place of
the US entity’s R&D team and that the CUT valuation did not manage to
reflect this value.
1348. Ruling in 148 TC No. 8, p. 103.
406
IP category (ii): Marketing IP (Amazon name, trademarks, trade names,

domain names and trade dress, including for the significant French,
German and UK markets)
The lion’s share of the IRS’s reassessment value was attributable to these
marketing intangibles.1349 The parties also agreed here that the pricing
should be CUT-based,1350 but disagreed on the applicable royalty rate, the
useful lives of the transferred IP (the revenue base of which the royalty
rate should be applied to), as well as the appropriate discount rate. The two
most significant issues were (i) the determination of the useful lives; and
(ii) whether the buy-in value should be reduced because the Luxembourg
entity already owned portions of the marketing IP.
First, with respect to useful life, the Court found that the marketing IP
items (which had strong positions in the relevant markets at the time of
transfer) had longer lives than the software, although not indefinite. The
Luxembourg entity was responsible for marketing development efforts in
Europe. As the value of the marketing IP would be strongly correlated
with the value of the other Amazon IP – which the Luxembourg entity co-
owned through the CSA – and as the Luxembourg entity, through its local
marketing efforts, would build the future value of the European marketing
IP, the Court found that no third party under such circumstances would be
willing to pay royalties for the use of the marketing IP forever. The Court
concluded that the marketing IP had a useful life of 20 years, a period long
enough to capture most of the present value asserted by the IRS on this
point.
Second, with respect to whether the buy-in should be reduced to take into
account prior ownership, the claim from the taxpayer was that the Lux-
embourg entity already owned part of the marketing intangibles (most
prominently, the editorial content of the Amazon websites in France, Ger-
many and the United Kingdom, as well as Amazon trademarks and domain
1349. The taxpayer valuation, using royalty rates ranging from 0.125% to 1%, a useful
life between 10-15 years and a discount rate of 18%, was in the interval of USD 251-
312 million, minus a reduction of 50% in order to account for prior ownership by the
Luxembourg entity of certain marketing IP, resulting in a value between USD 114-165
million. The IRS’s valuation, using a royalty rate of 2%, a useful life between 10-15
years and a discount rate of 18%, concluded with a value for the marketing intangibles
of USD 3.13 billion, not allowing a reduction for any previously owned marketing in-
tangibles of the Luxembourg entity.
1350. With royalty rates extracted from transactions between independent parties
from the ktMINE database, available at https://www.ktmine.com/ (accessed 10 Sept.
2015).
407
names for these jurisdictions) at the time at which the CSA was entered
into, and that the valuation amount therefore should be reduced by 50%.1351
The IRS opposed this, based on the assertion that the US entity should be
deemed the “true equitable owner” of the marketing IP, as it had made
all investments and taken all risks associated with building the marketing
value of the IP in Europe.
The court observed that the applicable section 482 regulations required that
intangible ownership should be determined with reference to legal owner-
ship, provided that this would not be contrary to the economic substance
of the underlying arrangement.1352 European group entities were registered
locally as owners of relevant marketing intangibles,1353 and the Court found
that there were sound business reasons for such practices (in some instanc-
es, required by local law). The European group entities had also used the
relevant marketing intangibles in their local business activities. The Court
therefore found the European marketing IP ownership to have economic
substance, and was thus unwilling to set aside legal ownership.
With respect to the IRS’s assertion that the US entity bore the costs and
risks of the European marketing efforts, the Court simply noted that be-
cause the European group entities only received sales commissions, they
incurred business risks (i.e. no guarantee that they would earn a profit).
The author finds the Court’s assessment on this point to be superficial. The
question was not whether the European legal owner entity incurred normal
business risks, but whether it took the risk of developing the value of the
marketing IP locally i.e. whether it incurred the marketing costs. If it in
fact was the US entity that bore the European marketing costs, that should
clearly indicate that the economic substance of the marketing arrangement
between the US and Luxembourg group entities indeed was that the US
entity was the owner and that the Luxembourg entity functioned only as a
title-holding agent. If so, the US entity should have been entitled to the re-
sidual profits from the European exploitation of the marketing intangibles.
There would then be no basis upon which to reduce the buy-in price for pre-
viously owned IP. The Court should have assessed this more thoroughly.
The result on this issue was a reduction of the buy-in price with 25% for
1351. Ruling in 148 TC No. 8, p. 147.

1352. Treas. Regs. § 1.482-4(f)(3)(ii)(A).
1353. There was a reorganization of the European part of the group around the time
at which the CSA was formed, in which marketing IP registered to other European
entities in the Amazon group was transferred to the Luxembourg entity. Thus, for the
purpose of assessing European ownership of the marketing IP, the court had to take into
account prior practices between the US entity and the other European group entities.
408
prior ownership, still leaving a considerable 75% for the buy-in value al-
locable to the US entity.
IP category (iii): Customer information
The US entity contributed information on European customers to the CSA,

thereby, in effect, referring its existing European customers to the Luxem-
bourg entity. Also here, it was agreed that the CUT method should be ap-
plied, on the basis of internal CUTs on third-party referral fees to Amazon.
The IRS claimed that this did not capture all residual value inherent to the
customer information. The Court agreed, in principle, but found that the
residual value was short-lived and had been sufficiently taken into account
in the valuation of the marketing IP.
14.2.6. Concluding comments on Veritas and Amazon.com
Through the decades, high-value, US-developed intangibles have been

transferred outbound through CSAs to group entities resident in ju-
risdictions with favourable taxation of IP income without triggering
taxation of the “full” IP value on the way out, made possible through
low buy-in valuations. Eventually, the IRS developed the “realistic
alternatives”-based CIP approach to buy-in valuation to counter these
practices. The CIP doctrine was not incorporated into the cost-sharing
regulations until 2009 (through the income method; see the discussion
in section 14.2.8.3.), but the IRS nevertheless started applying the doc-
trine to prior income years even if it had not yet been anchored in the
regulations as a specified pricing method. Therein lies the heart of the
problem in both Veritas and Amazon.com: the Court did not find room
in the previous-generation (1995) cost-sharing regulations to apply the
CIP approach.
While it is true that there was no specified income method that the Court
could have used,1354 it would have been, in the author’s view, natural to
apply one of the available one-sided methods to remunerate the foreign
cash-contributing entity (e.g. the CPM, with some adjustments). While the
tunnel-vision, item-by-item valuations under the CUT method in both Veri-
tas and Amazon.com may come across as straightforward applications of
1354. Avi-Yonah (2017), at p. 4 suspects that the result would have been the same in
Amazon.com had the case been tried under the current cost-sharing regulations (which
contain the specified income method). The author does not share this view.
409
buy-in methodology set out in the previous-generation cost-sharing regula-

tions, there should be no doubt that this CUT approach failed to capture
significant intangible value transferred out of the United States.
There should also be no doubt that the CUT method results do not respect
the acceptable price boundaries dictated by the realistic alternatives prin-
ciple (which had a clear legal basis in the 1994 regulations).1355 In both
Veritas and Amazon.com, the US entity, through its cost-sharing contri-
bution, in reality transferred its right to future residual profits to a group
financing entity that contributed nothing apart from cash to the ongoing
R&D activities carried out (by the US entity) under the CSA. The US entity
would simply be better off by continuing its ownership position and then
licensing out the results to other group entities. A true arm’s length buy-in
price should, of course, reflect this fact. No third party would be willing to
settle for less. It is not as the Tax Court claimed in Amazon.com that “re-
alistic alternatives” pricing would be equal to a restructuring of the actual
controlled transaction. The CSA would stand as such, and the only con-
sequence of applying the IRS’s pricing approach would be the allocation
of more income to the US entity. Thus, the price of the CSA contribution
would go up, but everything else would remain the same. This is pricing,
not restructuring.
It should also be noted that the IRS was too aggressive in both cases. The
valuation in Veritas contained errors. The Amazon.com valuation clearly
contained value that was attributable to functions and assets already pre-
sent in the Luxembourg entity before the CSA was formed, and thus not
caused by the IP transferred from the US entity. This value should therefore
have been kept outside of the buy-in price. One can hope that the IRS will
not repeat such mistakes in its future valuations under the income method.
With respect to its unwillingness to accept the CIP approach in Amazon.

com, the author finds it peculiar that the Court did not reflect more thor-
oughly on the appropriateness of remunerating the Luxembourg entity
as a routine value chain contributor under a one-sided methodology. In
this also lies the observation that the taxpayer structure was tax-driven.
The Luxembourg entity, while far from being a simple cash box, seems
to have functioned – among other things – as an IP holding entity in
Europe, and its royalty income was likely not taxable by much in Lux-
embourg.
1355. See supra n. 1329.
410
In the end, the result of Amazon.com was not entirely negative for the IRS.
Yes, the battle over the CIP approach was lost (again), but the Tax Court
did recognize that there was considerable value inherent to the transferred
marketing IP (which accounted for the lion’s share of the reassessment). On
this point, the IRS achieved a relatively significant victory, as the valuation
parameters applied by the Court (the useful life of 20 years, the royalty
rate of 1% and only a 25% reduction due to prior ownership) generally
yielded a higher valuation than asserted by the taxpayer. Further, while
the CUT-method valuation assessments of the Court may possibly provide
some guidance for future cases, the effect of this will be negligible, due to
the comprehensively high threshold set in the current-generation cost-shar-
ing regulations for applying this methodology.1356 The current cost-sharing
regulations, with the specified income method and bypass of the pre-2018
version of the IRC section 936 IP definition allowing for the capture of re-
sidual intangible value in the buy-in amount, will ensure that the IRS does
not face the same struggles going forward as it did in Veritas and Amazon.
com.
Why, then, would the IRS go through the trouble of relitigating Veritas as
it did with Amazon.com? The author assumes that the litigation risk must
have been apparent to the IRS. Even if the Tax Court would be willing to
accept the CIP approach as a pricing methodology as such, the IRS would
still be facing the problem that the pre-2018 version of the section 936 IP
definition rejects the taxation of outbound residual intangible value. The
portion of the reassessment value in Amazon.com attributable to work-
force in place and other residual value could therefore, in any case, not
stand. The significant income at stake undoubtedly incentivized the IRS to
try. Also, the IP structure established under the Amazon CSA will deprive
the United States of (likely) significant residual profits from intangibles
developed under the agreement going forward, perhaps over decades, even
though the intangible value will be created in the United States through
ongoing R&D there (and the US entity’s R&D costs will even be tax-de-
ductible in the United States). That has likely been a difficult pill to swal-
low for the IRS.1357
1356. See sec. 14.2.8.2. on the CUT method for buy-in pricing.
1357. The IRS appealed the Amazon.com ruling to the US Court of Appeals for the
Ninth Circuit on 29 September 2017. At the time of writing this book, the content of the
IRS’s legal arguments in support of the appeal is not known.
411
14.2.7. Key concepts under the current cost-sharing

regulations
The cost-sharing regulations make use of some key concepts with rather
particular terminology. In order to analyse how operating profits are allo-
cated pursuant to the current buy-in pricing methods, it is necessary to first
present these concepts.1358 The author does so in sections 14.2.7.2.-14.2.7.5.
before discussing the buy-in pricing methods in section 14.2.8.
14.2.7.2. Cost-sharing transactions
All controlled participants in a CSA must commit to and engage in cost-

sharing transactions, in which they make payments to each other so that
in each taxable year, each controlled participant’s share of the IDCs is in
proportion to its respective reasonable anticipated benefit (RAB) share.1359
Intangible development costs include all costs, in cash or in kind (including
stock-based compensation),1360 in the ordinary course of business incurred
after the formation of a CSA, which are directly identified with or are rea-
sonably allocable to the intangible development activity.1361
14.2.7.3. RAB share
The concept of RAB shares is important, as they determine how intangi-

ble development costs and ownership interests in the intangibles developed
1358. For a walkthrough of the main concepts, see also Brauner (2010), at pp. 3-5. For
a discussion of the concepts under the 1995-generation cost-sharing regulations (which
are still partly relevant), see, e.g. Benshalom (2007), at p. 653.
1359. Treas. Regs. § 1.482-7(b)(1)(i). A reasonably anticipated cost-shared intangible
is any intangible (within the meaning of § 1.482-4(b)) that the controlled participants
intend to develop under the CSA; see Treas. Regs. § 1.482-7(d)(1)(ii).
1360. The cost-sharing regulations include elaborate provisions pertaining to stock-
based compensation in Treas. Regs. § 1.482-7(d)(3) that were implemented in the wake
of the Xilinx case. See also the 2015 Tax Court ruling in Altera v. CIR (145 TC 91),
invalidating the requirement to include stock-based compensation in the intangibles
development costs pool. On cost allocation for CSAs, see, in particular, Dix (2010).
1361. Treas. Regs. § 1.482-7(d)(1)(iii). This includes costs incurred in attempting to
develop cost-shared intangibles, regardless of whether the costs ultimately lead to suc-
cessful development of the anticipated intangibles, other intangibles developed unex-
pectedly or no intangibles.
412
under the CSA (cost-shared intangibles) are allocated among the CSA par-
ticipants.
RABs are measured either on a direct or indirect basis.1362 A direct basis

measurement is the estimated benefits generated through the exploitation
of the cost-shared intangibles, typically based on additional revenue plus
cost savings less any additional costs incurred. An indirect basis measure-
ment is made by reference to certain parameters that reasonably may be
assumed to relate to future benefits (e.g. units used, produced or sold, sales
or operating profits). A controlled participant’s share of RABs is equal to
its own RABs divided by the sum of the RABs.1363
14.2.7.4. Non-overlapping (ownership) interests in intangibles

developed under the CSA
Each controlled participant must receive a non-overlapping interest in the

cost-shared intangibles without further obligation to compensate another
controlled participant for such interest (divisional interests).1364 Thus, for
transfer pricing purposes, each participant will be considered the owner
of its respective interest in the developed intangible, and therefore, as the
point of departure, entitled to the residual profit from it.
The interests can be divided among the CSA participants (i) on a territorial
basis (the world market must then be divided into two or more non-overlap-
ping geographic territories);1365 (ii) based on all uses to which cost-shared
intangibles are to be put (all anticipated uses must be identified, each con-
trolled participant must be assigned at least one such use and, in the aggre-
gate, all of the participants must be assigned all of the anticipated uses);1366
or (iii) on some other basis (e.g. the site of manufacturing), provided that a
set of criteria is met.1367
1362. Treas. Regs. § 1.482-7(e)(2)(ii), Example 1. If there is a significant divergence

between projected benefit shares and benefit shares adjusted to take into account any
actual benefits to date (adjusted benefit shares), the IRS may use adjusted benefit shares
as the most reliable measure of RAB shares and adjust the intangibles development cost
shares accordingly; see Treas. Regs. § 1.482-7(i)(2)(ii). See also Treas. Regs. § 1.482-
7(i)(2)(ii)(D), Examples 1-4.
1363. Treas. Regs. § 1.482-7(e)(1)(i).
1364. Treas. Regs. § 1.482-7(b)(1)(iii).
1365. Treas. Regs. § 1.482-7(b)(4)(ii).
1366. Treas. Regs. § 1.482-7(b)(4)(iii).
1367. Treas. Regs. § 1.482-7(b)(4)(iv).
413
14.2.7.5. Platform contributions
All controlled participants must commit to and engage in platform con-

tribution transactions (PCTs), to the extent that there are such contribu-
tions.1368 In a platform contribution transaction, each controlled participant
is obligated to make arm’s length payments (buy-in) to each controlled
participant that provides a platform contribution. The entity that renders
the platform contribution (typically, a unique, pre-existing intangible) is
called the PCT payee, while the entity that must pay the platform contri-
bution transaction payment (buy-in amount) is called the PCT payer. The
author will, for the sake of simplicity, try to avoid these rather cumbersome
terms when possible.
The regulations define a platform contribution as:

[…] any resource, capability, or right that a controlled participant has devel-
oped, maintained, or acquired externally to the intangible development ac-
tivity (whether prior to or during the course of the CSA) that is reasonably
anticipated to contribute to developing cost shared intangibles.1369
The regulations have a relatively relaxed view as to which rights, capa-

bilities and resources may constitute a platform contribution, both with
respect to the degree of causality required between the contribution and the
intangible development, as well as with respect to the type of contribution.
First, with respect to causality, it is not required that the contribution itself
is subject to further development, only that it may aid in developing cost-
shared intangibles. For instance, if the aim of a CSA is to develop and test
new pharmaceutical compounds and a patented research tool to develop
the compounds is contributed to the CSA, the research tool will consti-
tute a platform contribution because it is anticipated to contribute to the
research activity covered by the CSA. This is true even if the CSA activity
is not anticipated to result in the further development of the research tool
itself or in new patents based on the research tool.1370
Second, with respect to the type of contribution, development inputs that

traditionally have not been considered to qualify as intangibles under the
IRC section 936(h)(3)(B) definition (residual intangibles, such as the value
1368. Treas. Regs. § 1.482-7(b)(1)(ii).

1369. Treas. Regs. § 1.482-7(c).
1370. Treas. Regs. § 1.482-7(c)(4)(ii), Example 3. See also Treas. Regs. § 1.482-7(b)(1)
(iv), Example 2.
414
of a workforce or research team in place) may also be considered platform

contributions.1371 For instance, if a US parent commits a research team with
a successful track record to a CSA for the development of a vaccine, the ex-
pertise and existing integration of the research team is regarded as a unique
resource or capability that is reasonably anticipated to contribute to the
development of the new vaccine. This represents a regulatory bypass of the
pricing restrictions inherent to the section 936 IP definition1372 and is nec-
essary for the IRS in order to capture residual intangible value in the buy-in
price. Commentators have – and rightly so – criticized the platform contri-
bution concept for being in conflict with the section 936 IP definition.1373
Any right to exploit an existing resource, capability or right without further

development, such as the right to make, replicate, license or sell existing
products (make-sell rights), does not constitute a platform contribution,
unless exploitation without further development is reasonably anticipated
to contribute to developing a cost-shared intangible.1374 For instance, if a
US parent contributes the rights to further develop version 1 of a software
program to a CSA with a foreign subsidiary and, at the same time, licenses
make-sell rights to the software to the subsidiary, only the former rights
will be considered platform contributions.1375 The ordinary transfer pricing
rules must be applied to determine the arm’s length consideration in con-
nection with the make-sell licence arrangement.1376
A platform contribution is presumed to be exclusive, in the sense that the

resource, capability or right is not reasonably anticipated to be commit-
ted to any business activities other than the CSA activity.1377 This is an
1371. Treas. Regs. § 1.482-7(c)(5), Example 2. See also the analysis of the US IP defi-
nition in sec. 3.2.
1372. See the discussion of the US IP definition in sec. 3.2.
1373. Id.
1374. Treas. Regs. § 1.482-7(c)(4).
1375. Another thing entirely is that the platform contribution and the make-sell rights
may be valued in the aggregate. See the discussion of the income method in sec.
14.2.8.3. See also the comments on the US best-method rule in sec. 6.4. and on the ag-
gregation of controlled transactions in sec. 6.7.2.
1376. Treas. Regs. § 1.482-7(c)(4)(ii), Example 1.
1377. Treas. Regs. § 1.482-7(c)(2). The 2005 proposed CSA regulations (70 FR
51116-01) introduced the concept of a reference transaction designed to benchmark
the buy-in profit allocation, which was criticized for being overly broad. The concept
was scrapped in the 2009 temporary regulations (74 FR 340-01) and replaced with
the presumption that a party with a pre-existing intangible would provide resources,
capabilities or rights to a CSA on an exclusive basis, i.e. the external contributions were
reasonably expected to contribute only to the CSA and not to other business activities.
This presumption had pricing implications, as the value of an exclusive contribution
generally would be higher than the value of a non-exclusive contribution.
415
important assumption, as the pricing of the platform contribution will then

encompass the entire value from the relevant use of the contribution. Nev-
ertheless, the taxpayer has the possibility of rebutting the exclusivity pre-
sumption.1378 For instance, if the platform contribution is a research tool,
the controlled participants could rebut the presumption by establishing
that, as of the date of the PCT, the tool is reasonably anticipated to contrib-
ute to the CSA activity and also be licensed to an uncontrolled taxpayer.1379
14.2.8. The buy-in pricing methods
The final 2011 cost-sharing regulations contain a range of buy-in pricing

methods.1380 The methods are meant to be – and are – comprehensively
restrictive.1381 As the IRS answer to the profit shifting practices of multi-
nationals, pursuant to which US-developed intangibles have been migrated
for artificially low buy-in amounts, the pricing methods now ensure that
residual profits are allocated in a manner that reflects where the underlying
intangible value creation took place.
Two fundamental assumptions underlie these methods. First, the pricing

must be aligned with the upfront contractual terms and risk allocation,
with respect to anticipated benefits and obligations, over the entire antici-
pated term of the CSA. This is the so-called “investor model principle”,
pursuant to which the profits allocable to each participant in the CSA must
be appropriate for the risks connected to the CSA activity carried out by
the participant. The regulations illustrate the principle through an example
that rejects taxpayers’ CSA practices in which the buy-in is based on an
1378. Treas. Regs. § 1.482-7(c)(2)(ii).

1379. The value of the contributed resource from uses other than the CSA activity
should not be included in the platform contribution transaction (PCT) payment; see
Treas. Regs. § 1.482-7(c)(2)(iii). If there are such other applications of the resource, the
value of the PCT payment must be prorated in proportion to the relative economic value
reasonably anticipated from the platform contribution to the CSA activity, compared to
the value reasonably anticipated from the platform contribution to other business activi-
ties; see Treas. Regs. § 1.482-7(c)(2)(iii)(B).
1380. Treas. Regs. § 1.482-7(g)(1). See also Brauner (2010), at p. 8 for a discussion of
the CSA transfer pricing methods. The discussion in sec. 14.2.8. pertains to the pricing
methodology. It should be noted that the regulations also contain (flexible) rules on the
form of payment for buy-ins; see Treas. Regs. § 1.482-7(h).
1381. The transfer pricing methodology set out in the US cost-sharing regulations is
generally perceived as highly technical and complicated, which was also the case under
the previous-generation (1995) regulations. See Benshalom (2007), at p. 666.
416
unspecified version of the RPSM and the royalties paid to the US parent
are rapidly ramped down after the formation of the CSA.1382 The implica-
tions of the investor model principle, as an independent legal norm, seem
unclear.1383
Second, and most importantly, the buy-in pricing must ensure consistency
with the realistic alternatives of the controlled party that makes the plat-
form contribution (the PCT payee).1384 This condition is not met when the
anticipated present value from participation in the CSA is less than that
which could be achieved through an alternative arrangement realistically
available.1385 The income method (discussed in section 14.2.8.3.) opera-
tionalizes the realistic alternatives principle in the context of a buy-in valu-
ation. The IRS’s position taken in the 2009 temporary regulations is that
the principle supplements the best-method rule to help determine which
method will provide the most reliable measure of an arm’s length result for
the buy-in calculation.1386 The author finds it doubtful whether the principle
in and of itself can play any significant role for pricing purposes alongside
the specified pricing methods of the current CSA regulations.1387
In sections 14.2.8.2.-14.2.8.7., the author will tie some comments to the

available buy-in pricing methodology.
14.2.8.2. The CUT method
The CUT method may be applied to evaluate whether the buy-in amount
is at arm’s length by reference to the amount charged in a CUT.1388 Appli-

1383. E.g. if (even though unlikely) a genuine CUT is found upon which the buy-in
valuation is based and this results in a generous allocation of income to a foreign cash-
box entity, could it be argued on the basis of the investor model that the foreign entity
has been allocated too high of a return? Such an argument should, in the author’s view,
be rejected.
1384. See also Brauner (2010), at p. 8 on this point. See the 2011 budget proposal (JCS-
3-11 No. 11) for a convincing discussion of the realistic alternatives principle.
1385. On the use of projections in the CSA regulations, see Brauner (2010), at p. 9.
1386. See the preamble to the 2009 temporary cost-sharing regulations (74 FR 340-
01). On the 2009 regulations, see Kochman et al. (2009); and Dickenson (2010).
1387. See also the comments on this issue in the 2005 proposed CSA regulations (70
FR 51116-01).
1388. Treas. Regs. § 1.482-7(g)(3). In this context, the CUT method refers both to
§ 1.482-4(c) and § 1.482-9(c). That value must then be multiplied by each PCT payer’s
respective RAB share in order to determine the arm’s length PCT payment due from
each PCT payer. The reliability of a CUT that yields a value for the platform contribu-
tion only in the PCT payer’s division will be reduced to the extent that the value is not
417
cation of this method is contingent on the identification of a CUT, which

must be similar to the controlled CSA with respect to contractual terms,
the degree to which allocation of risks is proportional to the RABs from
exploitation of the R&D results and the profit potential of the R&D efforts.
Given the strict comparability criteria, this method should rarely be rel-
evant. Further, the profit allocation that results from applying this method
must align with the results that would follow from a valuation based on the
estimated future operating profits from the intangible developed under the
CSA. This restriction will effectively ensure that CUT-based buy-in pric-
ing is consistent with the results of the income method and the PSM, which
both rely on DCF analysis.
14.2.8.3. The income method
The income method represents a further development and regulatory im-

plementation of the unspecified 2007 aggregated CIP approach to buy-
in valuation as a specified transfer pricing method. The methodology is
founded on and operationalizes the realistic-alternatives-available princi-
ple in the context of CSA buy-in valuations. The core idea is that the buy-in
cannot result in a lower amount of operating profits to the US entity than
what would have been the result for this entity if it instead had chosen to
not enter into the CSA and gone for its best realistic alternative.1389 The in-
come method is potent, in particular, because its application of the realistic
alternatives principle is combined with the IRC section 936 IP definition
bypass, which is the platform contribution concept.1390 This will capture, in
the buy-in amount, also residual intangible value (goodwill, workforce in
place and going concern value), even if such value does not qualify under
the section 936 definition.1391 The income method will be the go-to method
in typical CSA scenarios,1392 in which a US group entity with self-devel-
oped unique and valuable intangibles and an R&D team in place enters
into a CSA with a foreign cash-box group entity that will only contribute
consistent with the total worldwide value of the platform contribution multiplied by the
PCT payer’s RAB share. For a discussion of the CUT method in the context of CSAs,
see, e.g. Wittendorff (2010a), at p. 576.
1389. It must clearly be assumed that third parties would not have been willing to enter
into a CSA for less. See also Brauner (2010), at p. 17.
1390. See sec. 14.2.7.5. for a discussion on platform contributions.
1391. See the discussion of the US IP definition in sec. 3.2.
1392. Treas. Regs. § 1.482-7(g)(4). For a general discussion of the income method, see,
e.g. Wittendorff (2010a), at p. 577.
418
funding to the joint development.1393 The method is designed to allocate all

residual profits to the US entity in such cases.
The underlying methodology and profit allocation pattern of the income

method is akin to that of the CPM. It requires the tested party (the PCT
payer) to pay a buy-in amount that makes the present value of its participa-
tion in the CSA equal to the present value of its best realistic alternative,1394
which is to license make-sell rights to the fully developed intangibles,
thereby not committing to share any of the R&D costs and risks. Con-
versely, the best realistic alternative of the PCT payee (the US entity that
renders platform contributions to the CSA) would be to undertake the com-
mitment to bear the entire R&D risk and license the resulting intangibles to
an uncontrolled licensee. As for the CPM, the scope of the income method
is confined to scenarios in which the tested party only contributes routine
development inputs. Otherwise, if the tested party also contributes non-
routine development inputs (pre-existing intangibles or R&D), the RPSM
will be the relevant method.1395
The buy-in amount determined by the income method is the difference

between the net present value for the tested party of the CSA and licensing
alternatives. The cost-sharing alternative is the actual CSA, pursuant to
which the tested party must bear cost contributions throughout the duration
of the CSA and make the buy-in payment.1396 The value of this alternative
is the present value of the divisional profit minus operating cost contribu-
tions minus cost contributions and PCT payments over the duration of the
CSA.1397 The licensing alternative is the same, apart from the cost contribu-
tions and buy-in payment being replaced with an obligation to pay an arm’s
length royalty rate.1398
1393. The income method does not allow the PCT payer a return on its contributions
to the intangibles development funding. The author finds this aggressive allocation of
income in favour of the United States to be in conflict with the arm’s length principle;
see the comments in secs. 14.2.8.3., 22.4.11. and 26.3.
1394. Treas. Regs. § 1.482-7(g)(4). The present values of the cost-sharing and licensing
alternatives should, in general, be determined by applying post-tax discount rates to
post-tax income; see Treas. Regs. § 1.482-7(g)(4)(i)(G)(1).
1395. Treas. Regs. § 1.482-7(g)(4)(1)(D). The RPSM is discussed in sec. 14.2.8.6. A
profit split will also be used if the foreign subsidiary has mature operations and pos-
sesses significant operating intangibles (e.g. local marketing IP) that will contribute to
the profit from exploiting the cost-shared intangibles or from existing make-sell rights.
1396. Treas. Regs. § 1.482-7(g)(4)(i)(B).
1397. Treas. Regs. § 1.482-7(g)(4)(ii).
1398. Treas. Regs. § 1.482-7(g)(4)(1)(C).
419
The arm’s length royalty rate charged to the tested party in the licensing
alternative may be determined on the basis of rates extracted from CUTs
under the CUT method1399 or, more practically, indirectly under the CPM
through the allocation of a normal market return for the tested party’s rou-
tine contributions (the residual profits will be equal to the royalty pay-
ment). In the latter alternative, the tested party is completely stripped of
residual profits.
The present value of the CSA and licensing alternatives must be deter-
mined using an appropriate discount rate.1400 When the market-correlated
risks in the alternatives are not materially different, it may be possible to
use the same discount rate.1401 Normally, however, the discount rates should
be different. The tested party will incur less risk as a licensee, as it will
not contribute any cost contributions, and therefore will not share the R&D
risks, resulting in a lower discount rate in the licensing alternative than in
the CSA alternative.1402 The determination of discount rates has proven to
be a source of contention and was at issue in both Veritas and Amazon.
com.1403
A closely related and problematic issue is the determination of the useful

life of the platform intangible (the horizon over which the buy-in valuation
shall be performed). The 2007 aggregate CIP approach for buy-in valu-
ation effectively treated platform intangibles as having perpetual useful
lives, as argued in Veritas and Amazon.com. When the income method
was introduced in the 2005 proposed regulations,1404 it was criticized for
presuming that the platform contribution would have a perpetual useful
life. Indeed, the proposed regulations did imply that for pricing purposes,
the pre-existing intangible contribution would be valued using techniques
1399. Treas. Regs. § 1.482-7(g)(4)(ii); see also § 1.482-4(c)(1) and (2). If the licens-
ing alternative is evaluated using the CUT method, the additional comparability and
reliability considerations stated in § 1.482-4(c)(2) pertaining to the CUT method may
apply; see Treas. Regs. § 1.482-7(g)(4)(vi)(C) and § 1.482-4(c)(2).
1400. See Treas. Regs. § 1.482-7(g)(2)(v) and § 1.482-7(g)(4)(vi)(F).
1401. Treas. Regs. § 1.482-7(g)(4)(i)(F).
1402. Conversely, self-development will be riskier for the licenser than entering into a
CSA (as the latter alternative would partly relieve the licenser of its funding burden),
necessitating a higher discount rate.
1403. The IRS’s assertion is that taxpayers inappropriately seek to reduce the PCT
payment through the use of a discount rate for the CSA alternative that is significantly
higher than the discount rate used for the licensing alternative, while taxpayers argue
that the fixed cost profile of R&D development costs under the CSA alternative (as
opposed to the variable royalty costs under the licensing alternative) warrants a signifi-
cant discrepancy between the discount rates; see Finley (2016).
1404. 70 FR 51116-01.
420
that presumed the contribution to be perpetual. The idea was that the main
bulk of value came from core technology, which would be the basis, or the
platform, for future improvements, the value of which was derived from
and contingent on the core technology. The 2009 temporary regulations
modified this view by recognizing that, depending on the facts and circum-
stances, the value of the technology invested by the PCT payee in the plat-
form contribution could have a “finite, not a perpetual, life”.1405 The IRS
now seems to have abandoned its prior infinite-life assertion in favour of
relatively long useful lives, typically ranging from 25-40 years.1406 Taxpay-
ers typically argue for useful lives based on short-term product life cycles.
Different approaches to discount rates and useful lives will likely yield
significant discrepancies between IRS and taxpayer buy-in valuations. This
may trigger a potential for future controversies over the application of the
income method. This will, however, only be relevant if taxpayers going
forward at all choose to structure their outbound transfers of US-developed
IP through CSAs, in light of the powerful buy-in pricing methodology now
available to the IRS through the income method.
The author will go through two of the income method examples contained
in the regulations, as they illustrate key aspects of the methodology. The
first example pertains to a US parent that has developed version 1 of a
new software application and enters into a CSA with a foreign cash-box
subsidiary to develop future versions.1407 The agreement assigns the US
exploitation rights to future versions of the software to the parent, while
the subsidiary is assigned rights for the rest of the world. Version 1 and the
parent’s R&D team are reasonably anticipated to contribute to the develop-
ment of future versions. Both inputs are deemed to be platform contribu-
tions that require compensation.1408
The example determines and compares the net present values for the sub-
sidiary from participating in the CSA and from its best alternative of en-
tering into a licence agreement with the parent when the parent alone has
fully developed future versions of the software. The projected sales and
operating costs are the same under both alternatives for the 15-year lifes-
pan of the software.
1405. 74 FR 340-01.
1406. See Finley (2016).
1408. The subsidiary will not perform any R&D and it does not furnish any platform
contributions to the cost-sharing arrangement, nor does it control any operating intan-
gibles at the inception of the CSA.
421
The difference lies in the cost contribution payments under the CSA alter-
native and the royalty payments under the licensing alternative, as well as
in the discount rates applied to calculate the net present values. The arm’s
length royalty payments are determined to be 35% of the subsidiary’s sales.
The CSA is deemed to be more risky than the licence agreement, thus jus-
tifying a higher discount rate. The discount rate is set to 15% in the CSA
alternative and 13% in the licensing alternative.
The net present value for the subsidiary for participating in the CSA is
889, and 425 for the licensing alternative. Thus, for the subsidiary to be
allocated the same amount of profits under the CSA alternative as in the
licensing alternative, it must make a buy-in payment to the parent equal to
the difference between 889 and 425, which is 464. This ensures that the
subsidiary is allocated a normal market return for its routine contributions,
while all residual profits are allocated to the US parent.
The next example illustrates the application of the income method with the
addition of a terminal value.1409 The example pertains to a US parent that
has developed technology Z and enters into a CSA with a foreign cash box
subsidiary. The parent is assigned the right to future applications of tech-
nology Z in the United States, while the subsidiary is assigned the rights
for the rest of the world.
Both the R&D team of the parent and the rights to further develop and
exploit future applications of technology Z are reasonably anticipated to
contribute to future applications of technology Z and are thus regarded as
platform contributions. The subsidiary does not perform any R&D, does
not furnish any platform contributions to the CSA and does not control any
operating intangibles at the inception of the CSA that would be relevant
to the exploitation of either current or future applications of technology Z.
The subsidiary determines that entering into the CSA is a riskier alterna-
tive than a licence agreement in which the parent develops future applica-
tions of technology Z. Further, the appropriate discount rate for the licens-
ing alternative, based on the discount rates of comparable uncontrolled
companies undertaking similar licence agreements, is 13%, while the rate
for the CSA alternative is 14%.
Income projections are drawn up individually for the first 8 years of the
CSA. After year 8, the subsidiary expects its sales of all products based on
422
the exploitation of technology Z in the rest of the world to grow at a rate of

3% per year in the future. Neither the parent nor the subsidiary anticipates
cessation of the CSA activity with respect to technology Z at any determi-
nable date.
The subsidiary estimates that its manufacturing and distribution costs for
exploiting technology Z will equal 60% of gross sales from technology Z
from year 1 onwards, and anticipates cost contributions of 25 per year for
years 1 and 2, 50 for years 3 and 4 and 10% thereafter.
Had the parent further developed technology Z on its own, it would have
charged an uncontrolled licensee a royalty of 30% of the licensee’s rev-
enues. In light of the expected sales growth and the fact that the CSA ac-
tivity will not cease at any determinable date, the buy-in valuation must
include a terminal value calculation.1410
The subsidiary determines that its present value under the CSA (divisional
profits minus operating cost contributions and other cost contributions) is
1,361, while the present value of the licensing alternative is 528. In order
for the present value of the cost-sharing alternative to equal that of the
licensing alternative, the buy-in payment must be 833. This payment en-
sures that the subsidiary is allocated a normal market return for its routine
contributions, while the residual profits in their entirety are allocated to the
US parent.
It may be observed that the present value of the income estimates of the
individual first 8 years of the CSA represents approximately only 38% of
the total present value in both the cost-sharing and licensing scenarios.
The rest of the buy-in amount (62%) is due to the terminal value. This
illustrates the importance of the assumptions made in the terminal value
calculation, which was also a key issue in Veritas.1411
A fundamental assumption in this example is that the cash flows from both
the cost-sharing and licensing alternatives will continue in perpetuity. In
other words, the terminal value represents the present value of the cash
flows from and including year 9 and into eternity. This will often be unreal-
istic, as the cost-shared intangibles will likely have limited useful lives. An
1410. See the discussion of useful lives, growth rates and terminal values in sec.
13.3.3.
1411. Veritas Software Corporation & Subsidiaries v. CIR, 133 T.C. No. 14 (U.S.
Tax Ct. 2009), IRS nonacquiescence in AOD-2010-05. See the discussion in sec.
14.2.4.
423
intangible may eventually become technically or commercially obsolete,

which was the view of the Tax Court in Amazon.com with respect to the
transfer of pre-existing software technology.1412 The perpetuity assumption
may, in some cases, be unsuitable if the valuation pertains to a specific
R&D process, which will normally relate to a concrete product or group
of products.1413 Platform contributions may, however, in practice, often en-
compass several different items of IP and resources transferred together as
a package (e.g. workforce in place, know-how and pre-existing manufac-
turing and marketing intangibles). That was the case in both Veritas and
Amazon.com. Such package transfers may be akin to business transfers,
making the perpetuity assumption comprehensively more suitable.1414
Also, the growth assumptions made in the terminal value calculation are
significant,1415 as seen in Veritas. The larger the growth rate, the larger the
terminal value will be. The example assumes that the subsidiary’s “sales
of all products based upon exploitation of Z in the rest of the world [will]
grow at 3% per annum for the future”.1416
Due to the impact of the growth rate on the terminal value calculation,
there should be concrete – and well-founded – reasons for choosing a
particular rate. The fact that terminal values often are responsible for a
significant portion of the total present value and that the calculation of
it is dependent on assumptions of how long profits will be generated and
on annual growth rates makes these DCF-based valuations susceptible to
speculative assessments from both taxpayers and tax administrations and
should be reviewed critically.
In conclusion, the income method, like the CPM, will be relevant where
only the US parent contributes non-routine development and exploitation
contributions to the value chain that generates the residual profits (e.g.
unique, pre-existing IP and workforce in place). This will typically be the
case if the foreign subsidiary is a cash-box entity that only provides intan-
gible-development funding.
1412. 148 TC No. 8. See also the discussion in sec. 14.2.5.

1413. The weight of this argument is somewhat reduced by the fact that the further into
the future the cash flows lie, the less impact they will have on the present value due to
a high discount factor.
1414. Perpetuity assumptions are commonly used in DCF business valuations where
the value of an entire enterprise as a going concern is estimated.
1415. See the discussion of useful lives, growth rates and terminal values in sec. 13.3.3.
1416. Id.
424
If, however, the foreign entity is an established enterprise (as the case was
to some extent in Amazon.com) that has developed valuable local market-
ing intangibles (customer lists, know-how, goodwill, etc.) that contribute
to the residual profits, the income method should not be applied. In these
cases, the residual profits should be split between the US and foreign enti-
ties pursuant to the relative values of their non-routine development and
exploitation contributions under the PSM.1417
14.2.8.4. The acquisition price method
The acquisition price method evaluates whether the buy-in amount is at

arm’s length by reference to the amount charged for the stock or asset in an
uncontrolled transaction (acquisition price).1418 The method is ordinarily
used where substantially all of the acquired target’s non-routine contribu-
tions are covered by a buy-in.
The regulations illustrate the methodology in an example in which a US

parent enters into a CSA in year 1 with a foreign subsidiary to develop
group Z products. The parent is assigned the US rights to exploit the group
Z products, while the subsidiary is assigned the rights for the rest of the
world.1419 The RAB shares of the parent and subsidiary are 60% and 40%,
respectively. In year 2, the parent acquires company X for USD 110 mil-
lion in cash.1420 The value of company X consists of workforce in place and
technical intangibles of USD 100 million and tangibles of USD 15 million,
minus liabilities of USD 5 million.
The workforce and technical intangibles are reasonably anticipated to con-

tribute to the development of group Z products and are therefore platform
contributions for which the subsidiary must make a buy-in. As the subsidi-
ary’s share of the anticipated benefits is 40% and the value of the platform
contributions is 100, the platform contribution payment is 40. The scope of
application of this method is narrow due to its reliance on highly particular
CUTs and will therefore likely rarely be relevant. Further, the profit al-
location that results from applying this method must align with the results
that would follow from a valuation based on the estimated future operat-
1417. See the discussion of the RPSM in sec. 14.2.8.6.

1418. Treas. Regs. § 1.482-7(g)(5). For a discussion of the acquisition price method,
see, e.g. Wittendorff (2010a), at p. 582.
1419. Treas. Regs. § 1.482-7(g)(5)(v), Example.
1420. The parent and company X are treated as one taxpayer for income tax purposes
in the United States.
425
ing profits from the intangible developed under the CSA. This restriction
will effectively ensure that CUT-based buy-in pricing is consistent with the
results of the income method and PSM, which both rely on DCF analysis.
14.2.8.5. The market capitalization method
The market capitalization method evaluates whether the buy-in amount

is at arm’s length by reference to the average market capitalization of the
controlled participant that provides platform contributions to the CSA and
whose stock is regularly traded on an established securities market.1421
The US regulations illustrate the methodology in an example in which a

US parent enters into a CSA with a foreign subsidiary to develop soft-
ware.1422 The parent owns software under development that is not yet ready
for the market. The CSA assigns the exclusive right to exploit the software
in the United States to the parent, and the foreign rights to the subsidiary.
The parent’s RAB share is 70%, and it is 30% for the subsidiary. The par-
ent’s assembled team of researchers and in-process software are reasonably
anticipated to contribute to the CSA development,and therefore deemed as
platform contributions for which compensation is due from the subsidiary.
The average market capitalization of the parent is 205 at the time at which
the CSA is entered into. The value of tangibles and other assets is 5, and
there are no liabilities. Aside from these elements, the value of the parent
consists of only the value of the research team and the in-process software.
The value of the platform contributions of the parent is therefore 200. As
the RAB share of the subsidiary is 30%, the buy-in payment is 200 × 0.3
= 60.1423
The method can only be used when substantially all of the participants’
non-routine contributions are encompassed by the CSA. As this restric-
tion comes in addition to the ordinary CUT requirements, it is safe to say
that the method is limited to a narrow field of application.1424 Further, the
profit allocation that results from applying this method must align with the
1421. Treas. Regs. § 1.482-7(g)(6). For a discussion of the market capitalization meth-
od, see, e.g. Wittendorff (2010a), at p. 583.
1422. Treas. Regs. § 1.482-7(g)(6), Example 1.
1423. For an extension of the example, see Treas. Regs. § 1.482-7(g)(6), Example 2.
1424. The significance of the restriction is illustrated in an example in which the parent
has significant non-routine assets that will be used solely in a separate business divi-
sion that is unrelated to the CSA and are not deemed platform contributions and are not
covered by the buy-in; see Treas. Regs. § 1.482-7(g)(6), Example 3.
426
results that would follow from a valuation based on the estimated future
operating profits from the intangible developed under the CSA. This re-
striction will effectively ensure that CUT based buy-in pricing is consistent
with the results of the income method and PSM, which both rely on DCF
analysis.
Also, this method will likely capture all intangible value contributed to the
CSA, including residual intangible value (workforce in place, goodwill and
going concern value) that does not qualify as IP under the pre-2018 ver-
sion of the IRC section 936 IP definition, due to the platform contribution
bypass of the definition.1425
14.2.8.6. The RPSM
The RPSM evaluates whether the buy-in amount is at arm’s length by

reference to the relative value of each CSA participant’s non-routine de-
velopment and exploitation contributions to the operating profits from the
cost-shared intangibles.1426 The method may not be used where only one
controlled participant makes significant non-routine contributions. In those
cases, the income method will, in practice, be the relevant method. The
income method allocates the entire residual profits from the exploitation of
the cost-shared intangibles to the single CSA participant that renders non-
routine contributions, while the RPSM divides the residual profits among
several participants that contribute non-routine inputs.
The residual profits to be split, pursuant to the relative values of the non-
routine contributions, equal the present value of the operating profits from
the value chain to which the cost-shared intangibles are connected over
the duration of the CSA, minus market returns for routine contributions,
operating cost contributions and cost contributions, using an appropriate
discount rate.1427 As under the general PSM, the relative values may be
1425. See the discussions of the US IP definition and platform contributions in secs.
3.2. and 14.2.7.5., respectively.
1426. Treas. Regs. § 1.482-7(g)(7). The relationship between the general PSM in
§ 1.482-6 and the RPSM for CSA buy-ins is that the former shall only apply to CSAs to
the extent provided (and as modified) in § 1.482-7(g). For a discussion of the general US
PSM, see ch. 9. See also Wittendorff (2010a), at p. 584 for a discussion of the RPSM in
the context of CSAs.
1427. Market returns for routine contributions include market returns for operating
cost contributions and exclude market returns for “pure cash”cost contributions; see
Treas. Regs. § 1.482-7(g)(7)(iii)(B).
427
measured by external market benchmarks that reflect fair market value or

through the capitalized cost of developing the non-routine contributions.1428
The regulations contain an example that illustrates the key points of the
methodology.1429 It pertains to a US parent that has partially developed a
nano-disc technology and enters into a CSA with a foreign subsidiary to
further develop the technology for commercial exploitation. The subsidi-
ary has developed significant marketing intangibles outside of the United
States (customer lists, ongoing OEM relationships and established trade-
marks). The CSA assigns the US and foreign rights to exploit the nano-disc
technology to the parent and the subsidiary, respectively.
The parent’s technology is reasonably anticipated to contribute to the de-

velopment of commercially exploitable nano-discs and thus is deemed to
be a platform contribution for which arm’s length buy-in compensation is
due. The subsidiary does not contribute any inputs to the intangible devel-
opment, but contributes its local marketing intangible to the subsequent
exploitation of the cost-shared intangible. Thus, it is only in the territory of
the subsidiary that both the parent and the subsidiary make significant non-
routine contributions (to the development and exploitation, respectively).
The relative value of the respective contributions to the residual divisional
profit attributable to the cost-sharing activity must therefore be determined
only in the territory of the subsidiary.1430
The determination of the present value of the residual profits in the terri-
tory of the subsidiary is founded on the following assumptions:1431
− revenues: no sales in year 1 and sales of 200 in year 2, with a 50%
increase per year through year 5. The annual growth rate is then ex-
pected to decline to 30% per year in years 6 and 7, to 20% per year
in years 8 and 9 and to 10% in years 10 and 11, and then 5% per year
thereafter;
1428. See the discussion of the general US PSM in ch. 9. If the non-routine contribu-
tions are also used in other business activities, the value must be allocated among all
of the business activities in which they are used in proportion to the relative economic
value of the relevant business activity as the result of such non-routine contributions;
see Treas. Regs. § 1.482-7(g)(7)(iii)(C)(2).
1429. Treas. Regs. § 1.482-7(g)(7(v), Example 1.
1430. The parent is entitled to the entire residual profits from exploitation of the cost-
shared intangible in its territory, as it alone contributes non-routine value chain inputs
there.
1431. The value of the residual profits in the territory of the subsidiary for each year of
the CSA is determined as the following (with all elements only from the subsidiary’s
territory): revenues minus routine costs and market returns for routine contributions.
428
− routine costs are anticipated to equal 40 in year 1, 130 in year 2, 200 in

year 3 and 250 in year 4;1432
− total operating expenses connected to product exploitation (including
operating cost contributions) equal 52% of sales per year;
− the subsidiary undertakes routine distribution activities in its mar-
kets that constitute routine contributions. It is estimated that the total
market return on these contributions will amount to 6% of the routine
costs; and
− the subsidiary expects its cost contributions to be 60 in year 1 and 100
in years 2 and 3, decline to 60 in year 4 and thereafter equal 10% of
revenues.
The appropriate discount rate is set to 17.5%. This yields a present value of
the residual profits from the exploitation of the cost-shared intangible in the
subsidiary’s territory of 1,395. The parent and subsidiary determine that
the value of the parent’s nano-disc technologies relative to the value of the
subsidiary’s marketing intangibles is approximately 150%. In other words,
the parent is entitled to 60% of the present value of the residual profits,1433
which are 837, in the form of a buy-in payment from the subsidiary.1434
In conclusion, it may be observed that the RPSM will generally be the rel-
evant pricing method when the foreign subsidiary is an established enter-
prise, as it may have developed valuable local marketing intangibles (cus-
tomer lists, know-how, goodwill, etc.) that contribute to the residual profits
generated by the exploitation of the cost-shared intangibles in its market.
This was, in effect, the view of the Tax Court in Amazon.com with respect
to the transfer of marketing intangibles, as it reduced the buy-in value to
take into account prior ownership of the foreign group entity.1435
Thus, when the foreign subsidiary contributes valuable operating intangi-

bles, there will be no room to apply the income method. In such cases, the
residual profits shall be split under the RPSM between the US and foreign
entities pursuant to the relative values of their non-routine development and
exploitation contributions.
1432. Routine costs are costs other than cost contributions, routine platform and oper-
ating contributions and non-routine contributions.
1433. 150 ÷ (150 + 100) = 0.6.
1434. The second RPSM example pertains to a scenario in which the residual profits
from both the parent and subsidiary markets are split; see Treas. Regs. § 1.482-7(g)(7)
(v), Example 2.
1435. See the discussion of the ruling in sec. 14.2.5.
429
14.2.8.7. Unspecified methods
The cost-sharing regulations allow the use of unspecified methods.1436 Such

methodology must be based on the realistic alternatives of the controlled
parties and provide information on the prices or profits that the controlled
participants could have realized by choosing a realistic alternative to the
CSA.1437 As with any method, an unspecified method shall not be applied
unless it provides the most reliable measure of an arm’s length result under
the best-method rule.1438 In light of the highly developed specified pricing
methods available under the current cost-sharing regulations and the fact
that the go-to income method and RPSM can be used in the absence of
CUTs, there should not be much need to apply unspecified methods. If an
unspecified method has been used, its resulting profit allocation should be
reviewed critically.1439
14.3. Buy-in pricing under the OECD TPG
The OECD TPG have historically been sparse with respect to the alloca-
tion of operating profits through CSAs.1440 At the time at which the 1979
OECD Report on Transfer Pricing and Multinational Enterprises was writ-
ten, CSAs were not seen as posing any particular risk of base erosion.1441
The Report discussed how the right to deductions for R&D costs should be
delimited in the context of CSAs,1442 but did not directly touch upon how
1436. Treas. Regs. § 1.482-7(g)(8). See Wittendorff (2010a), at p. 586 for a discussion
of the use of unspecified methods in the context of CSAs.
1437. See the discussion of unspecified transfer pricing methods in ch. 12.
1438. See the discussion of the US best-method rule in sec. 6.4.
1439. See sec. 14.2.3. for a discussion on the use of unspecified methods under the pre-
2009 CSA regulations.
1440. See also Brauner (2010), at p. 10 and the critical comments in Brauner (2016), at
p. 111, where it is noted: “Perhaps the least intellectually defensible part of the OECD
TPG was that this chapter had not enjoyed much analysis or any other attention pre-
BEPS; however, the importance of cost sharing, the idiosyncratic United States rule
that had inspired the discussion of CCAs, forced the OECD to try and pour some con-
tent into this formerly essentially empty bucket. It did not do that.” While the author
certainly shares some of these views, he will not go quite as far in his criticism.
1441. 1979 OECD Report, para. 109. The United States was apparently the only coun-
try with domestic legislation on the transfer pricing of CSAs. Other countries had only
limited experience in the field; see para. 110 of the 1979 OECD Report. Germany,
however, developed guidelines for cost sharing where R&D costs could only be valued
in the aggregate.
1442. 1979 OECD Report, paras. 102-124. The Report also took, in the author’s view,
the questionable position that cost contributions should contain a profit element; see
para. 119. The 1968 US IRC sec. 482 regulations (33 Fed. Reg. 5848) allowed deduc-
430
Buy-in pricing under the OECD TPG
the residual profits from exploitation of cost-shared intangibles should be

allocated.1443
The 1995 OECD TPG introduced a chapter on CSAs. It stated that any
contribution of pre-existing rights should be “compensated based upon an
arm’s length value for the transferred interests”.1444 No elaboration was of-
fered with respect to the acceptable pricing methodology.1445 Considerable
doubt was left as to which buy-in measurements would be acceptable.1446 In
1997, the OECD adopted a report on cost contribution agreements, which
did not entail significant changes to the 1995 text.1447 The guidance con-
tained in the 2010 OECD TPG was virtually, in all respects, the same as in
the 1995 OECD TPG.
The new 2017 chapter in the OECD TPG on CSAs, although not a vo-
luminous text, contains substantial restrictions on buy-in pricing.1448 The
principle approach is that CSA pricing is subject to the same analytical
framework for review as other intangibles transactions. The logic behind
this position is that parties performing activities under arrangements with
similar economic characteristics should receive similar expected returns,
irrespective of whether the contractual arrangement is classified as a CSA,
contract R&D agreement, etc. The 2017 guidance on risk,1449 intangible
ownership (the “important functions doctrine” and the guidance on the
tions for actual expenses without the addition of a mark-up. There should, in the au-
thor’s view, be no profit element in the cost contributions. The profits should be taxed
subsequently, when the cost-shared intangible is exploited and generates profits.
1443. According to the 1979 OECD Report, each CSA participant should bear its fair
share of costs and risks, and in return be entitled to its share of the results; see 1979
OECD Report, para. 103. The Report did not go further into the division of the results,
but stated that costs should usually be divided among the contributing parties in pro-
portion to the relevant entity’s use or return from the intangible developed under the
CSA; see para. 106.
1444. 1995 OECD TPG, para. 8.31.
1445. 1995 OECD TPG, ch. VIII. The need for additional buy-in guidance was empha-
sized; see 1995 OECD TPG, para. 8.1. Brauner (2010), at p. 10 argues that while the US
CSA regulations provided a firm taxpayer a safe-harbour solution, the OECD TPG only
offered more “loose” assurances for taxpayers (i.e. that tax authorities should refrain
from reassessments).
1446. See 1995 OECD TPG, para. 8.15, which stated that countries had experience
with both the “use of costs and with the use of market prices for the purposes of meas-
uring the value of contributions to arm’s length CCAs”.
1447. OECD 1997 Report on Cost Contribution Arrangements.
1448. See, however, Brauner (2014a), at p. 100, where he criticizes the OECD for not
putting enough effort into guidance on CSAs (given the significant focus on revising
the general intangibles guidance in the BEPS revision of the OECD TPG).
1449. OECD TPG, paras. 1.56-1.106. The new risk guidance is discussed in sec.
6.6.5.5. of this book.
431
determination of a risk-adjusted funding return) 1450 and hard-to-value in-

tangibles also apply in the context of CSAs.1451
Under the new guidance, the value of each participant’s CSA contribution
must be consistent with the value that independent enterprises in compa-
rable circumstances would have assigned to it.1452 A distinction is drawn
between contributions of pre-existing value (e.g. unique intangibles) and
current contributions (e.g. ongoing R&D). If a pre-existing intangible is
contributed to the CSA, its value should be determined pursuant to the
transfer pricing methods and the new guidance on valuation.1453 The value
of current functional development contributions is not based on the po-
tential value of the resulting technology, but on the value of the functions
performed.1454 This value is determined under the transfer pricing methods,
the new guidance on intangibles and the guidance on intra-group services.
Thus, the rule for both categories of contributions is that the pricing should
reflect their value, not costs.1455 It is warned that a cost-plus-based com-
pensation with a “modest mark-up” will not reflect the anticipated value
of – or the arm’s length price for – the contribution of an R&D team in all
cases.1456 Further, the pricing of current development contributions at cost
will generally not provide a reliable basis for the application of the arm’s
length principle.1457
1450. OECD TPG, paras. 6.50-6.58 (the “important functions doctrine”); and paras.
6.60-6.64 (R&D funding remuneration). The “important functions doctrine” and R&D
funding remuneration guidance are analysed in secs. 22.3. and 22.4., respectively.
1451. OECD TPG, paras. 6.181-6.195. This guidance on hard-to-value intangibles is
discussed in sec. 16.5.
1454. Id.
1455. Measurement of current contributions at cost is allowed when this may be more
administrable; see OECD TPG, para. 8.27.
1456. Id. This statement should be read in light of the 2015 rejection of the previous
2009 OECD position that an R&D entity could be remunerated on a cost-plus basis
under a contract R&D agreement that allocated the residual profits to the cash-box en-
tity; see OECD TPG, paras. 7.4 and 9.26. See also the discussion in sec. 22.3.3.2., with
further references.
1457. OECD TPG, para. 8.28. See, however, Brekel et al. (2016), where it is argued that
concurrent R&D contributions could be remunerated using cost-plus, while the residual
profits are subsequently split among the same participants that contributed the ongoing
R&D inputs. As long as this would yield an allocation of residual profits similar to that
which would have been the case had the important functions doctrine (see the discus-
sion in sec. 22.3.2.) been applied to allocate the residual profits, the author would agree.
432
Buy-in pricing under the OECD TPG
It is, in the author’s view, clear that a group entity that renders important
development functions should be entitled to a portion of the residual profits
from the intangible developed under the CSA.1458 In other words, alloca-
tion of a concurrent normal return remuneration to this entity pursuant to
a one-sided pricing method will not satisfy the arm’s length standard. This
restriction is entirely necessary in order to avoid BEPS and to align profit
allocation with the underlying intangible value creation.
In the end, the author questions whether the distinction between pre-ex-
isting value and current contributions is at all useful, in particular, as the
new guidance states that both types of contributions should be priced at
value.1459 In practice, it will often be the same group entity that contributes
pre-existing intangibles and ongoing R&D, and an aggregated valuation
approach will presumably often be appropriate. In these cases, it is not
meaningful to distinguish between non-routine contributions, as they are
priced together. It therefore remains to be seen whether multinationals will
argue for low buy-ins based on the assertion that compensation for ongoing
R&D should be “carved out” of the buy-in amount and priced separately
on a cost-plus basis. Pre-existing intangibles and ongoing R&D should, in
the author’s view, be priced together as a whole, since the value of both
elements is reflected in the estimated future profits generated through the
exploitation of the intangible developed under the CSA. This is also the US
approach.1460
The new guidance contains an example that illustrates that a group entity
that only provides intangibles development funding to the CSA will be
allocated only a risk-adjusted return, while the residual profits will be al-
located to the CSA participant that contributes non-routine inputs.1461
The example pertains to Companies A and B, which enter into a CSA to

develop an intangible. Company B provides a pre-existing intangible that
forms the basis for the development, as well as the R&D team, while Com-
pany A only provides intangibles development funding (an annual 100 in
years 1-5).
1458. See the discussion of the OECD “important functions doctrine” in sec. 22.3.2.
1459. See also Storck et al. (2016), at p. 219, where it is made clear that the question of
whether current R&D efforts should also be subject to market-value buy-in under the
new OECD TPG has attracted attention and that some are of the view that cost alloca-
tion could be sufficient for such ongoing R&D efforts.
1460. See the discussion of the new US guidance on aggregated valuation of controlled
transactions in sec. 6.7.2. (on the effects of the new guidance on the best-method rule,
see sec. 6.4.).
1461. OECD TPG, ch. VIII, Example 4.
433
The value of Company B’s contribution is determined indirectly as the

residual of the total anticipated operating profits and the return allocable
to Company A’s funding contribution. It is estimated that the exploitation
of the cost-shared intangible will generate global profits of 550 per year in
years 6-15, with 60% of the profits generated in the market of Company A
and 40% in the market of Company B.
The controlled pricing allocates 330 in operating profits to Company A in

each year. This pricing is benchmarked against the best realistic alterna-
tive of Company A, which is to receive a risk-adjusted rate of return on its
funding capital (essentially the same pricing approach as that applied in
the US income method).1462 This return is determined to be 110 per year in
years 6-15, equal to an internal rate of return of 12%.1463 Thus, in order for
the pricing to be at arm’s length, Company A must surrender (through bal-
ancing payments) 220 per year to Company B so that the profits allocable
to Company A in each year is 110, equal to its best realistic alternative.1464
The example describes Company A as a typical cash-box entity that only

provides intangibles development funding to the CSA. In cases involving
similar entities, it is the author’s view that it must be assumed that a rate
around the 12% return on invested capital is the OECD-endorsed arm’s
length profit allocation.
The author observes that the examples in the new guidance on cost sharing
seems to indicate that the cash-box funding entity (Company A) will not be
entitled to any remuneration for its R&D funding contributions if it is not
deemed to control the financial risks associated with its funding.1465 Such a
solution would be contrary to the general approach towards the remunera-
tion of R&D funding that the OECD has taken in the new guidance on

1463. It should be noted that footnote 1 of the new CSA guidance does not explain how
the 110 in annual income to the funding investment in years 6-15 is derived. It is just as-
sumed that this level of profit represents an arm’s length return. It is further stated that
this result is shown just to demonstrate the principles of the example and is not meant
to offer any “guidance as to the level of arm’s length returns to participants in CCAs”.
While the author recognizes this reservation, he finds it difficult to see that the rate of
return used in the example should have no relevance in cases pertaining to pure cash-
box entities. After all, the rate of return selected is used to illustrate an arm’s length
return.
1464. Navarro (2017), at p. 269, finds that the result of the example is contrary to the
arm’s length principle. The author strongly disagrees with this assertion, based on
much of the same reasoning as for his objection to Navarro’s approach for remunerat-
ing group entities that carry out DEMPE functions (see infra n. 1988).
1465. OECD TPG, ch. VIII, annex, Example 5, para. 22.
434
The relationship between US and OECD buy-in pricing
intangibles, pursuant to which the funding entity is entitled to a risk-free

return in cases in which it does not control the financial risk.1466 Further,
given that the new guidance on cost sharing is based on the view that the
economic characteristics of the controlled parties’ activities should guide
the profit allocation, irrespective of the legal form in which the activity is
organized (CSAs, contract R&D agreements, etc.), as well as the fact that
paragraph 8.9 of the OECD TPG explicitly states that paragraphs 6.60-
6.64 apply for the purpose of remunerating R&D funding contributions,
it seems clear that paragraph 22 of Example 5 should be interpreted as
requiring the allocation of a risk-free rate of return for R&D funding con-
tributions from a cash-box entity that is deemed to not control the financial
risks associated with its funding contribution.
14.4. The relationship between US and OECD

buy-in pricing
On the basis of the discussions in this chapter, it may be observed that there
are clear similarities between the US and OECD approaches to buy-in pric-
ing.1467 While the new OECD guidance lacks the detailed and instructive
character of the US cost-sharing regulations,1468 as well as pricing meth-
odology specifically tailored to the CSA context, the realistic alternatives
principle will no doubt form an effective restriction on CSA pricing, as it
also does through the US income method.1469 The material core of the pric-
ing approaches is therefore essentially the same. Both the US and OECD
rules will strip residual profits from a cash-box entity that only provides
intangibles development funding to the CSA.1470

1467. See also, however, Brauner (2016), at p. 111, where it is emphasized that there are
significant differences between the US and OECD regimes, particularly with respect to
the US CSA regime being a transfer pricing safe harbour and a legal vehicle to avoid the
US partnership taxation rules, as well as the OECD regime encompassing more than
just joint development of IP. While the author agrees with this, his focus merely rests
on the buy-in pricing aspects of the US and OECD regimes, and on this point, there are
clear similarities, as the pricing must qualify as arm’s length under both regimes.
1468. Brauner (2016), at p. 123 notes that it may be difficult to avoid circumvention of
the OECD CSA regime without more detailed rules.
1469. See the discussion of the US income method in sec. 14.2.8.3.
1470. See also the critical comments on the structure and content of the OECD CSA
rules in Brauner (2016), at pp. 124-126. Brauner seems to attribute little in the way of
specific content to the OECD TPG CSA chapter, in particular due to the risk that the
(partly ambiguous) OECD provisions may be interpreted in diverging and relaxed ways
by jurisdictions. While the author, to some extent, agrees with this, he still finds the
core content of the OECD CSA buy-in guidance to be clear: it will strip cash-box enti-
435
The difference lies in the fact that the US regulations will allocate a CPM-
based normal market return to this entity, while the OECD TPG are more
generous, as they allow the cash-box entity to earn a risk-adjusted rate of
return on the capital that it has invested into the CSA R&D. While the
author finds the OECD approach to be the most principled of the two (i.e.
the OECD and US approaches), as capital by itself should indeed attract an
arm’s length return, it does trigger a risk of relatively substantial “leakage”
of operating profits from high-tax jurisdictions (where R&D typically is
carried out) to low or no-tax jurisdictions. The United States avoids this
risk through its “cut-off” approach to intangibles development funding un-
der the income method.1471
ties of residual profits. In reality, this OECD profit allocation approach is much akin to
that of the income method of the US CSA regulations.
1471. For a further analysis of the problems connected to these approaches to intangi-
bles development funding, the author refers to the discussions in secs. 22.4.11. and 26.3.
436
Chapter 15
Taxpayer-Initiated Compensating
Adjustments to Indirect IP Pricing
15.1. Introduction
The question in this chapter is to what extent a group entity may report in
its tax return a transfer price that is an arm’s length price for the controlled
transaction in question, but differs from the amount actually charged
throughout the income year, i.e. whether the controlled taxpayer may per-
form a compensating or year-end adjustment.1472 This stands in contrast
to the issue of whether tax authorities may carry out periodic adjustments
(which is the topic of analysis in chapter 16). Due in part to the practical
significance of year-end adjustments as a result of the dominant role of
the transactional net margin method (TNMM) in global transfer pricing
practice and in part to the legal uncertainty surrounding some aspects of
year-end adjustments, the author finds it necessary to carry out a fairly
broad analysis of the subject.1473
The author will first provide a lead-in to the topic of year-end adjust-
ments in section 15.2. The TNMM will be used as an illustration tool, but
what is said largely applies analogously for the resale price and cost-plus
methods.1474 The author will then analyse the authority to perform year-
end adjustments under US law and the OECD Transfer Pricing Guidelines
(OECD TPG) in sections 15.3. and 15.4., respectively. In section 15.5., he
discusses a recommendation from the EU Joint Transfer Pricing Forum
(JTPF). Lastly, he will analyse case law on year end-adjustments in sec-
tions 15.6.-15.10.
1472. These terms will be used interchangeably throughout the chapter.

1473. See De Simone et al. (2015) for a recent and interesting economic analysis of
how multinationals “shift” profits between jurisdictions after the year’s end so that the
reported profits in each jurisdiction conform to an arm’s length return.
1474. The relevant difference between the methods is that the transactional net margin
method (TNMM) sets a target net profit margin, while the resale price and cost-plus
methods set a target gross profit margin.
437
Chapter 15 - Taxpayer-Initiated Compensating Adjustments to Indirect
IP Pricing
15.2. A lead-in to compensating adjustments
Compensating adjustments are inherently linked to the application of one-

sided transfer pricing methods (see the analysis of the comparable prof-
its method (CPM) and TNMM in chapter 8). As the determination of the
amount of some of the elements in the tested party’s operating profits may
be contingent on external factors (customer demand, raw material market
prices, etc.), it will not be possible at the start of the year for a group to
adapt its transfer prices so that the arm’s length profit margin is reached.
Only in rare circumstances will the budgeted amounts align exactly with
the amounts actually realized. Without a year-end adjustment, to be carried
out when the actual results have become known (i.e. after the end of the
income year), it will simply not be possible for the taxpayer to achieve the
arm’s length transfer pricing result dictated by the TNMM.
Let us say, for instance, that a Norwegian group that designs, manufac-
tures and sells luxury yachts decides that its low-risk US distribution entity
shall be allocated a TNMM-based return for the income year 2008 go-
ing forward. The Norwegian parent owns all unique intangible property
(IP) employed in the value chain and carries out research and development
(R&D) and manufacturing. In the late autumn of 2007, the group performs
a transfer pricing analysis of comparable unrelated US distributors based
on financial accounting data from commercial databases.
Because audited financial statements are ready at the earliest some months
after the end of the accounting period to which the accounts refer, there is
a lag in the available third-party profit data. The most current data is for
the income year 2006. Based on the averages for the years 2004-2006, the
analysis shows an interquartile net profit margin range between 5.25% and
6.5%, with a median of 5.8%. The study therefore concludes that the net
profit return for the US distribution entity for 2008 going forward should
be set to 5.8%. This is the target TNMM return for the US distribution
entity for 2008.
The list (transfer) prices for yachts sold from the Norwegian parent to the
US subsidiary are then designed to realize the target return based on esti-
mates of what the external sales and operating expenses of the US distribu-
tion entity will be in 2008 going forward. The transfer prices will enable
the subsidiary to reap the annual 5.8% net profit return.
Because there is uncertainty connected to the estimates upon which the

prices are based, the controlled parties agree on a year-end adjustment pro-
438
A lead-in to compensating adjustments
vision. Let us suppose that the list prices for the Norwegian yachts are
based on the budget for the US distribution entity shown in table 15.1.
Table 15.11475
Parent Subsidiary Margins

(Norway) 1475 (United (subsidiary)
States)
Sales 17,500 2,500 -
Cost of goods sold (COGS) 12,600 1,800 -
Gross profit 4,900 700 28.0%
Operating expenses 3,892 556 -
Net profit 1,008 144 5.8%
The subsidiary’s sales and operating expenses are external. Given the budget-
ed sales of 2,500 and operating expenses of 556, a transfer price of 1,800 for
the yachts (COGS) will result in the target net profit margin of 5.8%. In 2008,
the financial crisis impacts the US market for luxury boats significantly.
By the end of 2008, the US subsidiary has generated the actual results
displayed in table 15.2.
Table 15.2

(Norway) (United (subsidiary)
States)
Sales 17,500 2,125 -
COGS 12,600 1,800 -
Gross profit 4,900 325 15.3%
Net profit 1,008 −231 −10.9%
Due to a 15% reduction in sales, the subsidiary was unable to reach the
5.8% target TNMM net profit return rate for 2008. Instead, it incurred a
loss (negative net profit margin of −10.9%).
1475. The data for the parent includes its worldwide transactions (also transactions
with entities other than the US subsidiary).
439
IP Pricing
This is not an arm’s length net profit margin. In order to align the actual
results with the profit allocation result prescribed by the TNMM, the Nor-
wegian parent carries out a compensating year-end adjustment by reducing
the sales prices charged for the boats sold to the US subsidiary by 355,
thereby lowering the COGS for the subsidiary from 1,800 to 1,445. This
brings the net profit margin of the US subsidiary in line with the TNMM,
as it then realizes the target net profit margin of 5.8% (shown in table 15.3).
Table 15.3

(Norway) (United (subsidiary)
States)
Sales 17,145 2,125 -
COGS 12,600 1,445 -
Gross profit 4,545 680 32.0%
Net profit 653 124 5.8%
In this example, the year-end adjustment is performed by reducing the

price of goods. Alternatively,1476 it could have been performed by lowering
royalties paid by, or administrative expenses allocated to, the subsidiary,
or simply by the parent reimbursing some of the subsidiary’s operating
expenses (e.g. marketing costs).1477
1476. An interesting issue is whether the taxpayer is free to adjust prices of goods that
have a clear market price (e.g. commodity goods such as sugar and oil). When the pri
cing is based on a one-sided method, it will be the profit margin allocated to the tested
party that determines whether the pricing is at arm’s length. The one-sided methods do
not ask what the arm’s length price is for any specific controlled exchange among the
controlled parties, but rather what the arm’s length net remuneration is that the tested
party must be left with as a result of all of its controlled transactions with the entrepre-
neur (i.e. an aggregated pricing approach). For this view to be valid, however, all con-
trolled transactions in the relevant value chain must be carried out between the same
taxpayers. This was not the case in GlaxoSmithKline Inc. v. R. (see sec. 6.7.4.), in which
the Canadian distribution entity of GSK purchased raw materials from Switzerland
and licensed manufacturing and marketing intangible property (IP) from the United
Kingdom. Further, even when all controlled transactions are carried out between the
same group entities, the classification of a year-end adjustment may be relevant for non-
transfer pricing purposes (e.g. VAT, customs and foreign exchange).
1477. See Barmentlo et. al. (2013), who argue that companies may not be able to de-
fend year-end adjustments to royalties and service charges due to difficulties in clas-
sifying payments and identifying comparables to support the charges applied. While
the author shares the Barmentlo’s view that it is important for year-end adjustments to
440
A lead-in to compensating adjustments
For the controlled parties, the compensating adjustment ensures that the
actual profit allocation, as reflected in their tax return positions, cor-
responds to the arm’s length result dictated by the applicable one-sided
method (here, the TNMM). The pre-adjustment profit allocation in the ex-
ample was not in accordance with the arm’s length principle. No unrelated
distributor would be willing to perform marketing and distribution ser-
vices for a negative net profit margin. The post-adjustment return, however,
corresponds to what comparable unrelated distributors would demand as
compensation.
For the involved jurisdictions, year-end adjustments may increase or de-

crease the tax base. When the actual net profit margin of the tested party is
less than the target arm’s length return and the taxpayer does not perform
a year-end adjustment, its residence jurisdiction is likely to carry out a
reassessment. Further, the jurisdiction of the non-tested party will likely
react if a compensating adjustment is performed, as it will reduce its tax
base, thus bearing the “cost” of the adjustment, as the case was in Vingcard
(discussed in section 15.6.).
In the reverse situation, where the tested party realizes a net profit that
exceeds the target return, a downward year-end adjustment should ensure
that the residual profit is allocated to the non-tested party. The jurisdiction
of the tested party may possibly want to perform a transfer pricing assess-
ment to test whether the TNMM has been applied correctly, as its tax base
will be reduced by the adjustment, as the case was in T.Srl (discussed in
section 15.7.).
It is important to bear in mind that compensating adjustments refer to

the process of correcting an actual result so that the controlled profit al-
location is brought in line with the arm’s length allocation required by
the TNMM. It is not a question of reassessing the arm’s length pricing as
such. The 5.8% target arm’s length net profit margin result in the example
is the same both before and after the adjustment. A year-end adjustment
be founded in written agreements that were in effect for the year to which the adjust-
ment refers, the author does not share the concern with respect to comparables. Most
year-end adjustments are carried out in order to align the results of the tested party with
the profit margin yielded by the one-sided methods. In such cases, there is no need for
a comparable transaction supporting that a year-end adjustment in and of itself may be
made. The margin yielded by the one-sided methods will, in itself, justify the adjust-
ment. The Norwegian Vingcard Elsafe AS v. Skatt Øst, Utv. 2012 p. 1191, reversing in
part, Utv. 2010 p. 1690 pertained to year-end adjustments carried out partly through a
price reduction for goods sold and partly through the reimbursement of guarantee and
marketing expenses. See the discussion in sec. 15.6.
441
IP Pricing
is therefore a matter of reconciliation: preliminary pricing is adjusted so

that the return position of the taxpayer reflects the arm’s length result.
This separates the issue from ordinary transfer pricing disputes in which
the problem is to determine the appropriate net profit margin to be used
under the TNMM. For instance, this could occur when the tested party’s
return position is that it should earn a 5.8% net profit margin while the tax
authorities claim that a correct application of the TNMM yields a net profit
margin of 7.2%. This is, in principle, a normative question, as the problem
is to determine whether the pricing reflects a correct interpretation and
application of the relevant transfer pricing method. However, the author
admits that the line between a reconciling year-end adjustment and a nor-
mative transfer pricing adjustment is thin. This is true in particular when
the year-end adjustment is not motivated by the desire to align controlled
prices charged throughout the year so that the total net profit margin cor-
responds with the arm’s length TNMM margin, but rather that the transfer
pricing study that formed the basis for the target margin is updated at the
time of filing the tax return to include a more recent accounting period.
This update may alter the arm’s length profit margin itself. Such a year-end
adjustment is, in practice, rather indistinguishable from a transfer pricing
adjustment.
Further, a year-end adjustment refers to the process of correcting an actual,

but unassessed, result. The adjustment is carried out before the taxpayer
files its tax return for the income year to which the adjustment refers.1478
The basic system of the OECD TPG (and US transfer pricing law for that
matter) is that transfer pricing should be based on the transactions actually
carried out by the controlled parties. This refers to the controlled transac-
tion as such, including all pricing mechanisms contained within it. The
concurrent transfer prices used throughout an income year may be single
components in a larger pricing system agreed to by the controlled parties.
For instance, the pricing mechanism in the distribution agreement in the
above example consists of two provisions, namely (i) the list prices used
throughout the year; and (ii) the adjustment provision for aligning the ac-
tual net profit of the tested party to the arm’s length TNMM margin. The
1478. Year-end adjustments may entail a range of non-income tax complications, par-
ticularly with respect to VAT, customs, foreign exchange, regulatory price control (e.g.
for pharmaceutical products) and financial accounting. Such issues are becoming in-
creasingly important in many OECD member countries; see the 2006 OECD compa-
rability report, Timing issues in comparability, para. 41. For further discussions, see
Pheiffer et al. (2015), at p. 302; and Cottani (2007).
442
Taxpayer-initiated adjustments under US law
list prices are preliminary in this context. The ultimate point of the p ricing
framework is to remunerate the US distribution entity for its marketing,
sales and distribution functions with an arm’s length net profit for its
efforts. Year-end adjustments can therefore be seen as a transfer pricing
documentation issue, as the taxpayer is not assessed based on its prelimi-
nary pricing, but on its return position. The focus should therefore be on
whether the group has sufficiently documented the controlled pricing pro-
visions and its efforts to align its preliminary pricing as closely as possible
with the prices that, at the beginning of the year, it estimated would result
in an arm’s length net profit for the tested party.
15.3. Taxpayer-initiated adjustments under US law
The US Internal Revenue Code (IRC) section 482 regulations allow tax-
payer-initiated year-end adjustments.1479 It is stated in the regulations that,
if necessary to reflect an arm’s length profit allocation, a controlled taxpay-
er “may report on a timely filed U.S. income tax return (including exten-
sions) the results of its controlled transactions based upon prices different
from those actually charged”.1480 (Emphasis added)
The provision applies regardless of the type (upwards or downwards) and

form (COGS, royalties, reimbursements, etc.) of adjustment. The quoted
provision was incorporated into the final 1994 US regulations, likely
making the United States the first jurisdiction to adopt a year-end adjust-
ment provision.1481 The introduction of the provision coincided with the
1994 regulatory introduction of the CPM, PSM, the best-method rule and
the commensurate-with-income standard and should be seen in light of
the fact that it would be difficult to properly apply the profit-based pricing
methods, in particular the CPM, without access to year-end adjustments.
1479. Even though year-end adjustments are allowed under the US Internal Revenue
Code (IRC) sec. 482 regulations, such adjustments may be barred or limited by VAT
and customs regulations. For instance, where “property [is] imported into the United
States in a transaction … between related persons”, the taxpayer is not entitled to report
a cost of goods sold (COGS) higher than that used for calculating the value for customs
purposes; see IRC sec. 1059A.
1480. Treas. Regs. § 1.482-1(a)(3). The 1993 temporary regulations (58 FR 5263-02),
section 1.482-1T(e)(2), spoke of “compensating adjustments” when the taxpayer made
adjustments (before filing the tax return) for reimbursements or other payments nec-
essary to ensure that the controlled transactions yielded an arm’s length return. This
terminology was scrapped in the final 1994 regulations (59 FR 34971-01).
1481. See Rosenbloom (2007).
443
IP Pricing
The taxpayer is not allowed to perform a downwards profit adjustment in

an untimely or amended return, but an upwards adjustment is allowed. An
example of the former may, for instance, be if a US taxpayer realizes that it
has made a transfer pricing error by allocating too little profits to a foreign
subsidiary, resulting in an underpayment of foreign tax and overpayment of
US tax. In this situation, there are two main alternatives. First, if the multi-
national files an amended return in the foreign jurisdiction, increasing the
profit allocation, and pays the additional tax, it would incur double taxa-
tion, as the incremental foreign profits also have been taxed in the United
States. This alternative is therefore unfavourable for the multinational.1482
Second, if the multinational does not file an amended return in the foreign
jurisdiction but rather awaits the tax assessment, it could seek relief from
the US competent authorities in the case of a reassessment in the foreign
jurisdiction.
A connected issue is whether the taxpayer may perform an adjustment of

its assessed return position at an even later stage, e.g. in the course of an
examination, during consideration by Appeals or in a docketed case. This
depends on an interpretation of the commensurate-with-income standard
expressed in the second sentence of IRC section 482.1483 Even though the
language of the provision itself may seem to allow otherwise, the author
finds it clear on the basis of the legislative discussions,1484 the White Pa-
per1485 and the IRS interpretation of IRC section 482, as reflected in the
current regulations, that the commensurate-with-income standard entitles
only the IRS (and not the taxpayer) to perform periodic adjustments. The
Office of Chief Counsel of the IRS concluded likewise in a 2007 memo-
randum.1486
The author is not convinced that this asymmetrical solution necessarily is

for the best. From a US perspective, the point of departure is that the tax-
1482. The US competent authority generally accepted cases of taxpayer-initiated ad-

justments if the foreign jurisdiction would grant reciprocity in similar cases, but that
position changed in 2010; see Elliott (2011). There are, however, signals that the US
Internal Revenue Service (IRS) is considering reverting to its previous position.
1483. The second sentence reads: “[I]n the case of any transfer (or license) of intan-
gible property (within the meaning of section 936(h)(3)(B)), the income with respect
to such transfer or license shall be commensurate with the income attributable to the
intangible.”
1484. 1985 House Report, pp. 425-426. The discussion is geared towards preventing
base-eroding outbound transactions of US-developed intangibles. See also the discus-
sion of the US periodic adjustment authority in sec. 16.2.
1485. See, e.g. White Paper (Notice 88-123), at p. 480: “Because Section 482 may be
applied only by the Service […].”
1486. IRS AM 2007-007, issue 1.
444
Taxpayer-initiated adjustments under US law
payer is afforded the right to perform a year-end adjustment, as long as it is

reflected in a timely filed return. The taxpayer also has the right to perform
an adjustment in an amended return, as long as the adjustment increases
US income. What remains is that the taxpayer is cut off from the right to
perform a subsequent adjustment that decreases US income.
The IRS would likely argue that the taxpayer is free to choose a contingent
payment form that will ensure that the profits allocated to the United States
are commensurate with the income generated by the transferred IP. If, on
the other hand, the taxpayer chooses a fixed-payment structure, it chooses
to accept the risk that the actual results may deviate from those projected
at the time at which the agreement was entered into, and the taxpayer must
stand by its original choice. Implicit in this logic lies the assumption that
the multinational will only wish to adjust the pricing if it is beneficial for
it to do so.
For instance, let us say that a US parent sells early-phase, unproven IP to

a foreign subsidiary for 100, which equals the present value of the profits
estimated at the time at which the transaction was entered into. If the actual
profits reaped by the foreign subsidiary from its subsequent exploitation of
the IP are higher than estimated, and thus indicate that the price should
have been 200, it is unlikely that the taxpayer would wish to adjust the
transfer price, as that would increase US income tax. The IRS, on the other
hand, would likely want to do so.
Conversely, if the actual profits are lower than expected and thus indicate
that the price should have been set to only 50, the taxpayer would likely
want to adjust the transfer price in order to lower its US income tax. The
IRS would likely have the taxpayer stand by its initial fixed payment form
“bet”. Thus, the asymmetry of the commensurate-with-income provision
lies in the fact that in the first scenario, the multinational does not wish to
adjust, but the IRS may have the right to do so. In the second scenario, the
multinational wishes to adjust, but it does not have the right to do so.
The author thinks that there is a lot of merit to the argument that the tax-
payer is free to choose the form of controlled compensation. The essence
is that if the taxpayer chooses a fixed-payment structure, it should stand
by it, even if the subsequent actual profits indicate that the fixed payment
should have been lower. However, had the situation instead been that the
actual results indicated that the price should have been higher, the author
finds it unlikely that the IRS would exercise restraint and not carry out a
periodic adjustment.
445
IP Pricing
As the author discusses in chapter 16 on periodic adjustments, the US regu-

lations set the bar high for respecting a fixed-payment structure when the
actual and projected profits diverge significantly. This indicates that the
IRS may make an adjustment even if the actual profits were not reason-
ably foreseeable at the time of the transaction, as long as the extraordinary
events were not entirely beyond the control of the related parties, e.g. if the
increased profits were due to unexpected demands caused by successful
marketing.
In the author’s view, the main problem with the asymmetrical structure
of the commensurate-with-income rule is that it will allow for non-arm’s
length profit allocations in which the taxpayer is not entitled to adjust its
transfer pricing. In the example above, the IP was migrated from the Unit-
ed States for 100, but the lower-than-expected profits indicate that the price
should have been 50. The United States may have been allocated more
income than an application of the transfer pricing methods – taking into
account the actual profit experience – would allow for.1487 In the context of
unique IP, this may entail that intangible value (residual profit) is not al-
located to the jurisdiction where it was created.
The author does not find it apparent that this asymmetry is aligned with
the arm’s length principle, as it purposefully blocks taxpayer-initiated in-
come adjustments that would reflect an arm’s length allocation of income.
He thinks that the achievement of arm’s length results should carry more
weight than the argument that the taxpayer is free to choose the form of
compensation and therefore must accept the risk that follows with it if a
fixed-payment form is adopted.1488
1487. E.g. assume that the original price was set based on the comparable price method
(CPM), designed to leave the transferee with only a normal return on its routine func-
tions, given that its profits turned out as projected. If the foreign transferee’s actual prof-
its turn out lower, but it still paid the original 100 transfer price based on the projected
profits, this will mean that the transfer price steals all, or at least portions, of the normal
return compensation that was intended for the transferee. In other words, the result is
not compatible with an application of the CPM. However, had a subsequent downward
adjustment of the original transfer price from 100 to 50 been performed, that would
have left more profit with the foreign transferee, thus ensuring that the actual result was
compatible with the CPM. A similar issue was raised in GlaxoSmithKline Holdings
(Americas), Inc. v. CIR, 117 TC 1 (see sec. 15.10., with further references).
1488. IRS AM 2007-007 also addressed the connected issue of whether a taxpayer
may apply the commensurate-with-income (CWI) standard to challenge an IRC sec.
482 adjustment in a situation in which the IRS did not itself apply the CWI provi-
sion. The memorandum position is that CWI results may be achieved by the taxpayer
through upfront contingent payment forms that comport with economic substance, and
may therefore not invoke sec. 482 to make a CWI adjustment that enables it to “walk
446
Year-end adjustments under the OECD TPG
15.4. Year-end adjustments under the OECD TPG
The OECD TPG introduced a discussion of year-end compensating ad-

justments in 2010.1489 The text refers to the fact that some OECD member
countries practice the arm’s length principle in a manner that embraces
year-end adjustments, while others do not. The practice of not accept-
ing year-end adjustments is referred to as the “arm’s length price-setting
approach”.1490 The taxpayer must then, in its transfer pricing documenta-
tion, demonstrate that it made reasonable efforts to comply with the arm’s
length principle at the time at which the controlled transaction was carried
out, based on information that was reasonably available at that point.1491
Conversely, the practice of accepting year-end adjustments is referred to
as the “arm’s length outcome-testing approach”.1492 When filing the tax
return, the taxpayer tests the outcome of its actual transactions to demon-
strate that it is consistent with the arm’s length principle.
The OECD TPG take no stance with respect to the priority of the two
approaches.1493 The text recognizes that if compensating adjustments
are permitted only in the residence jurisdiction of one of the parties to
the controlled transaction, double taxation may occur, as corresponding
away from a deal it struck for itself”. It may, however, be of aid to the taxpayer that
Treas. Regs. § 1.482-1(g)(4) provides the right to a setoff against an IRS sec. 482 reas-
sessment for another non-arm’s length transaction (primary transaction). The setoff
and primary transactions must be “between the same controlled taxpayers” and “in
the same taxable year”. Further, the taxpayer must establish “that the transaction that
is the basis of the setoff was not at arm’s length” and must establish “the amount of the
appropriate arm’s length charge” for such setoff transaction.
1489. OECD TPG, paras. 3.70-3.71 and 4.38-4.39. Para. 4.38 refers to the fact that at
least one OECD member country has a procedure under its domestic law for year-end
adjustments. Given the similarity between the wording of the OECD TPG on this point
and § 1.482-1(a)(3) of the US regulations, the OECD member country being referred to
is presumably the United States.
1491. This includes not only information on comparable uncontrolled transactions
(CUTs) from previous years, but also on economic and market changes that may have
occurred between those previous years and the year of the controlled transaction.
1493. OECD TPG, para. 4.39 makes the factual assertion that most OECD member
countries do not recognize compensating adjustments on the ground that the tax return
should reflect the actual transactions carried out. However, as discussed in sec. 15.5.,
the EU Joint Transfer Pricing Forum (JTPF) conducted a recent survey of the domestic
laws of EU Member States (of the 34 OECD member countries, 21 are EU Member
States), indicating that the domestic laws of many EU Member States do indeed allow
year-end adjustments, even if there is no specific legislation in place in most countries.
447
IP Pricing
adjustment relief may be unavailable if no primary adjustment has been

made.1494
The OECD TPG see compensating adjustments as a comparability issue.

The problem with this perspective is whether a year-end adjustment should
be performed in order to align the controlled profit allocation with updated
profit data from third-party comparable uncontrolled transactions that be-
come available after the end of the income year. As commented on in sec-
tion 15.2., this issue strongly resembles a genuine transfer pricing issue, in
contrast to the situation in which there is a need to align the actual con-
trolled profit allocation with the TNMM target return because the actual
results from external factors (e.g. the revenues or costs of the tested party
from its transactions with third parties) deviates from the estimates made
before the start of the income year and on the basis of which the transfer
prices were designed.
Information from contemporaneous third-party transactions is deemed to

be the most reliable information for benchmarking controlled profit alloca-
tions.1495 Such information reflects how independent parties behaved in an
economic environment similar to that in which the controlled transaction
took place.1496 Of course, this information will eventually become avail-
able, but, as recognized in the 2006 OECD comparability report, likely not
until after the taxpayer’s filing date.1497 Outcomes from third-party transac-
1494. The OECD TPG encourage competent authorities to use their best efforts to
resolve double taxation due to different country approaches to year-end adjustments.
Resorting to a mutual agreement procedure (MAP), however, will, in practice, be easier
said than done. Rosenbloom (2007) argues that arts. 9 and 25 of the US Model Income
Tax Convention (the transfer pricing and MAP provision, respectively) contain lan-
guage broad enough to encompass taxpayer adjustments. The same is argued in Rev.
Proc. 2006-54 (now superseded by Rev. Proc. 2015-40). The author is inclined to ap-
ply a similar view with respect to arts. 9 and 25 of the OECD Model Tax Convention
(OECD MTC). The arm’s length income taxed in one jurisdiction under art. 9 (1) shall
be relieved from taxation in the other jurisdiction under art. 9(2). There is no exemp-
tion for cases in which the arm’s length income taxed in the first jurisdiction comes to
be as a result of a taxpayer-initiated year-end adjustment. The broad wording of art. 25
also encompasses taxpayer adjustments, as even matters concerning “double taxation
in cases not provided for in the Convention” are covered; see OECD MTC, art. 25(3).
1495. The OECD has stretched the term “contemporaneous” information further than
the language itself indicates. The 2006 OECD Comparability Report, in the section on
timing issues, at para. 32, states that “contemporaneous” does not necessarily refer to
data from the same tax year. An example is provided in which a transaction that took
place in November 2001 could be more comparable to a transaction that took place in
January 2002 than one that took place in January 2001.
1497. 2006 OECD Comparability Report, section on timing issues, para. 30.
448
Year-end adjustments under the OECD TPG
tions in years prior to the relevant income year may therefore be the only
information available on which to base a TNMM transfer pricing study.
An example contained in the 2006 comparability report illustrates this. In

the late fall of 2004, the taxpayer designs the transfer prices for 2005 based
on third-party accounting data for 2002 and 2003. In the spring of 2006,
when preparing the tax return for 2005, the transfer prices are updated with
third-party data from 2004, which differs significantly from the data from
2002 and 2003 used to design the transfer prices.1498
The 2006 report draws up two alternatives for the taxpayer. The first alter-
native is to consider the 2005 prices actually charged as arm’s length prices
set on the basis of the information that was available at the time at which the
prices were designed (i.e. the third-party profit data from 2002-2003).1499
In this scenario, there will be no need for a year-end adjustment.1500 The
second alternative is to recognize that the transfer prices did not conform
to the arm’s length range based on the third-party information available at
the time of filing the 2005 tax return. A compensating adjustment should
then be carried out to bring the profit margin of the tested party within the
updated arm’s length range. Thus, all information available at the time of
filing the tax return is relevant (including the third-party profit data from
2004).1501
Even though the OECD TPG take no position on whether year-end adjust-
ments must be respected,1502 the transfer pricing methodology embraced
in the OECD TPG may arguably nevertheless insist on a year-end adjust-
ment. It will simply not be possible for a taxpayer to apply the one-sided
methods properly without the ability to perform year-end adjustments.1503
1498. Some years after the filing of the 2005 return, the tax authorities audit the return
on the basis of third-party data from 2005. The data indicates that the prices charged by
the taxpayer fall outside the arm’s length range. This is a transfer pricing problem (see
the discussion of periodic adjustments in sec. 25.5.), not an issue of year-end adjust-
ments.
1499. 2006 OECD Comparability Report, section on timing issues, para. 14.
1500. There might still be a need to perform a year-end adjustment if the actual results
from the tested party’s third-party transactions deviated from the results that were esti-
mated prior to the income year (and on which the transfer prices were based).
1501. 2006 OECD Comparability Report, section on timing issues, para. 18. This in-
formation will normally not include third-party accounting data for 2005, as such data
will typically not be available then; see para. 21.
1502. OECD TPG, paras. 3.67-3.71 and 4.38-4.39.
1503. See the lead-in discussion in sec. 15.2. At the time of filing the tax return, there
may be updated information available on CUTs indicating that the arm’s length range
must be modified.
449
IP Pricing
If a taxpayer performs a year-end adjustment in order to report a taxable

income in line with that which is dictated by an applicable OECD pricing
method, that adjustment should, in the author’s view, prevail with effect for
taxation in both jurisdictions. Otherwise, there could be double taxation of
the adjusted amount.
The author will add that the current OECD TPG wording on year-end
adjustments, limited as it is, cannot easily be reconciled with other, and
older, sections of the OECD TPG.1504 In particular, this was true for the
historical 1995/2010 guidance on periodic adjustments.1505 The traditional
persistence of the OECD with respect to restraining the pricing methods
to only taking into account information available at the time at which the
controlled transaction was entered into was diluted by the introduction of
the TNMM and profit split method (PSM) in 1995, but was still clearly
present when the guidance on compensating adjustments was penned. In
June 2012, the OECD issued a draft on timing issues, proposing revised
language on compensating adjustments.1506 The draft text was not signifi-
cantly different from the existing 2010 text. It reiterated the point of depar-
ture that transfer pricing should be based on data from contemporaneous
uncontrolled transactions, but opened the door for both the arm’s length
price-setting approach and the arm’s length outcome-testing approach.1507
The draft expressed fear that some OECD member countries could be less
willing to accept revisions of the 2010 OECD TPG text on uncertain IP
valuation (in the run-up to the BEPS revision of the intangibles chapter of
the OECD TPG) if the guidance on timing issues took a restrictive posi-
tion on which information could be used to test the reasonableness of ex
ante valuation profit projections. In light of the fact that the OECD finally
managed to reach consensus on new wording on periodic adjustments (in
particular, for hard-to-value intangibles) in the autumn of 2015,1508 the time
1504. See OECD TPG, paras. 3.72-3.73, as well as scattered statements on hindsight;
see paras. 2.130, 3.74, 6.32, 8.20, 9.56, 9.57 and 9.88. See also the annex to ch. IV, para.
49; and the annex to ch. VI (2010 version), para. 4.
1505. The 1995/2010 guidance on periodic adjustments was contained in paras. 6.28-
6.35. The 2015 guidance encompasses paras. 6.181-6.195.
1506. OECD 2012 Draft on timing issues relating to transfer pricing (2012 OECD tim-
ing draft). This draft was issued on the same day that the OECD released its 2012 Dis-
cussion draft Revision of the special considerations for intangibles in Chapter VI of the
OECD Transfer Pricing Guidelines and Related Provisions (2012D).The coincidence of
these drafts were not accidental, as timing issues were relevant to certain controversial
issues on the transfer pricing of intangibles (e.g. periodic adjustments).
1507. 2012 OECD timing draft, para. 3.68.
1508. OECD TPG, paras. 6.181-6.195. See the discussion in sec. 16.5. on the OECD’s
periodic adjustment provision.
450
Year-end adjustments in the European Union
is now ripe for a revision of the OECD TPG wording on compensating

adjustments to bring more clarity to the rules. A natural point of departure
for the OECD would be to embrace the work carried out by the EU JTPF
(discussed in section 15.5.).
15.5. Year-end adjustments in the European Union
In October 2002, the European Commission established the EU JTPF to

find pragmatic solutions to problems arising from the application of the
arm’s length principle within the European Union.1509
In 2011, the JTPF carried out a survey of the domestic tax laws of EU
Member States on year-end adjustments. The responses reflect the situa-
tion as of 1 July 2011 and present an interesting and fairly recent empiri-
cal picture of the domestic law in the European Union.1510 To provide a
brief overview, no EU Member States had legislation that prevents year-
end adjustments. Of the 27 respondents, 15 Member States generally ac-
cepted year-end adjustments.1511 All 15 Member States allowed compen-
sating adjustments to be carried out before the closure of the financial
statements.1512 Most Member States regarded a deviation from an arm’s
length result as a sufficient reason to perform a year-end adjustment. The
Member States were asked whether an intercompany agreement was nec-
essary for claiming a year-end adjustment, but no clear response was pro-
vided in the survey summary. The majority of Member States did not
have provisions explicitly obliging taxpayers to make a compensating
adjustment.1513
1509. The JTPF is an expert group, with one representative from each EU Member
State and 16 non-government members, chaired by an independent person.
1510. JTPF/019/REV1/2011/EN.
1511. These 15 states have legal guidance or administrative practice in place governing
year-end adjustments. Some of these states require year-end adjustments to be reflected
in the financial statements or only allow adjustments where third parties would have
been contractually obliged to carry out a year-end adjustment, or under exceptional
circumstances.
1512. In some states, the permissibility of year-end adjustments after the closing of the
financial statements depends on whether the adjustments result in lower or higher taxes.
Further, the majority of the 15 states allow compensating adjustments before the tax
return is filed, but only a few allow adjustments subsequent to the filing.
1513. Some EU Member States, however, see the performance of a year-end adjust-
ment as a necessary consequence of the general requirement to conform with the arm’s
length principle.
451
IP Pricing
A report on compensating adjustments was issued in November 2013.1514

The conclusion of the report was that a taxpayer-initiated year-end adjust-
ment, downwards as well as upwards, shall be accepted if the same prices
are reported by each of the Member States involved, and1515 (i) the taxpayer
makes reasonable efforts to achieve an arm’s length outcome before the
relevant transactions are carried out and describes this in the transfer pric-
ing documentation; (ii) the taxpayer makes the adjustment symmetrically
in the financial statements of both Member States involved; (iii) the same
approach is applied by the taxpayer over time; (iv) the adjustment is made
before filing the tax return; and (v) the taxpayer is able to explain why the
forecast did not match the actual result, when required by the internal leg-
islation of at least one of the Member States involved. The report takes the
stance that the adjustment should be, to the most appropriate point, in the
arm’s length range.1516
15.6. Case law: Vingcard (Norway, 2012)
At issue in the Norwegian Vingcard Elsafe AS v. Skatt Øst case was wheth-
er the taxable income of a Norwegian parent could be reduced through
year-end adjustments.1517 The reassessment disallowed the adjustments in
full. It was upheld by the first-tier Oslo City Court, while the Borgarting
Appellant Court upheld the reassessment for one of the two income years
1514. JTPF/009/FINAL/2013/EN. In June 2014, the European Commission invited

the European Council to endorse the report, and the EU Member States to implement
its solutions in their domestic laws (COM(2014) 315 Final, para. 3, second section).
The report was endorsed by the European Council in March 2015 (7249/15). The report
recommendations are applicable only to compensating adjustments that the taxpayer
has reflected in its financial statements and are elaborated on in its transfer pricing
documentation; see para. 8 of the report. They have no bearing on the ability of a tax
administration to make an audit adjustment at a later stage or in MAPs; see para. 19
of the report. The report recognizes that the different approaches of the EU Member
States (i.e. the price-setting versus outcome-testing approach) were often grounded in
disparate understandings of fundamental transfer pricing concepts, including timing is-
sues, information on contemporaneous CUTs, as well as what constitutes inappropriate
use of hindsight; see para. 6 of the report.
1515. JTPF/009/FINAL/2013/EN, para. 17. The decision of whether to oblige a tax-
payer to make an adjustment is left to the discretion of the EU Member States (para.
16).
1516. JTPF/009/FINAL/2013/EN, para. 18. See also OECD TPG, para. 3.55.
1517. Ruling by Borgarting Appellant Court dated 18 June 2012 (Utv. 2012 p. 1191),
reversing in part the ruling by Oslo City Court dated 29 November 2010 (Utv. 2010
p. 1690).
452
Case law: Vingcard (Norway, 2012)
at issue.1518 The case illustrates several interesting aspects of year-end ad-

justments. The author therefore finds an analysis justified. He will outline
the factual pattern of the case in section 15.6.2. The stances of the Oslo
City Court and the Borgarting Appellant Court on the use of comparables
are discussed in section 15.6.3., and their stances on the application of the
TNMM for allocating income are discussed in section 15.6.4.
15.6.2. The factual pattern of the case
The Norwegian company Vingcard Elsafe AS (VEAS) was leading in the

field of card-based security systems for hotels and cruise liners. It carried
out product development and owned all unique IP employed in the value
chain. 75% of its products were sold through local group distribution enti-
ties. The United States was the most important market for VEAS. Local
distribution was carried out by the subsidiary Assa Abloy Hospitality Inc
(AAH). The US hotel market had a downturn during the 2000s due to the
11 September 2001 attack, impacting the sales of AAH negatively.
The profit allocation among VEAS and AAH was based on official price
lists determined by VEAS, designed to provide its local distribution enti-
ties with an arm’s length gross profit margin pursuant to the resale price
method. This transfer pricing policy became difficult to realize without
year-end compensating adjustments. In early 2005, in connection with the
preparation of the 2004 financial statements, it became clear that AAH
would incur a significant loss due to reduced sales and increased costs. To
align its taxable result with the transfer pricing policy of the group, it was
decided that VEAS would, with effect for 2004, compensate AAH for the
difference between the budgeted and actual gross profit margin of AAH
(33% and 23%, respectively). As a result, VEAS went from positive to
negative taxable income in Norway, while the US income of AAH went
from a negative to a (relatively small) positive result.
For the income year 2005, the group adopted a TNMM-based profit alloca-
tion, based on the advice in a report by PricewaterhouseCoopers.1519 This
1518. The case was decided under the domestic arm’s length provision in sec. 13-1 of
the Norwegian Tax Act (NTA), for which the significant source of law for interpretation
in international transactions are the OECD TPG.
1519. The PwC report compared the net profit margin of the aggregated profits of
AAH with the net profit margin of the aggregated profits from the selected 18 European
distributors. Even though not found to be a methodological error in and of itself, the
Court found the comparison of blended net profit margins to add further uncertainty
453
IP Pricing
pricing guaranteed AAH an annual net profit margin between 1% and 5%,
ensuring that it would always generate a positive net margin. As the actual
2005 result of AAH was not sufficient to meet the agreed TNMM profit
interval, VEAS compensated AAH with approximately NOK 26.4 million,
bringing the 2005 net profit margin of AAH up from 3.8% to 1%.
15.6.3. The comparables supporting the taxpayer profit

allocation
One of the two main issues in the case was whether the reassessment
should be upheld, on the basis that the third-party profit data used under the
taxpayer’s application of the resale price method and TNMM did not meet
the comparability requirements of the respective methodologies. The Oslo
City Court rejected the profit data extracted from the unrelated distributors
selected by the taxpayer as comparables because the distributors operated
in other business sectors than AAH did. It had not been proven that the dis-
tributors were afflicted by a market downturn similar to that which struck
the hotel sector subsequent to 11 September 2001. As the taxpayer had not
performed any comparability adjustments that satisfied the Court, the year-
end adjustments were rejected.
The author does not share the view of the Court. First, there was undoubt-
edly a difference between the law as it was written in the 1995 OECD TPG
and the law as it was practiced by multinationals and tax authorities in the
mid-2000s. Subsequent to the adoption of the 1995 OECD TPG, global
use of the TNMM became widespread.1520 The practice of using blended
third-party profit margins as comparables was one of the reasons behind
the revision of the 1995 OECD TPG text on comparability and the profit-
based methods.1521 Thus, during the income years at issue, the OECD was
to the transfer pricing assessment. In the author’s view, the Court should have pointed
out that this was not a problem here, as AAH carried out only transactions relevant to
the transfer pricing assessment. The Court referenced para. 3.42 of the 1995 OECD
TPG, which required that “profits attributable to transactions that are not similar to the
controlled transactions under examination should be excluded from the comparison”.
However, the attorney general did not protest the examination on this ground. The
Court therefore accepted that the blended third-party net profit margins could be used
as the arm’s length range under the TNMM.
1520. This assumption is consistent with the business comments referred to in the 2006
OECD Comparability Report in the section on aggregation of transactions, at paras.
8-9. See also the 2008 OECD discussion draft, at para. 115.
1521. See the preface of the 2006 OECD Comparability Report, as well as the section on
putting a comparability analysis and search for comparables into perspective, at para. 8.
454
in the process of relaxing the comparability requirements relevant to the

TNMM. In this context, the wording of the 1995 OECD TPG should not
have been interpreted narrowly. Both multinationals and tax authorities
normally seek to adjust their transfer pricing positions based on signals
from the OECD, even if those signals have yet to materialize in adopted
consensus texts.
Second, AAH should only be entitled to a normal return for its routine
marketing, resale and distribution functions. A comparability adjustment
could have, in this case, resulted in negative margins for AAH, or at least
margins close to zero. Such a result would be unreasonable. No unrelated
party would be willing to perform marketing, resale and distribution ser-
vices close to free of charge or at a loss. Further, the overall differences
between AAH as the tested party and the selected comparables seem so
notable that it is not clear whether an adjustment for decreased revenues
would have provided a more reliable result. On top of the hotel market
decline, there were likely also other significant differences, e.g. different
geographical markets, different market types, different products and dif-
ferent relationships to suppliers. The comparability adjustment requested
by the Court as justification for its rejection of the comparables would, in
the author’s view, only be meaningful if the comparables had been Euro-
pean distributors of customs security systems for hotel chains and cruise
liners.
Third, the TNMM pricing was restrained by the interquartile arm’s length
range. This compensated for potential material differences between the
tested party and the selected comparables.
In contrast to the Oslo City Court, Borgarting Appellant Court found

the selection of TNMM comparables to be reasonably representative,
and thus upheld the TNMM pricing. For the reasons above, the author
agrees.
This leads to the next point. The concurrent pricing between AAH and
VEAS throughout 2004 and 2005 was based on the USD price list. The
TNMM was applied indirectly through the year-end adjustment, which
aligned the initial USD price list-based transfer pricing result to what the
profit allocation would have been had the TNMM been applied at the out-
set. When the Oslo City court disallowed the use of year-end adjustments,
it therefore effectively based the transfer pricing solely on the USD price
list, i.e. on the comparable uncontrolled price (CUP) method with the USD
455
IP Pricing
price list as the comparable transaction.1522 In light of the fact that the USD
and EUR price lists were standard terms used by VEAS for sales to related
and unrelated distributors,1523 the author questions whether this yielded a
more reliable pricing method than the TNMM with the 18 European dis-
tributors as comparables would have.
First, the prices charged to unrelated distributors were increased by 10-

16% (relative to prices charged to related distributors) to compensate for
marginal administration expenses and small-volume purchases. Converse-
ly, AAH was the exclusive distributor in the largest and most important
market for the VEAS products. It is unlikely that such a distributor would
be willing to purchase at the same prices used for marginal distributors.1524
If no adjustment were made for this fact, the prices should be rejected as
comparables.1525
1522. The Oslo City Court found the USD price list to be the best method to determine
the transfer prices. The Court specifically stated that it did not find relevant the question
of whether the price list qualified as a comparable under the comparable uncontrolled
price (CUP) method of the 1995 OECD TPG. The reason for this was that the status of
the 1995 OECD TPG as a source of law under sec. 13-1 of the NTA was, in principle,
only guiding and not binding. The Court found the basis for this understanding in the
2001 Norwegian Supreme Court ruling in Norsk Agip AS v. The State, Rt 2001 1265.
Subsequent to Agip, sec. 13-1(4) was introduced with effect from 2008, enhancing the
position of the OECD TPG as a source of law under domestic Norwegian law. With re-
spect to Vingcard, the author finds, for the reasons stated in the text, that the unadjusted
USD price list did not qualify as a comparable under the CUP method of the 1995
OECD TPG. It was therefore a significant breach of the OECD TPG when the Oslo
City Court allocated income to the controlled parties based on an erroneous application
of the CUP method. The author finds it clear that there was no legal basis in the Agip
ruling on which to apply a transfer pricing method under sec. 13-1 of the NTA that was
directly contrary to fundamental principles drawn up in the OECD TPG.
1523. The price lists were, however, mainly used for determining transfer prices for
goods sold within the group. Of course, the price terms used for group distributors are
irrelevant under the CUP method, as they cannot be regarded as uncontrolled transac-
tions.
1524. A useful approach would, in the author’s view, be to see the issue of adopting a
discount as a comparability adjustment to the USD list price with the 10-16% increase.
The question would simply be whether that comparable had to be adjusted for the fact
that AAH had a significantly greater bargaining position towards VEAS than any of
the unrelated distributors and whether VEAS would have had a clear self-interest in
keeping AAH afloat through a difficult period in order to retain its footing in the US
market. In the author’s view, the answer to this question is likely a yes. Para. 2.9 of the
1995 OECD TPG stated that every effort should be made to adjust the data so that it
may be used appropriately with the CUP method. As no comparability adjustment for
a discount was adopted, the author finds it highly doubtful as to whether the USD price
list could actually be regarded as a valid comparable under the CUP method.
1525. The reassessment would then be based on terms used in the controlled – not the
uncontrolled – agreements of the group. It is unacceptable to use a controlled transac-
tion as a comparable under the arm’s length principle.
456
Second, the CUP method resulted in lower-than-budgeted gross profits and

a negative net return for AAH.1526 Consistent negative returns over a 6-year
period, two of which were at issue, can hardly be considered realistic re-
turns for an unrelated distributor. Further, the Court failed to recognize
that the taxpayer’s TNMM allocation, on an overall basis, provided a more
reliable result than the CUP allocation. There was clearly functional com-
parability between AAH and the European distributors selected as compa-
rables. Further, application of the TNMM as a benchmark for determining
the size of the year-end adjustment provided AAH with a normal return.1527
The author finds it more plausible that an unrelated distributor would be
willing to undertake distribution functions for the low positive net result
yielded by the TNMM than for the negative net result yielded by the CUP
method.
The author’s impression is therefore that the Oslo City Court’s applica-
tion of the CUP method may have been based on insufficiently comparable
transactions and resulted in a non-arm’s length profit allocation.
15.6.4. The contractual risk allocation
The second issue of the case pertained to the allocation of risk among
VEAS and AAH. Even though the Oslo City Court rejected the TNMM
comparables, the bearing basis for its dismissal of the year-end adjustments
was that the business risks of AAH could not, with effect for taxation, be
contractually allocated to VEAS to the extent done in the controlled dis-
tribution agreement.
The Court stated that “when applying the TNMM it must be adjusted for
failures that are due to the circumstances of the buyer, which the com-
pensating adjustments did not take into account”.1528 Further, it stated that
1526. The negative results came from declining sales and increased operating expens-
es. AAH also generated negative results in 2001-2006, clearly indicating that the list
prices were too high.
1527. For 2005, the adjusted net profit margin of AAH was 1%. Such a modest return
should not be viewed as anything but an honest attempt by the taxpayer to allocate an
arm’s length result to the US group entity. Perhaps even this result is too low for an
unrelated distributor to be interested in distributing the VEAS products over a longer
period of time, but at least there is data suggesting that unrelated and functionally simi-
lar distributors earn a comparable net profit.
1528. Unofficial translation of the following Norwegian wording from the ruling: “Ved
bruk av TNMM må det korrigeres for resultatsvikt som skyldes kjøpers egne forhold,
noe priskorreksjonene ikke tok høyde for.”
457
IP Pricing
the responsibility for risks casually linked to AAH (e.g. bad management
and lack of control over spending) could not be contractually allocated to
VEAS. The Court asserted that AAH’s negative 2004 and 2005 results
were due to an increase in costs that were not directly related to products
purchased from VEAS, e.g. installation service costs and inventory write-
downs.1529 Further, it asserted that AAH should have borne the risk of the
market decline in the hotel sector,1530 and that a “floating guarantee” for
AAH to earn a positive net profit margin could not be established.
The author takes issue with this. First, the Court did not provide a legal
basis for its assertion. It almost seems as if it based its views on an analogy
for the principles of general contract law. This is not a sensible interpreta-
tion of the 1995 OECD TPG.1531 In the author’s view, the Court should
have recognized that a tested party under the TNMM would be allocated
a modest and fixed net profit margin every year. The tested party had no
risk of being allocated a negative result under the TNMM. The flip side
of this is that it had no real upside; its income would never exceed the
agreed arm’s length net profit margin. All residual profits were allocated
to the other party to the controlled agreement, i.e. the entrepreneur. That
income will vary from year to year, from potentially negative to sizeable.
The entrepreneur therefore incurs significant risks. A year-end adjustment
was therefore literally the price that VEAS had to pay for the privilege of
being the risk-bearer in the controlled agreement and for being entitled to
the residual profits.
There is asymmetry in the position of the Oslo City Court. On the one
side, it refused to accept that AAH was stripped of risks, entailing that
AAH could incur losses. On the other side, it was seemingly fine with ac-
cepting that AAH had no possibility of earning a profit above the normal
return of the agreed interquartile arm’s length net profit range and that
VEAS was allocated the residual profits in good years. This asymmetrical
solution entailed that AAH would not earn a taxable income reflective of
its risks.
1529. Based on the accounting data referred in the ruling, the author finds it clear
that the main cause of AAH’s negative results were its operating expenses (including
marketing, resale and distribution costs), as opposed to the price of products purchased
from VEAS.
1530. If so, the reasoning of the Court was inconsistent, as the drop in AAH’s sales was
due to a general market decline that also affected other distributors.
1531. It is unclear as to which norm the Court applied to distinguish the costs that
could be validly allocated to VEAS from those that could not.
458
The Borgarting Appellant Court sided with the taxpayer on the risk alloca-
tion issue,1532 based on a functional analysis in which VEAS was identified
as the entrepreneur of the group. The Court found that VEAS, in reality,
carried the significant value chain responsibilities, e.g. for installation and
maintenance services. However, as the TNMM-based distribution agree-
ment with year-end adjustments was not in place until 2005, the Court ac-
cepted the 2004 reassessment, in which the allocation of income between
VEAS and AAH was based on the USD price list, with no year-end adjust-
ment. The taxpayer allocation based on applying the USD price list with a
year-end compensating adjustment was upheld for 2005.1533
The controlled pricing structure in this case was beneficial for Norway. It
is easy to overlook this fact, as the tested party lost money in the income
years at issue. However, had the situation instead been that AAH made
significant profits, the TNMM would have ensured extraction of the re-
sidual profits from the United States to Norway, leaving the United States
with only a normal return from local sales. The pricing logic is that the
Norwegian entity was the risk bearer. It was precisely this entity’s risk
of incurring a negative result that materialized in the income years under
review. In the author’s opinion, the Norwegian tax authorities should not
have challenged the controlled profit allocation, based on the recognition
that the Norwegian entity indeed was entitled to residual profits in good
years.
Had the reassessment not been partially overturned by the Borgating Ap-
pellant Court, it would have been a win-win for Norway, i.e. residual profits
in good years and no risk of losses in bad years. Such legally unfounded
1532. The attorney general argued that the controlled risk allocation should be reject-
ed, as it would allow AAH to enjoy high profits in periods with positive results and to
“pass on the bill” to VEAS in periods with negative results. Apart from being positively
erroneous, the argument is of such low quality that it should not have been presented
before a court of law. The state further argued that an unrelated distributor would have
the risk of incurring negative results, while AAH would always end up with a posi-
tive net result. That argument is, in and of itself, correct, but of course, an unrelated
distributor with full risk would also have the possibility of earning a return above the
interquartile arm’s length net profit margin. That was not the case for AAH.
1533. Like the Oslo City Court, the Borgarting Appellant Court viewed the reimburse-
ment of installation, guarantee and general ledger costs as a contribution of capital.
The Appellant Court’s assessment seems influenced by the fact that there were some
ambiguities concerning the reimbursed costs that the taxpayer had not sufficiently at-
tempted to clarify.
459
IP Pricing
and unbalanced pricing assertions should be avoided. The reassessment

position of the Norwegian tax authorities also seems unacceptable from the
point of view of the United States. A local low-risk distribution entity that
consistently reports losses would likely attract US audit attention, on the
basis that such an entity should not incur losses at all.1534
15.7. Case law: ITCO (Italy, 2010)
A recent Italian case illustrates a downward profit adjustment for a tested

party,1535 a situation opposite of that in Vingcard. An Italian distribution
subsidiary purchased software products for local resale from a UK sister
company. On the last day of the 2004 fiscal year, the UK company sent
an invoice to the Italian subsidiary for the amount of GBP 947,456. This
was an upward price adjustment for products sold during the year. The tax
authorities disregarded the year-end adjustment under the rationale that the
transfer of income from Italy to the United Kingdom constituted tax avoid-
ance under the domestic general anti-avoidance rule.1536
15.8. Case law: H1 A/S (Denmark, 2010)
A Danish 2010 Western High Court ruling pertained to rent payments

from a subsidiary to a parent in connection with the lease of a discotheque
in 2001 and 2002.1537 In the original agreement, dated 22 March 2001,
the monthly rent was set to DKK 300,000. During 2001, the subsidiary
incurred expenses to decorate the locales and purchase inventory. On
1534. This point is underlined by the consistent losses of the US distribution entity in
the period of 2001-2006.
1535. Ministry of Finance (Tax Office) v. ITCO, Case No. 11949 (Supreme Court of
Italy, 2012). The case was tried as an anti-avoidance matter under the local general anti-
avoidance rule. The legal arguments are therefore not directly relevant to the discussion
in this chapter.
1536. The argument was founded on the fact that (i) the adjustment was performed on
the last day of the fiscal year; (ii) it was an adjustment of prices actually charged; and
(iii) the adjusted prices deviated from the average purchase price for goods paid by the
subsidiary. The taxpayer appealed the reassessment to the Provincial Tax Court, where
it prevailed. The Lombardy Regional Tax Court rejected the appeal of the tax authori-
ties. The case was then appealed by the tax authorities to the Supreme Court regarding
a burden-of-proof issue, which found that this particular case pertained to the issue of
proving the existence and relevance of intra-group costs. For that question, the burden
of proof rested with the taxpayer.
1537. H1 A/S v. Skatteministeriet, SKM2010.455.VLR (Vestre Landsret, 2010). For
comments on the case, see Wittendorff (2010d); and Pedersen (2010).
460
Case law: H1.1.1 A/S (Denmark, 2012)
2 January 2002, it was agreed that the total rent for 2001 should be set
to DKK 975,000, a reduction of approximately 64% relative to the origi-
nal agreement, due to market conditions and the level of competition. The
agreement was also amended so that the concurrent rent payments should
be regarded as preliminary payments and that the final annual rent should
be determined on a yearly basis upon the consideration of market condi-
tions. It was not disputed that the post-adjustment rent was at arm’s length.
With respect to 2001, the Court found that there was no legal basis for
the year-end adjustment. With respect to 2002, for which the amended
agreement was in effect, the Court found, under a concrete interpretation,
that the wording of the agreement was too imprecise to justify a year-end
adjustment. The reasoning of the Court has, and justly so, been subject to
criticism.1538
15.9. Case law: H1.1.1 A/S (Denmark, 2012)
The author will also briefly mention a 2012 ruling by the Supreme Court
of Denmark pertaining to the 1999 and 2000 tax assessments of a Danish
insurance subsidiary.1539 At issue was the transfer pricing treatment of a
debt instrument used by the subsidiary as part-financing for the purchase
of shares from its US parent.1540 The tax authorities decreased the 1999
income of the subsidiary though imputed interest expenses on the deemed
loan, calculated from 1 July to 15 October, and increased its 2000 income
by the same amount, effectively moving the deductions from 2000 to 1999.
The Danish subsidiary was unable to utilize the deductions in the 1999
income period. The Supreme Court upheld the reassessment. The evidence
did not support the taxpayer’s assertion that a binding agreement for com-
pensation in the form of a zero-coupon note was entered into before 15 Oc-
tober 1999.
1538. Id.
1539. H1.1.1 A/S v. Skatteministeriet, SKM2012.92.HR (Supreme Court of Denmark,
2012). For comments on the case, see Wittendorff (2012b); and Wittendorff (2012c).
The case primarily turned on an interpretation of the statute of limitations for making
transfer pricing adjustments under Danish law. Wittendorff sees the case as a retroac-
tive transfer pricing adjustment. The author does not fully agree, as the Court relied on
a rather concrete interpretation of the controlled agreements. Further, the case does not
pertain to the issue of aligning the actual results of the taxpayer with the result dictated
by a transfer pricing method.
1540. In June 1999, it was clear that the purchase would be partly loan-financed. The
form of compensation was not final at that time. It first became clear on 15 October
1999 that the loan should be in the form of a zero-coupon note.
461
IP Pricing
15.10. The US GlaxoSmithKline settlement (2006)
In part 3 of the book (on IP ownership), the author will discuss the nota-
ble US case of GlaxoSmithKline Holdings (Americas), Inc. v. CIR, which
ended in a settlement in 2006.1541 Among the issues in that case was the al-
location of profits to a US distribution subsidiary (GSK US) in the GlaxoS-
mithKline group with respect to the distribution and sale of the blockbuster
drug Zantac. The group allocated only a normal market return to the US
distribution entity based on the assertion that it only provided routine func-
tions to the value chain and that all unique IP was owned by the UK parent
(GSK UK).1542 The profits of the US entity were reduced in the income
years at issue through what were essentially year-end adjustments in order
to ensure full extraction of the residual profits from the US to the United
Kingdom.
In order to reduce the profits of the subsidiary to what the GSK group
deemed to be “appropriate gross margins”, a licence agreement was in-
troduced.1543 To achieve concurrent alignment of the gross profit margins
actually earned by the subsidiary with the appropriate gross profit margins,
the royalty rates and prices were adjusted, presumably on an annual basis.
The adjustments increased the prices for goods sold, as well as royalties
for the licensing of manufacturing and marketing intangibles charged, by
the UK parent to the US subsidiary. GSK commenced its US distribution
activities in 1978. By 1987, GSK US had become immensely profitable.
The taxpayer’s view was that without adjusting the prices for goods and of
the royalties charged by GSK UK, the US subsidiary would have earned
residual profits, for which there was no transfer pricing justification. The
IRS took issue with this on the basis that no increase in the value of the
licensed patents or other manufacturing IP had occurred that justified such
upward adjustments of the royalty rates under the licence agreement.1544
1541. GlaxoSmithKline Holdings (Americas), Inc. v. CIR, 117 TC 1 (Tax Ct. Docket
No. 5750-04, 2004).
1542. See the discussion in sec. 19.2.5.2.
1543. See the taxpayer petition dated 2 April 2004 (available at Tax Analysts, doc.
2004-7600), at p. 20 (sec. x).
1544. The Notice of deficiency disregarded the periodic upward adjustments of the
royalty rates because they were “inconsistent with dealings between arm’s length enti-
ties and because there was no increase in the value of the licensed patents and other
manufacturing intangibles or in the income attributable to these intangibles”; see the
Notice of deficiency, dated 6 January 2004 (included as exhibit A in the taxpayer peti-
tion dated 2 April 2004 (id.)), explanation of adjustments, at sec. b) (Royalties). GSK,
on the other hand, argued that the CWI standard “required moderation of Petitioner’s
profit levels to reflect the income commensurate with Glaxo Group’s intangible con-
462
The US GlaxoSmithKline settlement (2006)
GSK argued that the income allocated to the US entity should be consist-
ent with its functions performed, assets used and risks incurred. In other
words, as a routine distribution entity, it should be stripped of all residual
profits and left with only a normal market return. This allocation pattern
was achieved by applying a tailored “resale minus pricing” method, under
which the parent charged the subsidiary a percentage of the end market
price for the particular product,1545 designed to leave it “with a gross profit
margin for the product portfolio sufficient to cover the costs of its activities
and earn at least an appropriate profit”.1546 Thus, parallel to the allocation
pattern of the CPM, the subsidiary was treated as the tested party and allo-
cated a return based on the earnings of purported comparable distributors.
The author finds it unlikely that the taxpayer and the IRS were completely
accurate in their assertions. The most interesting question was not wheth-
er adjustments to the royalty rate could be made, but the extent to which
they could be made, i.e. whether the pricing that resulted from the adjust-
ments was at arm’s length. Further, the US marketing and sales activity
performed by, as well as the sales and assets of the US entity, gradually
expanded. In the author’s view, this indicated that GSK, for instance, by
applying the PSM, would be entitled to allocate more residual profits to the
parent company, given that the cause of the increased profits indeed were
the manufacturing intangibles owned by the parent. For the income years
1989-1996, the return position of GSK was to allocate 31% of the total
residual profits from US sales of USD 12.7 billion to the subsidiary. This
approximate 30/70 position in favour of the United Kingdom was based
on the view that most of the residual profits were due to patents and other
manufacturing intangibles owned in the United Kingdom.
Based on the information contained in the court documents, it seems clear

that the parent created the patent for Zantac and other relevant pharmaceu-
tical products through its own R&D efforts in the United Kingdom. The
subsidiary only carried out secondary manufacturing and did not provide
any unique inputs in this respect. The author therefore finds it likely that
the parent was entitled to the entire residual profits from the manufactur-
ing IP licensed to the subsidiary. In other words, the resale-minus method
advocated by GSK, or any other method that treated the subsidiary as the
tributions to U.S. sales”, that “Zantac is exactly the type of product contemplated by
Congress when I.R.C. section 482 was amended in 1986” and that “the 1987 License
achieved the objective of giving additional returns to Glaxo U.K., the inventor of the
intangible”; see taxpayer petition dated 2 April 2004 (id.), at p. 20 (sec. z).
1545. See taxpayer petition dated 2 April 2004 (id.), at p. 18.
1546. Id.
463
IP Pricing
tested party for that matter, could be applied to set the royalty rate for
these intangibles. However, as the US distribution subsidiary likely owned
unique marketing intangibles, it could not be remunerated based on a one-
sided method. A split of the residual profits would be necessary, entailing
application of the PSM. In principle, a year-end adjustment in the form
carried out by GSK should then not be acceptable. However, in the author’s
view, there was good reason to adjust the royalty charged to the US subsidi-
ary, provided that its increase in profits was due to an increase in the value
of the UK manufacturing intangible relative to the US marketing IP.
464
Chapter 16
Periodic Adjustments of Controlled IP Transfer Pricing
16.1. Introduction
In this chapter, the author will discuss the periodic adjustment provision
of the US regulations and the OECD Transfer Pricing Guidelines (OECD
TPG), both with respect to intangibles transfers in general and in the spe-
cific context of cost-sharing arrangement (CSA) buy-in payments.1547 “Peri-
odic adjustment” refers to the question of whether a controlled intangibles
transfer with a fixed-price term that was at arm’s length at the outset of the
agreement can be adjusted in later income periods if subsequent events
cause the profit allocation to no longer be at arm’s length. Without a pe-
riodic adjustment provision, such transfers would be shielded from reas-
sessments. The result of a periodic adjustment is simply to ensure that the
controlled profit allocation is at arm’s length for the income periods under
review.
The author begins by commenting on the development of the US com-

mensurate-with-income standard in section 16.2. The periodic adjustment
provision under the current US regulations is discussed in section 16.3. The
application of this provision to lump-sum payments is discussed in section
16.4. The author then turns to the OECD’s periodic adjustment provision in
section 16.5. Periodic adjustment of CSA buy-ins is discussed in sections
16.6. and 16.7. for the US regulations and the OECD TPG, respectively.
16.2. The development of the US periodic adjustment

concept
The author will tie some comments to the legislative development that
eventually resulted in the periodic adjustment provision of the current 1994
1547. For early commentaries on periodic adjustments, see, e.g. Levey et al. (1987),
at p. 636; Lashbrooke (1989), p. 189; McSchan (1989); Ungerman (1989), at sec. II.B.;
and Clark (1993), at sec. II. For relatively recent commentaries on periodic adjustments,
see, e.g. Navarro (2017), at p. 242; Wittendorff (2010a), at pp. 675-694; Bullen (2010),
at pp. 326-333; Martinez (2010); and Andrus (2007), at pp. 647-650.
465
Chapter 16 - Periodic Adjustments of Controlled IP Transfer Pricing
US regulations. He will comment on the 1985 House Report and the 1988
White Paper in section 16.2.2., on the relationship between the transfer
pricing methods and the periodic adjustment provision in the White Paper
in section 16.2.3. and on the exceptions from the White Paper’s periodic
adjustment provision in section 16.2.4.
16.2.2. The 1985 House Report and the 1988 White Paper
The 1985 House Report recognized the problem of multinationals deliber-

ately shifting intangible income out of the United States in order to lower
their effective tax rates. At the time, typical BEPS transactions included
“roundtrips” and early-stage transfers of unproven US-developed intangi-
bles for a low price, where the multinationals subsequently would claim
that it was not possible to foresee the profit potential of the intangibles at
the time of the transfer. The House Report concluded as follows:
The committee does not intend, however, that the inquiry as to the appropriate
compensation for the intangible be limited to the question of whether it was
appropriate considering only the facts in existence at the time of the trans-
fer. The committee intends that consideration also be given the actual profit
experience realized as a consequence of the transfer. Thus, the committee
intends to require that the payments made for the intangible be adjusted over
time to reflect changes in the income attributable to the intangible. (Emphasis
added) 1548
Thus, the conclusion of the 1985 House report was that payments for mi-
grated intangibles should be commensurate with the actual profits attribut-
able to the transferred intangibles. The legislative discussion resulted in
the following sentence being added to US Internal Revenue Code (IRC)
section 482 in the 1986 tax reform: “In the case of any transfer (or license)
of intangible property (within the meaning of section 936(h)(3)(B)), the
income with respect to such transfer or license shall be commensurate with
the income attributable to the intangible.”
No detailed discussion of the periodic adjustment authority was provided

until the 1988 White Paper addressed the issue. The White Paper found,
with reference to the 1985 House Report, that an intangible transfer that is
commensurate with the income attributable to the intangible must reflect
“the actual profit experience realized as a consequence of the transfer”.1549
1548. 1985 House Report, at pp. 425-426.

1549. White Paper (Notice 88-123), at p. 63.
466
The development of the US periodic adjustment concept
The core of the periodic adjustment discussion in the White Paper per-
tained to adjustments of controlled intangibles transfers for non-contingent
lump-sum royalty or sale payments (in other words, fixed-price arrange-
ments). According to the White Paper, periodic adjustments were neces-
sary to reflect substantial changes in the income stream from a transferred
intangible, taking into account the activities performed, assets employed
and risks borne by the related parties.1550 Substantial changes in the income
stream referred to the level of discrepancy between the income projected at
the time of transfer and the actual income subsequently realized.
The White Paper found that periodic adjustments were consistent with the
arm’s length principle for two reasons.1551 First, unrelated parties generally
provide some mechanism to adjust for change in the profitability of trans-
ferred intangibles. Second, the actual profit experience is generally the best
indication available of the anticipated profit experience that unrelated par-
ties would have taken into account at the outset of the arrangement.
The White Paper tied its discussion of periodic adjustments to two base
scenarios. The first base scenario pertained to a long-term licence with
a fixed royalty, which yielded an arm’s length allocation of income at the
outset of the agreement. Due to subsequent alterations in the functions per-
formed, assets contributed and risks incurred, the allocation in later stages
was not at arm’s length (the French scenario). This discussion must be
seen in the historical context of the White Paper, and particularly in light
of French, in which the Tax Court judge took the questionable stance that
it was “inappropriate to view 1963 and 1964 in isolation. [R]oyalties were
not necessarily paid for value received during those particular years”.1552
The logic behind this reasoning was that the judge in French viewed all
income years under the licence as a coherent whole for pricing purposes.
The decisive factor was thus whether the income derived under the en-
1550. The US Internal Revenue Service (IRS) in general will not reallocate income
if it would only result in minor changes relative to the taxpayer’s return position. See
also 1985 House Report, at p. 426, where it is stated that “adjustments will be required
when there are major variations in the annual amounts of revenue attributable to the
intangible”.
1551. Periodic adjustments were intended to be prospective, and thus only apply for the
income years under review and subsequent years; see White Paper (Notice 88-123), at
p. 67.
191022 (IRS TAM, 1977), distinguished by 1979 WL 56002 (IRS TAM, 1979) and
distinguished by 1992 WL 1354859 (IRS FSA, 1992). See the analysis of the ruling in
sec. 5.2.3. of this book.
467
tire duration of the agreement was balanced. In French, this meant two
decades. The legal consequence of the ruling was that, prior to the 1986
revision of IRC section 482, controlled licence agreements with fixed pric-
ing were shielded from subsequent reassessment, as long as they provided
arm’s length results at their inception. The introduction of the commen-
surate-with-income standard and the periodic adjustment authority was a
legislative rejection of French.
The White Paper contains an example in which a periodic adjustment was

carried out to reflect changes in functions in an outbound licensing struc-
ture.1553 In the example, a US parent manufactured and marketed widgets
domestically in addition to performing research and development (R&D).
A European subsidiary produced and marketed widgets for the European
market and also maintained an R&D department. The R&D activity of the
parent company resulted in promising technology. A licence agreement
was entered into with the subsidiary, under which the subsidiary was to
further develop the technology.
At the time at which the agreement was entered into, it was envisaged that
it would take 4 years and considerable effort from the subsidiary’s staff
of researchers to ready the technology for mass production. Due to those
assumptions, the licence agreement fixed a 50/50 profit split between the
parent and subsidiary. It became apparent that the process of preparation
for mass production of the technology was not as demanding as originally
assumed, and it was completed 5 years ahead of schedule. The contribu-
tions from the subsidiary were fewer and less complex than planned.
The White Paper approached the example by recognizing that the brief
duration of development, as well as the fact that the subsidiary did not
contribute any unique intangibles to the development, necessitated an ad-
justment to the fixed-price structure. The solution prescribed by the White
Paper was to switch pricing methods from the profit split method to the
basic arm’s length return method (BALRM). The result was to go from an
equal division of residual profits among the controlled parties to a scenario
in which the US parent was allocated the entire residual profits.1554
1553. White Paper (Notice 88-123), appendix E, Example 13.

1554. The author finds the example somewhat opportunistic. The basic arm’s length
return method (BALRM) was intended for cases in which one of the controlled parties
contributed purely routine functions to the value chain. It is not entirely clear that this
was the case in the example. Even if the subsidiary’s functions were modest and less
complex that planned, were they routine? The solution of this outbound example is also
not easily reconcilable with another example presented in the White Paper, pertaining
468
The second base scenario discussed in the White Paper pertains to struc-
tures in which unproven high-profit-potential intangibles were transferred
to a foreign group entity at a low price (a “cherry-picking” scenario). Such
strategies were common among multinationals at the time. The transferred
intangibles would typically be ongoing R&D, or completed R&D that was
not yet commercialized. It would be challenging to ascertain an arm’s length
consideration for such transfers at the time at which the controlled agree-
ment was entered into, not only due to the unique character of the intangi-
bles, but because there would often not be any commercial experience avail-
able on which to base the price assessment on. The 1986 introduction of the
commensurate-with-income standard made it clear that actual profits would
be given priority in assessing whether the initial pricing was at arm’s length.
The White Paper touches upon this issue in an example that deals with
periodic adjustments to reflect changes in profitability indicators.1555 In
the example, a US pharmaceutical parent performed R&D and patented
a new drug, which it licensed to a foreign subsidiary for manufacturing
and worldwide marketing. At the time at which the agreement was entered
into, it was assumed that the product would attain profitability akin to that
of existing products. Uncontrolled licence agreements between the parent
and third parties for existing products were thus used as comparables, all of
which were based on an assumption that the product would achieve a 15%
market share. The actual market share of the new product proved to be 8%
in year 2, 16% in year 3 and 21% in year 4.
The parent decided that its original pricing based on the 15% market share
assumption was still valid. Then, in year 5, the market share went up to
50%. The White Paper commented that the 50% market share was far
beyond what was envisioned in the comparable uncontrolled transactions
(CUTs). Further, because the product proved to be more profitable than
assumed, the comparables would no longer be valid. Due to the unique in-
tangible at play, no new comparables were found. The White Paper solution
was to switch the transfer pricing method for year 5 from the CUT method
to the BALRM.
to an inbound licensing structure (see White Paper (Notice 88-123), Appendix E, Ex-
ample 12). The US subsidiary in the inbound example also performed marketing and
manufacturing functions, and here, the White Paper argued for a profit split solution,
not the BALRM. In the author’s view, a more suitable result in the outbound example
would have been to adjust the profit split in favour of the US entity, due to the change
in functions, but not strip all residual profits from the European subsidiary by applying
the BALRM.
1555. White Paper (Notice 88-123), Example 14.
469
16.2.3. The relationship between the transfer pricing

methods and the periodic adjustment provision
in the White Paper
The relationship between the transfer pricing methods for intangibles pro-
posed in the White Paper (the BALRM and the BALRM with profit split)
and the concept of periodic adjustments is not immediately clear. It is stated
that “periodic adjustments are easier to analyze for the arm’s length return
method than for the methods involving comparable transactions”.1556 The
impression given by the White Paper is that the solutions provided under
the BARLM and the periodic adjustment concept are more or less similar.
Also illustrative of the almost invisible line between the profit-based trans-
fer pricing methods and the concept of periodic adjustments is the clarifi-
cation of the White Paper that even a substantial increase in the residual
profits allocable to a foreign entity in a licence agreement would not neces-
sarily lead to an adjustment of the taxable income of the US entity.1557 If
the intangible-related income increased “solely due to the efforts of the
transferee”, no part of the increased income should be allocated to the
US transferer. This should not be regarded as an exception to the periodic
adjustment provision as such, but as a result of applying the profit-based
transfer pricing methods.1558 In this scenario, the functions performed, as-
sets used and risks incurred would show that it was the increased efforts
of the licensee that created the additional residual income, and that the
licensee therefore should be entitled to it.
The question is then whether the periodic adjustment concept was neces-
sary on top of the new pricing methods introduced by the White Paper. The
answer to this question lies in the legislative rejection of French through
the 1986 incorporation of the commensurate-with-income standard in IRC
section 482. There is no doubt that the White Paper envisioned that real-
locations of income by way of periodic adjustments should be carried out
using either the BALRM or BALRM with profit split. However, in order
to carry out such adjustments in the first place to fixed-pricing intangibles
transfers that yielded arm’s length results at the outset, special authority
1556. Id., at p. 102.

1557. Id., at p. 65.
1558. Presumably this would only apply to the BALRM with profit split. If the
BALRM was applied to a foreign entity as the tested party, it should only be allocated
a normal return and be stripped of any excess profits, regardless of whether the profits
were due solely to the efforts of the foreign entity.
470
was needed, due to the immunity from subsequent allocations granted to

such agreements in French. Such authority was provided with the com-
mensurate-with-income standard. Thus, in the context of the White Paper,
the periodic adjustment provision should be viewed as a prerequisite for
applying the proposed profit-based methods to reallocate income in subse-
quent periods under a fixed-price intangibles transfer agreement.
16.2.4. Exceptions from the White Paper’s periodic

adjustment provision
Two exceptions to the periodic adjustment rule were envisioned in the

White Paper. First, the existence of exact CUTs would preclude periodic
adjustments.1559 The exception would presumably not be of much signifi-
cance in practice, as the CUT would have to pertain to a transfer of the
same intangible as transferred in the controlled transaction under substan-
tially similar conditions, but could potentially play a role for make-sell
rights that were licensed also to third parties. The White Paper had a scep-
tical attitude towards the use of inexact CUTs.1560 This is highlighted by the
fact that if a royalty rate was based on an inexact comparable, the taxpayer
would have to use the BALRM to justify the royalty rate. Thus, inexact
comparables did not provide a safe harbour from periodic adjustments.1561
The second exception applied where the unexpected profits could not have
been foreseen at the time at which the agreement was established. There
were several requirements for this exception to apply.1562 First, the con-
trolled agreement could not contain an adjustment clause under which un-
related parties would have adjusted the royalty. Second, unrelated parties
would not have included an adjustment provision that covered the devel-
opment causing the unexpected profitability. Due to concerns of abusive
behaviour, the White Paper envisioned a high standard of proof for this
exception, requiring clear and convincing evidence that the subsequent
profitability could not have been anticipated.
1559. An exact comparable is described in the White paper (Notice 88-123), at p. 106,
as an uncontrolled transfer of the same intangible under substantially similar circum-
stances as in the controlled transaction.
1560. An inexact comparable is described in the White paper (Notice 88-123), at
p. 106, as an uncontrolled transfer of an intangible similar to that transferred in the
controlled transaction, with definitive and ascertainable differences, so that reliable
comparability adjustments can be performed.
471
16.3. The US periodic adjustment provision
In the following sections, the author will discuss the periodic adjustment
provision for controlled intangibles transfers under US law. The main rule
is analysed in section 16.3.2., and the exceptions from the authority are
discussed in section 16.3.3.
16.3.2. The main rule: The profit allocation must be

commensurate with income
Under IRC section 482, the reported income from controlled intangibles
transfers must be aligned with the actual profit experience. This commen-
surate-with-income principle is operationalized through the periodic ad-
justment rule finalized in the 1994 regulations.1563 The rule states that if
an intangible is transferred under an arrangement that covers more than
1 year, the consideration charged in each taxable year may be adjusted to
ensure that it is commensurate with the income attributable to the intan-
gible. In determining whether to make such adjustments, all relevant facts
and circumstances throughout the period in which the intangible is used
may be considered. The determination in an earlier year that the amount
charged was at arm’s length will not preclude the adjustment of an amount
charged for the intangible in a subsequent year.1564 Periodic adjustments are
relevant for controlled agreements with fixed pricing structures that were at
arm’s length at the outset.
Before commencing his interpretation, the author will tie some comments
to the reasoning that underlies the periodic adjustment rule. It should be
observed that the rule will ensure an arm’s length allocation of income in
every period under a multiple-year licence agreement. The simple reason
for this is that the actual income will be allocated among the controlled
parties according to the transfer pricing methods. Without a “true-up”
rule, a fixed-price, multiple-year transfer agreement for intangibles will be
shielded from subsequent adjustments, even if it results in non-arm’s length
1563. Treas. Regs. § 1.482-4(f)(2).

1564. A periodic adjustment under the US commensurate-with-income requirement
may be made in a subsequent taxable year without regard to whether the taxable year of
the original transfer remains open for statute-of-limitation purposes.
472
The US periodic adjustment provision
allocations of income in subsequent income periods. Such an outcome in

general does not seem sensible to the author.
Critics of the commensurate-with-income standard and the periodic adjust-

ment authority will generally argue that fixed-price arrangements should
be shielded, as long as the original pricing was based on the best informa-
tion available at the time of the transfer, as unrelated parties generally must
adhere to their agreements regardless of whether the deal in hindsight turns
out to be “good” or “bad”.
The author finds this argument to be of limited relevance. It is based on

the fundamental assumption that third-party agreements are comparable to
controlled intangibles transfers. The argument is thus closely linked to the
CUT method. In the vast majority of cases, however, there will be no CUTs
available upon which to base the pricing of unique and valuable intangi-
bles. The transactional net margin method (TNMM) or residual profit split
method (RPSM) will typically be applied to price the controlled transfer.
The fundamental assumption behind the argument will thus generally not
hold.
The debate on whether periodic adjustments are justified seems to hinge on

a question of priorities, specifically whether one wants a transfer pricing
system that shields non-arm’s length profit allocations under the banner of
“information available at the time of transfer” or a system that ensures that
the actual profit allocation conforms to the arm’s length pricing methods
(in practice, the TNMM and RPSM) in every period covered by the con-
trolled agreement. The author finds that a sensible transfer pricing system
should not accept non-arm’s length allocations of income, unless there is
clear evidence that unrelated parties in fact would have adopted a similar
initial fixed price.
In other words, periodic adjustments should, in the author’s view, ideally

only be cut off in cases in which the initial pricing is based on a “genu-
ine” CUT that pertains to the same intangible transferred in the controlled
agreement. If no such CUT is available, there should, in the author’s view,
only be narrow exceptions (if any) to the periodic adjustment provision.
Accepting the shielding of fixed-priced agreements in these cases could
defy the core idea that the allocation of income in controlled agreements
should mirror that which would have been agreed by third parties. The
closest one can get to an estimate of what third parties would have agreed
on in cases in which there is no CUT available are the results yielded by
the accepted transfer pricing methods (in practice the TNMM and RPSM).
473
Of course, from a US domestic perspective, this debate is purely theoreti-

cal. The United States made its choice with the 1986 commensurate-with-
income standard, which allows adjustments based on actual results, with
some narrow exceptions. The author will interpret the periodic adjustment
rule of the US regulations.
The content of the rule itself seems clear: reassessments may be based
on the actual profits allocable to the transferred intangible, regardless of
whether this information was available at the time at which the controlled
agreement was entered into. Of course, the original taxpayer pricing will
be based on an application of a transfer pricing method using information
on projected income available at the time of the transfer. A subsequent
US Internal Revenue Service (IRS) periodic adjustment, however, will be
based on an application of a transfer pricing method using information on
actual profits that first becomes available after the controlled agreement
is entered into. The periodic adjustment rule is fundamentally a one-sided
“true-up” rule, in the sense that only the IRS is entitled to reassess the
controlled pricing based on the actual profit experience. Comparatively,
taxpayers are afforded the right under US law to perform year-end adjust-
ments based on actual results.1565
Interestingly, the IRS does not seem to fully agree with the author’s view
on the content of the periodic adjustment rule. In 2007, the Office of Chief
Counsel released an IRS memorandum addressing the issue of whether
the word “income” in the second sentence of IRC section 482 referred to
past, projected or actual profits.1566 The memorandum states that the arm’s
length standard requires that a controlled transaction be priced so as to
realize results consistent with those that uncontrolled taxpayers would have
realized if they engaged in the same transaction under the same circum-
stances.1567
Further, a key consideration in determining the transfer price for an in-

tangible under the arm’s length standard is the profit potential that uncon-
trolled taxpayers would have reasonably anticipated as of the time of the
transfer. The memorandum refers to statements in the legislative history of
the commensurate-with-income standard, as well as in the White Paper,
expressing concerns about the disadvantage that the IRS faces in examin-
ing the profit potential that could reasonably be expected at the time of
1565. See the discussion of taxpayer-initiated adjustments under US law in sec. 15.3.
1566. IRS AM 2007-007, at p. 10 (issue 3).
1567. See also Treas. Regs. § 1.482-1(b)(1).
474
controlled intangibles transfers.1568 The position of the memorandum is that

the periodic adjustment authority allows the IRS:
[…] in its discretion, provisionally to treat the income actually resulting from
the transferred intangible as evidence of what should have been projected at
the time of the transfer and to make periodic adjustments to reflect the pricing
had such results been projected at such time. (Emphasis added) 1569
The memorandum asserts that the legislative history does not suggest
that the actual profit experience should be “determinative”, but rather that
“the intention is to give the IRS a presumptive basis for making periodic
adjustments”.1570 The memorandum supports its assertion with reference to
statements in the White Paper claiming that the actual profit experience,
in the absence of comparables, generally is the best indication available of
the anticipated profit experience that arm’s length parties would have taken
into account at the outset of the arrangement.1571 Further, the regulations
allow taxpayers to rebut this assumption by showing that the actual profit
results were beyond the control of the taxpayer and could not reasonably
have been anticipated at the time of the transaction.
The memorandum therefore concludes that the word “income” in the sec-
ond sentence of IRC section 482 generally should be construed as operat-
ing profits attributable to the intangible that the taxpayer would “reasonably
and conscientiously have projected at the time it entered into the controlled
transaction”,1572 or in other words, projected profits. The memorandum
then adds that the IRS, examining a transaction only after the fact, is in-
herently at a disadvantage in assessing whether the pricing was supported
by upfront, reasonable and conscientious evaluation of projected operating
profits attributable to the transferred intangible.
1568. The 1985 House Report, at p. 424, stated: “Taxpayers may transfer such intangi-
bles to foreign related corporations or to possession corporations at an early stage, for
a relatively low royalty, and take the position that it was not possible at the time of the
transfers to predict the subsequent success of the product. Even in the case of a proven
high-profit intangible, taxpayers frequently take the position that intercompany royalty
rates may appropriately be set on the basis of industry norms for transfers of much less
profitable items.” The White Paper echoed these concerns when it observed that tax-
payers often looked solely at the “purportedly limited facts” known at the time of the
transfer to justify an inappropriately low charge for the intangible, or when it noted that
“[p]eriodic adjustments will also obviate the need for the often fruitless inquiry into
the state of mind of the taxpayer and its affiliate at the outset”; see White Paper (Notice
88-123), at p. 47 and footnote 173.
1569. IRS AM 2007-007, at p. 11.
1570. Id.
1571. See White Paper (Notice 88-123), at p. 71 (under conclusions, no. 2).
1572. IRS AM 2007-007, at p. 12.
475
The author’s impression is that the memorandum downplays the role of

actual profits under the commensurate-with-income standard. First, clear
statements in the legislative history and in the White Paper indicate that ac-
tual results shall play a rather decisive role in IRS reassessments.1573 Also,
the current regulations themselves, as will be discussed in sections 16.3.3.-
16.3.5., seem to place a more central role on actual results than recognized
in the memorandum.
Second, the main assertion of the memorandum is that the role of actual
profits under the periodic adjustment rule is that of the “best evidence”
of the results that were reasonably foreseeable at the time at which the
controlled transaction was entered into, as opposed to being “determina-
tive”. In practice, the author assumes that the line between these two roles
is rather ambiguous. It is indeed difficult to imagine that actual profits,
given that they are deemed as the best evidence of the profits that should
have been projected, will not be determinative of the profit allocation in a
reassessment.
Third, the memorandum repeatedly states that the taxpayer is entitled to

rebut the presumption that the actual profits are indicative of the profits that
should have been projected at the time of the transaction. The problem here
is that the current regulations set the bar very high. An exception is made
from the periodic adjustment authority for actual profits due to extraordi-
nary events that were beyond the control of the taxpayer and that could
not reasonably have been anticipated at the time at which the controlled
transaction was entered into. The regulations give an example of such an
extraordinary event: an earthquake. A similar exception is made for peri-
odic adjustments in the context of cost-sharing arrangements (CSAs), in
which an example refers to the “severe, unforeseen disruptions in their sup-
ply chains”.1574
Fourth, the decisive role played by actual profits run through the US regu-
lations as a red thread. In all other cases in which the initial controlled pri
cing is based on a “genuine” CUT, a periodic adjustment will be performed
if the actual results deviate more than +/- 20% from the profits projected
1573. 1985 House report, at p. 425: “The committee does not intend, however, that
the inquiry as to the appropriate compensation for the intangible be limited to the
question of whether it was appropriate considering only the facts in existence at
the time of the transfer. The committee intends that consideration also be given to
the actual profit experience realized as a consequence of the transfer.” (Emphasis
added)
1574. Treas. Regs. § 1.482-7(i)(6)(vi)(A)(2).
476
at the time at which the controlled agreement was entered into. The role of
actual profits in the context of periodic adjustments for CSAs is even more
pronounced. Under the current regulations, a periodic adjustment exami-
nation will be triggered if the actual profits of a foreign CSA participant
exceed 150% of the participant’s investment. Thus, in this case, the actual
profits are not even compared to the profits that were projected at the time
at which the buy-in contribution was rendered, but rather benchmarked
against a pre-set absolute limit for allowed returns.1575 It can therefore be
concluded that the US regulations in general do not tolerate actual profits
that move outside of certain pre-determined ranges. If they do, periodic
adjustments will be triggered.
One may speculate whether the interpretation of the memorandum is influ-

enced by the treaty obligations of the United States. In any case, it is dif-
ficult to reconcile the memorandum’s interpretation of the second sentence
of IRC section 482 with the legislative history, the White Paper, as well
as the plain language of the 1994 final regulations. In conclusion, it seems
relatively clear that the regulations indeed do allow periodic adjustments
to be carried out based on actual profits, regardless of whether they were
foreseeable at the time of the transaction, provided that none of the narrow
exceptions from the periodic adjustment rule are triggered.
The author will now tie some comments to the further guidance in the reg-
ulations. The periodic adjustment rule is illustrated in an example in which
a US parent has developed a new drug that is expected to gain a dominant
market share and command a premium price.1576 The parent licenses make-
sell rights to the drug to its European subsidiary for 5 years. The royalty
rate is based on projections of annual sales and profits from the subsidiary’s
exploitation of the licensed rights. Due to the significant profit potential of
the drug, the parent is unable to locate a CUT, and therefore concludes that
the CUT method will not provide a reliable measure of an arm’s length
result. The comparable profits method (CPM) is applied, yielding a royalty
of 3.9% as compensation to the parent, which enables the subsidiary to
earn an arm’s length return for its manufacturing and marketing functions.
1575. A periodic adjustment of a CSA buy-in will be based on information known

(including actual results) as of the determination date. With respect to the remainder of
the cost-sharing arrangement (CSA) period, the projected results, as determined by the
IRS as at the determination date, will replace the results projected by the taxpayer at
the outset of the CSA.
1576. Treas. Regs. § 1.482-4(f)(2)(iii), Example 1.
477
In year 5, it is examined whether the royalty rate is commensurate with

the income attributable to the drug. In particular, the projected and actual
profits attributable to the drug are compared. They are as follows:
Year Projected Actual Actual/projected

profits profits
1 200 250 125%
2 250 300 120%
3 500 600 120%
4 350 200 57.1%
5 100 100 100%
Total 1400 1450 103.6%
In making this determination, it is considered whether any of the excep-

tions to the periodic adjustment rule are applicable.1577 Based on these
amounts, it is concluded that the actual total profits earned through year
5 were not less than 80%, nor were they more than 120% of the profits
that were projected when the licence agreement was entered into. If it is
determined that the other requirements of the exception are met,1578 no ad-
justment will be made. A twist is then added to the example, where the ac-
tual profits of the European subsidiary are much higher than the projected
profits, as follows:1579
Year Projected Actual Actual/projected

profits profits
1 200 250 125%
2 250 500 200%
3 500 800 160%
4 350 700 200%
5 100 600 600%
Total 1400 2850 203.6%
1577. The exceptions are discussed in sec. 16.3.3.

1578. See the discussion in sec. 16.3.3.4. of the periodic adjustment exception that ap-
plies when the original pricing was based on a method other than the CUT method.
1579. Treas. Regs. § 1.482-4(f)(2)(iii), Example 2.
478
In this case, the profits earned through year 5 are more than 120% of the
projected profits. The exception to the periodic adjustment rule can there-
fore not be applied. A periodic adjustment is therefore performed.1580 The
result of the periodic adjustment is that the subsidiary is stripped of all
profits above those that must be allocated to it in order to achieve the target
CPM normal market return on its routine value chain contributions. This
is done by increasing the royalty payable to the parent, in effect allocating
all residual profits to it.
16.3.3. Exceptions to the periodic adjustment rule
In light of the 1986 amendment to IRC section 482 and the 1988 White Pa-
per, it was expected that the final 1994 regulations would include a periodic
adjustment rule. The bulk of the controversy surrounding the rule did not
pertain to the main principle itself, but rather to the exceptions to the ad-
justment provision. The periodic adjustment rule (including 3 exceptions)
was first introduced in the 1992 proposed regulations, with a clear link
to the CPM.1581 Overall, the 1992 proposed exceptions were broadened,
compared to the principal lines drawn up in the 1988 White paper, but
nevertheless remained narrow.1582
1580. For purposes of determining whether an adjustment should be made to the roy-
alty rate in year 5, the IRS aggregates the actual profits from years 1-4 (even though
they are closed under the statute of limitations) with the profits of year 5. An adjustment
will only be made with respect to year 5.
1581. 1992 Prop. Treas. Regs. (57 FR 3571-01) § 1.482-2(d)(6). Two of the proposed
exceptions pertained to cases in which the profits did not move outside the comparable
profit interval (CPI) (the range of acceptable comparable price method (CPM)-based
returns allocable to the tested party); see supra n. 528 and n. 931). The last exception
pertained to when the profits moved outside of the CPI due to economic conditions
that were “beyond the control” of the group and that could be neither “anticipated nor
reasonably foreseeable”.
1582. See the comments in sec. 16.2.4. on the exceptions in the White Paper (Notice
88-123). Also, closely connected to the 1992 proposed periodic adjustment provision
was the concept of sound business judgement, introduced in the same regulations. Un-
der this concept, it was to be assessed “whether uncontrolled taxpayers, each exercising
sound business judgment on the basis of reasonable levels of experience (or, if greater,
the actual level of experience of the controlled taxpayer) within the relevant industry
and with full knowledge of the relevant facts” would have agreed to the contractual
terms corresponding to those under examination; see 1992 Prop. Treas. Regs. (57 FR
3571-01) § 1.482-1(b)(1). This also highlighted that integrated intra-group transactions
could be examined as a whole, e.g. roundtrips, in which US-developed intangibles are
479
The 1992 proposed regulations attracted considerable criticism for this,

most notably from the 2nd OECD Task Force Report, which reiterated the
recommendation of the original Task Force Report that periodic adjust-
ments should be confined to “truly abusive cases”.1583 The response of the
1993 temporary regulations was to modify the exceptions. According to its
preamble, the scope of the exceptions was “broadened”. The author finds
that characterization debatable. The 1994 final regulations added three ex-
ceptions. In sections 16.3.3.2.-16.3.3.5., the author will discuss the 1994
final exceptions to the periodic adjustment rule.
16.3.3.2. First exception: Initial fixed pricing based on CUTs

involving the same intangible (“genuine” CUT exception)
If the same intangible is transferred to an uncontrolled taxpayer in a fixed

pricing CUT under substantially the same circumstances as in the con-
trolled transaction, the CUT may serve as the basis for applying the CUT
method to test the allocation of income in the first taxable year in which
substantial periodic consideration is required to be paid.1584 If the amount
paid in that first year is at arm’s length, no allocation will be made in a
subsequent year.
There is no doubt that this exception, which was not introduced until the
final 1994 regulations, even though substantially the same as drawn up in
the White Paper,1585 is narrow. It will typically be relevant when a multi-
national licenses rights to the same intangible to both controlled and un-
controlled parties located in similar markets. For instance, if a US parent
licenses make-sell rights to the same intangible to a subsidiary in Norway
and to an unrelated party in Denmark, the latter licence agreement may
arguably by used to test the pricing in the former agreement, as the Norwe-
gian and Danish markets have similar traits.
As discussed in section 16.3.2., critics of the periodic adjustment provi-

sion normally assert that only information available at the time at which
the controlled transaction is entered into should be relevant in a transfer
pricing reassessment. This exception is the answer of the US regulations
licensed to a foreign subsidiary that manufactures the tangible using the licensed intan-
gible and then sells the product back to the United States.
1583. 2nd Task Force Report, para. 2.23.
1584. Treas. Regs. § 1.482-4(f)(2)(ii)(A).
1585. See the discussion of the periodic adjustment exceptions in the White Paper (No-
tice 88-123) in sec. 16.2.4.
480
to this criticism. Periodic adjustments are precluded if the pricing is based

on a “genuine CUT”, i.e. an uncontrolled transaction pertaining to the
same intangible under substantially the same circumstances. The author
agrees that an application of the periodic adjustment provision in this con-
text would conflict with the arm’s length principle, simply because a truly
comparable unrelated transaction exists in these rare cases and thus illus-
trates that unrelated parties indeed would fix the pricing of the intangibles
transfer at the outset.
16.3.3.3. Second exception: Initial fixed pricing based on CUTs

involving a comparable intangible (inexact CUT exception)
If the fixed price of the controlled transaction is determined by applying

the CUT method, where the third-party reference transaction pertains to
the transfer of a comparable intangible under comparable circumstances,
the US regulations take a more sceptical stance. This is made clear by
the fact that the regulations will not tolerate the aggregate actual profits
of the controlled transaction being less than 80% or more than 120% of
the prospective profits foreseeable at the time at which the comparability
of the CUT was established. This range of allowed profits is absolute, in
the sense that a periodic adjustment will be triggered if it is breached. The
following criteria must be fulfilled in order for the inexact CUT exception
to be applicable:1586
(1) The controlled taxpayers entered into a written agreement that pro-
vided for an amount of consideration with respect to each taxable year
covered by the agreement, this consideration represented an arm’s
length amount for the first taxable year in which substantial periodic
consideration was required to be paid and the agreement remained in
effect for the taxable year under review.
(2) There is a written agreement setting out the terms of the CUT relied
upon to establish the arm’s length consideration that contains no pro-
visions that would permit any change in the amount of consideration,
a renegotiation or a termination of the agreement, in circumstances
comparable to those of the controlled transaction in the taxable year
under review (or that contains provisions permitting only specified,
non-contingent, periodic changes to the amount of consideration).
1586. Treas. Regs. § 1.482-4(f)(2)(ii)(B)(1)-(6).
481
(3) The controlled agreement is substantially similar to the CUT, with re-
spect to the time period for which it is effective and the adjustment
provisions described in criterion (2).
(4) The controlled agreement limits the use of the intangible to a specified
field or purpose in a manner that is consistent with industry practice
and any such limitation to the CUT.
(5) There were no substantial changes in the functions performed by the

controlled transferee after the controlled agreement was executed, ex-
cept for changes required by events that were not foreseeable.
(6) The aggregate profits actually earned or the aggregate cost savings
actually realized by the controlled taxpayer from the exploitation of
the intangible in the year under examination and all past years are not
less than 80% or more than 120% of the prospective profits or cost sav-
ings that were foreseeable when the comparability of the uncontrolled
agreement was established.
For the purpose of analysing this exception, the author finds it useful to
sort away the criteria that pertain to the more formal sides of the arrange-
ment. This goes for the requirements for written, long-term controlled and
uncontrolled licence agreements with fixed pricing and without adjust-
ment clauses (see criteria 1 (partly), 2, 3 and 4). These criteria are clearly
important, but in the author’s opinion, they are more prerequisites than
substantial material requirements. The real hurdles for the taxpayer are,
in particular, criteria 1 (partly), as well as 5 and 6, which the author will
comment on further in this section.
The requirement for an arm’s length payment for the first taxable year
in which substantial periodic consideration is required to be paid under
the agreement (criterion (1)) must presumably be satisfied by confirm-
ing that the controlled transaction provided a royalty payment equal to
that paid under the CUT for the relevant period. The strict comparability
requirements of the final regulations, in particular, the profit potential
criterion (also introduced in 1994),1587 severely limit the relevance of the
exception.
1587. See the discussion of the profit potential comparability criterion for the CUT
method in sec. 7.2.3.1. See supra n. 1208 for further references with respect to the profit
potential criterion.
482
Further, if there were substantial changes in the functions performed by the

related parties under the controlled agreement (criterion (5)), the exception
will not apply, unless the changes are due to unforeseeable events. The
likelihood of significant changes may vary depending on where the trans-
ferred intangible is in its lifecycle at the time of transfer. The intangible
may be unproven and only partially developed, or proven and established
and generating profits at the time of the transfer. The situation may also, for
that matter, lie somewhere in between. For instance, the controlled transac-
tion may be a licence agreement pertaining to a fully developed intangible,
but manufacturing know-how will be gradually developed and exchanged
between the controlled parties, akin to the factual pattern in French.1588
Many of the tax structures used by multinationals to shift intangible in-

come out of the United States in the decades prior to the 1994 regulations,
as reflected in, for instance, Eli Lilly and Bausch,1589 pertained to proven
and established intangibles. In these “roundtrip” transactions, the structure
of functions to be performed by the controlled parties was typically fixed
at the time of the agreement and was unlikely to change afterwards. The
whole point was to allocate certain assets and functions abroad, typically
migrating manufacturing intangibles and connected functions, while re-
taining distribution functions in the United States.
For transfers of unproven and partially developed intangibles, the situation

may be different. The White Paper example on periodic adjustments in the
context of a shift of functions pertains to a process that was under devel-
opment and would not reach commercialization for several years after the
agreement was entered into.1590 The agreement therefore had similarities to
a CSA in some ways.
A key element in the example was that the royalty rate was based on the
premise that the subsidiary would contribute complex functions to the fur-
ther intangible development, which turned out not to be necessary, thereby
triggering a need to adjust the controlled pricing. This example will likely
provide some guidance as to whether “substantial changes in the functions”
191022 (IRS TAM, 1977), distinguished by 1979 WL 56002 (IRS TAM, 1979) and
distinguished by 1992 WL 1354859 (IRS FSA, 1992).
affirmed in part, reversed in part by 856 F.2d 855 (7th Cir., 1988); and Bausch & Lomb
Inc. v. CIR, 92 T.C. No. 33 (Tax Ct., 1989), affirmed by 933 F.2d 1084 (2nd Cir., 1991).
1590. See White Paper (Notice 88-123), Appendix E, Example 13; and the discussion
in sec. 16.2.2. on the 1985 House Report and the White Paper.
483
have occurred. The line will presumably be drawn for changes subsequent
to which the original pricing can no longer be defended. For instance, this
can occur when the RPSM is originally chosen to price the agreement but
later developments make it inapplicable, thus necessitating a shift of pric-
ing method to the CPM, or vice versa.
The exception for changes due to events that were “not foreseeable” is pre-
sumably not very practical for partially developed intangibles, as these in-
tangibles may entail such high risks that a considerable span of different
outcomes may be equally probable at the time of the transfer. This will
often depend on whether the R&D is “blue sky” or whether the develop-
ment is based on a pre-existing intangible. There will generally be less
risk involved in developing second and subsequent generations of an estab-
lished intangible.
The last criterion (criterion (6)) is that the aggregate actual profits allocable
for the transferred intangible do not deviate from the aggregate profits esti-
mated at the time at which the CUT was deemed to be comparable by more
than +/- 20%. This establishes a range of allowed profits. Under the 1993
temporary regulations, the profits subject to comparison were those for all
open years. Under the 1994 final regulations, the comparison applies to all
years, thus making the pool of data potentially larger. The view of the final
regulations is that it therefore will be less likely that the actual profits fall
outside the range of projected profits solely due to timing differences, thus
making period adjustments less frequent.
The +/- 20% criterion represents a link to the profit potential comparability
requirement of the CUT method.1591 The immediate observation is that if
there is a discrepancy of some magnitude between (i) the estimated profits
of the intangible that, at the inception of the agreement, had to be based
on those expected from the CUT; and (ii) the actual profits under the con-
trolled agreement, this could indicate that the controlled and uncontrolled
transactions indeed were not comparable in the first place. If so, the excep-
tion would not apply at all, as the CUT method would be inapplicable.
Further, it does not seem entirely clear to the author that a +/- 20% de-
viation between projected and actual results is useful for licence agree-
ments. Typically, a licence agreement does not set fixed amounts to be
paid, only a fixed royalty rate. Thus, the licenser and licensee will share
1591. See sec. 7.2.3.1. on the profit potential comparability criterion for the CUT meth-
od (as well as supra n. 1208 for further information on the profit potential criterion).
484
risk. If the licensee does not generate sales, the licenser will not receive
royalty payments. Because only the division of actual profits, and not abso-
lute amounts, is set at the time of the agreement, there is no obvious logic
in operating under a concept of actual profits deviating from those foresee-
able at the time of the agreement.
The author does, however, admit that this is a theoretical argument. What
the regulations clearly are concerned about in these scenarios is that the
(normally foreign) subsidiary will be left with larger profits than intended
under the original transfer pricing if the royalty rate is not adjusted up-
wards to take into account the increased profitability of the subsidiary. For
instance, if a 5% royalty of the reference CUT was set based on a compa-
rable price method (CPM) analysis in which the subsidiary would be left
with a normal return on its routine functions after paying the royalty, and
actual sales are larger than the projected sales, this will leave the subsidi-
ary with some portion of the residual profits from the transferred intan-
gible. It is this profit that will be stripped from the subsidiary if its actual
profits are outside of the allowed +/- 20% range.
The logic behind the +/- 20% criterion therefore seems sound in cases in
which the initial controlled pricing was determined based on a one-sided
method. It does not seem obvious that the same can be said for cases in
which the initial pricing was based on the CUT method. After all, a CUT
may allocate profits differently than the other pricing methods, depending
on the bargaining positions of the unrelated parties.
For instance, there is no guarantee that a low-risk distributor will not be

allocated residual profits in a CUT, even though economic reasoning would
argue otherwise. Nevertheless, the +/- 20% criterion refers to the profits
foreseeable at the time at which the comparability between the controlled
transaction and the reference CUT was established. There is no “check” as
to whether the income allocation under the CUT conforms to the result that
would have been yielded by the CPM or RPSM. The decisive parameter is
solely whether the actual returns deviate by more than +/- 20% from the
CUT-based projections, whatever they were.
The 2nd Task Force Report observed that where the original pricing was
CUT-based, it would be contrary to the arm’s length standard to perform
periodic adjustments if the CUT remained comparable for the entire pe-
riod and there was no change in the functions performed by the controlled
485
parties.1592 This argument refers to a situation in which there is no lack of

comparability, but the actual profits from the transferred intangible simply
move outside the allowed +/- 20% range of profits.
The author must admit that he finds this argument rather convincing. It is
difficult to find a periodic adjustment justified in these cases, especially
considering the strict comparability requirements for applying the CUT
method. After all, if the CUT remains comparable for the entire period,
chances are that it may also have experienced the same increase in prof-
itability as the controlled transaction. The final regulations, however, do
not accept this argument. This is a clear expression of the US scepticism
towards CUT-based pricing of unique intangibles transfers.
In summary, the author’s impression of the CUT-based exception is that it

is narrow, ensuring wide application of the periodic adjustment rule. The
exception does not allow for relatively high profit fluctuations (outside the
+/- 20% range), which may typically be the case for unproven intangibles.
Then again, the CUT method will likely not be applicable to such transfers
due to the obvious risk of lack of comparables, reducing the relevance of
the discussion significantly. Pricing of unproven intangibles will also be
challenging for multinationals to carry out properly, likely necessitating a
range of highly discretionary comparability adjustments to the purported
CUT.
This raises obvious questions: Why then adopt a fixed-price structure?

Why not include a provision to adjust the pricing in light of the actual
results? These questions no doubt underlie the scepticism towards trans-
fer pricing based on fixed-price-structure CUTs that is expressed through
the +/- 20% criterion. Nevertheless, it does allow for relatively low profit
fluctuations (inside the +/- 20% range), which may more likely be the case
when proven intangibles are transferred, as future income may be more
predictable than for unproven intangibles. This also makes it more plausi-
ble that no price adjustment clauses were included in the CUT.
Overall, the author finds the +/- 20% criterion useful, as a deviation be-
tween the actual and projected profits outside of this range raises serious
questions as to whether the fixed-price reference CUT was sufficiently
comparable in the first place.
1592. 2nd Task Force Report, para. 2.29.
486
16.3.3.4. Third exception: Initial fixed pricing based on methods

other than the CUT method
If the initial fixed arm’s length transfer price is determined under any
method other than the CUT method, no allocation will be made under the
periodic adjustment rule if the following criteria are fulfilled:1593
− the controlled taxpayers entered into a written agreement that provided
for an amount of consideration with respect to each taxable year cov-
ered by the agreement, and the agreement remained in effect for the
taxable year under review;
− the consideration called for in the controlled agreement was an arm’s
length amount for the first taxable year in which substantial periodic
consideration was required to be paid, and relevant supporting docu-
mentation was prepared contemporaneously with the execution of the
controlled agreement;
− there were no substantial changes in the functions performed by the
transferee since the controlled agreement was executed, except for
changes required by events that were not foreseeable; and
− the total profits actually earned or the total cost savings realized by the
controlled transferee from the exploitation of the intangible in the year
under examination and all past years are not less than 80% or more
than 120% of the prospective profits or cost savings that were foresee-
able when the controlled agreement was entered into.
For this exception to apply, the non-CUT-based transfer pricing has to be

supported by relevant contemporaneous documentation. Similarly to the
CUT-based exceptions, this exception demands that an arm’s length pay-
ment is carried out in the first year in which a substantial periodic consid-
eration is required to be paid under the controlled agreement. The taxpayer
will thus have to document that the royalty paid in that period, typically
pursuant to the CPM or RPSM, represented an arm’s length consideration.
The testing of the asserted arm’s length consideration will likely be more
intensive when more residual profits are allocated to the foreign entity.
If the consideration is deemed arm’s length, a periodic adjustment can be

made if there are substantial changes in the functions performed by the con-
tracting parties subsequent to the formation of the agreement. As in the CUT-
1593. Treas. Regs. § 1.482-4(f)(2)(ii)(C). As this exception applies when the initial
pricing is based on methods other than the CUT method, the requirements under the
CUT-based exceptions pertaining to the terms of the CUT and similarities to the con-
trolled agreement are not relevant. The requirement to specify the licensed use of the
intangible in the controlled agreement was, for some reason, removed.
487
based exception discussed above, if the alternation in functions performed is

caused by unforeseeable events, a periodic adjustment cannot be performed.
In order to avoid a periodic adjustment, the actual intangible profits for all
open income years under review must also here lie within the acceptable
range of +/- 20% of the projected profits. The author refers to his discus-
sion of this criterion above under the CUT-based exception and will only
comment on two particularities here.
First, if the initial taxpayer profit allocation was determined using the
CPM, for instance, to achieve a target return for the tested party of 8% on
the basis of projected earnings, there must be a subsequent adjustment that
strips or adds profit from or to the tested party, depending on whether its
actual profits prove to be higher or lower than projected. This is because it
will be unlikely that the actual return of the tested party achieves precisely
the target return due to variance in external factors, such as sales to and
costs of third parties. In practice, the taxpayer will often have performed a
taxpayer-initiated compensating adjustment (true-up) in these cases.1594 If
so, the issue will not be left for the IRS to resolve.
Second, if the RPSM was used to set the original transfer price, the royalty
was deemed arm’s length in the first year in which a substantial payment
was made and there were no other changes apart from the actual profits
being larger than estimated, the author does not quite see the rationale for
making an adjustment. After all, the relative split of the profits will be the
same as it was in the year of the first substantial arm’s length payment.
However, also here, the intention is likely to seek to limit the return al-
locable to the foreign entity. That return will, of course, become greater as
the profits grow beyond what was foreseeable at the time of the transaction,
regardless of whether the relative split is the same. If there were no changes
in the functions performed, assets contributed or risks incurred by the par-
ties, however, there will be no reason under the RPSM to restrict the profits
allocable to the foreign entity.
16.3.3.5. Fourth exception: Extraordinary events
The last exception pertains to extraordinary events.1595 No periodic adjust-

ment will be carried out even if the aggregate actual profits fall outside the
allowed +/- 20% range of projected profits due to extraordinary events that
1594. See the analysis of taxpayer-initiated compensating adjustments in ch. 15.

1595. Treas. Regs. § 1.482-4(f)(2)(ii)(D).
488
could not reasonably have been anticipated and the rest of the requirements
of US Treasury Regulations (Treas. Regs.) § 1.482-4(f)(2)(ii)(B) (pertain-
ing to CUT-based original pricing) or § 1.482-4(f)(2)(ii)(C) (pertaining to
the exception for original pricing based on methods other than the CUT
method) are satisfied. In order for the exception to be triggered, the follow-
ing criteria must be met:
− due to extraordinary events that were beyond the control of the controlled
taxpayers and could not reasonably have been anticipated at the time at
which the controlled agreement was entered into, the aggregate actual
profits or aggregate cost savings realized by the taxpayer are less than
80% or more than 120% of the prospective profits or cost savings; and
− all of the requirements of the second or third exception are otherwise
satisfied (see sections 16.3.3.3.-16.3.3.4.).
The exception is illustrated by an example in which a foreign parent licens-

es a new air-filtering process to its US subsidiary. The licence runs for 10
years, and the profits are projected to be 15 per year. The fixed royalty rate
is based on an inexact CUT, and the requirements under the CUT-based
exception are fulfilled, apart from the +/- 20% criterion.1596
In examining year 4 of the licence, the IRS determines that the aggregate
actual profits earned by the US subsidiary through year 4 are 30, i.e. less
than 80% of the projected profits of 60. However, the subsidiary establishes
that this is due to an earthquake that severely damaged its manufacturing
plant. Because the difference between the projected and actual profits is
due to an extraordinary event beyond the control of the US subsidiary and
could not reasonably have been anticipated at the time at which the licence
agreement was entered into, the requirements under the “extraordinary
events” exception have been met, and no periodic adjustment is made.
Based on the relatively strong language of the exception, requiring “ex-

traordinary events beyond control”, as well as the earthquake example,
there is little doubt that the threshold for triggering the “extraordinary
events” exception is set very high. Qualifying events must likely be akin to
typical force majeure occurrences, such as war and natural disasters. The
IRS argues in the aforementioned (see section 16.3.2.) 2007 memorandum
1596. Specifically, the parent and subsidiary have entered into a written agreement that
provides for a royalty in each year of the licence, the royalty rate is considered arm’s
length for the first taxable year in which a substantial royalty is required to be paid,
the licence limits the use of the process to a specified field, consistent with industry
practice, and there are no substantial changes in the functions performed by the US
subsidiary after the licence agreement is entered into.
489
that periodic adjustments should not be made on the basis of actual profits
that could not have been reasonably anticipated at the time at which the
controlled transaction was entered into, but rather that actual profits pro-
vide the IRS with a presumptive basis for making periodic adjustments.1597
Further, taxpayers have the ability to rebut that presumption by showing that
the actual results were beyond the control of the taxpayer and could not rea-
sonably have been anticipated at the time of the transaction. On this point, the
2007 memorandum refers to the “extraordinary events” exception.1598 The au-
thor finds the 2007 memorandum to be rather speculative, as it seems to pre-
sent the “extraordinary events” exception as being broader than it actually is.
While the author agrees with the IRS that an earthquake certainly is an
extraordinary event and beyond the control of even multinationals, he does
not agree that such events are the only circumstances that will not be fore-
seeable for a taxpayer at the time at which a controlled intangibles transfer
is agreed to. For instance, a range of market conditions, such as the intro-
duction of new competing products, new competitors or infringement of
patent rights, may cause actual results to deviate from the projected re-
sults. Yet, the “extraordinary events” exception will likely not encompass
the majority of such events. The author’s view is therefore that the ability
of the taxpayer to rebut the presumptive basis of the IRS is limited indeed.
16.3.4. Five-year cut-off rule
If the requirements of the exception for initial CUT-based pricing or the

exception for initial pricing based on methods other than the CUT method
are met for each year of the 5-year period, beginning with the first year in
which substantial periodic consideration was required to be paid, no peri-
odic adjustment will be made in any subsequent year.1599
16.3.5. Concluding remarks on the exceptions
The exceptions to the periodic adjustment rule are rather convoluted, nar-
row and not particularly practical. The criteria for triggering the exceptions
1597. IRS AM 2007-007, at p. 11 (with further references to the White Paper [Notice
88-123] in its footnote 19).
1598. IRS AM 2007-007, at p. 11 (see also its footnote 20).
1599. Treas. Regs. § 1.482-4(f)(2)(ii)(E).
490
Periodic adjustments to lump-sum IP transfers under US law
are many and strict. Whether the +/- 20% range of allowed profits is suit-
able is a matter of opinion. The exceptions are, of course, intended to be re-
strictive, seeing as the commensurate-with-income language in IRC section
482 requires the profit allocation in controlled intangibles transfers to be
commensurate with the actual profits allocable to the transferred intangi-
ble. In order to attain this alignment of projected and actual profits, the ad-
justment provision must be broad. Realistically, the exceptions therefore do
not seem to encompass scenarios in which there will be any practical need,
as seen from the point of view of the IRS, to perform subsequent pricing
adjustments to controlled intangibles transfers with fixed-price structures.
16.4. Periodic adjustments to lump-sum IP transfers

under US law
The White Paper was sceptical towards the use of fixed lump-sum pricing
in intangibles transfers, including deployment of intangibles through CSAs
(buy-ins). Two problems were identified with respect to periodic adjust-
ments of such transactions. First, an adjustment could be barred by the
statute of limitations. Second, if the adjustment was made “mid-stream”,
it would not be possible to foresee the last part of the income stream with
certainty. Thus, the adjustment could be too high or too low. The solution
of the White Paper was to treat the transaction as open. The lump-sum
payment was regarded as a pre-payment for the intangible. No adjust-
ment would be carried out before the aggregate of a calculated adjustment
amount exceeded the prepayment.
Building on the initial thoughts presented in the White Paper, the 1994
regulations include a provision that addresses the use of fixed lump-sum
payments.1600 It provides that fixed lump-sum payments are subject to peri-
odic adjustments to the same extent as licence agreements that provide for
periodic royalty payments. The regulations determine that if an intangible
is transferred in a controlled transaction for a lump sum, that amount must
be commensurate with the income attributable to the intangible.1601
This criterion will be met if the equivalent royalty rate in a taxable year is
equal to an arm’s length royalty rate. The equivalent royalty rate for a tax-
able year is the amount determined by treating the lump sum as an advance
1600. The 1992 proposed (57 FR 3571-01) and 1993 temporary (58 FR 5263-02) regu-
lations did not include this rule, but the issue was reserved.
1601. Treas. Regs. § 1.482-4(f)(6).
491
payment of a stream of royalties over the useful life of the intangible (or the
period covered by the controlled agreement, if shorter), taking into account
the projected sales of the licensee as of the date of the transfer. In order to
determine the equivalent royalty amount, a present value calculation must
be performed, based on the lump sum, an appropriate discount rate and the
projected sales over the relevant period.
This is illustrated in an example in which a US parent licenses make-sell

rights to a patented kitchen appliance to its European subsidiary.1602 The li-
cence covers 5 years, and a single lump sum charge of 500,000 is paid at the
outset of the agreement. The projected sales of the subsidiary are as follows:
Year Projected
sales
1 2,500,000
2 2,600,000
3 2,700,000
4 2,700,000
5 2,750,000
The equivalent royalty amount for the licence is determined by deriving

a rate equal to the lump-sum payment divided by the present value of the
projected sales of the subsidiary. Based on the riskiness of the kitchen ap-
pliance business, an appropriate discount rate is determined to be 10%.
The present value of the year 1-5 sales is approximately 10,000,000, yield-
ing an equivalent royalty rate of 5%. Therefore, the equivalent royalty
amounts for each year are as follows:
Year Projected Equivalent

sales royalty
amount (5%)
1 2,500 000 125,000
2 2,600 000 130,000
3 2,700 000 135,000
4 2,700 000 135,000
5 2,750 000 137,500
1602. Treas. Regs. § 1.482-4(f)(6)(iii).
492
The OECD periodic adjustment provision
The equivalent royalty amount is subject to periodic adjustments to the

same extent as an actual royalty rate would be. No periodic adjustment is
made if any of the exceptions from the general periodic adjustment rule
(as discussed in section 16.3.3.) apply. Thus, the +/- 20% allowed range
between projected and actual profits, as well as the other criteria of the
exceptions, apply also in the context of lump-sum payments.
Thus, with respect to the above example, if the equivalent royalty amount
is determined not to be at arm’s length in any of the 5 taxable years, a
periodic adjustment may be carried out. The adjustment will be equal to
the difference between the equivalent royalty amount and the arm’s length
royalty for that taxable year. The determination of an arm’s length royalty
will be carried out by applying the transfer pricing methods. The CPM
would likely be applied to determine the arm’s length royalty payment in
the example, as only the US parent provides unique value chain inputs.
The lump-sum payment provision should therefore be seen as a technical

provision pertaining to how one shall benchmark a lump-sum payment
against an arm’s length royalty rate in order to assess whether there should
be an adjustment of income in any of the taxable years during a licence
agreement that applies a fixed lump-sum payment structure.
16.5. The OECD periodic adjustment provision
The OECD voiced its opinion on periodic adjustments already in the 1993
Task Force Report, commenting on the 1992 proposed US regulations. The
report conceded that there was a need for periodic adjustments in some
cases.1603 The position was that they should take into account only profits
that could reasonably have been foreseen at the time of the transaction and
that the use of actual profits to inform a review of the ex ante controlled
pricing could be inconsistent with the arm’s length principle outside of
“truly abusive cases”.1604 2 years later, the 1995 OECD Discussion Draft
proposed to include guidance on periodic adjustments in the OECD Trans-
fer Pricing Guidelines (OECD TPG),1605 resulting in a slightly more fo-
1603. Task Force Report, para. 3.64.

1604. Id., at para. 3.65.
1605. Under the heading “periodic adjustments”, the 1995 OECD discussion draft stat-
ed that no adjustments should be carried out if independent parties would have agreed
to a similar fixed compensation structure without adjustment clauses while facing a
comparable level of valuation uncertainty (see para. 39). A periodic adjustment could
only be carried out if the tax authorities had “no other recourse to determine an appro-
493
cused consensus text later that same year.1606 The 2010 OECD TPG made
no changes to the 1995 text.
The question addressed in the 2010 text was whether a controlled intangi-
bles transfer with a fixed-priced structure, which yielded an arm’s length
royalty payment at the outset, could be adjusted in later income periods if
the profit allocation then proved not to be at arm’s length. It was recognized
that independent enterprises in some cases could deem the projected profits
allocable to the transferred intangible sufficiently reliable to fix the pricing
at the outset.
If this was the situation in the controlled transfer, the 2010 text did not
permit periodic adjustments.1607 In other scenarios, the projected profits
could be so uncertain at the outset that independent parties either would
only have (i) used fixed pricing in short-term agreements; or (ii) in the case
of long-term fixed-price licence agreements, included price adjustment or
renegotiation clauses. The 2010 text allowed periodic adjustments to be
carried out in such cases.1608
The problem was separating the cases in which periodic adjustments were
barred from the cases in which they were allowed. For this purpose, two
criteria were drawn up. First, if independent enterprises would have “in-
sisted on a price adjustment clause” in comparable circumstances, a tax
administration was permitted to determine the pricing based on such an
imputed adjustment clause.1609 Second, if independent enterprises would
priate transfer price”. The draft argued that the tax authorities should seek to determine
a price that would not require future adjustments. The periodic adjustment provision
was limited to exceptional cases in which independent parties would have insisted on
adjustments; see 1995 OECD discussion draft, para. 40.
1606. Without explanation, the 1995 OECD discussion draft heading “periodic adjust-
ments” was replaced in the 1995 OECD Transfer Pricing Guidelines (OECD TPG)
with “arm’s length pricing when valuation is highly uncertain at the time of the trans-
action”, presumably in order to distance the guidance from the obvious influence of
the US regulations, likely making it easier to reach consensus on the final text on this
controversial issue. Several of the problematic criteria from the 1995 OECD discussion
draft were removed. Focus was on what unrelated parties would have agreed to, given
valuation uncertainty at the time of transfer; see 1995/2010 OECD TPG, paras. 6.28
and 6.32. Plainly put, tax authorities were entitled to adjust fixed-price structures if
unrelated parties would not have adopted a fixed payment form under comparable cir-
cumstances. The guidance provided for this hypothetical assessment was modest. The
final 1995 text, as opposed to the 1995 OECD discussion draft, included examples on
the application of the periodic adjustment authority.
1607. 1995/2010 OECD TPG, para. 6.32.
1608. 1995/2010 OECD TPG, para. 6.34.
1609. Id.
494
have considered “unforeseeable subsequent developments so fundamental

that their occurrence would have led to a prospective renegotiation” of the
fixed price, a periodic adjustment was allowed.1610
Compared to the US periodic adjustment provision, the 2010 OECD rule

was structured in a reverse manner. The point of departure under the US
rule is that periodic adjustments shall be carried out in all cases to ensure
that the profit allocation is commensurate with the income generated by it,
to which specific and narrow exceptions are made.
The practical question under US law is not to ascertain whether a periodic

adjustment is allowed at the outset, but whether that provision is cut off in a
concrete case due to one of the exceptions being triggered. The point of de-
parture under the 2010 OECD rule, on the other hand, was that no adjust-
ment was allowed unless independent enterprises would have “insisted” on
including a price adjustment clause in the fixed-price licence or subsequent
developments were so “fundamental” that they would have agreed to a re-
negotiation of the fixed-pricing terms.
The OECD rule hinged on a hypothetical assessment, contingent on the

degree of uncertainty surrounding the projected profits. It will be very dif-
ficult, on an after-the-fact basis, to ascertain the degree of uncertainty that
actually surrounded the projected profits at the time at which the agree-
ment was entered into, not to mention what level of uncertainty would en-
tice multinationals to adopt variable pricing structures in the absence of a
CUT.
The core point of the pricing structure of a controlled licence agreement

will normally be to provide the licensee with an arm’s length return, typi-
cally determined under the TNMM (e.g. a normal market return to routine
entities in the principal model).1611 If the projected profits are so uncertain
that it is likely to be difficult to attain the target return for the tested party
using a fixed-price structure, it seems sensible to assume that multination-
als would not adopt such structures.1612 This logic speaks in favour of a
1610. Id.
1612. E.g. if an arm’s length target return for a low-risk distributor subsidiary is set to
5% of the total operating costs, and a fixed royalty rate is set to 40% of sales based on
projected sales of 1,000. Only if the actual sales of the subsidiary prove to be 1,000 will
the target 5% return be achieved. If it is equally likely that the subsidiary’s sales will be
500, 750, 1,250 or 1,500, it will make little sense to fix the pricing at the outset, as it is
likely that the target return of 5% will not be met using a fixed-price structure. Instead,
the agreement should include a price adjustment clause so that the royalty rate of 40%
495
relatively low threshold for periodic adjustments under the 2010 text. How-
ever, this interpretation runs afoul of the restrictive 2010 language, which
indicated a high threshold.
Regardless of how the 2010 text should be understood with respect to the
adjustment threshold, in the end, it seems hard to get around the fact that
the best practical indicator of the degree of uncertainty that surrounded
the projected profits at the time at which the transaction was entered into
is the degree of discrepancy between the projected and actual profits. If
it is significant, it will likely provide a convincing argument that the pro-
jected profits were surrounded by such a large degree of uncertainty that
independent enterprises would not have opted for a fixed-price structure.
Contrary to the US regulations, no specific range of allowed discrepancy
between the projected and actual profits was indicated in the 2010 text.1613
On that basis, it can be observed that the periodic adjustment provision

of the 2010 OECD TPG was ambiguous. While it was clear that periodic
adjustments in some cases were allowed, the threshold for triggering the
provision was not. Undoubtedly, this was due to the historical resistance
among some OECD member countries against periodic adjustments.
An obvious problem with the 1995/2010 approach of tying the periodic

adjustment provision to an assessment of the degree of uncertainty sur-
rounding projected profits, combined with a lack of guidance on the use of
actual profits, is that it would likely be possible to argue convincingly for a
wide range of solutions, resulting in a significant potential for controversy.
The author would assume that this made the 2010 rule difficult to apply in
practice, possibly resulting in double taxation.
The 2017 OECD TPG introduce some new language on periodic adjust-
ments while retaining the basic point of departure of the 2010 text. It is
recognized that unrelated parties may base the pricing of an intangibles
can be adjusted depending on actual sales in order for the subsidiary target return of
5% to be reached. In practice, the taxpayer will often perform year-end adjustments to
transactional net margin method (TNMM)-based pricing structures to adjust prices
in light of actual results so that the tested party’s target return can be reached; see the
analysis of year-end adjustments in ch. 15. There will then, in theory, be no need for a
periodic adjustment rule for tax authorities, as the taxpayer will already have adjusted
its prices to an arm’s length result.
1613. Further, the relationship between the 2010 OECD periodic adjustment rule and
the transfer pricing methods was problematic. The 2010 text did not distinguish the
periodic adjustment provision depending on which transfer pricing method was used to
set the initial transfer price.
496
transfer on profit projections. If these are sufficiently predictable, the

parties may choose to fix the pricing at the time of the transaction.1614
However, major unforeseeable events may occur, or events that, in prin-
ciple, were foreseeable at the time of the transaction, but had low proba-
bilities of materializing. This may justify a renegotiation of the price.1615
Further, in cases in which the financial projections are uncertain, third
parties may adapt by choosing short-term agreements, including price
adjustment clauses, or basing the price on contingent payment terms in
which the quantum or timing of payments rely on future events (pro-
gressive royalties, additional payments at pre-determined thresholds,
etc.).1616
On this basis, the OECD TPG state that if unrelated enterprises in com-
parable circumstances would have agreed on the inclusion of a contractual
mechanism to address the valuation uncertainty (e.g. a price adjustment
or renegotiation clause or a contingent payment structure), the tax admin-
istration should be entitled to impute such a mechanism against which to
benchmark the controlled pricing. Further, if third parties would have con-
sidered subsequent events so fundamental that their occurrence would have
led to a renegotiation of the fixed pricing, the tax administration may adjust
the price to an arm’s length level following such events.
This is essentially in line with the 2010 consensus text, apart from the new
reference to contingent payment terms as a mechanism to address valua-
tion uncertainty.1617 The significant new addition in the 2017 text is that the
“old” 2010 text is now linked to new guidance on hard-to-value intangibles
(HTVIs).1618
1614. OECD TPG, para. 6.182.

1615. OECD TPG, para. 6.184.
1616. OECD TPG, para. 6.183.
1617. The emphasis on contingent payment arrangements is likely influenced by the
US intangible property (IP) ownership provisions, pursuant to which the IRS may im-
pute contingent payment arrangements aligned with economic substance to allocate
residual profits; see the analysis in sec. 21.3.3.
1618. The examples used in the 2010 text are now, thankfully, scrapped. The examples
were ambiguous and did not provide any useful guidance. The first example pertained
to a controlled, 3-year licence of proven make-sell rights to a drug with a fixed royalty
rate (see 2010 OECD TPG, annex to ch. VI, Example 1). The example did not properly
explain why unrelated parties would have insisted on a price adjustment clause, but
simply stated that one should “assume there is evidence that independent enterprises
would have insisted”. The valuation uncertainty referred to in the example pertained to
typical market risks that will normally exist for transfers of unique intangibles. It does
not seem entirely obvious that independent parties would have included adjustment
clauses due to such typical risks. In the second example, a parent company licensed
497
HTVIs are defined as follows:

[I]ntangibles or rights in intangibles for which, at the time of their transfer
between associated enterprises, (i) no reliable comparables exist, and (ii) at
the time the transaction was entered into, the projections of future cash flows
or income expected to be derived from the transferred intangible, or the as-
sumptions used in valuing the intangible are highly uncertain, making it dif-
ficult to predict the level of ultimate success of the intangible at the time of
the transfer.1619
Transactions involving HTVIs typically include (i) transfers of partially

developed intangibles (R&D); (ii) intangibles that are not expected to be
exploited commercially until years after the transfer; (iii) an intangible that
is not in and of itself an HTVI but is integral to the development of other
intangibles that qualify as HTVIs; (iv) intangibles that are expected to be
exploited in a novel manner and lack a track record; (v) an intangible that
is transferred for a lump sum; or (vi) an intangible that is contributed to or
developed under a CSA.
It is the author’s view that the HTVI definition in most cases will encom-
pass all unique and valuable intangibles that are capable of generating re-
sidual profits. There will likely be significant uncertainty pertaining to fac-
tors such as the useful life of an intangible (risk of technical obsolescence),
discount rates, etc. This is amply illustrated in Veritas, which pertained
to the transfer of intangibles connected to established, high-selling soft-
ware. The fact that the definition is intended to encompass a broad range
of intangibles is further indicated by the comprehensive listing of typical
transactions that fall within the definition.
The point of the HTVI guidance is that information asymmetry will ex-
ist between multinationals and tax authorities with respect to the val-
uation of HTVIs. A multinational will possess the technical business
knowledge necessary to critically estimate potential profits and associ-
ated risks. Tax authorities generally lack this information, placing them
in a precarious situation with respect to reviewing the ex ante taxpayer
pricing.
make-sell rights to a microchip to its subsidiary for 5 years at a fixed royalty rate of 2%
(see 2010 OECD TPG, annex to ch. VI, Example 3). The example was, in the author’s
view, unfit to provide guidance on the periodic adjustment provision, as in reality, it
pertained to an ordinary transfer pricing review of an agreement that was not at arm’s
length at the outset, as opposed to a periodic adjustment.
1619. OECD TPG, para. 6.189. See also Navarro (2017), at p. 242, on this topic.
498
To compensate for this information asymmetry, the new 2017 guidance

entitles tax administrations to use ex post profit data to inform the determi-
nation of whether the ex ante controlled pricing was arm’s length.1620
The link between the old 2010 language and the new HTVI guidance is as
follows: provided that none of the exceptions from the periodic adjustment
provision are triggered, ex post profit data may be used to determine the ex
ante pricing pursuant to the new HTVI guidance. The profit allocation re-
sult that follows from the reassessed ex ante price is regarded as that which
would follow from the price adjustment or renegotiation clauses adopted
by third parties pursuant to the old 2010 language. Thus, the new HTVI
guidance will, in practice, inform both as to when a periodic adjustment
can be carried out (i.e. when the exceptions from the periodic adjustment
provision are not triggered) as well as the extent of it (the data on actual
profits will, in practice, be determinative of the price).
This represents a significant clarification of the threshold for applying the

OECD periodic adjustment provision and puts an end to the comprehen-
sive ambiguity that has plagued the periodic adjustment guidance since its
inception 2 decades ago.1621 The 2017 text expresses the opinion that the
HTVI guidance is consistent with the arm’s length principle, as it does not
entail the use of hindsight when ex post financial results are taken into ac-
count without considering whether the information could or should reason-
ably have been known or considered at the time at which the transaction
was entered into. This is sought to be ensured through the four exceptions
to the HTVI adjustment provision.
1620. OECD TPG, para. 6.192. Ex post actual financial results provide presumptive
evidence as to the arm’s length character of the controlled pricing, and may be rebutted
by the taxpayer. See the discussion below in this section of the exceptions to the provi-
sion to use ex post results to assess ex ante pricing. The OECD released a discussion
draft on 23 May 2017 containing proposed guidance on the practical implementation of
the hard-to-value intangible (HTVI) language contained in the 2017 OECD TPG. See
also Brauner (2016), at p. 109, where he notes that the new guidance on HTVIs does not
go beyond the confides of the arm’s length principle. The author agrees. However, see
also Fedusiv (2016), at p. 488, where it is argued that the HTVI guidance goes beyond
the arm’s length principle and thus may result in double taxation, essentially based on
the view that the guidance cannot be reconciled with the wording of art. 9 of the OECD
Model Tax Convention.
1621. The HTVI guidance should, in the author’s view, be seen as a success for the
OECD. It was a long time coming, being among the last items in the intangibles project
to reach consensus. The guidance offers generous authority to perform periodic adjust-
ments.
499
First, no adjustment may be made if the taxpayer provides:1622

− details of the ex ante projections used at the time of the transfer to de-
termine the profit allocation (including how risks were accounted for,
e.g. probability weighted, the consideration of reasonably foreseeable
events and other risks and the probability of occurrence); and
− reliable evidence that any significant difference between the financial
projections and actual outcomes is due to (i) unforeseeable develop-
ments or events occurring after the determination of the price that
could not have been anticipated by the associated enterprises at the
time of the transaction; or (ii) the playing out of the probability of oc-
currence of foreseeable outcomes and evidence that these probabilities
were not significantly overestimated or underestimated at the time of
the transaction.
An example is provided in which the controlled pricing was based on sales

projections of 100, and actual sales turned out to be 1000. If the higher
sales were due to an exponentially higher demand for the products due to a
natural disaster or some other unexpected event that was clearly unforesee-
able at the time of the transaction or if they were appropriately given a very
low probability of occurrence, the ex ante pricing should be recognized as
arm’s length.1623 This establishes a comprehensively high, force majeure-
like threshold for applying the exception, akin to the US “extraordinary
events” exception.1624
Second, no adjustment may be carried out if the transfer of the HTVI is

covered by a bilateral or multilateral advance pricing arrangement in ef-
fect for the period in question between the countries of the transferer and
transferee.1625
Third, no adjustment may be carried out if the actual compensation for the
HTVI does not deviate more than +/- 20% from the financial projections
that formed the basis for the controlled pricing at the time of the transac-
tion.1626 This exception is akin to the +/- 20% allowed range of discrepancy
between the projected and actual profits under the US regulations.1627
1622. OECD TPG, para. 6.193, i).

1623. OECD TPG, para. 6.194.
1624. See the discussion of the US exception in sec. 16.3.3.5.
1625. OECD TPG, para. 6.193, ii).
1626. OECD TPG, para. 6.193, iii).
1627. See the analysis in secs. 16.3.3.3. and 16.3.3.4.
500
Fourth, no adjustment may be carried out if a commercialization period of

5 years has passed following the year in which the HTVI first generated
unrelated party revenues for the transferee, and during this period, the dif-
ferences between the projected and actual outcomes were not greater than
+/- 20%.1628
In summary, it seems clear that the 2017 revision of the periodic adjust-
ment provision was heavily influenced by the solutions drawn up in the
US regulations. As the case is for the US regulations,1629 it may be ques-
tioned as to whether the OECD periodic adjustment provision in reality
is limited to taking into account actual profits only based on information
that could or should reasonably have been known or considered at the time
at which the transaction was entered into. Actual profits will be used in
all cases in which the deviation from projected profits exceeds +/- 20%
and the profits were not caused by force majeure-like events. As the author
sees it, this may also encompass information that not necessarily could or
should reasonably have been taken into account at the time of the trans-
fer. His view is that this is necessary in order to ensure effective periodic
adjustments.
A systemic difference is that the OECD exceptions, as opposed to those

of the US regulations, are not categorized pursuant to which method the
taxpayer applied in order to price the controlled intangibles transfer. A pe-
culiarity is that the OECD exceptions do not include an exception to the pe-
1628. OECD TPG, para. 6.193, iv).

1629. The key role played by actual profits run through the US regulations as a red
thread. When the initial pricing is not based on a “genuine” CUT, a periodic adjust-
ment will be performed if the actual results deviate more than +/- 20% from the profits
projected at the time at which the controlled agreement was entered into, given that
the “extraordinary events” exception is not triggered. The role of actual profits in the
context of periodic adjustments of CSA buy-ins is even more pronounced. Under the
current regulations, a periodic adjustment examination will be triggered if the actual
profits of a foreign CSA participant exceed 150% of that participant’s CSA investment.
The “extraordinary events” exception prevents reallocations here also. See also the
2007 IRS Coordinated Issue Paper (CIP) (LMSB-04-0907-62), at p. 9, which instructs
the IRS to “decline to make a periodic adjustment to a royalty on the basis of outcomes
that could not be reasonably anticipated at the time the intangible transfer was entered
into. The regulations clearly reflect the intent that the IRS exercise restraint in making
periodic adjustments based on only the upfront reasonable expectations and not based
on subsequent events which could not be reasonably anticipated”. This indicates that
the IRS will not take full advantage of the periodic adjustment authority bestowed upon
it in the regulations. The author is sceptical of whether this stance will be adhered to
in practice if the question is truly put to the test. To do so could, in the author’s view,
arguably be contrary to the legislative intent behind the commensurate-with-income
standard.
501
riodic adjustment provision when the initial controlled pricing is based on

a “genuine” CUT. While likely not of significant practical importance due
to the general lack of CUTs pertaining to unique and valuable intangibles,
it does, in principle, seem unacceptable to perform a periodic adjustment in
these cases. The CUT will represent a “guarantee” that third parties would
indeed operate with fixed pricing here.
In the author’s view, it is clear that an exception to the adjustment provi-

sion must apply in these cases. This follows from the overall purpose of the
arm’s length principle, i.e. to provide parity in the taxation of related and
unrelated parties. Such an exception may also be anchored in paragraph
6.185 of the OECD TPG, as it is clear that in cases in which the taxpayer
pricing is supported by a “genuine” CUT, it will be established that third
parties would not have agreed on the inclusion of a “mechanism to address
high uncertainty in valuing the intangible”.
Further, under both the US and OECD provisions, tax authorities will –
apart from when the initial pricing is based on a “genuine” CUT (and when
there is an advance pricing agreement) – be entitled to perform a periodic
adjustment in cases in which there is a discrepancy between projected and
actual profits exceeding +/- 20%, unless these are caused by a force ma-
jeure-like event (war, natural disasters, etc.).
Perhaps the most interesting aspect of the 2017 HTVI guidance is the peri-
odic adjustment exception for cases in which the HTVI transfer is covered
by a bilateral or multilateral advance pricing agreement. Such an agree-
ment represents a safe harbour from reassessments. Given the limited
scope of the other periodic adjustment exceptions, this will likely be an
attractive option for multinationals. It will be interesting to see whether
this will lead to “profit bargains” with local tax authorities and how this
may affect international profit allocation. For instance, it may be that a
jurisdiction in which a purported routine entity is resident will not accept
a TNMM-based return at the lower end of the arm’s length range but will
require application of the profit split method based on the assertion that
there is, for instance, a local marketing intangible.
Such source state profit allocation assertions may conflict with typical tax
structures employed by multinationals, such as the centralized principal
model. This may trigger a dilemma for multinationals: they can either
pursue typical “minimum profit allocations” to source states based on the
502
Periodic adjustments of buy-in pricing under the US regulations
TNMM and face the risk of potential reassessments, or they can enter into
an advance pricing agreement with a more favourable profit allocation to
the source state in order to be shielded from the uncertainty associated
with potential subsequent periodic adjustments.
Multinationals may benefit from utilizing this exception, thereby gaining

predictable profit allocations and avoiding potentially protracted contro-
versy over historical transactions, while at the same time being able to ad-
vocate their profit allocation assertions in local negotiations. Tax authori-
ties will benefit from advance pricing agreements by avoiding the use of
scarce resources to audit and prepare reassessment cases, along with the
potential of agreeing to a somewhat higher profit allocation to the source
state in exchange for the safe harbour from reassessments that the advance
pricing agreement offers to the multinational.

the US regulations
16.6.1. Introductory remarks
CSAs have traditionally been popular instruments for migrating US-de-

veloped intangibles to jurisdictions with more favourable taxation of in-
tangible income. The contribution of a pre-existing intangible to a CSA,
which triggers a buy-in requirement, is substantially similar to a partial
sale of the intangible, as significant portions of the ownership of intangi-
bles developed under the CSA is allocated to foreign CSA participants.
Thus, when a US entity contributes a pre-existing intangible to a CSA,
it relinquishes its right to the parts of the profit potential connected to
the new versions of the intangible developed under the CSA that are al-
locable to the ownership shares of the other CSA participants. It is this
loss of profit potential that the buy-in payment shall compensate the US
entity for.
The commensurate-with-income standard applies to all intangibles trans-

fers, including contributions of pre-existing intangibles to CSAs. However,
without detailed guidance on how to apply the standard in the context of
CSAs, it would be difficult to enforce the standard, as CSAs are complex
instruments and the initial transfer pricing of buy-ins itself may offer plen-
503
ty of technical challenges. This emphasizes the need for a periodic adjust-

ment provision tailored to the context of CSA buy-ins.
The current cost-sharing regulations have introduced such provisions.1630

The provisions will ensure that the profit allocation dictated by the CSA
transfer pricing methods are abided by, in the sense that the actual profit
allocation conforms to the allocation pattern intended by the methods. In
line with its general stance on periodic adjustments, the IRS has, on several
occasions, maintained that the focus of the periodic adjustment authority
for CSAs is to ensure that the initial buy-in pricing was at arm’s length and
that no adjustment will be made if the under-pricing was not knowable at
the outset.1631 Further, if the above-market returns of a foreign CSA par-
ticipant are due only to uncertainty materializing in a favourable way, no
adjustments will be made.1632
In sections 16.6.2.-16.6.5., the author will discuss the periodic adjustment

authority for buy-ins under the current cost-sharing regulations.
16.6.2. A periodic adjustment is triggered if the AERR

is greater than the PRRR
As touched upon earlier, the terminology of the final cost-sharing regula-

tions is highly specific. The author refers to section 14.2.7. for an overview
of the essential terminology.
1630. The 2005 proposed CSA regulations (70 FR 51116-01) coupled the introduc-
tion of the buy-in investor model with a periodic adjustment provision tailored specifi-
cally to buy-in payments. Its preamble referred to the 1986 Congressional scepticism
concerning migration of US-developed, high-profit intangibles for relatively insignifi-
cant lump sums or royalties, which placed all the development downside risk with the
transferring US entity and the upside profit potential with the foreign transferee; see
2005 Prop. Treas. Regs. (70 FR 51116-01) § 1.482-7, preamble, sec. E.3. A tailored
buy-in periodic adjustment provision was deemed necessary to enable the IRS to reas-
sess when the controlled pricing did not appropriately reflect the profit potential of
the transferred intangible (typically when the actual profits realized indicate that the
buy-in was mispriced), thereby realizing the legislative intention behind the 1986 com-
mensurate-with-income standard that controlled profit allocation to “reasonably reflect
the relative economic activities undertaken by each”; see 1986 Committee Report, at
p. 1015.
1631. See the comments of M. McDonald (international economist at the US Treas-
ury), in Sheppard et al. (2006).
1632. See the comments of M. McDonald (international economist at the US Treas-
ury), in Nadal (2009).
504
In terms of context, multinationals may seek to migrate US-developed in-

tangibles by way of contributing them to CSAs for the purpose of further
research so that the bulk of the ownership of future versions of the intan-
gibles will be by foreign group entities, enabling allocation of the residual
profits to these foreign entities. In these cases, a US group entity may enter
into a CSA with a foreign cash-box subsidiary resident in a low-tax juris-
diction that only provides funding to the R&D performed under the CSA.
The income method is the go-to method for determining the buy-in amount
in these typical, tax-driven CSA cases.1633
Thus, one of the more practical questions pertaining to the application of

the periodic adjustment authority in the context of CSAs is likely whether
the actual return experienced by a foreign cash-box subsidiary proves
to be larger than the CPM-based normal market return intended to be
allocated to it under the income method. If so, and provided that no ex-
ception to the adjustment provision is available, the periodic adjustment
provision will see to it that the foreign entity is stripped of all residual
profits. Indeed, the periodic adjustment authority for CSAs is a potent
instrument.
If no exception is available, periodic adjustments of buy-ins may be carried

out if a controlled participant that was obligated to pay a buy-in amount has
realized an actually experienced return ratio (AERR) that lies outside of
the periodic return ratio range (PRRR).1634 A periodic trigger occurs if the
AERR is outside the PRRR.
The PRRR consists of return ratios that are not less than 0.667 or more
than 1.5.1635

1634. Treas. Regs. § 1.482-7(i)(6)(i).
1635. Treas. Regs. § 1.482-7(i)(6)(ii). Alternatively, if the CSA participants do not
comply with the CSA regulations’ documentation requirements, the periodic return
ratio range (PRRR) will consist of return ratios that are not less than 0.8 or more than
1.25. In the 2005 proposed CSA regulations (70 FR 51116-01), the range was set to
actual return ratios of no more than 2 and no less than 0.5, in practice, entailing that an
examination would be triggered in all cases in which the foreign transferee experienced
an actual return that exceeded a mark-up of 200% of his CSA investment. The only
significant change to the periodic adjustment authority in the 2009 temporary CSA
regulations (74 FR 340-01) was that the threshold for triggering a periodic adjustment
examination was lowered, as the PRRR was set to a range between 0.667 and 1.5, mak-
ing the periodic adjustment rule stricter. Thus, if the foreign transferee’s profits were
larger than 150% of its investment into the CSA, an examination (and likely a realloca-
tion) would be triggered.
505
The AERR is the present value of total profits (PVTP) divided by the pre-
sent value of investments (PVI). The present value of total profits is the
present value, as of the CSA start date, of the platform contribution trans-
action (PCT) payer’s actually experienced “divisional profits or losses”
from the CSA start date through the end of the adjustment year. Divisional
profits or losses are defined as operating profits before CSA-related expen-
ditures and taxes.1636
The present value of investment is the present value, as of the CSA start
date, of the PCT payer’s investment associated with the CSA activity, de-
fined as the sum of its cost contributions and buy-in payments from the
CSA start date through the end of the adjustment year.1637 The PVTP and
PVI are calculated using the applicable discount rate (ADR) and all in-
formation available as of the date of the determination. The ADR is the
general discount rate pursuant to Treas. Regs. § 1.482-7(g)(2)(v).1638
In the event of a periodic trigger, the Commissioner is entitled to carry out

periodic adjustments with respect to all buy-in payments between all PCT
payers and PCT payees for the adjustment year and all subsequent years
for the duration of the CSA activity. A periodic adjustment may be made
for a particular taxable year without regard to whether the taxable years
of the trigger buy-in remain open for statute-of-limitation purposes. Also,
the Commissioner may consider whether the outcome as adjusted more
reliably reflects an arm’s length result under all of the relevant facts and
circumstances, including any information known as of the determination
date.
1636. Divisional profits or losses are defined as the operating profits or losses sepa-
rately earned or incurred by each CSA participant in its division from the CSA activity,
determined before any expense (including amortization) on account of cost contribu-
tions, operating cost contributions, routine platform and operating contributions and
non-routine contributions (including platform and operating contributions), as well as
taxes (see Treas. Regs. § 1.482-7(g)(4)(iii)). It is not clear to the author as to which costs
are left to deduct from the sales in order to produce an operating profit after all of the
CSA-related costs are excluded.
1637. For the purpose of computing the present value of investments (PVI), PCT pay-
ments are all platform contribution transaction (PCT) payments due from a PCT payer
before netting against PCT payments due from other controlled participants.
1638. Treas. Regs. § 1.482-7(i)(6)(iv)(A). An exception from this rule applies for pub-
licly traded companies, for which the applicable discount rate (ADR) is the PCT payer’s
weighted average cost of capital as of the date of the trigger PCT (see Treas. Regs.
§ 1.482-7(i)(6)(iv)(B)).
506
16.6.3. Making a periodic adjustment: Applying the adjusted

RPSM
The periodic adjustment shall be carried out pursuant to the adjusted

RPSM.1639 The adjusted RPSM is the specified RPSM for determining the
initial buy-in pricing in Treas. Regs. section 1.482-7(g)(7) applied to deter-
mine the present value – as of the date of the trigger buy-in – of the buy-in
payments, with some modifications to adapt the method to the periodic
adjustment context, which the author will further explain.1640
First, the results projected by the taxpayer at the time at which the buy-in
payment was determined are replaced with the actual results up through
the determination date. Second, projected results for the remainder of the
CSA period, as estimated by the IRS at the time of the reassessment, shall
replace the original projected results, as estimated by the taxpayer at the
time at which the buy-in payment was determined. Third, the adjusted
RPSM, in contrast to the specified RPSM, is applicable even if only one
of the parties to the controlled agreement furnishes unique intangibles.1641
However, if only one party contributes unique inputs to the CSA, the ad-
justed RPSM will allocate the entire residual profits solely to that party.
The profit allocation is therefore similar to the CPM and, in particular, the
income method. The author therefore finds it rather misleading to label
the allocation method as a form of the RPSM. On the other side, if more
than one controlled party to the CSA contribute unique inputs, the adjust-
ed RPSM will allocate the residual profits among those participants, thus
earning its label as a profit split method.
However, as the lion’s share of audited CSAs presumably pertains to mi-

gration cases in which the US entity, as the only participant, furnishes pre-
existing intangibles to a CSA while funding comes from a foreign cash-
box entity, the adjusted RPSM can, for practical purposes, be seen as an
extension of the income method. This, as the adjusted RPSM reassessment,
ensures that the actual returns experienced by the foreign subsidiary are
equal to those that the income method would have allocated to it through
the discounted-cash-flow-based buy-in amount at the outset of the CSA,
had the actual results been known at that time and incorporated into the
income-method buy-in calculation.
1639. Treas. Regs. § 1.482-7(i)(6)(v).

1640. Treas. Regs. § 1.482-7(i)(6)(v)(B).
1641. Treas. Regs. § 1.482-7(i)(6)(v)(B)(3).
507
A comment on the 2006 temporary regulations included the question as to

whether it was consistent with the arm’s length standard to require the use
of the adjusted RPSM, as the arm’s length standard calls for the best-meth-
od rule to be applied in all circumstances, and the adjusted RPSM may not
be the best method in every circumstance. The US Treasury Department
and the IRS replied that this issue was sufficiently addressed in the regula-
tions, which require periodic adjustments to be administered in accordance
with the arm’s length standard.
Specifically, in determining whether to make periodic adjustments, the

IRS may consider whether the outcome as adjusted more reliably reflects
an arm’s length result under all relevant facts and circumstances. In the
author’s view, the question is interesting, but ultimately, not of much rel-
evance. The reason for this is that the adjusted RPSM in reality encom-
passes the allocation patterns of both the specified CPM and RPSM, as
it is applicable regardless of the number of participants that contribute
unique contributions to the CSA. It is therefore difficult to imagine a sce-
nario in which the adjusted RPSM cannot be applied to yield an arm’s
length result.
16.6.4. The procedure for making a periodic adjustment

to a buy-in payment illustrated
If it is concluded that a periodic trigger has occurred and that none of the
exceptions to the periodic adjustment provision are applicable, the periodic
adjustment to a CSA buy-in is determined using the steps below.1642 The list
of steps may come across as overwhelming and abstract, but the author will
go through an example of their application.
(1) determine the present value of the buy-in payments pursuant to the
adjusted RPSM, as of the date of the trigger buy-in;
(2) convert the first step’s result into a stream of contingent payments on
the base of reasonably anticipated divisional profits over the entire du-
ration of the CSA activity, using a level royalty rate. The conversion is
based on all information known as of the determination date;
(3) apply the second step’s rate to the actual divisional profit for taxable
years preceding and including the adjustment year to yield a stream
of contingent payments for these years, and convert the stream into a
present value as of the CSA start date;
1642. Treas. Regs. § 1.482-7(i)(6)(v)(A).
508
(4) convert any actual buy-in payments up through the adjustment year
to a present value as of the CSA start date, then subtract such amount
from the third step’s result. Determine the nominal amount in the ad-
justment year that would have a present value as of the CSA start date
equal to the present value determined in the previous sentence to de-
termine the periodic adjustment in the adjustment year;
(5) apply the second step’s level royalty rate to the actual divisional profit
for each taxable year after the adjustment year up to and including the
taxable year that includes the determination date to yield a stream of
contingent payments for such years. The second step’s rate applied to
a loss will yield a negative contingent payment for that year. Then,
subtract from each such payment any actual buy-in payment made for
the same year to determine the periodic adjustment for such taxable
year;
(6) for each taxable year subsequent to the year that includes the determi-
nation date, the periodic adjustment equals the second step’s rate ap-
plied to the actual divisional profit for that year. The second step’s rate
applied to a loss for a particular year will yield a negative contingent
payment for that year; and
(7) if the periodic adjustment for any taxable year is a positive amount, it
is an additional buy-in payment owed from the PCT payer to the PCT
payee. If the periodic adjustment for any taxable year is a negative
amount, it is an additional buy-in payment owed by the PCT payee to
the PCT payer.
The author will now go through an example of the periodic adjustment

provision.1643 It pertains to a US parent that enters into a CSA with a foreign
subsidiary to develop mobile phone technology. The parent contributes the
rights to an in-process piece of technology, for which compensation is due
from the subsidiary. The subsidiary has no platform contributions and no
operating contributions. The parent and subsidiary agree to fixed buy-in
payments of 40 in year 1 and 10 per year for years 2 through 10. At the
beginning of year 1, the weighted average cost of capital of the group is
15%. In year 9, the Commissioner audits years 5 through 7 of the CSA and
considers whether any periodic adjustments should be made.
The foreign subsidiary experiences the actual results during the first 7
years as shown in table 16.1.1644
1643. Treas. Regs. § 1.482-7(i)(6)(vii), Example 1.

1644. All cash flows are assumed to arrive at the beginning of each year.
509
Table 16.1
Year Sales Non-CC* CC PCT Investments Divisional AERR

costs payments (PVI) profit/loss
1 0 0 15 40 55 0 0
2 0 0 17 10 27 0
3 0 0 18 10 28 0
4 705 662 20 10 30 46
5 886 718 22 10 32 168
6 1,113 680 24 10 34 433
7 1,179 747 27 10 37 432
PV through 970 846 69 69 138 124 0.90
year 5
PV through 1,523 1,184 81 74 155 340 2.20
year 6
PV through 2,033 1,507 93 78 171 526 3.09
year 7
* CC = cost contribution
All present values in table 16.1 are calculated as of the start date of the
CSA. In any year, the present value of the cumulative investment (PVI)
represents the present value of the cost contributions and PCT payments.
The present value of the cumulative investments into the CSA through year
7 is 171. The present value of the cumulative divisional profit (PVTP) rep-
resents the present value of sales minus all costs apart from cost contribu-
tions and PCT payments. The present value of the cumulative divisional
profit through year 7 is 526.
The AERR is the present value of the divisional profit (PVTP) divided by
the PVI, which is 526 ÷ 171 = 3.09. This is a periodic trigger, as the AERR
for the subsidiary falls outside the PRRR of 0.67 to 1.5.1645 Thus, the PRRR
allows a maximum return of 150% on the investments into a CSA before
an examination is triggered.
At the time of the audit, it is determined that the first adjustment year in
which a periodic trigger occurred was in year 6, when the AERR of the
subsidiary was 2.20 (340 ÷ 155). The periodic adjustments should there-
fore be made using year 6 as the adjustment year. The arm’s length buy-in
payment from the subsidiary to the parent is determined for each taxable
1645. Treas. Regs. § 1.482-7(i)(6)(ii).
510
year using the adjusted RPSM. Periodic adjustments shall be made for each
year to the extent that the buy-in payments actually made differ from the
arm’s length buy-in payment calculation under the adjusted RPSM.
The question then turns to how the buy-in payment calculation is deter-
mined under the adjusted RPSM. The cost-shared intangibles will be ex-
ploited through year 10. The CPM-based return on the subsidiary’s routine
value chain contributions is set to 8% of the non-cost-contribution costs.
The residual profit from the cost-shared intangibles is calculated in the
manner expressed in table 16.2.1646
Table 16.2
Year Sales Non-CC Divisional CCs Routine Residual

costs profit/loss return profit
1 0 0 0 15 0 -15
2 0 0 0 17 0 -17
3 0 0 0 18 0 -18
4 705 662 43 20 53 -30
5 886 718 168 22 57 89
6 1,113 680 433 24 54 355
7 1,179 747 432 27 60 345
8 1,238 822 416 29 66 321
9 1,300 894 406 32 72 302
10 1,365 974 391 35 78 278
Cumulative 3,312 2,385 927 124 191 612
PV through year
10 as of the CSA
start date
The cumulative present value of the divisional residual profit through year
10 is 612.
The periodic adjustments are calculated in a series of steps set out in Treas.
Regs. § 1.482-7(i)(6)(v)(A).
The first step is to determine a lump sum for the buy-in payment using the
adjusted RPSM. Because only the parent made non-routine contributions
1646. Residual profits = sales − non-cost-contribution costs − cost contributions − rou-

tine return
511
to the CSA, the entire residual divisional profit constitutes the buy-in, al-
locable to the United States. This strips the entire residual profits from the
subsidiary, which only receives an 8% normal market return on its routine
contributions to the CSA.
The second step is to convert the lump-sum buy-in payment that the sub-
sidiary is obliged to make into a level royalty rate based on the reasonably
anticipated divisional profit from the exploitation of the cost-shared intan-
gibles. This is accomplished by selecting a royalty rate that will transfer
all residual profits from the subsidiary to the parent over the lifetime of the
CSA. The net present value of the divisional profits from the cost-shared
intangibles over the 10-year life is 927. As seen in table 16.2, the present
value of the residual profits over the 10-year life of the cost-sharing agree-
ment is 612. Thus, in order to extract all residual profits from the total
divisional profits, a royalty rate of 66% (612 ÷ 927) is required.
The third step is to apply the 66% royalty rate to the actual results achieved
by the subsidiary through year 6 and then determine the aggregate present
value as of the start date of the CSA, as shown in table 16.3.
Table 16.3
Year Divisional Royalty Nominal royalty Nominal Periodic

profit rate due under payments adjustment
adjusted RPSM made
1 0 66% 0 40
2 0 66% 0 10
3 0 66% 0 10
4 43 66% 28 10
5 168 66% 111 10
6 433 66% 286 10
Cumulative PV 224 74 302
as of year 1
The present value of the cumulative nominal royalty due under the adjusted
RPSM through year 6 is 224. The present value of the cumulative buy-in
payments actually made through year 6, as agreed in the controlled trans-
action, is 74. The difference between the present value of the cumulative
nominal royalty and the present value of the actual buy-in payments is 150.
This difference, representing a net present value as of the CSA start date,
is then converted into a nominal amount as of the adjustment year, using
512
the 15% discount rate. That nominal amount is 302,1647 which is also the
periodic adjustment in year 6.
Thus, through this periodic adjustment, the actual buy-in payments “catch
up” to the arm’s length level of payments that should have been made in
years 1-6.
The next step is to determine the royalties due from the subsidiary to
the parent for the years subsequent to the adjustment year, i.e. for year 7
through year 9 (the year including the determination date). These assess-
ments are made for years 8 and 9 after the divisional profit for those years
materializes. For each year, the periodic adjustment is a buy-in payment
due in addition to the 10 buy-in payment required under the CSA. This
periodic adjustment is calculated as the product of the step-two royalty rate
and the actual divisional profit, minus the 10 that was otherwise paid for
that year, in the manner expressed in table 16.4.
Table 16.4
Year Divisional Royalty Royalty PCT Periodic

profit rate due payments adjustment
otherwise made
7 432 66% 285 10 275
8 416 66% 275 10 265
9 406 66% 268 10 258
Under the last step, the periodic adjustment for year 10 (the only year after
the year containing the determination date) is determined by applying the
step-two royalty rate to the actually experienced divisional profit. This pe-
riodic adjustment is a buy-in payment from the subsidiary to the parent and
replaces the payment of the 10 otherwise due (see table 16.5).
Table 16.5
Year Divisional Royalty Royalty PCT Periodic

profit rate due payments adjustment
otherwise made
10 391 66% 258 10 (not paid) 258
1647. 302 = 150 × 1.155
513
The periodic adjustment for year 10 is therefore 258.1648
16.6.5. Exceptions to the periodic adjustment provision
In sections 16.6.5.2.-16.6.5.6., the author will discuss the exceptions from

the CSA periodic adjustment authority. No adjustment will be made to a
buy-in payment if the taxpayer establishes that one of the exceptions is ap-
plicable.
16.6.5.2. Exception: The initial buy-in pricing is based on a CUT

involving the same platform contribution
The first exception to the periodic adjustment rule applies where the initial
buy-in pricing was based on a CUT involving the same platform contribu-
tion as in the trigger buy-in, i.e. a “genuine” CUT.1649 Presumably, it will
be highly rare that this exception is relevant, as the pre-existing intangible
contributed to the CSA then would have to be used as the platform for
further research in two parallel and independent R&D processes, i.e. the
controlled CSA and an uncontrolled CSA with third parties.
16.6.5.3. Exception: Extraordinary events
Second, an exception is made if the periodic trigger is due to extraordinary

events beyond the control of the CSA participants that could not reason-
1648. Lastly, the regulations provide an example of a periodic adjustment in which a

US parent enters into a CSA with three subsidiaries to develop version 2.0 of a com-
puter program (see Treas. Regs. § 1.482-7(i)(6)(vii), Example 3). Only the parent makes
a platform contribution, i.e. version 1.0 of the program, for which compensation is due
from the three subsidiaries. The example applies the adjusted residual profit split meth-
od to allocate all residual profits from the three subsidiaries to the parent, using the
exact same methodology described in the first example (see Treas. Regs. § 1.482-7(i)
(6)(vii), Example 1, discussed in sec. 16.6.4.).
1649. Treas. Regs. § 1.482-7(i)(6)(vi)(A)(1). Under this CUT-based exception, the
same pre-existing intangible has been contributed to an uncontrolled taxpayer under
substantially the same circumstances as those in the controlled CSA and with a similar
form of payment, typically a fixed compensation structure. This CUT serves as the
basis for applying the CUT method in the first year and all subsequent years in which
substantial buy-in payments are required to be paid.
514
ably have been anticipated as of the date of the buy-in.1650 An example is

provided in the US CSA regulations in which the facts are the same as
in the example discussed previously,1651 with the exception that the first
adjustment year in which a periodic trigger occurred was year 6, when
the AERR of the subsidiary was 2.73.1652 Upon investigation as to what
caused the high return in the market of the subsidiary in years 4 through
6, it becomes clear that competitors experienced severe and unforeseen
disruptions in their supply chains, resulting in a significant increase in the
market share of the controlled parties. Without this unforeseen event, the
periodic trigger would not have occurred. It is therefore determined that no
adjustments are warranted.
16.6.5.4. Exception: Reduced AERR does not cause a periodic

trigger
The third exception is rather technical and applies in the scenario in which
a reduced AERR is not deemed to cause a periodic trigger. The exception
came to life due to comments received on the 2006 temporary regulations,
in which the argument was made that the definition of “divisional profits
or losses” is overly broad and thus attributes too much value to the concept
of the AERR, thereby making the numerator in the periodic adjustment
calculation trigger disproportionately large relative to the denominator, and
thus too easily triggered.1653 Essentially, the exception states what shall be
included in the numerator (PVTP) in the calculation of the AERR.
The exception applies when the periodic trigger would not have occurred
if the tested parties’ (PCT payers) divisional profits or losses used to cal-
culate its PVTP had taken into account both expenditures for operating
cost contributions and routine platform contributions and excluded profits
or losses attributable to its routine contributions to the exploitation of cost-
shared intangibles, non-routine contributions to the CSA activity, operat-
ing cost contributions and routine platform contributions.1654
While the author finds it clear that operating cost contributions and routine
platform contributions shall be deducted from the operating profits, he finds
1650. Treas. Regs. § 1.482-7(i)(6)(vi)(A)(2).

1652. Treas. Regs. § 1.482-7(i)(6)(vi).
1653. See the preamble to the final 2011 cost-sharing regulations (76 FR 80082-01).
1654. Treas. Regs. § 1.482-7(i)(6)(vi)(A)(3). With respect to divisional profit or loss,
see supra n. 1636.
515
it peculiar that “profits or losses attributable to the … routine contributions

… non-routine contributions … operating cost contributions and routine
platform contributions” shall be excluded. In order to comply with this, prof-
its or losses must be allocated for (i) operating cost contributions; (ii) routine
platform contributions; (iii) routine operating contributions; (iv) non-routine
platform contributions; and (v) non-routine operating contributions.
No guidance is provided on how to separately allocate profits for these

items. Even further, if profits or losses due to these items are excluded
from the divisional operating profits or losses, it does not seem clear that
there would be much left, resulting in a relatively small PVTP, and thus a
relatively small AERR. That would again raise the risk of a periodic trig-
ger occurring due to an AERR below the lower threshold of the PRRR.
This exception should be clarified on this point, as the current text seems
ambiguous.
16.6.5.5. Exception: Increased AERR does not cause a periodic

trigger
Fourth, an exception is made for the scenario in which an increased AERR

is not deemed to cause a periodic trigger. The exception applies when the
periodic trigger would not have occurred if (i) the divisional profits or loss-
es of the PCT payer used to calculate its PVTP had included its reasonably
anticipated divisional profits or losses after the adjustment year from the
CSA activity, including from its routine contributions, operating cost con-
tributions and non-routine contributions to that activity; and (ii) the cost
contributions and PCT payments of the PCT payer used to calculate its
PVI had included its reasonably anticipated cost contributions and PCT
payments after the adjustment year.
These reasonably anticipated amounts are determined based on all infor-

mation available as of the determination date.1655 Information will then be
available on the actual results for periods between the adjustment year and
the determination date. The author assumes that the aim of the exception
is to provide the taxpayer with an opportunity to prevent an adjustment in
cases in which the results subsequent to the adjustment year have proven
1655. Treas. Regs. § 1.482-7(i)(6)(vi)(A)(4). The CSA participants may assume that
the average yearly divisional profits or losses for all taxable years prior to and including
the adjustment year in which there was substantial exploitation of cost-shared intangi-
bles will continue to be earned in each year over a period equal to 15 years minus the
number of exploitation years prior to and including the determination year.
516
Periodic adjustments of buy-in pricing under the OECD TPG
to not be as great. This may be relevant when the periodic trigger occurs
“mid-stream”. The rationale is likely that the CSA may, over a longer pe-
riod, yield lower returns that, over the life of the CSA, will bring the overall
CSA return down to a level that lies within the PRRR.
16.6.5.6. Exception: Cut-off rule
The fifth exception is a cut-off rule. If the AERR is within the PRRR for
each year of a 10-year period beginning with the first taxable year in which
there is substantial exploitation of cost-shared intangibles resulting from
the CSA, there will be no periodic adjustment in the following years. Fur-
ther, there will be no adjustment if the AERR falls below the lower bounds
of the PRRR in any year of a 5-year period beginning with the first taxable
year in which there is substantial exploitation of cost-shared intangibles.1656

the OECD TPG
The 2017 OECD TPG do not provide any specific guidance on period-
ic adjustments of CSA buy-in payments, but state that the general HTVI
guidance shall be applied analogously.1657 CSAs are complex instruments
that historically have been used to migrate unique intangibles of consider-
able value out of high-tax jurisdictions. It may be difficult to ascertain, on
an after-the-fact-basis, whether the initial pricing valuation was at arm’s
length. It seems rather obvious that the lack of guidance on how to apply
the OECD’s periodic adjustment provision to buy-in payments may cause
significant controversy among treaty partners. The core message of the
new 2015 CSA guidance is that a cash-box funding entity participating in
a CSA should only be allocated a risk-adjusted rate of return on its funding
contribution so that the residual profits are allocated to the jurisdiction in
which the intangible value was created (where R&D is carried out).
In order to enforce this intention, it may be necessary to carry out periodic

adjustments to strip the foreign funding entity of any excess profits on top
of the risk-adjusted rate of return that may accumulate to it as a result of its
actual profits being higher than the profits projected at the time at which
1656. Treas. Regs. § 1.482-7(i)(6)(vi)(B)(1); see also § 1.482-7(i)(6)(vi)(B)(2).

517
the buy-in was calculated. It is not clear how this shall be accomplished by
applying the HTVI guidance analogically. It may seem that the only possi-
ble solution is to align the buy-in price with the actual profit data in cases in
which the discrepancy between the projected and actual profit data exceeds
the allowed +/- 20% range.
Given the amount of complexity in calculating a buy-in amount and tak-

ing into consideration the projected R&D costs and the risk-adjusted rate
of return, as well as the time that may pass from the formation of the CSA
and recognition of the buy-in amount until the cost-shared intangibles are
fully developed and may be exploited, the author questions whether the
periodic adjustment authority will be practicable in the context of CSAs
without more detailed guidance. There is a risk that a lack of guidance
may lead to diverging practices among treaty partners, and thereby double
taxation. The OECD should address this issue as soon as possible, as effec-
tive operationalization of the buy-in pricing methods is reliant on proper
enforcement through periodic adjustment rules, akin to those of the US
regulations.
518
Chapter 17
The Allocation of Residual Profits to

a Permanent Establishment
17.1. Introduction
In order to gain a representative impression of how intangible operating
profits are allocated under international tax law, it is imperative that arti-
cles 7 and 9 of the OECD Model Tax Convention (OECD MTC) are viewed
together, as they express the same allocation norm, i.e. the arm’s length
principle. Both articles allocate business profits between two jurisdictions,
albeit on different levels within a multinational. More specifically, arti-
cles 7 and 9 allocate profits between two parts of the same legal entity and
two different group entities, respectively. Article 7 avoids juridical double
taxation by way of determining the maximum amount of profits taxable
in the source jurisdiction, and thus the maximum amount that a residence
state is obliged to provide relief for.1658 Article 9 avoids economic double
taxation by way of allocating profits between two residence states. To the
extent that an amount of profit is allocable to one of the jurisdictions, the
other must make a corresponding adjustment to exclude the same amount
from taxation there.
In this chapter, the author will discuss how profit allocation issues that
arise when a permanent establishment (PE) provides input to an intangi-
ble value chain shall be resolved under the OECD MTC.1659 Brief com-
ments will also be tied to the treatment under the UN MTC. The question
of how IP ownership shall be assigned to PEs under the OECD MTC,
however, will be analysed first in chapter 25 so that the discussion can be
seen in light of the analysis of the 2017 OECD Transfer Pricing Guide-
lines (OECD TPG) on intangible property (IP) ownership in part 3 of the
book.
1658. See also the OECD Commentary on Article 7, para. 27.

1659. For fairly recent discussions of the allocation of profits to permanent establish-
ments (PEs), see, e.g. Russo (2005); Schön (2007); and Pereira (2009). See also Edgar
(2005); Malherbe (2010); Sharma (2008); Baker et al. (2006); Bennett et al. (2005);
and Kobetsky (2011).
519
Chapter 17 - The Allocation of Residual Profits to a Permanent Establishment
17.2. Historical background
Even though the PE concept originates from the model agreement devel-
oped by the League of Nations in the 1920s, the rules for allocating op-
erating profits among the head office and a PE remained pronouncedly
ambiguous and divisive until 2010.1660 No uniform allocation pattern was
adhered to in practice. Multinationals likely exploited this ambiguity to
support modest profit allocations, resulting in growing discontent among
OECD jurisdictions. In light of this, the OECD, in the 1990s, embarked on
a project to revise the basic methodology for allocating profits to PEs. This
led to a notable discussion draft in 2001,1661 which observed that “there is
no consensus amongst the OECD Member countries as to the correct inter-
pretation of Article 7. This lack of a common interpretation and consistent
application of Article 7 can lead to double, or less than single taxation”.1662
The project initially embraced the paradigm that the article 7 allocation
rules should not be influenced by the transfer pricing concepts developed
under article 9. The explanation for this approach is to be found in the his-
torically rooted way of thinking that PEs were markedly different from re-
lated group companies, even though both articles 7 and 9 pertain to the al-
location of business profits.1663 A significant methodological breakthrough
was reached when the OECD acknowledged that the article 9 transfer pric-
ing principles should be applied by analogy for the purpose of allocating
profits to a PE.1664 This thinking resulted in the OECD’s landmark 2008
Report.1665 This progressive report diverged significantly from the material
content of the allocation rules contained in the Commentary on Article 7 at
1660. For a historical overview, see the 1992 IFA General Report in Maisto (1992), at
pp. 62-72. For an informed discussion of the attribution of profits to PEs under the pre-
2010 article 7 regime with an emphasis on agent schemes, see Vann (2006).
1661. Discussion draft on the attribution of profits to permanent establishments
(OECD 2001) [hereinafter the 2001 OECD PE Draft].
1662. 2001 OECD PE Draft, preface, para. 2.
1663. For the views of a proponent of the historical school of reasoning, see Vann
(2003), at p. 163, as well as pp. 143 and 157.
1664. On the gradual gravitation of the OECD towards applying a transfer pricing ap-
proach to allocate business profits to PEs, see the (critical) comments in Vann (2003), at
pp. 137-139. For critical comments on the authorized OECD approach transfer pricing
analogy, see Schön (2010a), at p. 252.
1665. OECD, Attribution of Profits to Permanent Establishments (OECD 2008) [here-
inafter 2008 Report]. It was found necessary to amend the commentaries to reflect the
new principles drawn up in the Report. This proved problematic, as the wording of
art. 7 of the OECD Model Tax Convention (OECD MTC) itself formed a restriction on
how far the commentaries could go in adapting the material solutions contained in the
Report.
520
Historical background
that time, also with respect to the allocation of residual profits from unique
IP. The historical OECD position on IP was that the principles developed
under the OECD TPG “cannot be applied in respect of the relations be-
tween parts of the same enterprise”,1666 and that “it may be extremely dif-
ficult to allocate ‘ownership’ of the intangible right solely to one part of
the enterprise and to argue that this part of the enterprise should receive
royalties from the other parts as if it were an independent enterprise”.
These dismissive statements made it difficult to allocate residual profits to

a PE, regardless of the extent to which it had contributed to the develop-
ment of the IP that generated the profits.1667 Further, the 2008 commentar-
ies seemingly allowed asymmetrical allocations, in the sense that costs
– but not income – resulting from IP development could be attributed to a
PE. No rationale was provided as to why this should be sensible. The profit
allocation result of the 2008 commentaries was likely that residual profits
from intra-group-developed IP were more or less always extracted from the
source jurisdiction. It goes without saying that this was detrimental to an
equitable distribution of operating profits among jurisdictions.
In 2010, a new version of article 7 was introduced, allowing full incorpo-

ration of the allocation principles developed in the 2008 Report.1668 The
report itself was updated in 2010 to incorporate references to the new lan-
guage of article 7,1669 but its material content was not altered from the 2008
version.
1666. See 2008 OECD Commentary on Article 7 (as it read before the 22 July 2010
revision, available in the 2008 update to the OECD Model Tax Convention), para. 34.
1667. It could be argued that these 2008 commentaries were in conflict with the word-
ing of OECD MTC art. 7(2) itself, expressing the fundamental separate entity ap-
proach. It seems fairly obvious that there must be scenarios in which a PE should be
entitled to residual profits, as unrelated enterprises would demand compensation for
their intangible development contributions.
1668. For the purpose of interpreting tax treaties based on the pre-2010 version of
art. 7, the revised 2008 commentaries must be applied. While the 2008 Report serves
as a background, there are limits as to how far it may influence the interpretation.
Its guidance on IP, services and internal debt (the core of the 2008 Report) is clearly
incompatible with the allocation paradigm of the pre-2010 version of art. 7 and the
revised pre-2010 commentaries. Over time, this problem will eventually become irrel-
evant, as new treaties will likely be based on the 2010 version of art. 7. There are new
treaties that have adopted the 2010 art. 7 approach to profit allocation, e.g. the Conven-
tion between the Kingdom of Norway and the United Kingdom of Great Britain and
Northern Ireland for the Avoidance of Double Taxation and the Prevention of Fiscal
Evasion with Respect to Taxes on Income and on Capital Gains (14 Mar. 2013), Trea-
ties IBFD. More will hopefully follow soon.
1669. OECD, Report on the attribution of profits to permanent establishments (OECD
2010) [hereinafter 2010 Report].
521
The discussion here is limited to the 2010 version of article 7. This is the
only version relevant to future treaties based on the OECD model. For this
version, the 2010 Report reigns as the supreme source of interpretation.1670
The report emphasizes that its transfer pricing approach is based on apply-
ing the OECD TPG by analogy.1671 It is specifically stated that, to the extent
that the OECD TPG are modified in the future, the 2010 Report should be
applied by taking into account the guidance in the OECD TPG as modified
from time to time.1672 This is a key point, in particular with respect to IP,
as the 2017 OECD TPG on the issue of IP ownership and transfer pric-
ing of IP depart significantly from the solutions drawn up in the previous
1995/2010 versions of the OECD TPG. Thus, to the extent that there is a
conflict between the allocation guidance contained in the 2010 Report and
the 2017 OECD TPG (which there is), the latter prevails.
17.3. The relationship between articles 7 and 9
The author would like to point out two noteworthy differences between
articles 7 and 9 with respect to profit allocation.1673 First, an article 9 alloca-
tion is potentially more attractive for profit shifting purposes than an arti-
cle 7 allocation. A residence state will normally opt to provide relief under
article 7 by way of credit for taxes paid in the source state as opposed to
the exemption method. This will effectively ensure that the multinational is
only relieved of residence taxation to the extent that is taxed on its source-
state profits. There is no similar mechanism under article 9. If an amount of
profit is allocable to residence jurisdiction 1, there will be a requirement for
residence jurisdiction 2 to exclude the amount from taxation, regardless of
whether it has been taxed in residence jurisdiction 1. In effect, relief under
article 9 is limited to an exemption method.
Second, there is a “cliff effect” associated with article 7, which lacks a par-
allel under article 9. Only if a PE is deemed to exist will allocation under
article 7 be triggered. Profit allocation is thus contingent on there being a
PE. The PE threshold is due to the fundamental system of the distributive
rules of the OECD MTC, which provide exclusive taxation rights for the
residence jurisdiction in the absence of a specific exception allowing the
source state to tax a particular item of income (in this case, the identifica-
1670. See OECD Commentaries, art. 7, para. 9.

1671. 2010 Report, preface, para. 10.
1672. Id.
1673. See also Vann (2003), at p. 151, on the relationship between profit allocation
under arts. 7 and 9.
522
The article 7 profit allocation system
tion of a PE).1674 This two-tiered character of the PE rule has caused ten-
sion between article 7(1) (governing whether a PE exists) and article 7(2)
(governing the profit allocation) in the sense that a taxpayer will be able
to entirely avoid profit allocation to a source state as long as a PE is not
deemed to exist. There has been widespread exploitation of this “cliff”
by multinationals, as amply illustrated through the use of commissionaire
structures to avoid PE status.1675
Thus, now close to a century after the inception of the PE concept in inter-
national tax law, it can be observed that the principles for allocating oper-
ating profits between a head office and a PE under article 7 of the OECD
MTC represent loyal applications of the basic transfer pricing concepts
developed under article 9. The substantial difference between the transfer
pricing rules under article 9 and the PE rules under article 7 lies not in the
material content of the allocation rules, but in the fact that the PE threshold
must be surpassed in order to trigger profit allocation to the source juris-
diction under article 7, while no similar threshold applies under article 9.
17.4. The article 7 profit allocation system
Article 7(2) of the OECD MTC determines the following:

[A PE shall be allocated the operating profits] it might be expected to make, in
particular in its dealings with other parts of the enterprise, if it were a separate
and independent enterprise engaged in the same or similar activities under
the same or similar conditions, taking into account the functions performed,
assets used and risks assumed by the enterprise through the permanent estab-
lishment and through the other parts of the enterprise.
To operationalize this approach, the 2010 Report assigns economic owner-

ship of assets, capital and risks to the PE, resulting in income and deduc-
1674. See art. 1 OECD MTC, which sets out that the model only applies when one of the
persons involved is resident in one of the contracting states; and art. 21, para. 1 OECD
MTC, which states that any income of a resident that has not specifically been allocated to
the source state (through the distributive rules) is taxable only in the residence state. Art. 7
bestows upon the source state the right to tax business profits earned there if the multina-
tional “carries on business … through a permanent establishment situated therein”.
1675. See OECD, Preventing the Artificial Avoidance of Permanent Establishment
Status – Action 7: 2015 Final Report (OECD 2015), International Organizations’ Doc-
umentation IBFD [hereinafter 2015 Action 7 Report].
523
tions. The underlying chain of reasoning is that assets give rise to risks,
which then must be allocated capital, parts of which will be interest-
bearing and give rise to interest deductions in the profit calculation for the
PE.1676 Additional profits are allocated on the basis of assigning related and
unrelated transactions to the PE, as well as recognizing its dealings with
the head office. To frame the discussion of the IP transfer pricing issues for
PEs (as well as the IP ownership issue discussed in chapter 25), the author
will tie some comments to the 2010 profit allocation system.
17.4.2. Assignment of assets, capital and risk to the PE
The concept of “significant people functions” plays a fundamental role in

the profit allocation regime drawn up in the 2010 Report.1677 Its primary
purpose is to assign economic ownership of assets and risks to a PE, in-
cluding economic ownership of IP, both internally developed and exter-
nally acquired.1678
While it is clear that the significant-people-functions concept pertains to

the identification of functions of elevated importance for profit allocation
purposes, the 2010 Report ironically makes no attempt to define the term,
1676. 2010 Report, part 1, paras. 21-26.

1677. See also Schön (2007), sec. III.A.1; and Oosterhoff (2008). For critical com-
ments on derivative applications of the concept for the purpose of allocating risk among
group entities, see Schön (2014), at pp. 20-22. See also the discussion in sec. 6.6.5.5.2.
1678. 2010 Report, part 1, para. 18. In contrast, the parameter for assigning economic
ownership of tangible assets is “use”, i.e. where the asset is physically located (see 2010
Report, part 1, para. 194). The profits allocable to a PE as a result of its ownership of
tangibles should, in theory, be unaffected by whether the use or significant people func-
tions comprise the allocation key. Assume, for instance, that an oil platform is used by
a PE, while the significant people functions are performed at the head office. If use is
the allocation key, the PE is deemed the owner and is allocated depreciation deductions
and interests expenses on external debt financing. Conversely, if the significant people
functions are applied as the allocation key, the PE will be deemed to lease the platform
and will be allowed deductions for lease payments. The view of the OECD was that,
over time, the depreciation and interest deductions would equal the deductible lease pay-
ments. Nevertheless, the short-term profits allocable to the PE will differ, with more
deductions initially in the owner scenario. Both approaches may be difficult to apply in
practice. In the ownership scenario, it may be difficult to determine the basis for depre-
ciating relatively “unique” tangibles and attribute capital to the PE. The lease scenario
may be challenging with respect to assets that have no clear significant people func-
tions attached, e.g. assets that rely on) “stop/go” orders. Of course, the location where
an asset is used may coincide with the place where the significant people functions are
performed, as typically will be the case for financial assets; see the 2010 Report, para.
20.
524
under the reasoning that the significant functions will vary among busi-
ness sectors.1679 Nevertheless, the author finds it clear that the core of the
concept is to allocate profits to the part of the enterprise where the sig-
nificant financial risk-taking functions are carried out. Over two thirds of
the pages of the 2010 Report pertain to profit allocation to PEs of banks,
insurance companies and global trading enterprises. Significant functions
for such PEs are referred to as “key entrepreneurial risk-taking” (KERT)
functions.1680 These functions “require active decision-making with regard
to the acceptance and/or management (subsequent to the transfer) of indi-
vidual risks and portfolios of risk”.1681 The OECD drafted part 2 of the 2010
Report on profit allocation to PEs of banks before the general guidance in
part 1 was penned. An unfortunate consequence of this seems to have been
that the KERT-functions concept was uncritically carried over to govern
also the assignment of economic ownership of assets to PEs outside the
financial sector, even though financial risks do not necessarily drive value
creation to the same extent in other sectors as in the financial industry.1682
1679. 2010 Report, part 1, para. 16.

1680. See 2010 Report, part 2, paras. 8-11. See also Vann (2006), at p. 378, on key
entrepreneurial risk-taking (KERT) functions.
1681. 2010 Report, para. 8. The OECD Discussion draft on the attribution of profits to
permanent establishment – Part I: General considerations (OECD 2004) [hereinafter
the 2004 PE draft] applied the term “KERT functions” (as opposed to the concept of
significant people functions) to assign economic ownership of assets (including inter-
nally developed IP) also outside of the financial, insurance and global trading sectors;
see 2004 PE draft, para. 228. It can be observed that the 2004 description of KERT
functions in the context of IP development is almost identical to the description of
significant people functions in the 2010 Report. The factual pattern in InverWorld,
Inc. v. CIR, T.C. Memo. 1996-301 (Tax Ct., 1996), seems to illustrate “typical” KERT
functions. At issue was whether a Cayman Islands-incorporated company, Inverworld
Limited (LTD), was taxable in the United States through its wholly owned US subsidi-
ary, Inverworld Inc (INC), under US domestic law. LTD was not afforded treaty protec-
tion, as there was no tax treaty between the United States and the Cayman Islands. The
question was whether LTD was engaged in trade or business within the United States. If
so, it would be taxable on income effectively connected to the conduct of such trade ac-
cording to IRC sec. 882(a)(1) and Treas. Regs. § 1.864-4(c)(5)(i). LTD advised and sold
financial services to external clients and received client funds to invest. LTD carried
out a range of functions and transactions in the United States through INC, including
receiving funds, placing loans, monitoring interest rates, valuing portfolios, dispersing
dividends and interests and selling and purchasing financial instruments. LTD also sup-
plied funds to INC, and INC acted almost exclusively for LTD and LTD’s clients. Had
the case been assessed under art. 7 of the OECD MTC, the author finds it likely that the
KERT functions would indeed have been identified at the level of the US PE.
1682. Precisely why the key functions for being assigned economic ownership of assets
for non-financial PEs were renamed from “KERT functions” in the 2004 PE draft to
“significant people functions” in the 2010 Report is unclear. Perhaps it was to stress the
distinction between financial and non-financial businesses; see Schön (2007), at sec.
525
The part of the enterprise that is assigned economic ownership of IP will

be entitled to “the income attributable to the ownership of the asset, such
as royalties; the right to depreciate a depreciable asset; and the potential
exposure to gains or losses from the appreciation or depreciation of the
asset”.1683
Thus, in order for a PE to be allocated residual profits, it must first be

assigned economic ownership of the IP in question (pursuant to the sig-
nificant-people-functions doctrine). This IP ownership issue is analysed
in depth in chapter 25. At this point, the author will just make one ob-
servation that relates to the significant-people-functions doctrine. In the
context of assigning economic ownership to internally developed IP, the
2010 Report specifically states that the main risk is losing the research and
development (R&D) investment.1684 This point of view originates from the
2004 OECD discussion draft on the allocation of profits to PEs, in which
it was stated that “performance of the development function(s) does not
of itself determine the legal ownership. Rather the key issue is which en-
terprise acts as the entrepreneur in deciding both to initially assume and
subsequently bear the risk associated with the development of the intangi-
ble property”.1685
Aligned with this, the 2010 Report allocates economic ownership to intra-
group developed IP to the part of the enterprise that assumes or manages
the financial development risks. As the author will revert to in chapter 25,
this 2010 position stands in stark contrast to the position taken by the OECD
on IP ownership of manufacturing IP in the 2017 OECD TPG.1686 The 2010
focus on R&D-funding facilitated BEPS, as multinationals placed funding
risk in low-tax jurisdictions. Also, the 2010 Report ignored that R&D is the
dominating IP value driver in a transfer pricing context and should thus be
determinative for the allocation of residual profits.1687
Further, once the PE has been allocated economic ownership of IP and

other assets, the question will be how much capital should be assigned to
the PE. The 2010 Report states that the PE should have “sufficient capital
III.A.1. However, as the alteration pertained solely to the “labelling” of the relevant
functions as opposed to the underlying material guidance, the author finds that strange.
A more likely explanation may be that there were concerns that the “KERT” terminol-
ogy could influence the interpretation of the PE threshold; see Bell (2007).
1683. See 2010 Report, footnote 4.
1684. 2010 Report, para. 89.
1685. See 2004 PE draft, para. 227.
1686. See the analysis in ch. 22.
1687. See the discussions in secs. 21.2. and 22.2. on value drivers in IP development.
526
to support the functions it undertakes, the assets it economically owns and

the risks it assumes”.1688
Capital is attributed to a PE based on the value of its assets.1689 The valua-

tion may be based on a variety of parameters, including financial account-
ing value, R&D costs, purchase price and market value.1690 Valuations of
unique IP tend to be subjective and may serve to increase the total amount
of assets (and thereby the interest-bearing capital) allocable to the PE. In
order to avoid speculative tax planning, the valuation of externally ac-
quired and internally developed IP should arguably be restricted to their
purchase price and R&D costs, respectively.1691
The 2010 Report requires additional free capital to be allocated to the PE

if it incurs entrepreneurial risks, for instance, through the development of
marketing IP in the source state without reimbursement of expenses from
the head office.1692 The allocation of capital to the PE cannot, in these cas-
es, be based on asset value alone. The additional free capital requirement
applies also in the context of the internal development of manufacturing
IP, to support the risk of the R&D efforts failing,1693 as well as for risks
connected to the subsequent IP exploitation (e.g. product liability risks).
The capital allocated to the PE cannot, in its entirety, be interest-bearing.

A portion of it must be “free” so that source-state operating profits are
not fully extracted through the payment of interest expenses to a foreign
head office.1694 The 2010 Report does not recognize interest expenses on
internal debt from the PE to the head office.1695 Thus, the issue of allocating
interest-bearing debt to the PE pertains only to the question of how much
of the enterprise’s external loans may be allocated to the PE.1696 The 2010
1688. 2010 Report, para. 28.

1689. 2010 Report, para. 109.
1690. 2010 Report, paras. 109-111.
1691. It should be noted that a multinational would likely not want to inflate its IP
values in the source jurisdiction for the purpose of maximizing its local interest deduc-
tions. A high value may be problematic if it later wants to migrate the IP to a different
jurisdiction, as the income tax transfer charge will be bigger.
1692. 2010 Report, para. 111.
1693. 2010 Report, para. 89. Relevant research and development (R&D) costs will be
the salaries of the researchers, costs for property and equipment and costs for the use of
pre-existing IP owned by unrelated or related parties or by the head office.
1694. See 2010 Report, part 1, para. 32.
1695. See 2010 Report, paras. 157-158.
1696. The fact that deductions for internal “interest payments” are completely cut off
under art. 7 is a surprisingly progressive solution. When source states have domestic
thin capitalization rules that allow interest deductions on internal debt within certain
527
Report provides an ambiguous answer to this question, as the OECD mem-

ber countries did not reach a clear consensus on the methodology. The 2010
Report accommodates a range of methods.1697
The 2010 Report also leaves considerable leeway for determining the in-
terest rate on the debt portion of the PE’s capital. As this pertains to ex-
ternal loans, the rate itself is at arm’s length. The problem, however, is to
determine which of the enterprise’s external loans should be deemed to be
included in the debt of the PE.1698 This is a matter of causality, and not a
simple one. The multinational and the source state will be incentivized to
allocate loans with high and low interest rates to the PE, respectively.
17.4.3. Assignment of transactions and dealings to the PE
The last part of the article 7 profit allocation system pertains to the as-
signment of profits from transactions and dealings to the PE.1699 First, a
PE shall be allocated operating profits from transactions between the en-
terprise and unrelated and related enterprises as are properly attributable
to the PE.1700 This allocation hinges on the identification of “those of the
limits (e.g. 30% of taxable EBITDA), the source and residence jurisdictions take on
more or less burdens in the form of more or less interest deductions, respectively, than
intended by the 2010 Report. It is the domestic law prerogative of source states to do so.
1697. The capital allocation approach assigns a portion of the free capital of the enter-
prise to the PE based on the proportion of assets allocated to the PE relative to the total
assets of the enterprise (if the PE is allocated 10% of the assets, it will receive 10% of
the free capital); see the 2010 Report, paras. 121-127. The economic capital approach
allocates free capital based on a wider range of risks (e.g. R&D risks); see the 2010
Report, para. 128. The thin capitalization approach requires a PE to have the same
amount of free capital as a comparable independent enterprise; see the 2010 Report,
paras. 129-133. A safe-harbour method is also allowed (and other methods particular to
insurance enterprises); see the 2010 Report, para. 139 and part IV. See also the discus-
sion in the 2010 Report in paras. 99-149, as well as the statement in para. 147. There
seems to be tension between this methodological freedom and the premise that the PE
should have the same creditworthiness as the enterprise as a whole (implying a similar
capital structure); see the 2010 Report, part 1, para. 30 and paras. 99-101.
1698. Two basic methods are used to determine the interest rate on PE debt funding.
The first is the tracing method, pursuant to which all internal movements of funds
between the head office and the PE are traced back to the unrelated lender. The PE is
then allocated the same interest rate as the lender charged. The second is the fungibil-
ity approach, which disregards the actual movement of funds and allocates a portion
of the actual interest expenses on the external debt of the enterprise based on a pre-
determined allocation key.
1699. 2010 Report, para. 44.
1700. 2010 Report, para. 98.
528
Transfer pricing of the IP transactions and the dealings of a PE
enterprise’s transactions with separate enterprises which should be hypoth-

esised to have been entered into by the PE”.1701 The 2010 Report finds that
it should be possible to determine this by way of a functional analysis of
the enterprise, taking into account the functions, assets and risks of the PE.
Second, a PE shall be allocated operating profits from dealings between

the PE and other parts of the enterprise, i.e. a form of “deemed transac-
tions”. With respect to dealings in which IP is transferred between the head
office and the PE, an analytical challenge will be the lack of a controlled
agreement that transfers the rights to the IP. It will therefore not be clear at
the outset as to which particular rights are purportedly conveyed to the PE.
It is, however, important not to exaggerate the significance of this issue. In
substance, the same problem is present in the context of the transfer pricing
of IP under article 9, because one cannot take for granted that the specifica-
tion of transferred rights in a written controlled agreement is aligned with
the economic substance of the actual behaviour of the controlled parties.
Only a thorough and careful functional analysis will reveal whether that
is the case. If it is not, the terms that may be deduced from the actual be-
haviour will prevail for pricing purposes. For example, it may be possible
to observe how the multinational has organized its distribution in other
markets and to interpret the dealing in light of this, e.g. it may be that it
uses local subsidiaries in common-law countries to distribute, market and
sell products, which license rights to marketing intangibles. In civil-law
countries, it may use commissionaire structures. To the extent that any
PEs are triggered in the civil-law countries, it may logically be assumed
that they have similar rights to the relevant marketing intangibles as the
local subsidiaries in the common-law countries, provided that the PEs and
subsidiaries contribute similar value chain inputs.
17.5. Transfer pricing of the IP transactions and

the dealings of a PE
The allocation of profits to a PE from its participation in transactions can
be computed directly in the case of third-party transactions, and through
article 9 of the OECD MTC in the case of related-party transactions.1702
With respect to the dealings of the PE, profit allocation shall, under the au-
thorized OECD approach, be based on “a comparison of dealings between
1701. Id.
1702. Id.
529
the PE and the enterprise of which it is a part, with transactions between

independent enterprises”.1703
Such benchmarking shall be carried out through analogical application of

the OECD transfer pricing methodology set out in the 2017 OECD TPG.
The author refers to his earlier discussions of the transfer pricing methods,
in particular the transactional net margin method (TNMM) and the profit
split method (PSM),1704 for a detailed analysis of how IP income shall be
allocated among the related contributors to an intangible value chain. The
following discussion is limited to comments on the particular profit alloca-
tion issues touched upon in the 2010 Report, with respect to dealings that
distribute rights to use IP among different parts of an enterprise.
In order to extract as much profit from the source state as possible, multi-
nationals will typically assert that the head office is the economic owner
of IP employed in the value chain for a product that is sold in the source
state. The key issue will then be to identify which rights to the IP have
been made available to the PE. It may, for instance, be that the PE begins
to make use of a marketing intangible developed by the head office. If
this qualifies as a dealing, it must be priced as a deemed “licence”.1705 An
arm’s length notional royalty rate for the licence should then be deducted in
the calculation of the profits of the PE.1706 Source-state deduction for such
deemed payments from the PE to the headquarters is problematic in the
views of some jurisdictions.1707 These states are concerned that such deduc-
tions can trigger the issue of whether the source state should be allowed
to levy withholding tax on the deemed payments, in order to ensure equal
treatment of PEs and subsidiaries and to avoid base erosion.
For the purpose of determining the amount of the notional royalty deduc-
tion, an important issue will be to determine whether the rights conveyed
to the IP are exclusive or not. This will affect the pricing, as an exclusive
right will normally be more valuable than a non-exclusive right. The fact
that the head office made IP available to a PE does not in itself imply that
1703. 2010 Report, para. 183.

1704. See chs. 8 and 9 for an analysis of the transactional net margin method and the
profit split method, respectively.
1705. 2010 Report, para. 206.
1706. 2010 Report, para. 209. The 2010 Report uses the term “might”, but the author
cannot see any reason as to why there should be uncertainty surrounding the ability to
deduct such royalty payments for the purpose of calculating the profits allocable to the
PE, provided that it has been established that the PE should be deemed a “licensee”.
1707. See para. 29 of the OECD Commentaries on the Article 7 of the OECD MTC
(2010).
530
Transfer pricing of the IP transactions and the dealings of a PE
that the PE has obtained an exclusive right to the IP.1708 In cases in which
the PE carries out manufacturing or distribution in the source state and
there are no other parties that make or sell the same products in this geo-
graphical market, it should likely be assumed that the PE has carried out
a dealing with the head office, pursuant to which it “licenses” exclusive
rights to make or sell the products through the use of the related manufac-
turing and marketing intangibles. This will reduce the profits allocable to
the source jurisdiction relative to what would be the case if the conveyed
rights to the intangible were deemed to be non-exclusive.
Conversely, when the PE is deemed to be the economic owner of the IP, it

shall be allocated all residual profits from the IP in question. If the intan-
gible economically owned by the PE is being used in a value chain where
there are no other unique inputs, the TNMM may be used to allocate a
normal market return to the head office and other group entities that pro-
vide routine inputs. The residual profits are allocable to the PE. If there are
other unique inputs in the value chain, the PSM should be applied to split
the residual profits among the intangible owned by the PE and the other
unique intangibles, pursuant to their relative values.
Further, it may be that the enterprise has licensed an intangible from a third
party. The economic ownership to the licensed right may be allocated to
one part of the enterprise, which then makes it available to another part. In
these cases, the dealing must be priced, either as an outright transfer of the
ownership or as a licence.1709 Should the PE be deemed to have acquired
the intangible or an interest in it, the PE should be entitled to depreciate the
cost base, subject to source country rules on depreciation.1710
Further, the 2010 Report contains guidance on the treatment of dealings

akin to cost-sharing arrangements (CSAs) between an enterprise and its
PE. Chapter VIII of the 2017 OECD TPG must be applied analogically to
allocate profits from such “CSA” dealings.1711
As a PE is not a separate legal entity, the identification through a functional

and factual analysis of a “CSA” between the enterprise and its PE will be
a purely notional construct for the purpose of aiding in the allocation of
profits to the PE.1712 The 2010 Report uses the term “economic CCA” (cost
1708. 2010 Report, para. 207.

1709. 2010 Report, para. 210.
1710. 2010 Report, para. 208.
1711. See 2010 Report, preface, para. 10.
1712. See 2010 Report, para. 213.
531
contribution arrangement).1713 A typical situation will be that in which IP

that has been developed by one part of the enterprise shall be further devel-
oped by the enterprise as a whole.1714 Two relatively clear instructions can
be read in the 2010 Report.
First, if one part of the enterprise has developed pre-existing IP and con-
tributes it to an R&D arrangement with other parts of the enterprise (the
purpose of which is to further develop the pre-existing IP), the other parts
must make a buy-in payment.1715 Thus, the other parts of the enterprise are
treated as having acquired an interest in the pre-existing IP. The buy-in
amount shall be set so that it allocates the residual profits to the part of the
enterprise that carries out important R&D functions, and limits the return
to the part of the enterprise that only contributes IP development funding
to a risk-adjusted rate of return.
Second, if the PE is found to have acquired only make-sell rights to the

pre-existing IP (as opposed to a beneficial interest in the IP itself), no buy-
in requirement is triggered. In these cases, a notional royalty paid from
the PE to the head office for the use of the IP may be attributed based on
applying the transfer pricing methods by analogy. The TNMM should be a
likely candidate to allocate a normal market return to the PE in such cases.
17.6. Allocation of operating profits to a dependent agent PE
A source state will only be entitled to tax the profits of a non-resident enter-
prise if it has a PE there.1716 A PE will not be deemed to exist if the enter-
1713. 2010 Report, para. 215. A range of practical issues will necessarily arise in these
“deemed” cost-sharing arrangement (CSA) scenarios. As contracts will be absent be-
tween parts of the same enterprise, countries may, in order to accept the existence of
deemed CSAs, require a multinational to meet a “significant threshold” in order to
provide reliable evidence of its position that a notional CSA is established between the
head office and a PE; see the 2010 Report, paras. 211 and 214-215. Further, when there
is no contemporaneous documentation available to support the claim, an enterprise
cannot claim the existence of a notional CSA after the fact.
1714. 2010 Report, para. 211.
1715. Id.
1716. OECD MTC, art. 7(1). The allocation of profits to a PE under US law relies on
the “effectively connected” provisions of the IRC; see IRC sec. 882(a)(1). Newer US
treaties contain language that allows the profit allocation to be based on the approach
taken in the 2008 Report. The United States will likely not allow the approach under
older treaties; see Zollo (2011), at p. 764. The question may be raised as to whether a
532
Allocation of operating profits to a dependent agent PE
prise merely carries on business through an agent of independent status.1717

If, however, a person other than an agent of independent status (i) acts on
behalf of the enterprise in the source state and has, and habitually exercis-
es, authority to conclude contracts in its name; or (ii) pursuant to the 2015
BEPS revision of article 5 of the OECD MTC, habitually plays the princi-
pal role leading to the conclusion of contracts that are routinely concluded
without material modification by the enterprise,1718 a PE will be deemed
to exist.1719 Such a PE is known as a dependent agent PE, the allocation of
operating profits to which will be the topic of discussion in this section.1720
The author is interested in this for two reasons. First, multinationals have
used certain agent schemes for distribution, marketing and sales of prod-
ucts based on unique IP in order to lower the amount of operating profits
allocable to the source state.1721 Second, there is a peculiar – and rather
ambiguous – relationship between article 9 and article 7 in these cases that
must be clarified, as it will often have consequences for the allocation of
operating profits from intangible value chains.
17.6.2. A lead-in to the discussion
When there is a dependent agent PE, the source country will have taxing
jurisdiction over two different entities. The first entity is the dependent
agent enterprise. Typically, this is a subsidiary resident in the source state
that has contracted to sell products for its foreign parent in return for a
local US R&D subsidiary will constitute a PE of the foreign parent. The R&D activ-
ity will generally not be seen as preparatory or auxiliary to the business operations of
the foreign parent. The subsidiary will also likely not be seen as an agent, as a typical
R&D entity will not habitually enter into contracts in the name of the foreign parent;
see Zollo (2011), at p. 774.
1717. OECD MTC, art. 5(6). For an interesting and thorough analysis of the historical
background of the agency PE, see Lüdicke et al. (2014).
1718. Further, the contracts are in the name of the enterprise, or for the transfer of
rights in property owned by the enterprise or that the enterprise has a right to use, or for
the provision of services by that enterprise; see OECD MTC, art. 5(5).
1719. OECD MTC, art. 5(5). This requires that the exceptions for preparatory and
ancillary activities under art. 5(4) are not triggered.
1720. See, in particular, Vann (2003), at p. 167; and Vann (2006) on this issue (under
the pre-2010 art. 7 regime). See also Toro (2009); Gazzo (2003); Kobetsky (2006);
Pinto (2006); Sheppard (2006); Russo et al. (2007); Innamorato (2008); Romano et al.
(2010); Gouthière (2010); Schoueri et al. (2011); Rosalem (2010); and Goede (2012).
1721. The author refers to the analysis of case law in sec. 25.6. See, in particular, the
Norwegian Supreme Court decision in Dell Products v. the State (Utv. 2012, s. 1),
where the tax authorities asserted that the local agent had contributed to the develop-
ment of a marketing intangible (ruling analysed in sec. 25.6.5.).
533
commission fee calculated on the basis of its local sales. The fee is subject
to article 9 transfer pricing.
The second entity is the dependent agent PE of the foreign parent. It must
be remunerated under article 7. This “article 5(5) PE” is a deemed PE, not
caused by a “fixed place of business” of the foreign parent in the source
state,1722 but due to the functions performed by the dependent agent enter-
prise.
The most pronounced “PE risk” will normally be related to this type of
PE, as a multinational will usually have more control over whether the PE
thresholds in article 5(1)-(4) are breached. In particular, source state tax
authorities may be inclined to assert that there is a dependent agent PE
subsequent to a business restructuring, e.g. when the source state entity
is stripped from a full-fledged manufacturer to a contract manufacturer
or from a full-fledged distributor to a buy-sell distributor or commission
agent.1723 Such PE assertions will normally aim to recoup the profits associ-
ated with the risks that were contractually stripped from the source state in
the business restructuring.
Dependent agent PE structures are controversial. The Action 7 of the BEPS

Action Plan states:
[I]n many countries, the interpretation of the treaty rules on agency-PE allows
contracts for the sale of goods belonging to a foreign enterprise to be negoti-
ated and concluded in a country by the sales force of a local subsidiary of that
foreign enterprise without the profits from these sales being taxable to the
same extent as they would be if the sales were made by a distributor. In many
cases, this has led enterprises to replace arrangements under which the local
subsidiary traditionally acted as a distributor by “commissionaire arrange-
ments” with a resulting shift of profits out of the country where the sales take
place without a substantive change in the functions performed in that country.
Similarly, MNEs may artificially fragment their operations among multiple
group entities to qualify for the exceptions to PE status for preparatory and
ancillary activities.1724
1722. See OECD MTC, art. 5(1).

1723. PE assertions from a source state following a “risk strip-down” conversion may
be based on the formerly resident group entity having a place of business in the source
state pursuant to OECD MTC, art. 5(1) if it maintains a certain amount of space “at its
disposal”; see the OECD Commentary on Article 5, para. 4. Alternative arguments are
that the formerly resident group entity has a place of management in the source state
pursuant to OECD MTC, art. 5(2) a), or that it constitutes an agent PE in the source
state under OECD MTC, art. 5(5).
1724. 2015 Action 7 Report.
534
Action 7 of the BEPS Action Plan was to “prevent the artificial avoidance
of PE status”.1725 As a result, article 5(5) and (6) were amended in the fall of
2015 as part of the BEPS package. The OECD found, as a matter of policy,
that when the activities that an intermediary exercises in a country are in-
tended to result in the regular conclusion of contracts to be performed by
a foreign enterprise, that enterprise should be considered to have a taxable
presence in that country unless the intermediary is performing these activi-
ties in the course of an independent business.1726 Article 5(5) and (6), as well
as their commentaries, were altered to implement this policy so that com-
missionaire (and similar) arrangements now trigger the existence of a PE.
No changes were made to the profit allocation rules in article 7 as a result

of the alterations to article 5.1727 The OECD’s view is that the revision of
article 5 does not require substantial modifications to the existing rules
on the allocation of operating profits to PEs, but that there nevertheless is
a need for some clarifying guidance on how the current profit allocation
rules should be applied in light of the revised article 5. The OECD issued
a 2017 discussion draft that seeks to address this.1728 The discussion in sec-
tions 17.6.3.-17.6.4. should be relevant for the ongoing OECD revision.
17.6.3. The OECD approach for allocating operating profits

to a dependent agent PE
Some transfer pricing schemes involving dependent agent PEs have result-
ed in insufficient allocation of income to source states when judged in light
of the economic activity carried out there. The allocation approach in the
2010 Report is a reaction to this.
Multinationals have traditionally embraced the “single-taxpayer approach”

with respect to the allocation of operating profits to a source state.1729 Un-
der this approach, an arm’s length article 9 fee to the dependent agent en-
terprise “fully extinguishes the profits attributable to the dependent agent
1725. Id.
1726. Subsequent to the 2015 alterations of art. 5(6), a person who acts exclusively (or
almost exclusively) on behalf of one or more enterprises to which it is closely related is
not considered an independent agent.
1727. Action 7 Report, para. 19.
1728. OECD, BEPS Action 7: Additional Guidance on Attribution of Profits to Per-
manent Establishments (OECD 2017) [hereinafter 2017DD]. This draft replaces (and
takes a new approach to the one adopted in) the July 2016 discussion draft on the same
matter.
1729. On this point, see Oosterhoff (2008), at p. 68.
535
PE”.1730 The logic is that the article 9 fee appropriately rewards the depend-
ent agent enterprise for all functions performed by it, as well as for assets
and risks that are contractually allocated to it. Thus, there should be noth-
ing left for which to remunerate the PE.
The 2010 Report asserts that there are three fundamental flaws in the sin-
gle-taxpayer approach. First, it does not result in a “fair division of taxing
rights” between residence and source states.1731 Assets and risks connected
to the source state activity are ignored solely on the basis of contractual
allocation to the non-resident enterprise. Second, for PEs other than the
dependent agent PE, assets that are economically owned and risks created
as a result of the PE’s functions will be assigned to it, regardless of con-
tractual allocation.1732 In other words, the single-taxpayer approach would
imply differing applications of the authorized OECD approach, depending
on which type of PE is involved.1733 The 2010 Report argues that this is in-
compatible with article 7 and the arm’s length principle. Third, the single-
taxpayer approach results in the curious outcome that there are no profit
consequences of identifying a dependent agent PE under article 5(5).1734
The 2010 Report finds this problematic in light of the principle of statutory
interpretation that provisions should generally not be interpreted in a man-
ner that renders them superfluous.1735 Thus, the single-taxpayer approach
remunerates the subsidiary, but not the PE.
The OECD position taken in the 2010 Report on the article 7 remuneration
of a dependent agent PE is motivated by the desire to “amend” the lack of
compensation for assets and risks under the article 9 remuneration. The
fundamental rule in the 2010 Report is that a dependent agent PE should
be remunerated pursuant to the same principles that are applied under arti-
cle 7 to attribute profits to other types of PEs.1736 In particular, the depend-
ent agent PE will be attributed profits that stem from the assets it economi-
1730. 2010 Report, para. 235.

1731. 2010 Report, para. 236.
1732. In these cases, there will not be a subsidiary that performs functions in the
source state for the non-resident enterprise, and therefore no transfer pricing under
OECD MTC, art. 9.
1733. 2010 Report, para. 237.
1734. The 2010 Report observes that source-state taxing rights will vary significantly
depending on whether or not there is a PE; see 2010 Report, para. 227. Otherwise, simi-
lar cases may result in widely differing profit allocations, depending on whether the
source state activities breach the PE threshold. The 2010 Report, expressing discontent
over the PE thresholds, finds that its profit allocation approach may mitigate some of the
consequences of this “cliff effect”.
1735. 2010 Report, para. 239.
1736. 2010 Report, para. 228.
536
cally owns, the risks it incurs and the capital it is assigned pursuant to the
“significant people functions” doctrine.1737 Alternatively, if the dependent
agent enterprise performs the significant people functions relevant to the
assignment of economic ownership of assets and the assumption and man-
agement of risk, the allocation of assets, risks and profits to the PE will be
reduced or eliminated.1738 The 2010 Report emphasizes:
[The] activities of a mere sales agent may well be unlikely to represent the sig-
nificant people functions leading to the development of a marketing or trade
intangible so that the dependent agent PE would generally not be attributed
profit as the “economic owner” of that intangible.1739
While the author agrees that this normally will be the case, there can be
no legal presumption for this outcome. A concrete assessment based on a
thorough functional analysis must be made in each case.
17.6.4. Is the dependent agent PE relevant?
The point of the 2010 Report’s approach is to ensure that the combined
allocation of profits under articles 7 and 9 provides the source state with in-
come that is commensurate with the functions carried out, assets used and
risks incurred by the multinational there.1740 What typically is at stake here
is the size of the normal market return allocable to the source state, nor-
mally not residual profits. Of course, it cannot be ruled out that there are
unique value chain inputs (IP) provided by the subsidiary or the PE in the
source state, in which case the source state should be allocated a portion of
the residual profits. The following discussion, however, is geared towards
value chain distribution structures that are designed to contractually strip
risks from the source state.1741 These are generally aimed at reducing the
level of normal return allocable to the source state. The more assets and
risks allocated to the source state, the larger the normal return must be.
The “default scenario” is that the local subsidiary purchases goods from a for-
eign group entity, typically resident in a low-tax environment. The subsidiary
1737. Id.
1738. 2010 Report, para. 233.
1739. Id.
1740. On this topic, see also Cottani (2016), in particular, at p. 180.
1741. On the risk-stripping of local marketing and distribution entities, see, in particu-
lar, Musselli et al. (2008a). See also Schön (2014); Vann (2003), at p. 153; and Vann
(2006). On contractual risk allocations in transfer pricing in general, see, in particular,
Schön (2014); and supra n. 752.
537
takes title to, stores and sells the goods to local customers in its own name. It
incurs a “normal” level of distribution risks, including that it may not be able
to sell the number of goods at the prices it targets (inventory risk) and that its
customers will not settle their credit purchases fully and on time (credit risk),
and may therefore incur a loss.1742 A “normal” distribution subsidiary should
earn a net operating margin similar to that earned by comparable unrelated
distributors (typically benchmarked under the TNMM). In the long run, this
must provide the subsidiary with a sensible return on its capital.
A buy-sell distributor scheme contractually strips the local subsidiary of

inventory risk. The subsidiary will, in these cases, buy products from a for-
eign group entity. The subsidiary owns the goods only momentarily, which,
in tax-planning vernacular, is known as “flash title”. The foreign group
entity warehouses the goods in the source state. This would normally fall
under article 5(4)(a), (b) and (f) of the OECD MTC, thus not triggering the
existence of a PE.1743 The sales will be carried out in the name of the sub-
sidiary, and it therefore incurs credit risk, but no inventory risk, due to the
“flash title”. A buy-sell distribution subsidiary should earn a net operating
margin similar to that earned by comparable unrelated distributors. This
will be less than the operating margin earned by a “normal” distribution
subsidiary, but still more than a pure agent.
Two schemes are popular for stripping local distribution subsidiaries of

both inventory and credit risk. First, there is the “agent scheme”, in which
the goods are contractually owned by the foreign entity and sales are car-
ried out by the local subsidiary in the name of the foreign entity. Second,
there is the “commissionaire scheme”, a frequently used variant of the
agent scheme, in which sales are formally not carried out in the name the
foreign entity, even though the local subsidiary does not take title to the
goods.1744 Both schemes allocate a minimum level of operating profits to
the source state. The agent will normally be allocated a steady, low-risk
return from year to year. Due to the comprehensive contractual stripping
of risks, an agent should not incur a taxable loss.
The attractiveness of the commissionaire structure relative to the plain

agent structure can be found on two levels. First, the former arrangement
1742. On the stripping of inventory and credit risk in particular, see Musselli et al.
(2008a).
1743. The 2015 BEPS revision of OECD MTC, art. 5(4) altered this; see 2015 Action
7 Report, paras. 10-13.
1744. On this, see, in particular, Vann (2003), at p. 153. See also further discussions on
the commissionaire structure in Verdoner (2011); and Angus (2012).
538
could be beneficial for business purposes, as local customers have a con-

tractual relationship only with the commissionaire, and from their point
of view, the commissionaire appears as an ordinary buy-sell distributor.
Second, the commissionaire scheme is resistant against source-state PE as-
sertions pursuant to article 5(5).1745 Due to the rather categorical language
applied in the pre-2015 article 5(5) and its commentaries,1746 it was prob-
able that local courts would be forced to respect commissionaire arrange-
ments based on the formal observation that the local sales agreements were
entered into in the name of the local subsidiary, not the foreign principal,
even if there was little or no substance to the arrangements. This was the
case in Dell (Norway), Zimmer (France) and Boston Scientific (Italy).1747
This second level is now eliminated through the 2015 BEPS revision of
article 5(5) and (6), which ensures that commissionaire (and similar) ar-
rangements trigger the existence of a PE.
The usefulness for multinationals of the buy-sell distributor and agent

schemes as legal vehicles for reducing the allocation of operating profits
to source states relies fully on the contractual risk-stripping that is given
effect for transfer pricing purposes under article 9. The 2010 Report in-
terprets the 2010 OECD TPG so that the contractual allocation of risks in
agent schemes should be respected. It states:
[U]nder the “single taxpayer” approach, those risks can never be attributed to
the dependent agent PE of the non-resident enterprise, whilst the authorised
OECD approach would attribute those risks to the dependent agent PE for tax
purposes if, and only if, the dependent agent performed the significant people
functions relevant to the assumption and/or subsequent management of those
risks. (Emphasis added).1748
1745. See, e.g. Musselli et al. (2008a), at p. 265.

1746. The wording of OECD MTC, art. 5(5) is that the person in the source state
must have “authority to conclude contracts in the name of the enterprise”. The OECD
Commentary on Article 5(5), para. 32 states that “only persons having the authority to
conclude contracts can lead to a permanent establishment”. Further, para. 32.1 requires
that the “agent … concludes contracts which are binding on the enterprise”.
1747. Under art. 5(5) b) of the UN MTC, however, a person who has no authority to
conclude contracts on behalf of the principal but has a stock of goods that he deliv-
ers on behalf of the non-resident enterprise could constitute a PE. Commissionaire
structures are reliant on the recognition of indirect representation by local private law.
Under common law, a properly authorized agent will bind the principal. Under civil
law, however, there is a distinction between direct representation, pursuant to which the
agent will bind the principal, and indirect representation, in which the principal is not
bound. This is why case law on PE assertions pertaining to commissionaire structures
have been limited to continental Europe.
1748. 2010 Report, para. 235.
539
Oddly enough, the 2010 Report does not question this treatment of risk.
The explanation likely partly lies in the fact that the aim of the Report is
to address allocation issues under article 7. However, the lack of critical
reflection on the article 9 transfer pricing is still peculiar, as the need for
article 7 in the context of a dependent agent PE is only triggered if the local
entity is not compensated for assets and risks under article 9. The question
is whether, in light of the 2017 OECD TPG on comparability and risk,1749
contractual stripping of a distribution subsidiary’s risks should be given
effect for pricing purposes under article 9.
A key point of the 2017 guidance is that contractual risk allocations should
not be respected if the risks are assigned to an entity that is not in control of
them.1750 Control is understood as the capability to make decisions to take
on the risk and whether and how to respond to the risk.1751 If there are dif-
ferences between the terms of the controlled agreement pertaining to the
assignment of risk and the conduct of the parties, the latter will generally be
taken as the best indication of the actual allocation of risk.1752 For instance,
if a foreign associated enterprise assumes all inventory obsolescence risk
by written contract, it must be ascertained where inventory write-downs
and decisions regarding production volumes and inventory levels are taken.
If the local distributor provides market feedback that is used for determin-
ing product specifications, the local entity could be deemed to share in the
product risk management.
Normally in agent schemes, the functions relevant to the assumption of

inventory and credit risks will be performed in the source state, simply be-
cause distribution, marketing and sales are carried out there. The activities
of the local entity will normally be influential on the amount of products
sold in the source state, which in turn drives inventory levels and customer
receivables. In Zimmer, for instance, the local subsidiary had the author-
ity to accept orders from clients, present offers and quotes, participate in
bid processes and negotiate discounts and payment terms. Because inven-
tory and credit risk are directly dependent on local sales functions, it must
presumably often be concluded that the local entity is in control of inven-
tory and credit risk, not the foreign principal that is contractually allocated
1749. See the discussion of the 2015 OECD guidance on controlled risk allocations in
sec. 6.6.5.5.2.
1750. OECD TPG, para. 1.98. In addition, it is a requirement for accepting controlled
risk allocations that the entity assigned the risk has the financial capacity to cover loss-
es, should the risks materialize; see OECD TPG, paras. 1.64 and 1.98.
540
these risks.1753 If so, the local subsidiary shall be allocated both inventory
and credit risks for the purpose of the transfer pricing of the commission
fee agreement with the foreign principal pursuant to article 9.1754
The article 9 remuneration of the local subsidiary must take into account
all functions performed, assets used and risks incurred. In particular, it
must be determined whether the local entity is in possession of unique
IP (e.g. marketing know-how and local goodwill).1755 If so, it shall be al-
located the residual profits from the IP. The TNMM (to allocate a normal
return to the foreign entity) or the PSM should then likely be applied, de-
pending on whether the IP owned by the local subsidiary is the only unique
contribution to the value chain. If the local entity is not in possession of
any unique IP, it shall only be allocated a normal market return. If the
TNMM is applied for this purpose, the third-party net profit data used
must be extracted only from full-fledged distributors that incur both inven-
tory and credit risks. Agents or third-party enterprises that act as buy-sell
distributors should be rejected as comparables because they will not reflect
the return allocable to an entity that is subject to both inventory and credit
risks. Source state tax authorities should be critical of any comparabil-
ity adjustments made to profit data extracted from buy-sell distributors or
agents that allegedly make the risk profile of the uncontrolled transactions
purportedly comparable to that of the controlled transaction.
In light of the 2017 OECD TPG on risk under article 9, it can no longer be
taken for granted that contractual allocations of inventory and credit risk
to foreign principal entities will be deemed to be in correspondence with
the economic substance of the actual behaviour of the controlled parties.
The relevant question is whether the local subsidiary is in control of the
inventory and credit risks. If so, it shall be allocated the operating profits
connected to these risks as a transfer pricing matter under article 9. The
criteria used to establish control in the new 2017 comparability guidance
on risk are akin to the “significant people functions” doctrine outlined in
the 2010 Report. The author finds it unlikely that these doctrines would
allocate inventory and credit risks differently.
1753. A different issue entirely is that no significant inventory risks will necessarily
exist in the value chain. The group may, for instance, have a “just-in-time” logistics
system in place, designed to minimize inventory risks and costs.
1755. This was the position of the Norwegian tax authorities in Dell Products v. the
State, Utv. 2012, s. 1 (ruling analysed in sec. 25.6.5.). The allocation issue was not as-
sessed by the Norwegian Supreme Court, as it held that the commissionaire scheme
used by Dell in Norway did not trigger the existence of a PE under art. 5(5).
541
If the local subsidiary is assigned a normal market return for its contribu-
tions of functions, assets and risks under article 9, this will be full com-
pensation to the source state, making the “dependent agent PE” assertion
superfluous. The end result for the source state of applying only article 9
pursuant to the 2017 OECD TPG on risks should be the same as the com-
bined application of (i) article 9 to remunerate the commission fee pursuant
to the 2010 OECD TPG on risk; and (ii) the authorized OECD approach
under article 7 to remunerate inventory and credit risk.
In light of this, the dependent agent PE under article 5(5) should no long-

er be relevant for the purpose of allocating operating profits to a source
state. Thus, there should be no profit allocation consequences of identi-
fying a dependent agent PE. This is a curious result. One would assume
that there would be consequences for identifying a PE under article 5(5).
Nonetheless, viewed in the context of the relationship between article 7
and article 9, as described in the 2010 Report, the primary purpose of the
authorized OECD approach was to “fix” the problem raised by the article 9
transfer pricing under the old (1995/2010) OECD TPG on comparability
and risk, pursuant to which the dependent agent enterprise was often not
allocated profits connected to inventory and credit risks if these risks were
contractually assigned to a foreign entity. As the 2017 OECD TPG on com-
parability and risk refuses to accept such formal risk allocations with effect
for transfer pricing, there is no longer any problem to fix.
The 2017 OECD discussion draft on the allocation of profits to PEs reaches
a similar conclusion, as it states the following:
[W]here a risk is found to be assumed by the intermediary under the guidance
in Section D.1.2 of Chapter I, such risk cannot be considered to be assumed by
the non-resident enterprise or the PE for the purposes of Article 7. Otherwise,
double taxation could occur in the source country through taxation of the
profits related to the assumption of that risk twice, i.e. in the hands of both the
PE and the intermediary.1756
Further, the discussion draft states:

In particular, when the accurate delineation of the transaction under the guid-
ance of Chapter I of the TPG indicates that the intermediary is assuming the
risks of the transactions of the non-resident enterprise, the profits attributable
to the PE could be minimal or even zero.1757
1756. 2017DD, at para. 18.

1757. 2017DD, at para. 19.
542
The UN approach for allocating profits to a PE
Thus, the core point of the OECD, as expressed in the 2017 discussion
draft, is that the source state shall be allowed to tax an arm’s length amount
of total income that reflects the return attributable to the functions per-
formed, assets used and risks incurred in the source state, regardless of
whether it is attributable to the dependent agent enterprise or the dependent
agent PE. The total income shall only be taxed once, and the draft allows it
to be taxed at the level of the dependent agent enterprise.1758
It can, in light of the above, be observed that the PE threshold under arti-
cle 5(5) becomes irrelevant with respect to whether the source state shall
be allocated the “additional profit” connected to inventory and credit risk,
as the income will be attributable to the local subsidiary (dependent agent
enterprise).
In the author’s view, this result is appropriate. The relaxed position of the
pre-2017 OECD TPG on controlled contractual risk allocations was highly
unsatisfactory. Third parties would not assume risks that they had no con-
trol over.1759 Further, giving effect to controlled allocations that dislocate
risks from the group entities responsible for the operational risk-generating
business activities could, in the long run, be damaging to the international
consensus on the separate entity approach and arm’s length pricing. Source
states could lose faith in the ability of the prevailing system to distribute
operating profits in an equitable manner that is aligned with economic real-
ity and value creation.
17.7. The UN approach for allocating profits to a PE
In sections 17.7.2.-17.7.3., the author will tie some comments to the allocation
of operating profits between a residence and source state pursuant to arti-
cle 7 of the 2011 UN MTC. The material content of the UN rule may diverge
significantly from article 7 of the OECD MTC, as the former is based on the
pre-2010 article 7 of the OECD MTC. The limited aim of this discussion is
merely to briefly indicate the main differences, as this may provide an inter-
1758. See 2017DD, paras. 26, 31 and 35.

1759. OECD, BEPS Actions 8, 9 and 10: Discussion draft on revisions to Chapter I of
the Transfer Pricing Guidelines (including risk, recharacterisation and special meas-
ures) (OECD 2014), at sec. D2, para. 38.
543
esting perspective on how intangible profits are allocated to the developing

countries that tend to use the UN MTC as a basis for their treaties.
17.7.2. The allocation norm under article 7 of the 2011

UN MTC
The UN rule for allocating operating profits earned by an enterprise in a

source state between that state and the jurisdiction where the enterprise is
resident is expressed in article 7(2) of the UN MTC. It states that the profits
attributable to a PE are “the profits which it might be expected to make if it
were a distinct and separate enterprise engaged in the same or similar activ-
ities under the same or similar conditions and dealing wholly independently
with the enterprise of which it is a permanent establishment”. The provision
is a reiteration of the pre-2010 version of article 7 of the OECD MTC.
The main focus of article 7 of the UN MTC is to limit the expenses that
are deductible in the calculation of the PE’s profits. The first sentence of
article 7(3), which is identical to the pre-2010 version of article 7(3) of the
OECD MTC, limits deductible items to “expenses which are incurred” for
the purposes of the business of the PE. Further, the provision specifically
excludes a range of income and expense items.1760 Article 7(3) denies de-
ductions for payments from the PE to the head office for “royalties, fees
or other similar payments in return for the use of patents or other rights”,
“commission” and “interests”.1761 Non-recognition of these items applies
symmetrically to the income side of the PE profit calculation. In other
words, the PE may not charge the head office for “royalties, fees, or other
similar payments”, “commission” or “interests”.1762
The reason why deemed royalty payments to and from a PE are not rec-
ognized is,1763 according to the sparse commentaries on article 7(3), that it
1760. These limitations were formulated by the former Group of Experts in 1980.
Some developing countries were of the opinion that UN MTC, art. 7(3) should clearly
reflect the treatment of relevant items in order to assist the application of the provision
and that a highlighting the exclusion of certain deductions would be instructive for
taxpayers; see para. 16 of the UN Commentary on Article 7.
1761. Only the first two items refer to the operating profits of the PE. Interests are
financial items that will not affect the operating profits as such (but will, of course,
influence the net taxable income of the PE).
1762. UN MTC, art. 7(3).
1763. The pre-2010 version of art. 7 of the OECD MTC also does not recognize
deemed royalty payments to or from a PE. While this is not explicitly indicated in the
wording of the provision, its commentary (see para. 34) is identical to that of art. 7(3)
544
would be difficult to assign IP ownership and IP development costs among

different parts of an enterprise.1764 Also, if deductions for deemed royalty
payments were to be allowed for the purposes of determining the profits of
the PE, this would trigger the issue of whether the source state should be
allowed to levy withholding tax on the deemed payments in order to ensure
equal treatment of PEs and subsidiaries and to avoid base erosion.1765 The
UN Committee of Experts rejected the article 7 authorized OECD profit
allocation approach in 2009, partially due to this problem.
It is not entirely clear to the author what the consequences of this are. For
instance, assume that there is a PE that distributes a pharmaceutical prod-
uct in the source state. The product is based on unique patents and trade-
marks owned by the enterprise. The PE “purchases” the products from the
head office and incurs some operating expenses to distribute, market and
sell the products locally. All sales are made to unrelated customers. The PE
realizes a significant residual profit in the source state. The question is how
this profit should be allocated between the residence state of the enterprise
and the source state of the PE under article 7 of the UN MTC.1766
This issue would have been relatively straightforward to resolve had trans-
fer pricing principles akin to the authorized OECD approach been appli-
cable. If the economic ownership of the unique IP had been assigned to
the head office, the TNMM would likely have been applied to allocate a
normal return on the routine distribution activities performed by the PE.
The operating profits of the PE after deduction for the normal return would
be allocable to the residence state as a deemed royalty payment for the
unique IP used by the PE.
This solution is likely unavailable because it results in the deduction of a

deemed royalty payment in the calculation of the profits of the PE in the
of the UN MTC. This is important because the 2008 revision of the OECD MTC art. 7
commentary (the amendment text is included as an appendix in the 2008 Report) set
out to include the profit allocation aspects of the 2008 Report that did “not conflict with
previous Commentary”; see the 2008 Report, preface, para. 8. Thus, the fact that the
commentaries on the point of deemed royalties were not altered in 2008 is a clear indi-
cation that the OECD approach for allocating operating profits from intangibles cannot
be applied under the old art. 7.
1764. See the Commentary on Article 7, of the UN MTC, para. 34. The focus seems to
be that the actual development costs should be allocated among all parts of the enter-
prise without any mark-up for profit or royalty.
1765. Compare para. 29 of the OECD Commentary on Article 7 of the OECD MTC
(2010).
1766. Or the pre-2010 OECD MTC, for that matter.
545
source state.1767 This conclusion, however, is far from certain. The word-
ing of article 7(3), or its commentaries for that matter, does not contribute
much to the interpretation. It is mentioned that administrative costs and
a share of IP development costs may be deducted, but such elements are
normally allocated to and included in the cost of goods sold (COGS) or
operating expenses.
The irony, as the author sees it, is that it will be easy to achieve the same
TNMM result simply by reclassifying the “deemed royalty” payment. For
instance, if an application of the TNMM indicates that the PE should make
a net operating margin of 5%, this can be achieved by adjusting the price
for the goods sold by the head office to an amount that would yield the
desired profit margin for the source state. This technique was, for instance,
used by the taxpayer in the Norwegian Vingcard Elsafe AS v. Skatt Øst
case in the context of article 9.1768 The economic result of an upward ad-
justment of the COGS to the PE would be the same as charging a deemed
royalty payment to the PE, i.e. the residual profits are extracted from the
source state. Alternatively, the COGS of the PE could be determined using
the resale price method. Given that it will normally be possible to reclas-
sify a deemed royalty to increased COGS or possibly also to operating
expenses allocable to the PE, it is not clear what the actual consequences of
the cut-off rule for deemed royalty payments in article 7(3) are.
In the author’s view, there is no question as to what the result should be

in the pharmaceutical distribution PE example used in this section. The
PE contributes only routine inputs to the value chain. The residual profits
should therefore not be taxable in the source state. It should not matter
whether the residual profits are classified as a deemed royalty payment
or as increased COGS. However, this obviously does matter under arti-
cle 7(3). The decisive question therefore seems to be whether article 7(3)
accepts the extraction of residual profits from the source state as long as
these payments to the head office are not classified as royalty payments.
1767. Neither the wording of UN MTC art. 7(3) or the UN Commentaries contribute

much to the interpretation. It is mentioned that administrative costs and a share of in-
tangible development costs may be deducted, but such elements are normally allocated
to and included in the cost of goods sold or operating expenses (see the analysis of the
concept of operating profits in sec. 6.2.).
1768. Ruling by Borgarting Appellant Court Ruling, dated 18 June 2012 (Utv. 2012
p. 1191), reversing in part the ruling by Oslo City Court, dated 29 Nov. 2010 (Utv. 2010
p. 1690). See the analysis of the case in sec. 15.6. in the context of taxpayer-initiated
year-end adjustments.
546
On the one hand, it will be easy to circumvent the cut-off rule if this is al-
lowed. On the other hand, there cannot be any question that the price for
which the products are sold from the head office to the PE would, at arm’s
length, reflect the value of the unique intangibles embedded in the products
if there was no separate licence agreement between the parties. Thus, the
extraction of residual profits from the source state through an increase in
the COGS yields an arm’s length result in the author’s example, which is
the overarching requirement indicated by the language in article 7(2).
On this basis, the author finds it unclear as to which extent article 7(2) of

the UN MTC allows the extraction of residual profits from source states
due to the cut-off rule for deemed royalty payments in article 7(3). There
are likely differing opinions on this matter among treaty partners. Never-
theless, the decisive factor should, in the author’s view, be that the language
in article 7(2) requires an arm’s length allocation of profits between the
residence and source states. This will simply not be possible to achieve
without making sure that residual profits are allocated to the part of the
enterprise that is deemed to be the economic owner of the unique IP used
in the value chain. If this were not adhered to, article 7(3) of the UN MTC
would represent a serious impediment to the allocation of intangible busi-
ness profits between residence and source states in a manner that is aligned
with value creation, third-party behaviour and fundamental economic rea-
soning.
17.7.3. The relationship between articles 7 and 9 of

the UN MTC
A curiosity is that article 9(1) and (2) of the UN MTC are reproductions of

articles 9(1) and (2) of the OECD MTC. The OECD TPG are used as the
basis for applying article 9 of the UN MTC.1769 This means that the interna-
tional allocation of business profits under UN MTC-based treaties may po-
tentially follow two bizarrely disconnected patterns, depending on whether
the multinational has organized its presence in the source country through
a PE or a subsidiary.1770 If organized through a PE, article 7 governs the
profit allocation. The allocation rules are ambiguous and do not recognize
deemed royalty payments to and from the PE. If, however, the source coun-
try’s presence is organized through a subsidiary, article 9 will govern the
1769. UN Commentary on Article 9, para. 3.

1770. This is also the situation for tax treaties based on the pre-2010 version of the
OECD MTC.
547
allocation of business profits. The calculation of the source state’s profits

should then follow the 2017 OECD TPG, including the important functions
doctrine for assigning economic ownership of IP, the guidance on risk and
comparability, revised guidance on the pricing methods, cost savings, lo-
cation-specific advantages, etc.
This lack of neutrality opens the door for tax planning and represents a
clear potential for BEPS. The author is not convinced that the asserted
practical advantages of the cut-off rule in article 7(3) for tax administra-
tions in developing countries can outweigh the obvious shortcomings of
this rule. Further, it may be that developing countries are missing out on
income due to this provision. For instance, if a PE has developed profitable
unique IP, the source state may not be entitled to tax the residual profits
under article 7 of the UN MTC.
548
Chapter 18
Introduction to Part 3
The topic for discussion in part 2 of the book was how operating profits
from intangible value chains shall be allocated among the different value
chain inputs under the transfer pricing methodology in US law and the
OECD Model Tax Convention (OECD MTC). The author will again refer
to the example used in the introduction to part 2, in which a Norwegian
subsidiary of a US-based multinational performs routine distribution func-
tions and licenses a patent from an Irish group entity in connection with its
local sales of a blockbuster drug. The role of the transfer pricing methods
here is to allocate the profits from the Norwegian sales among the value
chain inputs, i.e. the routine distribution functions and the foreign-owned
patent. Assume that the Norwegian profits are 100 and that the transac-
tional net margin method (TNMM) indicates that an arm’s length return
to the subsidiary is 20. The residual profits of 80 are thus allocable to the
foreign patent.
The next step is to allocate the residual profits. This is governed by the
intangible ownership provisions in the US regulations and the 2017 OECD
Transfer Pricing Guidelines (OECD TPG), which provide a link between
the determined residual profit amount and the group entities to which the
amount is to be allocated. The ownership rules are relevant for intangi-
bles developed within a multinational1771 when more than one group entity
was involved in the development.1772 The purpose of the rules is to ensure
1771. Conversely, when intangible property (IP) is externally acquired, the group will
control to which of its entities ownership is assigned. The acquisition will be carried
out by third parties. The price paid will thus be at arm’s length. To choose which group
entity – and thereby, which jurisdiction – the acquired intangible shall be assigned to is
the prerogative of the group, much analogous to the freedom a multinational has with
respect to where it wants to allocate its external debt financing. A nuance here is the
allocation of assets among the residence and source jurisdictions when there is a per-
manent establishment under art. 7 of the OECD Model Tax Convention (OECD MTC).
The allocation rules may entail that the assets and debt of the enterprise are allocated to
the source jurisdiction, even if they are legally owned by the enterprise in the residence
state; see the discussion in ch. 25.
1772. If all functions, assets and risks connected to the research and development
(R&D) (including enhancement and maintenance) of an item of IP are contributed by
one group entity alone and the development is carried out within its residence jurisdic-
tion, all income from the developed IP will be allocable there; see OECD Transfer Pri
cing Guidelines (OECD TPG), para. 6.71. The issues analysed in this part of the book
are not relevant in this (pronouncedly theoretical) scenario.
549
that intangible profits are allocated to the group entities that took part in
the intangible development in proportion to their value contributions, and
thereby to the jurisdictions where the intangible value was created. These
provisions will be analysed in this part of the book.1773
The analysis of the ownership rules should also be seen in the context of
the “principal model”.1774 In order for the model to work (i.e. yield an opti-
mally low effective tax rate for the multinational), ownership of internally
developed unique intangibles must not be assigned to the group entity resi-
dent in the (typically high-tax) jurisdiction in which it was created through
research and development (R&D), but to a group entity in a foreign (typi-
cally low-tax) jurisdiction. It is no secret that multinationals historically
have often been successful in doing so. The 2017 OECD TPG on intangible
property (IP) ownership are designed to eliminate these BEPS practices
and ensure that residual profits are allocated to the jurisdictions where the
intangible value was created (i.e. mainly to the R&D jurisdictions).1775
The ownership rules are, in principle, equally applicable to manufacturing

and marketing intangibles. The allocation assessments will, however, vary
among the two groups, because the respective development processes dif-
fer markedly. Manufacturing IP is created through a combination of R&D,
funding, pre-existing intangibles and the use of tangible assets, normally
emanating from only a few jurisdictions.1776
1773. For discussions of IP ownership under US law and the OECD MTC, see, e.g. Wit-
tendorff (2010a), at pp. 625-629; and Markham (2005), at pp. 48-53. On the US rules,
see, in particular, Andrus (2007), at pp. 634-637; and Levey et al. (2010), at pp. 117-
120. For less recent but relevant contributions, see Boykin (1996); Mentz (1999); Mentz
et al. (1997); Ossi (1999); and Przysuski et al. (2004a).
1774. See the discussion of the principal model in sec. 2.4.
1775. Assume that a functional analysis reveals that the Irish entity in the example
holds legal title to the patent and financed its development. All R&D necessary to cre-
ate the intangible was, however, performed by a US group entity. The new 2017 OECD
ownership provisions require that the group entity that financed the R&D efforts re-
ceive a risk-adjusted rate of return on its investment; see the analysis of intangible R&D
funding remuneration in sec. 22.4. The residual profits shall go to the group entity that
performed the important R&D functions; see the analysis of the “important functions
doctrine” in sec. 22.3. This may entail that the 80 in profits are divided so that, for
instance, 3 is allocated to the Irish entity as compensation for its administrative and
legal services in connection with its holding of legal title and 15 as a risk-adjusted rate
of return on its R&D financing contribution. The remaining residual profits of 62 are
allocated to the US R&D entity.
1776. For example, say that pre-existing IP is contributed by a group entity resident in
jurisdiction 1, ongoing R&D comes from an entity in jurisdiction 2 and R&D funding
comes from an entity in jurisdiction 3. The residual profits from the developed IP shall
be split among jurisdictions 1 and 2, while jurisdiction 3 is allocated a risk-adjusted rate
550
In contrast, marketing IP (e.g. the value of a global trademark) must be

developed in each single market through local marketing activities. Thus,
from a causality perspective, every market jurisdiction is involved in the
development of the local value of a marketing intangible.1777 More juris-
dictions may therefore feel entitled to a stake in the residual profits from
marketing IP than what is the case for manufacturing IP. Further, it will
normally be straightforward to distinguish between the development and
exploitation of R&D-based manufacturing IP. The same cannot often be
said for marketing IP, as promotional activities will both develop and ex-
ploit it. The ownership rules are, in this respect, more problematic for inter-
nally developed marketing IP than they are for manufacturing IP.
While the 2017 BEPS revision of the OECD IP ownership provisions was
geared towards internally developed R&D-based manufacturing intangi-
bles, it is important not to lose sight of the fact that the allocation of profits
of return on the funding contribution. This entails that the residual profits generated
through the worldwide exploitation of the IP are extracted from all market jurisdictions
in which the product based on the IP is sold, and they are allocated to jurisdictions 1,
2 and 3. There is a solid rationale (behind this allocation. As the IP value was created
in jurisdictions 1, 2 and 3 only, it seems reasonable that the profits should be allocated
there.
1777. Logically, this observation should indicate that the residual profits from the mar-
keting IP should at least partially be retained at source so that the value is taxed where
it is created. This reasoning is, however, too rudimentary when it comes to marketing
IP. The development of global marketing IP is, in contrast to R&D-based manufactur-
ing IP, mainly linked to one type of development contribution: marketing expenses.
A multinational is free to choose where to draw its funding from. Local marketing
expenses may therefore be reimbursed by a foreign funding entity, removing some of
the causality connection between the IP development and the source state. Neverthe-
less, while such reimbursements may facilitate foreign ownership of the local value of
marketing IP, and thereby extraction of the residual marketing profits from the source
state, they will come at a cost for the multinational, i.e. it will not be able to deduct the
expenses at source (as the local entity does not bear the expenses due to the reimburse-
ment). This may be problematic, for instance, if significant marketing expenses are
required to establish a new product in a competitive market. The risks involved may be
high, as the marketing may fail to yield the intended results. In these cases, it will likely
be essential to ensure local deductibility. If the marketing expenses deducted at source
are high enough, there is a risk both under the US and OECD rules that the source state
may claim entitlement to the profits. See the analysis of the US and OECD rules on
profit allocation in the context of internally developed marketing IP in ch. 23 and ch.
24, respectively. Further, the assignment of residual profits from internally developed
marketing IP is complicated by the possibility that the local distribution subsidiary
may have developed unique marketing IP that is distinct from the local value of the
foreign-owned trademark (goodwill, know-how, etc.). There may also be local market
characteristics (proximity to market, consumer preferences, etc.) that are separate from
marketing IP, but may yield incremental profits that must be allocated; see the analysis
of market-specific characteristics in ch. 10.
551
from internally developed marketing intangibles is an equally important

subject. The author has sought to reflect this in his analysis as he provides
extensive discussions of the ownership rules in the context of both types
of IP.
The outline of part 3 is as follows:

− the historical development of the US and OECD IP ownership rules is
discussed in chapter 19;
− a lead-in to the determination of IP ownership under the current US
and OECD rules is provided in chapter 20;
− the determination of IP ownership for intra-group-developed manufac-
turing IP under US law and the OECD MTC is analysed in chapter 21
and chapter 22, respectively;
− the determination of IP ownership for intra-group-developed market-
ing IP under US law and the OECD MTC is analysed in chapter 23 and
chapter 24, respectively;
− the determination of IP ownership in the context of permanent estab-
lishments (PEs) (i.e. among the head office and PE of an enterprise
under article 7 of the OECD MTC) is analysed in chapter 25; and
− concluding remarks are provided in chapter 26.
552
Part 3
Chapter 19
The Evolution of the US and OECD

Approaches to Intangible Ownership
19.1. Introduction
This chapter examines the structure and material content of the historical
US and OECD intangible property (IP) ownership rules, i.e. the rules that
determine how residual profits generated through the exploitation of IP de-
veloped intra-group shall be allocated among group entities. This analysis
will provide a valuable background for the interpretation of the current US
and OECD provisions governing IP ownership, which will be examined in
the following chapters.
The significance of IP ownership rules in transfer pricing has increased

over the years, as the profits of global enterprises have gradually become
more IP-driven. In the 1960s, many of the largest enterprises operated
within the car and oil industries, while they now can be found within typi-
cal IP sectors, such as the high-tech, software, pharmaceutical and retail
industries. In the 2015 BEPS revision of the OECD Transfer Pricing Guide-
lines (OECD TPG), IP ownership issues were among the most challenging
to reach consensus on, in particular with respect to intra-group-developed
manufacturing IP.
The United States introduced advanced and substance-based IP ownership

provisions already in the 1968 Internal Revenue Code (IRC) section 482
regulations. The OECD, however, was late to the party. The IP ownership
provisions of the historical OECD TPG were only moderately developed.
The main focus of this chapter will therefore lie on the historical US rules,
which have been a great influence on later thinking around IP ownership
in transfer pricing law.
The author will analyse the IP ownership provisions of the 1968 US regula-
tions in section 19.2., the 1992 proposed and 1993 temporary US regula-
tions in section 19.3. and the 1994 US regulations in section 19.4. Brief
comments will thereafter, in section 19.5., be tied to the approach of the
1995 and 2010 OECD TPG.
555
Chapter 19 - The Evolution of the US and OECD Approaches to Intangible
Ownership
19.2. Determination of IP ownership under

the 1968 US regulations
The 1968 regulations implemented a two-branched approach to the de-
termination of intangible ownership, which was based on whether the
joint development efforts were organized through a cost-sharing arrange-
ment (CSA).1778 If a CSA was entered into, the controlled parties could
share ownership of a jointly developed intangible, given that certain cri-
teria were fulfilled.1779 The operating profits generated by this intangi-
ble would then be split among the CSA participants pursuant to their
respective shares of the intangible development costs. If a CSA was not
entered into, the allocation of operating profits from a jointly developed
intangible was governed by the so-called “developer-assister” rule (DA
rule).1780 Its purpose was to protect the integrity of the CSA institute by
not allowing a split of residual profits in joint development arrangements
that did not satisfy the criteria set out in the 1968 regulations for qualify-
ing CSAs.1781
The DA rule was in effect for almost 3 decades and has influenced inter-
national transfer pricing jurisprudence on intangible ownership signifi-
cantly.1782 It forms an important backdrop for analysing the current own-
ership provisions of both the United States and OECD. The author will
therefore comment on the material content of the rule in section 19.2.2.
and discuss some possible limitations of the rule in sections 19.2.3. and
19.2.4. Two notable transfer pricing cases in which the application of the
DA rule to determine IP ownership was at issue will be discussed in sec-
tion 19.2.5.
1778. The predecessor to the 1968 regulations was Revenue Procedure 63-10, which
only addressed transactions between US-based multinationals and their Puerto Rican
subsidiaries. The rule was that such subsidiaries should be allocated profits from manu-
facturing intangibles that they owned, but not profits from marketing intangibles (trade-
marks, trade names and goodwill) used to distribute products in the United States; see
Revenue Procedure 63-10, sec. 4.01.
1779. 1968 Treas. Regs. (33 Fed. Reg. 5848) § 1.482-2(d)(4).
1780. 1968 Treas. Regs. (33 Fed. Reg. 5848) § 1.482-2(d)(1)(ii)(a).
1781. See Ossi (1999), at p. 993.
1782. For a discussion of the 1968 developer-assister (DA) rule, see, e.g. Wittendorff
(2010a), at p. 632.
556
Determination of IP ownership under the 1968 US regulations
19.2.2. The DA rule
The DA rule distinguished the developer of an intangible from other group

entities that assisted in the development.
For the developer, “no allocation with respect to such development activ-
ity” was to be made.1783 In other words, the developer was not entitled to
compensation for contributing to the development of his own intangible,
but was, as the owner, entitled to the residual profits from subsequent ex-
ploitation of the intangible. Under the DA rule, there could only be one
developer. Only if the developed intangible later would be made available
by the developer to other group entities through the transfer of complete or
limited rights to it would a transfer pricing allocation be justified.1784
Group entities that rendered assistance to the intangible development were

to receive an arm’s length compensation on a concurrent basis throughout
the development phase.1785 However, the DA rule was silent with respect to
the specifics of how this compensation should be determined.1786 Neverthe-
less, the author finds it reasonably clear that the 1968 US regulations did
not afford the assister any residual profits, regardless of the extent to which
he contributed to the intangible value creation.
The determination under the DA rule of which group entity was the devel-
oper relied on a broad assessment of the specific facts and circumstanc-
es.1787 The greatest weight, however, was placed on which group entity bore
the largest amount of the IP development costs (and incurred the connected
risks),1788 as well as the relative values of any intangibles made available
without compensation that were likely to substantially contribute to the de-
velopment. Other relevant factors included the location of development, the
capabilities of the various members to carry on the project independently
and the degree of control over the research and development (R&D) project
exercised by the respective group entities. The DA rule assessment was il-
1783. 1968 Treas. Regs. (33 Fed. Reg. 5848) § 1.482-2(d)(1)(ii)(c) (see (a)).
1784. 1968 Treas. Regs. (33 Fed. Reg. 5848) § 1.482-2(d)(1)(ii)(a).
1785. 1968 Treas. Regs. (33 Fed. Reg. 5848) § 1.482-2(d)(1)(ii)(b).
1786. Id. It only stated that “the amount of any allocation that may be appropriate with
respect to such assistance shall be determined in accordance with the rules of the ap-
propriate paragraph”.
1787. 1968 Treas. Regs. (33 Fed. Reg. 5848) § 1.482-2(d)(1)(ii)(c).
1788. Id. A group entity was not allocated the intangible development costs and risks
if the costs were not carried by the entity contemporaneously with their incurrence and
without regard to the success of the project. Thus, the recharging of, or contingent li-
ability for, development costs was insufficient.
557
Ownership
lustrated in an example in the 1968 regulations that pertained to company

X, which developed a product through a 4-year R&D process, in which
it invested USD 1 million each year, in addition to employing a valuable
intangible it owned.1789 Midway into the process, company Y contributed
four researchers who worked on the project at the facilities of X, at a cost
of USD 100,000, funded by Y. The example deemed X as the developer.
Further, X was deemed to have transferred the developed intangible to Y
when the latter registered the patent for the intangible upon completion.
Thus, the DA rule disregarded legal ownership for the purpose of deter-
mining which group entity should be regarded as the developer. Most of the
factors relevant in the DA assessment are efficient barriers against profit
shifting. Factors such as contributions of intangibles without proper com-
pensation, the location of the development activity, independent capabili-
ties to carry on the development and the degree of control over the R&D
project are all, albeit to varying degrees, challenging to circumvent.
The DA rule assessment bears similarities to the analogous determination

of which group entity is entitled to the residual profits from a jointly de-
veloped intangible under the 2017 OECD TPG.1790 However, in spite of the
comprehensive description of relevant factors, the DA rule was, in prac-
tice, interpreted so that the greatest weight was placed on the entity that
incurred concurrent R&D costs, mitigating the significance of the other
factors.1791 This separates the DA rule from the current OECD guidance,
1789. See 1968 Treas. Regs. (33 Fed. Reg. 5848) § 1.482-2(d)(1)(ii)(d), third example.
Two additional examples were also included. In the first, company X developed a new
machine to be used by company Y. The funding and intangibles required for the devel-
opment were provided by Y, which was deemed to be the developer. The example did
not provide guidance on the pricing of the assistance efforts of X. The second example
was a modification of the first, with the twist that company Y provided contingent de-
velopment funding, in the sense that Y would reimburse X for its development costs if
the machine was successfully developed. The example deemed X the developer. Later
access for Y to the new machine was made possible through a transfer from X, which
warranted a transfer pricing allocation.
1790. See the analysis of the OECD Transfer Pricing Guidelines (OECD TPG) on the
ownership of manufacturing intangible property (IP) in ch. 22.
1791. The factual pattern of Nestlé Co., Inc. v. CIR, T.C. Memo. 1963-14 (Tax Ct.,
1963) is interesting with respect to the DA rule (see the discussion of the ruling in sec.
5.2.3.). The US sales of Nestlé grew rapidly in the 1940s, due to both the quality of the
Nestlé products and the significant marketing efforts by the US entity. In fact, the mar-
keting expenses incurred by the US licensee in the period under review of 1947-1952
were approximately 33% larger than the royalties paid in the same period. Had the case
been tried under the 1968 DA rule, the US entity would likely have been regarded as the
developer of the US rights to the Nestlé trademark. The main reason is that the US en-
tity, over a longer period of time, incurred substantial marketing expenditures, as well as
558
pursuant to which there is a pronounced focus on restricting the amount of

profits allocable to intangible development-funding contributions in order
to limit profit shifting to “cash-box” group entities resident in low-tax ju-
risdictions (see the discussion in section 22.4.).
19.2.3. The DA rule was geared towards, but not limited to,

manufacturing IP
It has been claimed that the DA rule was mainly intended to determine
the entitlement to residual profits generated by manufacturing IP devel-
oped through joint R&D efforts as opposed to where a local distributor
increased the value of marketing IP legally owned by a foreign group en-
tity.1792 The author does agree that the rule was geared towards R&D. For
instance, when the rule spoke of assistance, an example pertaining to the
lending of laboratory equipment was used.1793 In principle, however, the
development of marketing IP was also encompassed by the broad wording
of the provision.
Nevertheless, the focus on R&D entailed a link to the CSA provision of

the 1968 regulations,1794 pursuant to which a US CSA participant would be
allocated residual profits from the subsequent exploitation of the developed
intangible in proportion to its share of the development costs. On the con-
trary, if the US entity was deemed developer under the DA rule, it would
be allocated the entirety of the residual profits from the jointly developed
intangible. The author therefore assumes that the DA rule may have func-
tioned as an incentive to organize joint R&D activities through a CSA so
that part of the residual profits could be shifted abroad for lower taxation.
For example, if a pharmaceutical product was developed through US R&D
while the funding came from a foreign subsidiary, the DA rule would likely
deem the foreign entity the developer.
From the perspective of the multinational, this would be positive in the

sense that the potential residual profits from a successfully developed and
commercialized drug would be allocated in full to the foreign entity and
taxed at a lower rate. The problem with this would be that none of the
the associated risk. The US entity should then have been entitled to the profits generated
through the exploitation of the US rights to the trademark, necessitating a downward US
Internal Revenue Code (IRC) sec. 482 adjustment of the outbound royalties.
1792. See Ossi (1999).
1793. 1968 Treas. Regs. (33 Fed. Reg. 5848) § 1.482-2(d)(1)(ii)(b).
1794. 1968 Treas. Regs. § 1.482-2(d)(4).
559
Ownership
R&D expenses could be deducted in the United States. For large and genu-
inely risky R&D projects, this would likely be problematic from the point
of view of the multinational. Alternatively, had the R&D been organized
through a CSA, the US entity would be able to deduct its share of the R&D
expenditures, with the drawback being that a commensurate portion of the
residual profits would be taxable in the United States. This relationship
between the DA rule and the CSA provision likely meant that it was more
attractive for multinationals to organize relatively high-risk R&D projects
as CSAs and relatively low-risk projects as non-CSA joint development ef-
forts governed by the DA rule.
19.2.4. Was the DA rule limited to the development of

entirely new IP?
It has been asserted that the DA rule was intended for cases in which
an entirely new intangible was developed, as opposed to where the value
of a pre-existing intangible was increased, through joint efforts.1795 This
assertion is based on the view that the Tax Court in Ciba purportedly
distinguished the development of a new intangible from the exploitation
of a pre-existing intangible.1796 In this case, a US distribution subsidiary
made and sold herbicides based on make-sell rights that it licensed from
its Swiss parent. The US Internal Revenue Service (IRS) disallowed de-
ductions for outbound royalty payments by asserting that the subsidiary
should be deemed the owner of the patent pursuant to an imputed CSA.
The Tax Court rejected this assertion and stated that “it is upon the de-
velopment of the triazine compounds, and not the exploitation of the U.S.
market for them, that we must keep our attention focused”. The significant
R&D functions were performed in Switzerland by the parent, which was
also the legal owner of the patent. The US distribution subsidiary only
carried out some auxiliary R&D functions connected to the products it
sold.
In the absence of evidence on the relative intangible development costs

and risks borne by the controlled parties, the Tax Court in Ciba focused
its DA rule assessment on the group’s R&D functions. It found that the
parent planned, performed, supervised and controlled the significant R&D
1795. See Ossi (1999).

mended by AOD-1987-21 (IRS AOD, 1987), and acq. 1987 WL 857882 (IRS ACQ,
1987); see the analysis of the case in sec. 5.2.3.4.
560
functions. Further, only the parent was capable of carrying on this R&D
independently. The routine R&D functions of the subsidiary, which were
characterized as “peripheral at best”, could have been outsourced to third
parties at a low cost. Also, the subsidiary had neither the required person-
nel nor the facilities to carry out any significant R&D. The parent was
deemed the developer. It was not at issue before the Court as to whether the
marketing and sales activities of the subsidiary contributed to or increased
the value of a separate marketing intangible. Thus, there is no merit to the
assertion that the Court distinguished the development of a new intangible
from the exploitation of a pre-existing intangible. In the author’s view, the
DA rule was applicable in all cases in which intangible value was created
by the joint efforts of group entities outside of CSAs.
19.2.5. Case law on the DA rule
In sections 19.2.5.2. and 19.2.5.3., respectively, the author will tie some
comments to two fairly recent cases in which the DA rule was applied to
determine IP ownership, namely GlaxoSmithKline Holdings (Americas)
v. CIR1797 and DHL Corporation v. CIR.1798 The factual patterns of these
cases, as well as the principal transfer pricing problems triggered by them,
are also highly relevant under the current US IP ownership rules (analysed
in chapters 21 and 23), as well as the current OECD IP ownership rules
(analysed in chapters 22 and 24).
19.2.5.2. GlaxoSmithKline Holdings v. CIR (2006)
19.2.5.2.1. Introductory comments
The UK-based multinational GlaxoSmithKline (GSK), one of the largest

pharmaceutical groups in the world, had a long-running dispute with the
IRS pertaining to the allocation of profits to its US distribution subsidiary
1797. GlaxoSmithKline Holdings (Americas), Inc. v. CIR, 117 TC 1 (Tax Ct. Docket
No. 5750-04, 2004). See Musselli et al. (2007a) for comments on the case.
1798. DHL Corp. v. CIR, T.C. Memo. 1998-461 (Tax Ct., 1998), affirmed in part, re-
versed in part by 285 F.3d 1210 (9th Cir., 2002). Several other problems apart from the
ownership issue were assessed in the case, e.g. the determination of common control
and valuation of a trademark.
561
Ownership
for the income years 1989-2005. The case was docketed in the US Tax
Court, scheduled for a February 2007 trial, but was settled in 2006.1799
Even though the case never reached trial and ruling, the problems it raised
are so significant that a discussion is warranted.1800 The main issue was
how much of the business profits generated by the US distribution subsidi-
ary, GSK US, owned by the UK parent, GSK UK, should be taxable in the
United States. One of the problems was whether GSK UK or GSK US
should be deemed owner of the US marketing intangibles and allocated the
residual profits from their exploitation.1801 The author will tie some com-
ments to this in sections 19.2.5.2.3.-19.2.5.2.5. However, in order to do so,
it is necessary to first draw up the background for the case.
19.2.5.2.2. A lead-in to the case
The UK parent was the legal owner of both patents and trademarks for
numerous pharmaceutical products, including Zantac, a global bestsell-
ing ulcer drug. Zantac was the result of lengthy and costly R&D carried
out in the United Kingdom from the mid-1970s to the early 1980s. The
global marketing strategy for Zantac was developed by the parent. The
drug was introduced in the United States in 1983 by the local marketing
and distribution subsidiary, GSK US, which licensed the relevant patents
and trademarks from the parent. It performed marketing and sales func-
tions and secondary manufacturing and provided assistance for approval
by the Food and Drug Administration. It grew rapidly by retaining and
investing its substantial earnings. By 1994, it was the second-largest US
pharmaceutical company, with annual sales of USD 3.7 billion and 6,500
employees.1802
1799. GSK agreed to pay the US Internal Revenue Service (IRS) approximately USD
3.4 billion to resolve the dispute. The payment was the largest settlement ever in a US
tax case. See the 2006 IRS press release, IR-2006-142. As the case did not result in a
ruling, not all details are available. The author based his discussion mainly on informa-
tion contained in the 2 April 2004 petition (which also contained the notice of defi-
ciency as an appendix) filed by GSK with the US Tax Court (available at Tax Analysts,
doc. 2004-7600).
1800. For comments on the case, see, e.g. Fris et al. (2006).
1801. There were two other main issues: (i) whether GSK could carry out year-end
adjustments (which is discussed in sec. 15.10.); and (ii) how the total US residual profits
(from the exploitation of both the UK manufacturing and US marketing intangibles)
should be split among the US subsidiary and the UK parent (which is discussed in sec.
19.2.5.2.5.).
1802. 2 April 2004 petition (available at Tax Analysts, doc. 2004-7600), at p. 11.
562
The transfer pricing dispute began during an IRS audit of GSK US in the
early 1990s, which GSK sought to resolve in 1999 by requesting relief from
the US IRS and Her Majesty’s Revenue and Customs (HMRC) in the United
Kingdom. Supposedly, negotiations between the United States and United
Kingdom collapsed, as the United Kingdom took the position (also held by
GSK) that no additional taxes were due to the United States.1803 In 2004,
the IRS issued a notice of deficiency covering 1989-1996 for approximately
USD 2.7 billion in taxes,1804 followed by additional notice in 2005 covering
the income years 1997-2000 for a further USD 1.9 billion.1805 The reassess-
ment covered the transfer pricing of a range of pharmaceutical products
sold by the US subsidiary, but the significant portion of the adjustments
pertained to sales of Zantac.
The lion’s share of the adjustments was carried out by adjusting the cost
of goods sold (COGS). Of the total increase in income, approximately
58% was due to downward adjustments of COGS on inputs purchased
from the UK parent. The notice of deficiency did not explain its calcula-
tion of the adjusted COGS. However, the author takes it that the COGS
were adjusted so that the parent could recoup its costs plus a profit margin
of 30%.1806 Even though it was not the most significant aspect of the case
from a fiscal point of view (of the total increase in income, approximately
25% was due to downward adjustments of royalties paid to the UK par-
ent), the matter of determining the arm’s length royalty for intangibles
licensed from the parent to the subsidiary has certainly been the most
debated.
19.2.5.2.3. The principal IRS argument that the US subsidiary

was the developer
In addition to charging the US subsidiary for royalties pertaining to its

deployment of manufacturing intangibles, the UK parent also charged for
marketing intangibles it legally owned. The IRS disallowed full deductions
claimed by the US subsidiary for royalty payments with respect to the UK-
1803. See the press release from GlaxoSmithKline dated 7 January 2004 (available at
Tax Analysts, doc. 2004-363).
1804. Available at Tax Analysts, doc. 2004-7600.
1805. See the description of the 2005 notices in GSK’s 12 April 2005 petition, at
pp. 4-6, available at Tax Analysts, doc. 2005-8873. See also the press releases from
GlaxoSmithKline dated 7 January 2004 and 26 January 2005 (available at Tax Ana-
lysts, doc. 2004-363 and doc. 2005-1656, respectively).
1806. Based on information stated on p. 6 of the 2 April 2004 petition (available at Tax
Analysts, doc. 2004-7600).
563
Ownership
owned marketing intangible. In the opinion of the IRS, the subsidiary was
not entitled to deduct royalties paid for trademarks and other marketing
intangibles, as it was “the owner for tax purposes of the trademarks and
marketing intangible purportedly licensed”.1807
The logic behind this assertion was that the US subsidiary was deemed to
be the “developer of said intangibles and because the economic substance
of … dealings … at the time the licensed drugs were first sold in the United
States establishes the existence of an imputed royalty-free license or other
transfer of the … marketing intangibles”.1808
In other words, the IRS held that the subsidiary was entitled to all residual
profits from the US exploitation of the marketing IP.1809
19.2.5.2.4. The secondary IRS argument that the US subsidiary

was the assister
Further, as a secondary argument, the IRS asserted that the subsidiary was
entitled to compensation for assistance rendered to the developer of the
marketing IP, i.e. a normal market return.
In terms of context, it should be mentioned that before Zantac was intro-

duced, the drug Tagamet was the largest-selling ulcer drug in the world.
Tagamet was manufactured and sold by the UK-based SmithKline Bee-
cham, a competitor of GSK. In 2000, GSK merged with SmithKline Bee-
cham. This likely gave GSK access to an advance pricing agreement (APA)
that Smithkline Beecham had entered into with the IRS in 1993 pertaining
to the allocation of income to the US distribution subsidiary that marketed
Tagamet for the income years 1991-1993. In that APA, the IRS accepted an
application of the resale price method in which the US subsidiary was the
tested party and allocated a gross profit margin based on what comparable
distributors earned.1810 Concretely, the APA deemed it sufficient that the
US subsidiary was allocated a 28% (gross) margin from its sales of Tag-
1807. See the 2004 notice of deficiency (available at Tax Analysts, doc. 2004-7600),
attachment: explanation of adjustments, point b).
1808. Id.
1809. Thus, the IRS asserted a profit split, with the residual profits from the manufac-
turing and marketing intangibles being allocable to the United Kingdom and United
States, respectively. On the IRS profit split assertion, see Musselli et al. (2008a), at
p. 270.
1810. For comments on this pricing, see Roberge (2013), at p. 225.
564
amet.1811 That margin was consideration for all promotional and distribu-
tion functions carried out by the subsidiary. Additional margins of 5% and
3% were allocated to the subsidiary as returns on the Tagamet trademark
and the SmithKline Beecham trade name, which the US subsidiary was
deemed to own under a CSA.
GSK applied for an APA for the transfer pricing of US sales of Zantac 1
year after SmithKline Beecham received its favourable APA. In the GSK
application, the US subsidiary was described as “a marketer of pharmaceu-
tical products”.1812 GSK requested approval of the resale price method for
allocating income to the US subsidiary as compensation for its promotion
and distribution functions. The application asserted that all significant in-
tangibles were developed and owned by the UK parent and other foreign
group entities, including the principal marketing intangibles for Zantac.1813
The IRS denied GSK’s application to use the resale price method, or any
other transfer pricing method under which the US subsidiary was treated
as the tested party.1814 Apparently, the IRS was not willing to issue an APA
unless the profit split method (PSM) was used for allocating income to
the subsidiary. GSK argued that not only had the IRS applied the solution
of the 1991-1993 SmithKline APA for an additional year, but the trans-
fer pricing method applied in that APA was also accepted for two other
SmithKline products for years 1999-2000. GSK further asserted that the
IRS had concluded at least one other APA for an inbound pharmaceutical
blockbuster product in which the US distribution subsidiary was treated as
the tested party and remunerated based on comparable data on operating
profits.
It was, however, mentioned that the result of that APA was supported by
a profit split determination, purportedly consistent with the SmithKline
APA. Even though an APA given to another taxpayer clearly was not rel-
evant for determining the allocation of income pursuant to IRC section
482 to GSK US (an APA is the result of individual negotiations between
the IRS and a taxpayer and thus not relevant for profit allocation for other
taxpayers),1815 GSK’s 2005 memorandum did provide interesting insights
into the functions performed by the US subsidiary. The 2005 memorandum
1811. See GSK’s memorandum dated 22 February 2005 in opposition to the IRS mo-
tion for partial summary judgment (available at Tax Analysts, doc. 2005-3635), at p. 13.
1812. Id., at p. 15.
1813. Id., at p. 16.
1814. Id., at p. 18.
1815. See also Verbeek (2005).
565
Ownership
claimed that the marketing and distribution functions of GSK US were

“substantially similar” to those performed by the US distribution subsidi-
ary in the SmithKline APA. Both relied primarily on brief, in-person pres-
entations to doctors, advertising in professional journals, seminars, etc.
The capabilities and qualifications of the sales staff of both companies
were also purportedly similar.
19.2.5.2.5. How would the case have been assessed under the DA rule?
Had the case gone to trial, it would have been decided partly on the basis
of the 1968 regulations and partly on the 1994 regulations.1816 Key ele-
ments in the 1968 DA rule assessment would have been (i) the incurrence
of significant, non-refundable marketing costs by the US subsidiary; (ii)
the use of local know-how, marketing expertise and relationships used to
market Zantac; (iii) the fact that the marketing was carried out in the Unit-
ed States; and (iv) the fact that such marketing could not have been carried
out by the parent from abroad.
Before the subsidiary became so profitable that it was capable of finan

cing the US marketing expenditures through its own earnings in 1986, the
UK parent reimbursed “substantial amounts of … marketing and launch
expenses”.1817 However, the subsidiary itself paid for all of the US market-
ing costs for the years under review. The subsidiary therefore bore the fi-
nancial risk of the marketing activities. Further, it seems likely that the US
subsidiary developed unique marketing intangibles in the form of customer
relationships, marketing infrastructure and goodwill that were separate
from the US value of the trademark. If so, such unique intangibles were
undoubtedly owned by the subsidiary and likely contributed positively to
the value-building process for the Zantac trade name in the United States.
As far as the author can see, this makes it likely that the US subsidiary
would have been deemed developer of the US value of the licensed trade-
mark under the 1968 DA rule. Thus, for the income years 1989-1994, all
residual profits from the licensed marketing intangibles should have been
allocated to the US subsidiary. For the subsequent income years 1995-
1996, the subsidiary would have likely, for the same reasons, been deemed
owner of the US marketing intangibles, either pursuant to the “economic
1816. The 1968 regulations (33 Fed. Reg. 5848) would apply to income years 1989-
1994, and the 1994 regulations (59 FR 34971-01) would apply to subsequent years un-
der review.
1817. 2 April 2004 petition (available at Tax Analysts, doc. 2004-7600), at p. 16.
566
substance” exception or the “multiple owners” exception of the 1994 regu-

lations.1818 It then also, for the last 2 years encompassed by the reassess-
ment, would have been entitled to all residual profits generated by the de-
veloped marketing intangibles. Thus, under US law, the UK parent should
not have been allocated any residual profits from the licensed marketing
intangibles.
The question would then be of whether article 9 of the 1975 UK-US tax
treaty would accept the US allocation.1819 The 1979 OECD report con-
tained no useful guidance with respect to the allocation of residual prof-
its from marketing IP.1820 The 1995 OECD TPG, which would have been
relevant at least for the last few of the reassessed years, contained only
limited guidance. Ultimately, the key issue for these last years would likely
have been, as indicated by the 1995 OECD TPG, whether the US distribu-
tion subsidiary incurred marketing costs that exceeded the level of costs
that would have been incurred by an unrelated distributor in comparable
circumstances. If so, the US allocation of residual profits from the locally
developed marketing intangibles should be acceptable under article 9 of the
1975 UK-US treaty.
19.2.5.3. DHL Corporation v. CIR (1998)
A key question in DHL was whether a US entity within the DHL group
should be considered the developer of, and thus entitled to the residual
profits from the exploitation of, international rights to the DHL trademark,
pursuant to the 1968 DA rule.1821
DHL’s international delivery business was built up gradually from the

1970s from a domestic US business into a substantial international opera-
1818. The economic substance exception was contained in the 1994 Treas. Regs. (59
FR 34971-01) § 1.482-4(f)(3)(ii)(A), and the multiple owners rule in the 1994 Treas.
Regs. (59 FR 34971-01) § 1.482-4(f)(3)(i). See the discussions in secs. 19.4.5. and
19.4.6., respectively.
1819. Convention between the Government of the United States of America and the
Government of the United Kingdom of Great Britain and Northern Ireland for the
Avoidance of Double Taxation and the Prevention of Fiscal Evasion with Respect to
Taxes on Income and Capital Gains (31 Dec. 1975), Treaties IBFD.
1820. See the comments in sec. 19.5.
1821. DHL Corp. & Subsidiaries v. CIR, T.C. Memo. 1998-461 (Tax Ct., 1998), af-
firmed in part, reversed in part by 285 F.3d 1210 (9th Cir., 2002). For comments on the
case, see, e.g. Brauner (2008), at p. 151; Schön et al. (2011), at pp. 202-204; Wittendorff
(2010a), at p. 659; Andrus (2007), at p. 637; Levey (2000); and Levey et al. (2002). See
also Casley et al. (2011), at p. 167.
567
Ownership
tion. The company DHL US carried out the courier business in the United
States, while the company DHLI carried out the international side of the
business through a network of local operating subsidiaries that functioned
largely autonomously. In the early 1990s, a consortium of international
investors, including Japan Airlines and Lufthansa, acquired a controlling
interest in DHLI. In connection with the acquisition, DHL US transferred
all rights to the DHL trademark to DHLI for a separate consideration. The
DHL trademark was the significant intangible asset in the DHL group and
was highly valuable. The IRS issued a reassessment premised on DHL US
owning the DHL trademark, increasing the transfer price.
It was not contested that DHL US owned the US rights to the trademark;
the question was whether it also, under the 1968 DA rule, owned the inter-
national rights.
DHL US stood as the legal owner of the trademark in a licence agreement

with DHLI. Other agreements and correspondence between the parties
also indicated that DHL US was the legal owner, but the agreements were
informal and imprecise. Also, in negotiations with investors, DHL US was
presented as the owner. However, the trademark was registered to DHLI in
all countries apart from the United States, and DHLI incurred all market-
ing costs with respect to the international rights.
The IRS asserted that, in the absence of a CSA, all rights to the trademark
were owned by DHL US. The arguments were that DHL US was the legal
owner of the trademark and that DHL’s marketing expenses were not above
an arm’s length level.1822 DHL, on the other side, argued that legal owner-
ship was irrelevant under the 1968 DA rule, And further that DHLI should
be regarded as the owner of the international rights because DHLI incurred
all marketing costs connected to those rights.
The Tax Court ruling was, in the author’s view, unfortunate for several
reasons. First, the court misguidedly focused on the licence agreement be-
tween DHL US and DHLI. It found that, under US intellectual property
law, a licenser was entitled to the value of a trademark even if that value
had been created through the efforts of a licensee.1823 Further, the validity
of a trademark licence was contingent on sufficient control by the licenser
1822. On this last aspect of the Tax Court ruling (i.e. on the “bright line test” with
respect to whether the marketing expenses exceeded an arm’s length level), see, in
particular, Levey et al. (2006), at p. 3.
1823. See Cotton Ginny, Ltd. v. Cotton Gin, Inc., 691 F.Supp. 1347 (S.D.Fla., 1988).
568
over the quality of services sold under the trademark by the licensee.1824
The Court found that the “unique relationship of the corporate entities”
was an adequate indication of the required control, even though there was
no legal basis in the agreement to constitute such control. The Court, in the
author’s view, disregarded the separate entity approach under IRC section
482, not to mention the fact that this entire problem was not relevant under
the 1968 DA rule.
Second, the Tax Court assessed the question of ownership and the determi-
nation of the developer under the 1968 DA rule separately. The two ques-
tions are the same and should have been determined as one under the DA
rule. In essence, the Court framed the issue under the DA rule as whether
the marketing efforts of DHLI justified joint ownership of the marketing
intangible, a notion contrary to the fundamental point of the DA rule, which
distinguishes a developer, who is entitled to residual profits, from assisters,
who are not. There could only be one developer under the 1968 DA rule.
Third, the Tax Court rejected DHL’s argument that the substantial market-
ing costs incurred by DHLI for establishing the DHL brand internationally
entitled DHLI to a stake in the value of the DHL trademark. The Court
dismissively stated:
[I]n answering the question of whether the ownership of the DHL trademark
was bifurcated between DHL and DHLI, we do not look to the section 482
regulations cited by petitioners. Although those regulations may have some
effect on our allocation decision, they are not relevant in deciding the owner-
ship of the trademark rights as a predicate for valuing the trademark.1825
Obviously, the 1968 regulations were not only relevant, but decisive.1826
Fourth, the Tax Court rejected that DHLI should be assigned ownership to
the worldwide trademark rights on the basis of its marketing expenditures,
as “there has been no showing that the costs incurred for registration and/
or enhancement of the trademark were more than a licensee would have
expended at arm’s length”.1827
1824. See 15 U.S.C. § 1055 and § 1127; and Haymaker Sports, Inc. v. Turian, 581 F.2d
257 (C.C.P.A., 1978). See also McCarthy (1996), at sec. 18.42, pp. 18-66. A lack of con-
trol provisions could entail that the licenser’s rights to the trademark are abandoned;
see Stanfield v. Osborne Industries, Inc., 839 F.Supp. 1499 (D.Kan., 1993), order af-
firmed by 52 F.3d 867 (10th Cir., 1995), certiorari denied by 516 U.S. 920 (U.S., 1995).
See also McCarthy (1996), at sec. 18.48, pp. 18-75.
1825. See T.C. Memo. 1998-46, at p. 43.
1826. See also Ossi (1999).
1827. See T.C. Memo. 1998-46, at p. 55.
569
Ownership
This was clearly not a relevant factor under the 1968 DA rule. The question
was simply that of which entity incurred the lion’s share of the development
costs. The author suspects that the Court’s erroneous interpretation of the
law on this point was influenced by the somewhat contradictory arguments
from the DHL group with respect to the significance of marketing expen-
ditures. On the one side, it was important for DHL to clearly convey that
DHLI incurred substantial worldwide marketing expenditures (worth ap-
proximately USD 340 million) in order to substantiate its claim that DHLI
should be considered the developer under the DA rule. On the other side,
DHL sought to defend its modest USD 20 million valuation of the trade-
mark transfer as an arm’s length price, a number grossly incommensurate
with the marketing expenditures incurred to build the value of the trade-
mark. The Tax Court’s opinion was, however, immoderate; DHLI was not
even afforded the status of assister under the DA rule.
On appeal, the Ninth Circuit reversed the Tax Court’s ruling on the issue of
the ownership of the international rights to the DHL trademark. The Ninth
Circuit found no legal basis to require DHL to demonstrate that DHLI’s
marketing expenditures exceeded the level that would have been incurred
in arm’s length dealings. It was found that the Tax Court had mistakenly
applied a rule in the 1994 regulations that governed the allocation of in-
come to assisters that stated that assistance did not include expenditures
of a routine nature that an unrelated party dealing at arm’s length would
be expected to incur under similar circumstances.1828 Of course, the 1994
provisions did not apply to the reassessed income years (1982-1992). The
Ninth Circuit found that even if it had been applicable, it would have been
difficult to apply, as there was no clear line between the development and
exploitation of a trademark. Further, DHLI, through its marketing efforts,
had developed the DHL trademark outside the United States, and it was
the service network of DHLI that was the basis for the value of the trade-
mark.1829
The Ninth Circuit found that the Tax Court’s reliance on legal ownership
for determining the allocation of profits was in conflict with the wording
of the DA rule. The Ninth Circuit found support for its interpretation in the
preamble of the 1994 regulations, where it was made clear that the 1968
regulations disregarded legal ownership for the purpose of determining
which entity was the developer under the DA rule.
1828. See 1994 Treas. Regs. (59 FR 34971-01) § 1.482-4(f)(3)(iii).

1829. See Levey (2000) for a discussion of the divide between the exploitation and
development of a trademark.
570
Determination of IP ownership under the 1992 proposed and 1993
temporary US regulations
Further, the Ninth Circuit Court assessed the four factors listed in the DA
rule.1830 It found that the first factor, pertaining to the relative development
costs and risks borne by each controlled party, clearly indicated that DHLI
was the developer of the foreign rights to the trademark. DHLI undertook
registration of the trademark in all foreign jurisdictions and bore the re-
lated expenses, it paid for all marketing campaigns (totalling USD 340
million), it bore costs for protecting the trademark against infringement
and handled all disputes. Further, all market-development activities out-
side of the United States were carried out by DHLI, and DHLI was found
to be better situated to carry on the development independently, given its
connections in international market jurisdictions. Finally, DHLI exercised
control over the advertising and development of the international rights to
the DHL trademark. Conversely, DHL US had not incurred any expenses
with respect to the foreign rights to the trademark.
The Ninth Circuit thus held that DHLI was the developer of the interna-
tional rights to the DHL trademark under the 1968 DA rule or, alterna-
tively, that DHLI provided assistance to DHL US’s development. If DHLI
was regarded as the developer, there would be no basis for allocating any
value from the worldwide rights to the trademark to DHL US. Alterna-
tively, if DHLI was regarded as an assister, DHL US would be obliged to
pay a consideration for that assistance, thereby completely setting off the
USD 50 million transfer pricing increase that followed from the IRS’s reas-
sessment. Thus, under both solutions, there would be no profit allocation
to DHL US with respect to the value of the worldwide rights to the DHL
trademark.

proposed and 1993 temporary US regulations
The 1992 proposed regulations carried over the DA rule, but added the
so-called cheese example, which would soon become infamous.1831 In the
period from the mid-1970s to the early 1990s, the IRS had garnered experi-
ence with the allocation of operating profits to US distribution subsidiaries
of foreign-based multinationals. The “cheese” example expressed the IRS’s
position on how such allocations should be performed.
1830. See the analysis of the 1968 DA rule in sec. 19.2.2.

1831. 1992 Prop. Treas. Regs. (57 FR 3571-01) § 1.482-2(d)(8)(iv), Example 4.
571
Ownership
The example pertained to a foreign parent that owned an established and

valuable trade name connected to a particular brand of cheese. The trade
name was not known in the United States prior to its introduction by a
newly established US distribution subsidiary. The subsidiary incurred mar-
keting expenses of USD 5 million, which were not reimbursed by the par-
ent company, and supervised local marketing efforts. The example deemed
the subsidiary the developer of the US rights to the trade name pursuant to
the DA rule. Some commentators argued that the example augmented the
DA rule,1832 but the IRS’s position was that it merely clarified it.1833
A problem with this initial version of the cheese example was, in the au-
thor’s view, its description of marketing expenses. No indication was pro-
vided as to the relative significance of the USD 5 million. Depending on
the size of the multinational in question and the relevant marketing budget,
the amount could be relatively large or small and lie below or above the
level of expenses that a comparable unrelated distribution company would
be obliged to incur without reimbursement, for instance, under a third-
party franchisee agreement. Thus, given that it would not be possible to
determine whether the expenditures were at arm’s length, the author finds
it speculative that the example nevertheless deemed the subsidiary the de-
veloper, thereby entitling it to the residual profits.1834

final US regulations
The 1994 US regulations introduced new rules for determining which

group entity should be assigned ownership, and thereby the residual profits,
from jointly developed intangibles. A key aspect of the 1994 IP ownership
approach was the partial relinquishment of the DA rule. The 1994 IP own-
ership provisions were heavily modified when the new intra-group service
rules were introduced in 2009, but the basic structure put in place in 1994
1832. It was also asserted that the “cheese” example was not compatible with the guid-
ance provided in the regulations for the application of the resale price to tangible trans-
actions.
1833. See the preamble to the 1993 regulations (58 FR 5263-02).
1834. The 1992 proposed regulations (57 FR 3571-01), including the “cheese” ex-
ample, were carried over to the 1993 temporary regulations (58 FR 5263-02) without
substantial changes. The DA rule was contained in § 1.482-4T(e)(3), and the cheese
example in § 1.482-4T(e)(3)(iv), Example 4.
572
Determination of IP ownership under the 1994 final US regulations
remains intact in the current IP ownership rules. In this sense, it can be said
that the 1994 provisions formed the foundation for the current US rules on
IP ownership that are analysed in chapter 21 and chapter 23. A thorough
understanding of the 1994 provisions will aid in the interpretation of the
current rules. The author therefore finds an analysis of the 1994 IP owner-
ship rules necessary.
The discussion is organized as follows. In order to properly frame the con-

text in which the 1994 IP ownership provisions were introduced, in sec-
tions 19.4.2. and 19.4.3., the author will comment on two key reasons for
why the United States replaced the DA rule. Further, the point of departure
for assigning IP ownership under the 1994 regulations will be discussed
in section 19.4.4. The “economic substance” exception is discussed in sec-
tion 19.4.5. The “multiple owners” exception is discussed in section 19.4.6.
Concluding remarks are provided in section 19.4.7.
19.4.2. Reason 1 for replacing the DA rule: Criticism against

its treatment of the legal owner
The DA rule was criticized for not giving effect to legal ownership for
profit allocation purposes. As far as the author can gather, this was founded
on the basic observation that the group entity deemed to be the developer
under the DA rule would not necessarily hold legal title to the IP. In this
direction, the preamble to the 1994 regulations stated that “at arm’s length,
the legal owner could transfer the rights to the intangible to another per-
son irrespective of the developer’s contribution to the development of the
intangible”.1835
This wording is ambiguous. Only the developer could transfer the intan-
gible at arm’s length under the DA rule. Another thing entirely is that the
group entity that held the legal title to the IP could legally transfer it to
another entity. The author therefore interprets the preamble to express that
practical complications could arise when the group entity that held the le-
gal title was not the entity deemed to be the developer under the DA rule.
While the author does not contest this assertion as such, his view is that it
1835. See the preamble to § 1.482-4(f)(3) in the 1994 regulations (59 FR 34971-01).
The sentence following the quote is as follows: “On the other hand, it would be unlikely
that at arm’s length an unrelated party would incur substantial costs adding value to an
intangible that was owned by an unrelated party, unless there was some assurance that
the party that incurred the expenses would receive the opportunity to reap the benefit
attributable to the expenses.”
573
Ownership
should have no bearing on the usefulness of the DA rule. In principle, there

should be no significant pricing consequences connected to controlled
transfers of the legal title alone.
Realistically, the true core of the criticism likely emanated from the desire
of multinationals to extract residual profits – particularly from marketing
IP – from the source by way of allocating them to a foreign legal owner, re-
gardless of the extent to which a US distribution subsidiary had contributed
to the intangible’s development. Multinationals were presumably behind
a 1994 draft of replacement provisions for the DA rule, penned by some
Washington lawyers.1836
The draft claimed that a legal owner in third-party relationships would

control the IP that it owned and therefore be entitled to the income from it.
The draft argued that the same should apply among related entities. Closely
connected to this was the comparable profits method (CPM)-akin pricing
argument that US distribution entities that only contributed routine inputs
to the value chain should not be entitled to residual profits, but should be
compensated with a normal market return (e.g. on a cost-plus basis). The
author does not find this chain of reasoning to be convincing. Intellec-
tual property law is concerned with the protection of a group’s IP from
third-party infringement, not with how the group decides to arrange the
registration of ownership of its IP internally among its own entities. The
allocation of operating profits among jurisdictions would be disconnected
from intangible value creation if it were to hinge on legal form. To claim,
as the draft did, that the allocation of operating profits pursuant to legal
title alone would comport with economic reality is, in the author’s view,
without merit.
Further, the draft claimed that there was inconsistency between the 1968
DA rule and the 1993 proposed regulations for tangible property. In par-
ticular, the allocation of operating profits described in an example in the
proposed regulations (Example 9),1837 illustrating the application of the re-
sale price method, was not compatible with the DA rule. Example 9 per-
tained to a US distribution subsidiary that sold a product manufactured by
its foreign parent. The product brand name at the outset was not widely
known in the United States and did not command a premium price. The
1836. The draft was written by W.P. McClure and four other lawyers with McClure,
Trotter & Mentz and was sent to the International Tax Counsel at the Department of the
Treasury in a letter dated 14 March 1994. See McClure (1994) for the letter and draft in
full.
1837. See 1993 Temp. Treas. Regs. (58 FR 5263-02) § 1.482-3T(c)(4), Example 9.
574
subsidiary and parent agreed that local US marketing should be carried

out to develop the brand name. The subsidiary incurred the marketing ex-
penses and supervised the marketing efforts. The parent bore the market-
ing costs by reducing the transfer price of the product sold to the subsidiary
for local resale. Thus, the subsidiary effectively recharged its marketing
expenses to the parent.
As a result of the marketing efforts, the product began to command a pre-

mium price in the United States. The example stated that this “value should
inure to the benefit of P since there has been no transfer to S of any intan-
gible rights in the brand name”. In other words, the residual profits due to
the US marketing intangible should be allocated fully to the parent. A fun-
damental criterion under the 1968 DA rule and the 1993 “cheese” example
(see section 19.3.) was that the developer bore the intangible’s development
costs.1838 This was not the case in Example 9. Thus, this assertion also
lacked merit.
19.4.3. Reason 2 for replacing the DA rule: OECD conformity
In the early 1960s, outbound foreign direct investment (FDI) from the
United States was approximately five times the amount of inbound FDI.1839
By the mid-1990s, the United States imported roughly the same amount of
capital as it exported. The increase in inbound FDI triggered transfer pric-
ing disputes.1840 Trade between the United States and Japan – at the time,
the two largest economies in the world measured by GDP – represented
one of the most significant commercial relationships globally.1841 A notable
1838. See 1993 Temp. Treas. Regs. (58 FR 5263-02) § 1.482-4T(e)(3), § 1.482-4T(e)(3)
(iv), Example 4.
1839. In the wake of the attention paid to roundtrip transactions in the late 1970s, the
IRS began investigating import transactions by foreign-based multinationals, focusing
on taxable losses due to marketing expenses. For an example of a case pertaining to
Japanese distribution subsidiaries in the United States, see US v. Toyota Motor Corp.,
569 F.Supp. 1158 (C.D.Cal., 1983); and U.S. v. Toyota Motor Corp., 561 F.Supp. 354
(C.D.Cal., 1983).
1840. The IRS issued audit statistics in the early 1990s. Total understatements for the
ten largest trading partners of the United States amounted to USD 1.3 billion. This
figure included understatements in the tax returns of 314 US subsidiaries of Japanese
multinationals, to which USD 507.8 million in adjustments were proposed. See Wartz-
man (1993); Polinsky (1993); Haas (1991); Turro (1990); and Dworin (1990).
1841. When negotiations for the revised Japan-United States treaty commenced in the
late 1960s, Japanese multinationals carried out relatively modest trade in the United
States. In the 1990s, that trade increased significantly. See Miyatake et al. (1994), foot-
note 25 for further references.
575
Ownership
disagreement arose between the competent authorities of the United States

and Japan with respect to the allocation of operating profits to US distribu-
tion subsidiaries of Japan-based multinationals.1842
The IRS asserted that the profits of US distribution subsidiaries were re-
duced as the result of profit shifting. Apparently, it was unofficially held
that Japanese multinationals used entities in low-tax treaty jurisdictions
(e.g. Singapore) to push US profits back to Japan without incurring tax.
These alleged schemes were made possible through a combination of ag-
gressive transfer pricing strategies, including the utilization of Singapore
tax holidays and the tax-sparing credit benefits of the 1971 Japan-Singa-
pore treaty.1843 The Japanese tax authorities held that Japanese enterprises
made relatively less operating profits than their US counterparts, and that
the profit margins of US distribution subsidiaries of Japanese multination-
als therefore could not meaningfully be compared to the margins of solely
domestic entities.1844 Japan held that its argument was supported by the fact
that the corporate income tax rate in Japan was higher than in the United
States and that it therefore did not make sense to shift taxable profits from
the United States to Japan.1845
Japan represented a resounding voice within the OECD, urging the United
States to conform its transfer pricing rules to the OECD consensus.1846 The
OECD’s views, as expressed in the 1992 and 1993 Task Force Reports,
caused movement on the US side towards moderation on several of the
positions taken in the 1992 proposed US regulations.1847
1842. For a thorough discussion of the Japan-US relationship, see Miyatake et al.
(1994). There were over 20 cases of competent authority negotiations between the
United States and Japan in 1996; see Yoost (1996), in some of which, there allegedly
were problems with obtaining effective double taxation relief due to reluctance from
both Japan and the United States to provide necessary refunds.
1843. See art. 21, nos. 1 and 4. See also Deloitte (2000); Tan (2001); and Miyatake et
al. (1994), footnote 82. Pursuant to art. 21 of the treaty, Japan was obliged to provide
credit for Japanese tax for tax paid on the same income in Singapore. The Japanese tax
credit included relief for the amount of Singapore tax that would have been payable if
not for a tax exemption or reduction under the special incentive provisions designed to
enhance the economic growth in Singapore.
1844. See also the comments on international reactions to the 1988 White Paper in sec.
5.3.3., in particular, supra n. 524.
1845. See Turro (1992).
1846. See the reported comments by officials of the Japanese Ministry of Finance and
National Tax Administration in Tax Management Transfer Pricing Report (TMTPR)
(1992).
1847. See supra n. 910 on the objections raised in the reports against the cost-plus
method.
576
The author assumes that the US relinquishment of the DA rule may have
been seen by the United States as a reasonable compromise, given the
OECD controversy over other provisions in the proposed regulations, in
particular the CPM and the periodic adjustment provision. The main US
concern was likely the avoidance of asymmetrical allocation of income and
expenses connected to US-developed IP, and in particular that US distribu-
tion subsidiaries of foreign-based multinationals should not be entitled to
deduct substantial marketing expenses if the resulting residual profits were
allocated to foreign group entities.
19.4.4. The point of departure under the 1994 IP ownership

rules: The legal owner (of IP subject to legal
protection) and the developer (of IP not subject to
legal protection) are entitled to residual profits
The 1994 regulations took a two-branched approach to the determination

of which group entity should be considered the owner of an intangible,
depending on whether the relevant intangible was subject to legal protec-
tion or not.
With respect to legally protected intangibles,1848 the legal owner of a right

to exploit such an intangible would “ordinarily” be considered the owner,
and thus entitled to the residual profits.1849 This is regardless of how and
to which extent other group entities contributed to the intangible devel-
opment.1850 Legal ownership included ownership acquired through law or
contract.
The 1994 regulations provided two exceptions to this rule. First, the IRS
was entitled to impute an agreement pursuant to which the residual profits
from the US exploitation of a foreign-owned intangible would be allocated
to the US entity if this aligned with the economic substance of the actual
conduct of the controlled parties.1851 Second, the right to exploit an intan-
1848. The intangibles of most significance in transfer pricing practice tend to fall with-
in this category. The lion’s share of the case law analysed in this book pertains to the
allocation of profits from patents and trademarks. The author therefore assumes that
the profit allocation consequences of the legal ownership rule (and its exceptions) were
relatively more significant than those of the ownership rules for intangibles not subject
to legal protection.
1849. 1994 Treas. Regs. (59 FR 34971-01) § 1.482-4(f)(3)(ii)(A).
1850. 1994 Treas. Regs. (59 FR 34971-01) § 1.482-4(f)(3).
1851. See the analysis in sec. 19.4.5.
577
Ownership
gible could be disaggregated into discrete legal units capable of separate

ownership.1852
If triggered, the exceptions would entitle a US group entity, under the ra-
tionale of deemed ownership, to the residual profits from an intangible that
it had contributed to the creation of, but did not hold legal title to.
Further, with respect to intangibles not subject to legal protection (e.g.

trade secrets and goodwill), the developer was considered the owner.1853
The regulations specifically stated that, outside of CSA contexts, when two
or more group entities jointly developed an intangible, only one of the enti-
ties would be regarded as the developer, with the remaining entities consid-
ered to be assisters.1854 Thus, for intangibles not subject to legal protection,
the DA rule in effect prevailed under the 1994 US regulations.
The 1994 regulations prescribed a broad facts-and-circumstances-based

assessment to determine which group entity should be deemed developer,
akin to that of the 1968 regulations. Key factors continued to be which
entity carried the largest portion of the development costs and which en-
tity contributed IP without adequate compensation. If this could not be
determined, the location of the development activities, the capacity of the
involved parties to carry on the development activities independently, the
extent to which each involved party controlled the development project and
the actual conduct of the parties had to be taken into account.
An assister was entitled to an arm’s length compensation “determined

in accordance with the applicable rules under section 482”.1855 However,
expenditures of a routine nature that an unrelated party dealing at arm’s
length would be expected to incur without compensation under circum-
stances similar to those of the controlled taxpayer would not command
remuneration.1856
An interesting question is whether an assister could be entitled to a por-

tion of the residual profits from the fully developed intangible pursuant to
the PSM, provided that it contributed unique inputs to the IP development
process. This may immediately seem somewhat counterintuitive. After all,

1853. 1994 Treas. Regs. (59 FR 34971-01) § 1.482-4(f)(3)(ii)(B).
1854. Id. Assistance included loans, services and the use of tangibles or intangibles as
contributions to the intangible’s development.
1855. 1994 Treas. Regs. (59 FR 34971-01) § 1.482-4(f)(3)(iii).
1856. Id.
578
if assisters could also be entitled to residual profits, why bother to separate

the developer from the assisters in the first place?
Nevertheless, the 1994 US regulations did refer to the transfer pricing

methods for determining the amount of operating profits allocable to an
assister, thereby, in principle, leaving the door open for applying the PSM.
The IP ownership provisions of the 1994 regulations were geared towards

marketing IP.1857 The second “cheese” example pertained to a US distribu-
tion subsidiary that incurred excessive marketing costs, but was deemed to
be an assister. It was allocated operating profits equal to “the fair market
value of the services” provided by it.1858 The wording could, at first glance,
seem to indicate that the subsidiary was entitled only to a normal market
return pursuant to the one-sided methods, but it does not rule out the pos-
sibility that the subsidiary could be entitled to a portion of the residual
profits.1859
In conclusion, the author’s impression is that assisters – if they contributed

unique development inputs – in principle could be allocated residual prof-
its under the 1994 regulations, as their remuneration was governed by the
pricing methods. However, if only marketing expenses were contributed,
the point of departure should be that a normal market return would suffice.
Thus, if a distribution subsidiary possessed local marketing intangibles
(e.g. goodwill, know-how or customer relationships) that it contributed to
the development of new, or to the enhancement of the value of existing,
marketing IP, the potential for a profit split could be triggered.
1857. The most significant contributions to the development of marketing intangibles

are marketing expenses. At least at the outset, marketing expenses should not be re-
garded as unique contributions that attract residual profits, but should rather be allo-
cated a risk-adjusted return. Compare the analysis of research and development (R&D)
funding remuneration under the OECD TPG in sec. 22.4.
1858. 1994 Treas. Regs. (59 FR 34971-01) § 1.482-4(3)(iv), Example 3.
1859. Langbein (2005), at p. 1076 noted that the regulations did not directly answer
the question of whether an assister should be allocated only a normal market return or
a portion of the residual profits. He found little support for either direction in the 1994
Treas. Regs. (59 FR 34971-01) § 1.482-2(b)(3), which governed profit allocation for
intra-group services, or in the provision pertaining to tangible rental in the 1994 Treas.
Regs. (59 FR 34971-01) § 1.482-2(c)(2). However, on the basis of the former provision,
Langbein concluded that “the best argument would seem to be that this provision im-
plies that the allocation should involve a ‘share’ of the profit, not just a marginal return”.
Further, he found that if the assister were to receive a portion of the residual profits,
it should be calculated on the basis of the assister’s intangible development costs. The
author take it that this implied that the residual profits should be allocated among the
legal owner of the intangible and the assister in relation to the intangible’s development
costs incurred by them.
579
Ownership
19.4.5. The first exception to the legal ownership rule:

Economic substance
The 1994 regulations authorized the allocation of profits from the exploita-
tion of legally protected IP to a group entity different from that which held
legal title to it, provided that the actual conduct of the parties indicated that
the former entity, in economic substance, was the owner.1860
The economic substance imputation provision was illustrated in an exam-

ple pertaining to a US distribution subsidiary that sold products under an
initially unknown trade name, which was legally owned by a foreign par-
ent.1861 The subsidiary incurred expenses in years 1-6 to promote the trade
name that were substantially larger than the marketing expenses incurred
by unrelated parties in comparable transactions. By year 7, the trade name
was established in the United States, and the product generated residual
profits. The contractual relationship between the entities was then struc-
tured so that the subsidiary became a commission agent, allowing it to earn
a normal market return only.
The example deemed it unlikely that the subsidiary, at arm’s length, would
incur such incremental expenses in years 1-6 without benefitting in year
7 onwards. It therefore allocated an “appropriate portion of the price pre-
mium” attributable to the trade name to the subsidiary.1862
The quoted wording may be interpreted as to allow a split of the residual

US marketing profits. That is somewhat peculiar, as the construction is that
the subsidiary, in economic substance, is deemed the owner. It therefore
seems more natural that the subsidiary would be entitled to the entirety of
the profits. However, a profit split would likely result in most of the profits
being allocated to the subsidiary in any case. The reason for this is that the
allocation would likely have to be based on an analogical application of the
PSM, under which the profits would be split pursuant to the relative value
of the respective intangible development contributions.1863 The legal owner
1860. 1994 Treas. Regs. (59 FR 34971-01) § 1.482-4(f)(3)(ii)(A), last two sentences;
see also § 1.482-1(d)(3)(ii)(B).
1861. 1994 Treas. Regs. (59 FR 34971-01) § 1.482-1(d)(3)(ii)(C), Example 3.
1862. Langbein (2005), at p. 1074 noted that it was unclear as to at which point the dis-
tributor would be deemed owner. The author agrees, in principle, but it may be that, in
practice, residual profits are generated subsequent to a relatively lengthy development
process, after which it could be clear that the distributor should be deemed owner.
1863. See the analysis of the US profit split method in ch. 9.
580
would likely not “contribute” anything apart from being registered as the
legal title holder, while all of the marketing functions and risks would be
provided by the subsidiary.
19.4.6. The second exception to the legal ownership rule:

The multiple owners rule (1994 cheese examples)
The 1994 regulations found that because the right to exploit an intangible
can be “subdivided in various ways”, a single intangible may have “mul-
tiple owners” under IRC section 482.1864 For instance, if the owner of an
intangible licensed exclusive exploitation rights for a specific geographical
area for a set period of time to a related entity, “both the licensee and the
licensor will be considered owners” with respect to their discrete exploita-
tion rights. This rule applied to both legally protected and non-protected
intangibles, but would likely be relevant mostly in the former context. The
author therefore views it as an exception to the legal ownership rule. He
will refer to this as the “multiple owners rule”.
The notorious 1992 cheese example (see the discussion in section 19.3.)
was now expanded into three variants in an attempt to illustrate both the
legal ownership rule and the multiple owners exception.1865 The examples
shed light on the degree to which the new approach yielded results that
deviated from the 1968 DA rule. In all three variants, the US subsidiary
incurred concurrent marketing expenses that were not refunded by its for-
eign parent.1866
In the base example,1867 the US subsidiary incurred marketing expenses

on a level comparable to that of uncontrolled distributors that marketed
similar products in the United States under a foreign trade name. In this
situation, no profits would be allocated to the subsidiary for its marketing
activity, as it would have incurred the same level of expenses, had it been
unrelated to its parent.
1864. 1994 Treas. Regs. (59 FR 34971-01) § 1.482-4(f)(3)(i).

1865. For an informed discussion of the “cheese” examples, see Culbertson et al.
(2003), at p. 118. See also the comments in Levey et al. (2006), at p. 2.
1866. The 1994 cheese examples were controversial. The IRS would remain focused
on clarifying them. See Bureau of National Affairs (BNA)(1998), reporting comments
by Associate Chief Counsel International Michael Danilack; and BNA (1999).
1867. 1994 Treas. Regs. (59 FR 34971-01) § 1.482-4(f)(3)(iv), Example 2.
581
Ownership
In the second variant,1868 the marketing expenses were significantly larger

than those incurred by independent distributors under similar circum-
stances. The expenditures were considered separate services rendered by
the subsidiary, which required fair market value compensation. The author
interprets the example to require only a normal market return (typically
cost-plus-based) to be allocated to the subsidiary, as opposed to residual
profits.1869
In the third variant,1870 the US subsidiary, while incurring marketing ex-

penses significantly above an arm’s length level, operated under a long-
term contract with its parent, pursuant to which it had the exclusive right to
resell cheese purchased from the parent at arm’s length.
In this case, the regulations simply regarded the subsidiary as the owner
of the trademark for US exploitation purposes. Consequently, the excess
marketing efforts of the subsidiary were not seen as services rendered for
the benefit of the foreign parent that should be compensated on a separate
basis. The subsidiary was entitled to a significantly greater reward: the
entire US residual marketing profits.1871 The same result would likely have
followed from the DA rule, had it been applicable.
The author is not convinced that the multiple owners rule was well found-
ed. First, its threshold for application seemed rather artificial. Both the
second and third cheese examples pertained to a situation in which the US
distribution subsidiary incurred marketing expenses significantly above an
arm’s length level. In the second example, the foreign legal owner pre-
vailed, resulting in extraction of the residual US marketing profits from
the source. Conversely, the US subsidiary was allocated the residual profits
in the third example. The author does not find any substantial differences
between the factual patterns in these two examples. Yes, the third example
specified that there was a long-term and exclusive licence agreement in
place. However, at least ostensibly, that description could, in substance, fit
just as well in the second example. Controlled distribution structures are
often exclusive in substance, regardless of whether they are formalized.

1869. See also Jie-A-Joen et al. (2007), under the section entitled “Arm’s length com-
pensation for (incremental) marketing activities”.
1871. See Langbein (2005), at p. 1076, who seemed to disagree. He questioned whether
the allocation of income to the US entity in this third example, in which the entity is
deemed the owner, would be significantly different from the allocation that it would
have received had it been seen only as a service provider (as the case was in the second
cheese example).
582
Second, when viewed as a profit allocation rule, the multiple owners provi-
sion seemed to break with the rationale behind the CPM and PSM meth-
odology introduced in the very same 1994 regulations. These new pricing
methods dictated that the allocation of operating profits should be founded
on an analysis of the functions performed, assets used and risks incurred,
and in particular, that only unique contributions should attract residual
profits. The multiple owners rule, however, instead allocated residual prof-
its based on a narrower assessment, focused on the incremental market-
ing costs incurred by the local distribution subsidiary. If these costs were
high enough, the subsidiary could be deemed the owner, provided that a
long-term, exclusive licence agreement was in place. In other words, the
subsidiary could be allocated residual profits even if it did not contribute
any unique development inputs.
A possible explanation for this inconsistency may lie in the basic observa-
tion that the multiple owners exception targeted marketing intangibles, the
development of which is mainly cost-driven1872 and less reliant on unique
inputs than what typically will be the case for R&D-based manufacturing
intangibles. Also, a distribution subsidiary that develops a local marketing
intangible may, in practice, own other unique intangibles (e.g. goodwill or
know-how) that are used in its marketing development. If so, the result of
the multiple owners rule may be easier to reconcile with the logic of the
new pricing methods.
Professor Langbein seemed to be wary of the consequences for the US

tax base of rules that attract residual profits based on costs (including the
economic substance exception and the multiple owners rule).1873 His point
seemed to be that outsourcing could place such costs outside the United
States, thereby extracting profits that otherwise would have been allocable
to the United States. The author agrees, in principle, but the point of depar-
ture for determining IP ownership, both under the US regulations and the
OECD MTC, was legal ownership. Had there been no economic substance
exception and the multiple owners rule, multinationals would be free to as-
sign legal ownership, and thus residual profits, to the group entity of choice
in the first place.
This is a theoretical observation, as it is difficult to imagine that a legal

ownership rule without exceptions could realistically be implemented. The
multiple owners rule was geared towards inbound transactions, where the
1872. See also 1979 OECD Report, para. 126.

1873. Langbein (2005), at p. 1078.
583
Ownership
practical concern was to attract residual profits to the United States when a
local distribution subsidiary incurred, and deducted, excessive marketing
expenses. The multiple owners rule likely sought to protect the US tax base
through a substance view akin to that of the 1968 DA rule, while also ac-
commodating the OECD to some extent by moving towards the more com-
monly accepted platform of legal ownership. Conversely, the new pricing
methods (at least the CPM) were more geared towards outbound contexts,
focused on taxing intangible value created in the United States. Ultimately,
however, there should, in principle, be no difference in the fundamental
logic applied for allocating operating profits, regardless of whether the con-
text is inbound or outbound. The concern in both scenarios is to provide
appropriate remuneration for intangible development contributions.
19.4.7. Concluding remarks on the 1994 US approach to

intangible ownership
The 1994 replacement of the 1968 DA rule for legally protected intangi-
bles with the legal ownership rule was not as dramatic as one immediately
could assume, due to the economic substance and multiple owners excep-
tions. The replacement was, however, in the author’s view, largely a step
backwards, from an outright substance assessment towards giving effect,
at least at the outset, to legal formalities for pricing purposes. As discussed
in sections 19.4.1.-19.4.6., the author suspects that the reason behind the
change was to accommodate the historical view of the OECD that residual
profits from intangibles should accrue to the legal owner. The 1994 revi-
sion likely entailed a certain relaxation compared to the results that would
have followed from the DA rule in that less operating profits would be
allocable to the United States in some contexts. For instance, a US distri-
bution subsidiary could incur excessive marketing expenditures, only to
be remunerated on a cost-plus basis as a service provider under the 1994
regulations,1874 while the same would normally not be possible under the
DA rule.1875 This was likely a dilemma for foreign multinationals with US
1874. See the comments on the second cheese example in sec. 19.4.6.
1875. This, however, did not mean that a foreign-based multinational could just re-
move the exclusivity clause from its distribution agreement with its US subsidiary and
thereby enable the extraction of US residual profits free of consequences under the 1994
regulations. The IRS apparently took the audit position that a US entity would not, at
arm’s length, have incurred incremental marketing expenses without being an exclusive
distributor. Unless there was an exclusive distribution agreement, the marketing costs
would not be incurred to the benefit of the US subsidiary, and therefore would not be
deductible. See Boykin (1996).
584
Determination of IP ownership under the historical OECD TPG
sales. On the one side, they could opt for exclusive distribution agreements,
but that would likely attract residual profits to the US subsidiary under
the multiple owners rule.1876 Alternatively, if the multinational opted for
non-exclusive US distribution, it would likely suffice to allocate a normal
market return to the US entity, but income tax deductions would then be
denied for marketing costs above an arm’s length level.
The choice between exclusive or non-exclusive distribution arrangements

was therefore likely puzzling when substantial marketing expenses over
consecutive income periods were expected to be necessary, followed by
sizable residual profits from US sales. This could, for instance, be practi-
cal when a new, high-profit-potential product had to be introduced in a
competitive market through extraordinary marketing activities in the ini-
tial sales years.
19.5. Determination of IP ownership under the historical

OECD TPG
The 1979 OECD Report did not provide much guidance on how residual
profits generated through exploitation of intra-group-developed IP should
be allocated among the group entities that contributed to the IP develop-
ment process. With respect to manufacturing IP developed through joint
R&D, it offered only a brief description of relevant contractual structures,
but no guidance on profit allocation was to be found. On marketing IP, it
noted that “the value for the licensor may increase as the result of efforts
and expenditure by the licensee”.1877 However, no guidance was offered on
how the arm’s length royalty rate should be established when the licensee
contributed to the intangible value. It merely indicated that the costs in-
curred by the licensee would affect the determination.
1876. Not only that, but the US subsidiary would then be regarded as owner of the
US marketing intangible. A buy-out charge could possibly be triggered on the value of
the intangible if the US subsidiary in a business restructuring later was contractually
stripped of assets and risks in order to accommodate a low-risk distributor classification
aimed at allocating less profits to the United States. As the 1994 regulations were issued
around the time during which multinationals were becoming interested in the possi-
bilities offered by tax-driven business restructurings to establish centralized principal
models (see supra n. 185), the author assumes that this was a relevant consideration.
585
Ownership
The rule of the 1995 and 2010 OECD TPG was that the legal owner should
be allocated the residual profits.1878 No particular guidance was offered
with respect to the allocation of residual profits from intra-group-devel-
oped manufacturing IP. With respect to marketing IP, however, the position
was that the distributor would not be “entitled to share in any return attrib-
utable to the marketing intangible” if the foreign legal owner reimbursed
the marketing costs.1879 The same would not necessarily be the case if the
costs were not refunded. A distributor under a long-term, exclusive distri-
bution agreement could receive sufficient “benefits from its investments in
developing the value of a trademark” from increased sales.
The author interprets this to entail that the distributor would not be entitled
to separate remuneration for his marketing activities. The logic was likely
that this scenario was analogous to franchising agreements, in which the
distributor is expected to incur marketing costs that benefit the value of a
trademark owned by the franchiser without any specific remuneration.
The 1995 OECD discussion draft was seemingly more lenient on this point.
It stated that the problem could be approached by looking at the “functions
that each party performs”.1880 This wording was scrapped in the consensus
text of the final 1995 OECD TPG and was replaced with a reference only
to the “substance of rights”.1881 The 1995 OECD discussion draft further
assumed that the distributor, in an arm’s length dealing, normally would be
entitled to “some additional profits” if those activities were successful.1882
The author interprets this as possibly entailing that the distributor was to
receive some share of the residual profits generated through the exploita-
tion of the intangible.
The 1995 OECD discussion draft further stated that, to the extent that the
distributor incurred greater expenditures than unrelated parties would have
incurred in similar circumstances, an appropriate compensation should be
received.1883 It is not clear whether this wording intended to convey that
a cost-plus remuneration would be sufficient for the incremental costs or
whether the distributor should be entitled to share in the residual profits.
1878. Conversely, see the 1995 OECD TPG, para. 3.36. See the discussion in sec.
19.4.3. on how the material content of the US IP ownership rules gravitated towards
this OECD consensus approach.
1879. 1995 OECD TPG, para. 6.37.
1880. See the 1995 OECD discussion draft, para. 43.
1881. See the 1995 OECD TPG, para. 6.38.
1882. See the 1995 OECD discussion draft, para. 44.
1883. Id.
586
Determination of IP ownership under the historical OECD TPG
The OECD position taken in the final 1995 consensus text on this profit
allocation issue for when a distributor incurred “extraordinary marketing
expenditures” without reimbursement from the owner of the marketing IP
was hesitant.1884 It is noted that the distributor “might” be allocated a por-
tion of the residual profits from the marketing intangible, “perhaps” either
through a decrease in the purchase price for the product or a reduction in
the royalty rate.1885 It is apparent that the OECD in reality was unable to
reach clear consensus on this issue.1886
1884. 1994 Treas. Regs. (59 FR 34971-01) § 1.482-4(3)(iv), Example 3. The discus-
sion is clearly inspired by the cheese examples of the 1994 US regulations, which, in
comparison, applied the similar terminology of “significantly larger than the expenses”
incurred by unrelated distributors.
1885. 1995 OECD TPG, para. 6.38.
1886. The guidance pertaining to the treatment of extraordinary marketing expendi-
tures was a delayed issue in the process towards the final 1995 consensus text. There
was no wording covering this issue in the 1995 OECD discussion draft.
587
Chapter 20
A Lead-In to the Determination of IP Ownership under

the US Regulations and OECD TPG: A Story about
Legal Ownership, Control and Economic Substance
20.1. Introduction
This chapter will serve as a lead-in to the topic of which group entity should
be deemed owner of, and thereby entitled to the residual profits from, in-
tangibles developed internally within a multinational under the current US
regulations and the OECD Transfer Pricing Guidelines (OECD TPG).1887
The group entity that holds legal ownership of an intangible will, in prac-
tice, be the entity that actually receives royalty payments from other group
entities that license the intangible, simply because the group has structured
its agreements like this. The legal owner therefore forms a practical point
of departure for identifying which entity should be allocated the residual
profits. If it is not possible to identify a legal owner, typically because the
intangible in question is not subject to legal protection, the group entity that
controls the intangible will form this point of departure. Neither the US
regulations nor the OECD TPG place decisive relevance on legal owner-
ship in and of itself for profit allocation purposes. In the context of transfer
pricing, legal ownership should be seen as a formality that is comprehen-
sively unsuitable to guide the allocation of operating profits among juris-
dictions. Instead, the United States and OECD rely on substance-based
rules to align the allocation of operating profits from internally developed
intangibles with the underlying development contributions to intangible
value.
Nevertheless, some aspects of the US and OECD provisions on legal

ownership require clarification. The author will therefore comment on
the provisions in sections 20.2. and 20.3., respectively, before some brief
common remarks on the use of control as a criterion for ownership are
made in section 20.4. The substance-based ownership rules that are deci-
sive for profit allocation under both the US and OECD regimes are ana-
1887. See also Pankiv (2017), at p. 80, who emphasizes that the concept of ownership
for transfer pricing purposes functions as a “label” for a group entity’s entitlement to
intangible property (IP) profits.
589
Chapter 20 - A Lead-In to the Determination of IP Ownership under the US
Regulations and OECD TPG: A Story about Legal Ownership
lysed in depth in chapters 21-25 of the book. The author will introduce
them in section 20.5., as they may be seen as exceptions to the “legal
ownership rule”.
20.2. The US regulations
The current US regulations give effect to legal ownership for profit alloca-
tion purposes only if doing so is compatible with the economic substance
of the controlled transaction.1888 The author will briefly comment on the
legal ownership rule in section 20.2.2. before discussing the treatment of
licensees in section 20.2.3.
20.2.2. Legal ownership
The legal owner of an intangible is the group entity designated as the

owner pursuant to the intellectual property law of the relevant jurisdic-
tion, or as the holder of the rights constituting an intangible under con-
tractual terms or other legal provisions.1889 This entity is considered the
“sole owner” of the respective intangible, and is thereby entitled to the re-
sidual profits from the intangible. This will, in practice, be the entity that
is registered as the owner in a central government registry, for instance,
the US Patent and Trademark Office, or designated as the legal owner in
controlled agreements.
1888. Treas. Regs. § 1.482-4(f)(3)(i)(A). The rule was proposed in 2003 (68 FR 53448-
01), affirmed in 2006 (71 FR 44466-01) and adopted without changes in the final 2009
service regulations (74 FR 38830-01). See also Treas. Regs. § 1.482-4(f)(3)(i)(B),
which delimits the scope of the legal ownership rule vis-à-vis intangibles developed
through cost-sharing arrangements (CSAs) that are governed by Treas. Regs. § 1.482-7.
Wittendorff (2010a), at p. 625 states that “Section 482 does not govern the ownership
of intangibles for transfer pricing purposes”. He also states, at p. 629, that “Article 9
(1) does not govern the tax ownership of intangibles”. The author finds these assertions
to be ambiguous. There is no doubt that the intangible ownership provisions of the US
regulations and OECD TPG are decisive for the allocation of operating profits from
internally developed intangibles among the group entities that contributed to the devel-
opment process. See Brauner (2008), at p. 125 for reflections on the US IP ownership
provisions.
1889. Treas. Regs. § 1.482-4(f)(3)(i)(A).
590
The US regulations
20.2.3. The treatment of licensees
The US legal ownership rule conceals a convoluted legislative stance with

respect to the treatment of specific rights to an intangible in particular li-
cences. It states that “the holder of rights constituting an intangible prop-
erty pursuant to contractual terms (such as the terms of a license) … will
be considered the sole owner of the respective intangible property”.1890
In other words, a licensee is considered the legal owner of the licence. In

this section, the author refers to the quoted provision as the “discrete owner
rule”.1891 Its implications for how operating profits from the US exploitation
of a licensed (unique) intangible should be allocated between a foreign
licenser (the legal owner of the intangible as such) and the US licensee (the
legal owner of the limited rights, i.e. the licence) are ambiguous.1892
The 1968 developer-assister (DA) rule took the position that a subsidiary
that was deemed owner of the licensed US rights should not be obligated to
pay any royalties.1893 In other words, the subsidiary was entitled to keep the
entirety of the residual profits from the locally developed trademark. This
was also the stance in the third 1994 “cheese” example (see the analysis in
section 19.4.6.). The question is whether the 2009 discrete owner rule takes
the same position.1894
First, the preamble to the 2003 proposed regulations stated that a purpose
of the new ownership provision was to replace the “multiple owners rule”
in the 1994 cheese examples with the identification of “a single owner for
each discrete intangible”. This gives the impression that the new approach
was to differ from the 1994 result. One would also logically assume this
1890. Id.
1891. The rule may conceivably also encompass other rights than those conveyed
through a licence agreement, but this is likely not practical. In any case, the regulations
focus solely on licence agreements in the context of the discrete owner rule.
1892. The problem is, in particular, relevant when a US subsidiary licenses a marketing
intangible and incurs an above-arm’s length level of marketing expenses to build the
local value of the trademark. See the analysis of this issue in sec. 23.4.3.
1893. See the analysis of the 1968 developer-assister (DA) rule in sec. 19.2.
1894. In terms of context, a foreign third-party licenser of a unique trademark may
be in a bargaining position that makes it possible to extract most of the residual profits
from a licensee’s local exploitation of the trademark. However, if the licensee contrib-
utes non-routine inputs to the value chain (goodwill, know-how, etc.), the licenser may
be placed in a relatively weakened bargaining position and forced to surrender more
of the residual profits to the licensee. It is apparently typical in third-party licensing
structures (e.g. franchising agreements) to oblige the licensee to perform a certain level
of marketing activities without remuneration.
591
to be the case, as there should be no reason to replace the 1994 multiple

owners rule otherwise. However, the replacement of a rule pertaining to
multiple legal owners of the same intangible with one identifying a single
legal owner for each legally protected facet of a single intangible does not
immediately seem to make much of a difference. The ambiguity of the
preamble makes it difficult to deduct arguments in either direction.
Further, the preamble specifically stated that it intended to replace the

multiple owners rule with a rule that gave effect to the legal ownership of
an intangible. The motivation behind this was that the application of the
notion of ownership under the 1994 regulations as an analytical tool for
allocating residual profits had been criticized.1895 The US Internal Revenue
Service (IRS) resented “black or white” income allocation patterns, as they
typically would not properly reflect the dynamic of the varying underlying
contributions to intangible development from the group entities involved.
The solution was to discard the idea behind the 1968 DA rule that only one
of the contributing entities should be entitled to the residual profits.
A goal of the 2009 approach was therefore to more clearly distinguish be-
tween ownership and transfer pricing rules. The fact that a licensee was not
deemed owner of an intangible that it contributed to the creation of did not
necessarily entail that it would not be entitled to a portion of the residual
profits, as such remuneration could result from an application of the trans-
fer pricing rules. Thus, the 2009 regulations abandoned the perceived “all
or nothing” approach of the 1968 and 1994 regulations.
For the 2009 approach to be meaningful, the first step, pertaining to the
determination of legal ownership, should not be clouded by allocation-mo-
tivated and disaggregated constructions, such as the one indicated by the
discrete owner rule. In fact, the author finds it difficult to reconcile the pur-
pose stated in the 2003 preamble with the discrete owner rule. After all, if
the aim were to distinguish between ownership and transfer pricing rules,
then why use the construction of discrete ownership at all? In the author’s
view, the logical step towards realizing the stated purpose would be to say
that, as the point of departure, the legal owner is allocated the residual
profits, while group entities that contribute to the intangible’s development
are entitled to an arm’s length consideration, which could – but not neces-
sarily would – attract a portion of the residual profits.
1895. 2003 Prop. Treas. Regs. (68 FR 53448-01), preamble, explanation of provisions,
no. 2.
592
The US regulations
Second, the wording of the provision indicates that a licensee is deemed

owner of the licensed rights. In order for the wording to have meaning, one
would assume that the owner is entitled to the operating profits generated
by the licensed rights. The wording of the discrete owner rule is so broad,
however, that the author takes issue with a literal interpretation.
For instance, shall a US trademark licensee be deemed owner of the US

rights even if it does not incur incremental marketing costs?1896 If so, the
licensee would likely be treated more favourably than comparable third-
party licensees that are obliged to pay an arm’s length royalty. This would
deviate from the arm’s length notion that there should be parity in the tax
treatment of controlled and third-party transactions. The author’s impres-
sion is therefore that an application of the discrete owner rule in cases in
which a US licensee incurs only an arm’s length level of marketing expen-
ditures is contrary to the arm’s length principle. Nevertheless, the wording
of the provision seems to also encompass these scenarios.
This touches upon a greater issue. What if the US licensee were not in-
volved in the intangible development at all? For instance, if a US subsidi-
ary licenses a proven patent from a Norwegian research and development
(R&D) performing parent, the generous wording of the discrete owner rule
may lend support to an argument that the subsidiary should be deemed
owner of the rights to exploit the patent in the United States. The next
logical step would be to claim entitlement to the residual profits generated
through the US exploitation.
Such a solution seems disconnected from the prevailing US Internal Rev-

enue Code (IRC) section 482 transfer pricing principles. In spite of the
broad wording, the 2009 examples focus on marketing intangibles, for
which a US licensee does contribute to the value of the trademark by carry-
ing out local marketing. Further, the ownership provisions are fundamental
norms that assign entitlement to profits to group entities because they were
involved in the intangible development and thus took part in the creation of
intangible value. In the author’s example, the US licensee played no part in
the development of the Norwegian patent. It therefore pertains to a pricing
issue, not the determination of ownership.
Third, the 2009 regulations provide two examples on the discrete owner
rule, which are relevant here. The first example pertains to a foreign
1896. See the analysis of the remuneration of intangible development contributions in

this context in sec. 23.4.2.
593
parent that holds legal title to the worldwide rights to a trademark and
is also registered as the legal owner in the United States.1897 It licenses
exclusive US rights to the trademark to a local subsidiary. The example
identifies the parent as the owner of the trademark pursuant to intel-
lectual property law, and the US subsidiary is identified as the owner of
the “license pursuant to the contractual terms of the license, but not the
owner of the trademark”. The example provides no indication as to the
income allocation consequences of deeming the US licensee the owner
of the licence.
The second example pertains to a foreign parent that carries out adver-
tising for trademarked athletic gear in the Olympics,1898 which also en-
hances the value of the licensed US trademark rights. The example obli-
gates the subsidiary to remunerate the parent for the services provided to
increase the value of the US rights, regardless of the fact that the parent
is the legal owner of the trademark as such. It seems reasonably clear
that the example deems the subsidiary the owner of the US trademark
rights. After all, if it were not, there should be no reason to compen-
sate the parent. The example does not precisely state that the subsidiary
shall be allocated the entire residual profits from its US exploitation of
the licensed trademark rights, but this is the logical consequence of the
obligation to remunerate the parent. This does not offer much clarifica-
tion, as it is not indicated as to which transfer pricing method the parent’s
remuneration shall be based on. If the marketing efforts of the parent are
seen as routine contributions, the cost-plus method or the transactional
net margin method (TNMM) will likely be applied to allocate a normal
market return to the parent. Alternatively, if the marketing efforts are
deemed unique contributions, the profit split method (PSM) will likely
be applied, resulting in a split of the US profits from the exploitation of
the US trademark rights.
Thus, both examples are ambiguous with respect to the extent to which the
US subsidiary, as the legal owner of the licence, is entitled to the residual
profits from the US exploitation of the licensed intangible, and therefore
do not lend much support to the interpretation of the discrete owner rule.
Fourth, if one were to interpret the discrete owner rule as allocating the en-
tire residual profits from the US exploitation of a licensed intangible to the
1897. Treas. Regs. § 1.482-4(f)(3)(ii), Example 1.

1898. Treas. Regs. § 1.482-4(f)(3)(ii), Example 6. See also the discussion under sec.
23.4.3.4.
594
The US regulations
licensee, this would, of course, be intrusive on the controlled allocation of

profits. As opposed to the owner of an intangible, a licensee is obligated to
pay royalties for its use of the licensed rights. This basic distinction makes
the construction of a licensee as the legal owner inherently problematic. If
the US licensee were deemed owner of the licensed US rights, one would
logically assume the result to be that the licensee should not pay any roy-
alties. This could be a sensible solution when the US subsidiary incurs
incremental marketing costs, but then based on the view that the subsidiary
should be deemed owner of the US rights pursuant to economic substance,
not legal ownership.1899
For these reasons, the author’s conclusion is that the discrete owner rule
should likely be interpreted as having no bearing on the allocation of re-
sidual profits from a licensed intangible. A US licensee, as the holder of
limited rights to an intangible, may, however, be entitled to some of the re-
sidual profits, but then as a result of the provisions on economic substance
or the transfer pricing methods.1900
It seems to the author that the notion of legal ownership here has been
“stretched” to include an artificial construction of licensees as deemed
owners of licensed rights. Why this has been done, when the 2009 owner-
ship provisions on economic substance and the transfer pricing methods
offer rich possibilities for allocating residual profits to the United States
when the facts and circumstances of a case make such an allocation ap-
propriate, is unclear.
An explanation of the discrete owner rule is likely to be found in the his-

torical development of the ownership provisions. The 1968 DA rule pro-
fessed a largely substance-driven assessment as to which entity was to be
regarded as the owner. In this light, the 1994 legal ownership rule repre-
sented a somewhat formalistic approach, but was supplemented with the
third cheese example, likely in order to at least ensure a substance-based
approach to a specific and practical factual pattern. The distinct owner rule
seems to be a continuation and further development of the multiple owners
rule of the third cheese example.
Compared to the 1994 approach, however, there is a significant differ-

ence in the legal context in which the discrete owner rule is to be applied.
1899. See the discussion in sec. 23.4.3.2.

1900. See the discussion of incremental marketing contributions to IP value in sec.
23.4.3.
595
The 2009 ownership provisions on economic substance offer a far more

substance-driven set of rules than the 1994 regulations. The IRS has sub-
stantial authority in these provisions to allocate income when a US licen-
see contributes to the development or value of a licensed intangible. Thus,
there should be no need or place for the discrete owner rule. The current
wording of the discrete owner rule should, in the author’s view, either be
clarified or taken out.
20.3. The OECD TPG
The legal owner of an intangible is the entity designated as such pursu-

ant to the applicable law or governing contracts.1901 The 2017 intangibles
guidance states that legal ownership by itself does “not confer any right
ultimately to retain returns derived by the multinational group from ex-
ploiting the intangible, even though such returns may initially accrue to
the legal owner as a result of its legal or contractual right to exploit the
intangible”.1902
The amount of profits allocable to the group entity that holds legal title to
an intangible developed intra-group depends on its intangible development
contributions in the form of functions, assets and risks.1903
Further, the current OECD guidance deems a licensee the legal owner of
its licence agreement. This provision seems to be a rather direct import of
the 2009 US regulations’ discrete owner rule.1904 The author sceptical of
the usefulness of viewing a licence agreement as a separate intangible for
profit allocation purposes. He refers to the discussion of the US discrete
owner rule in section 20.2.3.

1903. This is illustrated through an example that strips the legal owner of residual
profits when it does not contribute to intangible value creation; see OECD TPG, annex
to ch. 6, Example 1 (Premiere). The example allocates the residual profits to the parent
alone on the basis of its functions performed, assets used and risks assumed. The sub-
sidiary is allocated an arm’s length consideration for the patent administration services
and nothing more. The guidance contains two extensions of this example; see OECD
TPG, annex to ch. 6, Examples 2 and 3.
1904. Treas. Regs. § 1.482-4(f)(3). See also the discussion of the “discrete owners”
rule in sec. 20.2.3.
596
Ownership and economic substance
20.4. The determination of ownership of intangibles

not subject to legal protection
If the intangible in question is not subject to legal protection, the group
entity that controls it is considered the sole owner under both the current
US regulations and the OECD TPG.1905 This rule may be a practical point
of departure for assessing ownership of intangibles such as know-how and
local goodwill.
The US regulations’ use of control as a criterion for ownership represents a

departure from the 1994 approach, pursuant to which the DA rule was ap-
plied for determining ownership of intangibles not subject to legal protec-
tion, with intangible development costs being the key parameter. Thus, the
last remnant of the 1968 DA rule is now removed. The control provision is
illustrated in an example in the US regulations, in which a US manufactur-
ing and distribution subsidiary developed a valuable customer list.1906 The
example deems the subsidiary the owner of the list, as it “has knowledge
of the contents of the list, and has practical control over its use and dis-
semination”.
20.5. Ownership and economic substance
The fact that a group entity holds legal title to an intangible is, in itself, not
relevant for the determination of whether it should be allocated residual
profits generated by the exploitation of the intangible under the US regula-
tions and the OECD guidance.1907 Profit allocation hinges on substance, not
legal formalities.1908 The residual operating profits from the exploitation of
1905. Treas. Regs. § 1.482-4(f)(3)(i)(A); and OECD TPG, para. 6.40. The OECD TPG
describe control as the ability to make decisions concerning the use and transfer of the
intangible and the practical capacity to restrict others from using the intangible.
1906. See Treas. Regs. § 1.482-4(f)(3)(ii), Example 2.
1907. See also Pankiv (2017), at pp. 82 and 165; and Pankiv (2016), at p. 472. See also
Wilkie (2012); Wilkie (2014a); Lagarden (2014); and Schwarz (2015). Further, the sub-
stance requirements for IP ownership under the OECD TPG must be held completely
separate from the “beneficial owner” substance requirements for royalties under art. 12
of the OECD MTC, as the material content of the provisions differ substantially. On
this point, see Pankiv (2017), at p. 171. See also Monsenego (2014), at p. 17, where a
comparative analysis of the two concepts are carried out.
1908. See also Brauner (2016), at p. 106, with respect to the current OECD IP transfer
pricing provisions. See also Brauner (2015), at p. 78. See, however, seemingly opposing
viewpoints presented in Navarro (2017), at p. 230. In that direction, see also Kane (2014),
at pp. 311 and 314. For a recent, general and comparative overview of the “economic sub-
597
IP allocable to a group entity that purports to own the IP must reflect the
value of the functions, assets and risks contributed by it to the development
of the IP in question.1909
This is highlighted in the new IP ownership guidance of the 2017 OECD

TPG, which states that in order for the legal owner to retain all residu-
al profits generated through the exploitation of an intangible developed
within the multinational, the group entity holding legal title must have
performed and controlled all functions, provided all assets and bore and
controlled all risks related to the development.1910 The logic behind this
approach is that multinationals could easily shift their operating profits
anywhere if the allocation depended on legal ownership. Intangible profits
would then likely not be distributed to the jurisdictions where intangible
value was created.
Because legal ownership in and of itself does not determine the alloca-
tion of residual profits from the exploitation of unique IP under the US
regulations and the OECD TPG, it is not immediately clear to the author
how the relationship between the provisions on legal ownership (as dis-
cussed in sections 20.2.-20.3.) and the substance-based profit allocation
rules (which are the topic of discussion in chapters 21-24) should best be
portrayed.
On the one side, there is no doubt that legal ownership played a more central
role in transfer pricing in the past. It still serves as a “reference point” for
the profit allocation assessment.1911 Seen in this light, the substance-based
allocation rules could be viewed as exceptions from the “legal ownership
provisions”. This characterization would also align with the fact that the
stance” approach to IP ownership, see Rocha (2017), at p. 213. See also Wilkie (2016),
at p. 66, where the legal nature of the economic substance assessment is emphasized.
1909. In this sense, “IP ownership” can be seen as a label on a profit allocation result.
See also, in this direction, Wilkie (2012), at p. 238, where it is stated: “‘[O]wnership’[is
used] to connote what is necessary to explain the outcome of a functional analysis […].”
1910. OECD TPG, para. 6.71. This entails that if more than one group entity contrib-
utes to the creation of IP, the residual profits from the exploitation of that IP shall be
split among these group entities, i.e. there may be more than one “owner” of the IP
for transfer pricing purposes; see, e.g. Storck et al. (2016), at p. 217. See also Schwarz
(2015), where it is argued that the 2017 OECD TPG place too little weight on the legal
ownership of IP. Schoueri (2015), at p. 715 argues that the stripping of residual profits
from a group entity that only funds research and development (R&D) and holds legal
ownership of the developed manufacturing IP under the current OECD TPG is contrary
to the arm’s length principle. For the reasons expressed in the analysis in ch. 22, the
author does not agree.
598
substance-based rules may, but will not necessarily, allocate the profits to
a group entity different from that which holds legal title. Alternatively, the
substance-based allocation rules may be seen as prerequisites for giving
effect to legal ownership for profit allocation purposes, as the legal owner
of IP developed intra-group will only be allocated the residual profits if its
intangible development contributions in the form of functions, assets and
risks justify such an allocation.
Of course, this debate is only relevant for terminological purposes. Re-

gardless of the perspective chosen, the substance-based allocation rules
are decisive.
For the purpose of further discussions, the author chooses to characterize

the substance-based allocation rules as “exceptions” to the legal ownership
provisions. This is in spite of the fact that there really is no such thing as
a clear point of departure under the current US regulations or the OECD
TPG dictating that the legal owner should be allocated the residual profits.
The author does, however, find that his choice of terminology facilitates
subsequent discussions in a straightforward manner.
Moving on, what are these “substance-based allocation rules”? Both the
US regulations and the OECD TPG have traditionally (and still do) distin-
guished rules for determining which group entity should be considered the
owner of an intangible from rules that allocate operating profits for the IP
in question (i.e. transfer pricing methods that price subsequent transfers of
the intangible).1912 Both sets of rules are necessary in order to carry out a
complete profit allocation. While the rules on ownership determine which
group entities should be deemed entitled to the residual profits from the ex-
ploitation of unique IP to which they have contributed the development of,
the transfer pricing rules govern the allocation of operating profits among
the different intangible value chain inputs (and thus also for the unique IP
in question).
For instance, assume that the problem is to allocate a Norwegian distribu-

tion subsidiary’s operating profits of 100 from its local sales of a mobile
phone. A functional analysis reveals that there are two main value chain
inputs: (i) a patent licensed from a UK group entity; and (ii) routine sales
functions performed by the subsidiary. The transfer pricing rules (meth-
ods) determine how the profits of 100 shall be allocated among the value
chain inputs, i.e. the UK patent and the Norwegian sales functions. Assume
1912. The pricing rules are analysed in part 2 of this book.
599
that the TNMM is chosen to remunerate the subsidiary, and that an arm’s
length return is 10. That leaves a residual profit of 90 allocable for the UK
patent. The ownership rules determine which group entities are entitled to
the 90.
Assume that a functional analysis determines that legal ownership is held

by a UK entity, while the patent was developed fully through R&D carried
out by a US entity and financed by an Irish entity. An application of the
2017 OECD TPG on IP ownership determines that (i) the UK entity is en-
titled to 5 as compensation for administrative and legal services connected
to holding legal title; (ii) the Irish entity is entitled to 20 as a risk-adjusted
rate of return on its financing contribution; and (iii) the US R&D entity is
entitled to the residual profits of 65.
Even though the transfer pricing methods and the rules on intangible own-
ership are clearly distinct, serve different purposes and pertain to different
stages of the life of an intangible (namely the exploitation and development
phases, respectively), there is an undeniable convergence between these
two classes of provisions in the current generation of the US regulations
and the OECD TPG.
The ownership provisions are now essentially applications of the trans-

fer pricing methods in the specific context of intangible development. The
overarching idea is that a group entity should be allocated operating profits
generated by internally developed IP commensurately with the value of its
development contributions to that IP. This bears resemblance to the profit
split transfer pricing method,1913 albeit applied in the context of develop-
ment (as opposed to the context of subsequent exploitation, for which the
methodology was originally designed). Thus, there is no longer a signifi-
cant difference between the core material content of these two classes of
rules.
This convergence can be explained as follows. The US provisions on

intangible ownership were revised in connection with the introduction
of new transfer pricing methods for intra-group service transactions in
2009.1914 These new methods were meant to ensure that (i) the pricing
1913. See the analysis of the US and OECD profit split method (PSM) in ch. 9.
1914. The 2009 service methods mirror the general methods. See (i) the comparable
uncontrolled services price method in Treas. Regs. § 1.482-9(c), mirroring the com-
parable uncontrolled price and comparable uncontrolled transaction methods; (ii) the
gross services margin method in Treas. Regs. § 1.482-9(d), mirroring the resale price
method; (iii) the cost-of-services-plus method in Treas. Regs. § 1.482-9(e), mirroring
600
of intra-group services was harmonized with the commensurate-with-in-

come framework of the 1994 regulations for the transfer pricing of tangi-
ble and intangible transactions; and (ii) economically similar transactions
were treated similarly for transfer pricing purposes, regardless of their
formal classification. For example, if the effect of a service transaction
was to transfer an intangible, the profit allocation should not differ from
the allocation that would have been carried out had the transaction been
structured as an intangibles transfer. Pricing methods tailored to intra-
group service transactions had been a long time coming in the United
States.1915
The 2009 intra-group service pricing methods clearly influenced the con-
tent of the revised ownership rules. This is not surprising, as the allocation
of residual profits from unique and valuable IP developed intra-group, on
the one side, and the remuneration of controlled services that contribute to
the IP development, on the other side, are highly interrelated issues that are
not readily distinguishable.
the cost-plus method; (iv) the comparable profits method for services in Treas. Regs.
§ 1.482-9(f), mirroring the general CPM; and (v) the PSM in Treas. Regs. § 1.482-
9(g), mirroring the residual profit split method. The 2009 regulations also introduced
the services cost method in Treas. Regs. § 1.482-9(b), a safe harbour rule for routine
and low-yield “back office” services, pursuant to which the service transaction may be
priced at cost plus, and the US Internal Revenue Service (IRS) may only reassess if the
mark-up equals or exceeds 6%.
1915. Prior to the 1968 regulations (33 Fed. Reg. 5848), the courts applied a dis-
cretionary, facts-and-circumstances-based approach for pricing controlled service
transactions. The author refers to Plumb et al. (1963) and Surrey (1968). The 1968
regulations did not contain specified pricing methods for service transactions (thus
forcing taxpayers and the IRS to apply the methods for tangibles and intangibles
analogically), but set out the rule that reassessments could be made to reflect an
arm’s length charge for marketing, managerial, administrative, technical or other
services performed by one group entity for the benefit of, or on behalf of, another
group entity; see 1968 Treas. Regs. § 1.482-2(b)(1). The charge should be the amount
that would have been charged for similar services in unrelated transactions under
similar circumstances (see 1968 Treas. Regs. § 1.482-2(b)(3)), in practice, deemed
to be equal to the costs incurred by the renderer in performing such services, unless
the renderer established a more appropriate charge for services that were not an
“integral part of the business activity” of either the renderer or the recipient of the
services. The 1994 regulations introduced no new pricing methods for controlled
services. The 1968 transfer pricing rules for services were therefore first replaced
by the 2009 services regulations. For further on the services regulations, see IFA
(2004). See also Kirschenbaum et al. (2003); PG (2003); ACC (2003); API (2003);
B&M (2003); TEI (2003); SIA (2004); TEI (2004a); TEI (2004b); TEI (2004c); Wo-
losoff et al. (2004); ABA (2004); Birnkrant (2004); CIB (2004); Deloitte (2004); EY
(2004); FW (2004); MBRM (2004); MBRM (2005); Stewart (2009); and Murphy
(2010).
601
The influence of the 2009 service regulations is evident in the economic

substance IP ownership provisions,1916 as well as in provisions that govern
concurrent remuneration of group entities that contribute to IP develop-
ment.1917 The message of these provisions is that group entities that render
IP development contributions may be allocated a portion of the residual
profits if necessary to reflect an arm’s length charge.
Further, for a long time, the OECD guidance on IP ownership remained

conservative and sparse.1918 The new 2017 guidance seems heavily inspired
by the 2009 US service regulations in its sweeping demand that group
entities that contribute functions, assets and risks to the development of
an intangible must be remunerated commensurately with the value of their
development inputs, possibly entailing a portion of the residual profits.1919
The substance-based IP ownership rules that allocate residual profits

among group entities, which the author refers to in the following discus-
sions in chapters 21-24 as exceptions to the legal ownership rule, are sys-
temically arranged as follows:
– The US regulations: Two groups of provisions should be seen as one

coherent set of rules for the purpose of allocating residual profits from
the exploitation of unique and valuable IP developed intra-group. The
first group encompasses the comparability provisions on economic
substance.1920 The second group encompasses the provisions on the
remuneration of group entities that contribute to the value of intangi-
bles owned by other group entities.1921 The similarity with respect to
1916. Treas. Regs. § 1.482-4(f)(3). See also Treas. Regs. § 1.482-1(d)(3)(ii)(B), in par-
ticular, the examples in (C).
1917. Treas. Regs. § 1.482-4(f)(4).
1918. As discussed in sec. 19.5., the 1995 OECD TPG only provided modest guidance.
There were no changes from the 1995 OECD TPG to the 2010 OECD TPG with respect
to intra-group allocation of residual profits. In the aftermath of the 2007-2008 financial
crisis, however, the development of the OECD TPG on this point picked up consider-
able speed.
1919. On a more general level, while the OECD’s focus traditionally has been on
avoiding double taxation, the new 2017 guidance on ownership is pronouncedly geared
towards avoiding BEPS, i.e. ensuring that intangible profits are taxed where they are
created.
1920. Treas. Regs. § 1.482-4(f)(3). See also Treas. Regs. § 1.482-1(d)(3)(ii)(B), in par-
ticular, the examples in (C).
1921. Treas. Regs. § 1.482-4(f)(4).
602
the material content of these two groups is underlined by the fact that
certain examples pertain to similar factual patterns.1922
– The OECD TPG: Reference here is made to the section pertaining

to intangible ownership (with the core paragraphs being 6.47-6.68),
describing the significance of IP development contributions from dif-
ferent group entities in the form of functions, assets and risks for the
purpose of allocating residual profits from unique and valuable IP de-
veloped intra-group.
The author’s analysis of these substance-based exceptions is divided into

two main blocks on the basis of the type of intangible developed. He will
discuss the allocation of residual profits generated through the exploitation
of intra-group-developed IP as follows:
– manufacturing IP in chapter 21 and chapter 22 for the US regulations
and the OECD TPG, respectively; and
– marketing IP in chapter 23 and chapter 24 for the US regulations and
the OECD TPG, respectively.
The author finds it necessary to emphasize that the economic substance

exceptions of the US regulations and the OECD TPG in principle apply
equally to manufacturing and marketing intangibles. The reason why the
author has chosen to separate the discussions is that the factual patterns
relevant to the substance rules differ quite significantly, depending on
whether the internally developed intangible in question is a manufactur-
ing or marketing intangible. This is reflected in the fact that both the US
regulations and the OECD TPG, to a large degree, discuss the treatment of
the two types of intangibles separately.
Further, he would like to add that while there are strong similarities be-
tween the approaches taken by the United States and the OECD with re-
spect to the allocation of residual profits from marketing intangibles, their
approaches differ considerably when it comes to internally developed
(R&D-based) manufacturing intangibles.
More specifically, the reader will note that while the US economic sub-
stance provisions are heavily geared towards marketing intangibles and
inbound structures, they are conspicuously sparse with respect to manufac-
1922. These are the infamous “wristwatch” and “athletic gear” examples, which the
author will revert to in depth in chs. 21 and 23.
603
turing intangibles and outbound structures. This stands in contrast to the

IP ownership provisions of the 2017 OECD TPG, which are heavily geared
towards manufacturing intangibles developed intra-group and may be said
to constitute the core of the new anti-BEPS guidance on intangibles.
This is an important structural difference between the US and OECD ap-

proaches to IP ownership. The explanation behind it is not immediately
apparent, as the allocation of residual profits from internally developed
manufacturing intangibles is, of course, highly practical, also in the United
States.
The answer to this riddle likely lies in the fact that internal development
of manufacturing intangibles within a multinational is often – and particu-
larly in the United States (historically at least) – structured through specific
legal vehicles, such as contract R&D arrangements and, most importantly,
cost-sharing arrangements (CSAs). The US substance exceptions (as well
as the transfer pricing methods for services and IP) are relevant for ad-
dressing contract R&D arrangements, but the guidance on this particular
issue is modest.1923 With respect to CSAs, however, the US regulations of-
fer expansive, innovative and aggressive guidance. This takes “pressure”
off of – as well as relevance away from – the substance-based ownership
rules.
Thus, in order to form a representative view of how the US regulations

address profit allocation in the context of internally developed manufac-
turing intangibles, it is crucial that the substance-based exceptions to the
legal ownership rule are seen in light of the CSA regulations. When viewed
in isolation, the substance exceptions provide a misleading picture of the
depth and reach of the US regulations pertaining to the allocation of re-
sidual profits from internally developed manufacturing IP.
In contrast, while the new 2017 OECD guidance on intangible ownership

– in particular with respect to internally developed (R&D-based) manufac-
turing intangibles – is genuinely substance-driven and geared at effectively
putting a stop to the current profit-shifting practices of multinationals, the
new OECD guidance on CSAs, while certainly a step in the right direction,
pales in comparison to the potent and sophisticated US rules. This triggers
the question of what the consequences will be of these differing US and
OECD approaches to IP ownership on this point.
1923. See the discussion in sec. 21.3.7., with further references.
604
A multinational’s internal R&D will often be based on pre-existing intan-

gibles, R&D results and other relevant resources controlled by it (such as
R&D teams in place). The US regulations have therefore taken the posi-
tion that the allocation of profits from manufacturing intangibles developed
through CSAs is mainly a transfer pricing problem, in the sense that as
long as the contribution of the pre-existing intangible (normally developed
in the United States) to the CSA is priced at arm’s length, the “profit shift-
ing” with respect to the income generated through the subsequent exploita-
tion of future (second generation) intangibles developed through the CSA
to foreign group entities should be respected.
While this, on the surface, may seem as a comprehensively different ap-

proach to the OECD’s substance-driven ownership assessment, focusing on
the functions, assets and risks contributed to the intangible development,
it is not.1924 The US pricing of the pre-existing intangibles contributed to
CSAs is highly substance-based. Normally, the pricing will be based on the
income method, which represents an application of the principle of the best
realistic options available.1925
The author would therefore say that the difference between the US and
OECD approaches lies more in the methodologies applied than in the ma-
terial content of the rules (with respect to the required profit allocation
patterns). While it is not possible to gauge the extent to which the two ap-
proaches ultimately may yield differing profit allocation results, there is no
question that the ultimate aim of both paradigms is the same, i.e. to allocate
residual profits to the jurisdiction where the intangible value was created.
The US rules on CSAs were discussed in chapter 14 as a part of the transfer

pricing analysis. CSAs are clearly important for the purpose of allocat-
ing profits from internally developed manufacturing intangibles, and could
therefore just as well have been a component of this “ownership” part of
the book.
This disposition was chosen, however, because the material content of the
CSA provisions is more geared towards the issue of pricing than of owner-
ship. In particular, the main purpose of the US cost-sharing regulations
1924. The author will add that it might be questioned as to whether the OECD ap-
proach to determining ownership of internally developed manufacturing intangibles
places too little weight on traditional pricing assessments. The author will revert to this
issue when discussing the treatment of other unique intangible development contribu-
tions under the current OECD guidance in sec. 22.5.
605
is to provide guidance on the pricing of the contribution of pre-existing

intangibles to the CSA. A necessary consequence of this disposition is
that the reader should bear in mind, when going through the discussion of
the US economic substance exceptions pertaining to internally developed
manufacturing intangibles,1926 that the CSA provisions lurk in the back-
ground.1927
1926. The US economic substance rules that are applied in the context of internally
developed manufacturing intangibles are discussed in ch. 21.
1927. The US CSA regulations are analysed in ch. 14.
606
Chapter 21
The Distribution among Group Entities of Residual

Profits Generated through Exploitation of Internally
Developed Manufacturing IP under the US Regulations
21.1. Introduction
The question in this chapter is how the intangible property (IP) ownership
rules of the current US regulations allocate profits generated through the
exploitation of internally developed manufacturing IP among the group
entities that contributed to the development.
The core aspect of that question is – as always with respect to the de-
termination of IP ownership for transfer pricing purposes – which group
entity shall be deemed entitled to the residual profits generated by the
intra-group-developed IP. As discussed in chapter 20, the point of depar-
ture under the US regulations for determining which group entity should
be entitled to the residual profits generated through the exploitation of IP
developed intra-group is legal ownership. Nevertheless, legal ownership
will not be given effect for transfer pricing purposes if it is not consistent
with the economic substance of the actual behaviour of the controlled par-
ties (the economic substance exception).1928 The US regulations are sparse
and fragmented with respect to the application of the economic substance
exception to intra-group-developed manufacturing IP, in contrast to their
relatively extensive economic substance exception provisions for internally
developed marketing IP (which are analysed in chapter 23).
The author attributes this to two main factors. First, intra-group develop-
ment of manufacturing intangibles has historically often been organized
through cost-sharing arrangements (CSAs), which are now governed by
the comprehensive CSA regulations discussed in chapter 14. CSAs specify
that each participant in the agreement shall be entitled to a portion of the
residual profits that corresponds to the participant’s fraction of the intan-
gible development costs.1929 Thus, the agreements themselves resolve the
profit allocation issue. Second, the transfer pricing methods, applied to a
controlled transaction that is properly delineated to align with the econom-
1928. Treas. Regs. § 1.482-4(f)(3)(i)(A).

1929. Treas. Regs. § 1.482-7(b).
607
Chapter 21 - US Distribution among Group Entities of Residual Profits from
Exploitation of Internally Developed Manufacturing IP
ic substance of the actual conduct of the controlled parties, will ensure that
a group entity that contributes to the development of an intangible receives
an arm’s length compensation.1930
As the author will revert to, the US approach to determining ownership

of internally developed manufacturing IP is different from that of the
OECD.1931 The latter relies on a holistic substance approach, while the US
regulations are focused on some specific profit allocation aspects.
This chapter is divided into three sections. In section 21.2., the author dis-
cusses the value drivers behind manufacturing IP development. The bulk
of the chapter is dedicated to an analysis of the US economic substance ex-
ception applied to manufacturing IP in section 21.3. Lastly, in section 21.4.,
the author discusses the US stance on profit allocation in contract research
and development (R&D) arrangements in light of the new 2015 provisions
on the application of the arm’s length standard and the best-method rule
under US Internal Revenue Code (IRC) section 482, along with other code
provisions.
21.2. The value drivers in manufacturing IP development
Both the US and OECD rules on IP ownership fundamentally seek to al-

locate the profits generated by the exploitation of internally developed in-
tangibles according to the relative values of the underlying development
contributions from the involved group entities. The idea is that intangible
profits should be taxed where the intangible value was created. This focus
is particularly pronounced in the 2017 OECD TPG (discussed in chapter
22),1932 but also underlies the US rules.
In terms of causality, three main development inputs are, in general, as-

sumed to drive intangible value. First, the primary force behind the cre-
1930. Outside of cost-sharing arrangements (CSAs), the relevant methods will, in par-
ticular, be the comparable profits method and profit split method (PSM); see the analy-
sis in ch. 8 and ch. 9, respectively. In the context of CSAs, the relevant methods are,
in particular, the income method and the PSM; see the analysis in secs. 14.2.8.3. and
14.2.8.6., respectively.
1931. See the analysis of the new 2017 OECD Transfer Pricing Guidelines (OECD
TPG) on the ownership of manufacturing intangible property (IP) in ch. 22.
1932. The author further develops his viewpoints on IP development value drivers in
the context of the OECD TPG in sec. 22.2.
608
The value drivers in manufacturing IP development
ation of unique manufacturing intangibles is R&D.1933 This is the work

of clever, specialized and experienced minds, often working in concert.
These R&D functions represent relatively unique IP development inputs
that are not easily replaced. They are also pronouncedly immobile. There
are normally deep ties between a person and his home, and the typical
high-tax jurisdictions where these researchers work tend to be safe socie-
ties with good infrastructure. Thus, a multinational cannot easily motivate
its researchers to relocate to low-tax jurisdictions, even if generous com-
pensation packages are offered.1934
Second, unique manufacturing intangibles tend to be the result of lengthy,

costly and risky R&D processes that are entirely dependent on sufficient
access to risk-willing capital. Only a small fraction of R&D efforts achieve
commercially viable results, while fewer still generate residual profits.
However, projects that are successful may generate such extraordinary
profits that they not only sustain their own R&D costs, but also the R&D
costs of unsuccessful projects, in addition to providing satisfactory returns
on the total capital invested. Capital is highly mobile. Thus, it is easy for
a multinational to place R&D funding in a low-tax jurisdiction for profit
shifting purposes.
Third, the R&D carried out by a multinational often draws upon its histori-
cal research results, typically embodied in pre-existing unique intangibles,
such as when version 1 of a successful piece of software forms the basis for
ongoing R&D of version 2. In the context of IP development, pre-existing
intangibles should likely be viewed as relatively immobile development
inputs, as a transfer from the jurisdiction of origin may trigger a transfer
pricing charge equal to their full value. Due to this “lock-in” effect, a mul-
tinational may be unable to effectively shift the profits associated with the
pre-existing intangible abroad. For instance, if the buy-in amount for the
1933. This discussion focuses on the forces behind the creation of manufacturing IP
specifically. It should be remembered, however, that in order to create a successful
product based on manufacturing IP, significant marketing efforts (and marketing IP)
will normally also be necessary; see, in particular, Roberge (2013), at p. 228.
1934. This picture may, however, to some extent, be evolving. Many otherwise high-
tax jurisdictions are now offering IP regimes (the freedom to offer such regimes going
forward is restricted by the 2015 OECD nexus approach; see the analysis in sec. 2.5.)
or low general tax rates, in combination with local infrastructure normally not found
in typical low-tax jurisdictions. In combination with the current reliance of the OECD
allocation rules on substantial “core value” research and development (R&D) functions
may, over time, lead multinationals to base their “tax planning” on strategic human
resources (hiring polices etc.), in the sense that core R&D departments are located in
jurisdictions that offer favourable IP taxation, in order to ensure that residual profits are
not attracted to high-tax jurisdictions.
609
contribution of a pre-existing intangible to a CSA captures the entire value

of the intangible for future R&D purposes, there should be no tax gain in
migrating it.
21.3. The economic substance exception applied to

manufacturing IP
In sections 21.3.2.-21.3.7., the author will analyse how the US economic

substance exception shall be applied to determine ownership of internally
developed manufacturing IP, i.e. how the exception can be used to ensure
an arm’s length profit allocation to the group entities that contributed to the
development of the IP.
An important aspect of the economic substance exception is that it can

be applied to impute contingent payment terms that are consistent with
the economic substance of the controlled transaction.1935 This authority is
highly practical in the context of intra-group-developed manufacturing IP
– in particular, with respect to contract R&D arrangements – and the profit
allocation effects of applying the provision will therefore be the main fo-
cus in the following sections. Before the author begins that analysis, how-
ever, he will tie some comments to the relationship between the economic
substance exception applied to manufacturing IP and the transfer pricing
methods.
21.3.2. The economic substance exception should not replace

the transfer pricing methods
The economic substance exception can be applied to determine ownership

of intra-group-developed manufacturing IP, i.e. to determine how group
entities that contributed to the IP development shall be remunerated for
their efforts. In cases in which it will be meaningful to apply the economic
substance exception in this manner, there will be more than one group enti-
ty that has contributed to the IP’s development. If, however, the IP has been
developed by only one group entity (that has contributed all functions, as-
sets and risks necessary for the development), the ownership issue will be
1935. Treas. Regs. § 1.482-9(h)(i)(3) and § 1.482-1(d)(3)(ii)(B). See also the comments
on the general economic substance imputation authority in sec. 6.6.5.2.3.
610
The economic substance exception applied to manufacturing IP
clear: this entity is owner. However, if the developer entity then transfers
the completed IP to another group entity, the transfer pricing methods – not
the IP ownership rules – must be applied to ensure arm’s length pricing.
The current US regulations contain an example that touches upon this,

in which the legal ownership of internally developed manufacturing IP is
registered to a group entity that did not contribute to its development.1936
The example pertains to Company X, which performs R&D over a 4-year
period, resulting in a compound. In year 4, the data is made available to
Company Y, which uses it to register patents. Company Y, as the legal
owner of the patents, then enters into licence agreements with other group
entities using royalty rates, which enables it to extract all of the residual
profits from the local exploitation of the patents.1937 The example shows
that the performance of R&D and the transfer of results by Company X is
inconsistent with the registration and subsequent exploitation of the patent
by Company Y. An agreement consistent with the economic substance of
the controlled transaction is therefore imputed, pursuant to which an “ap-
propriate portion of the premium return from the patent rights” is allocated
to Company X.1938
It should be noted that this example applies the economic substance excep-
tion, as opposed to the transfer pricing methods, to remunerate the devel-
oper (Company X) for its IP transfer. The author takes issue with this profit
allocation approach, as well as the allocation result as such.
1936. Treas. Regs. § 1.482-1(d)(3)(ii)(C), Example 6. See also Langbein (2005), at

p. 1086. This example is one of three examples illustrating the application of the eco-
nomic substance exception in the context of intangible ownership, and the only example
that pertains to manufacturing IP. The other two examples pertain to marketing IP (in
cases in which both controlled parties actually played some part in the IP development)
and are analysed in secs. 23.3. and 23.3.3.
1937. In the initial version of the example contained in 2003 Treas. Regs. (68 FR
53448-01) § 1.482-1(d)(3)(ii)(C), Company X performed R&D over a 4-year period,
resulting in a compound. A different group entity, Company Y, was registered as the
legal owner of the patent. In the final version of the example, company X, in year
4, made a large amount of R&D data available to Company Y, which the latter used
to register patents. The 2003 version found it unlikely that Company X would have
engaged in R&D in the absence of contemporaneous compensation or a reasonable
anticipation of receiving a future benefit from the R&D activities. The example im-
puted an agreement under which a portion of the residual profits from the patent was
allocated to Company X, either as a deemed service agreement or as a transfer of the
developed compound. The 2006 version of the example, found in Temp. Treas. Regs.
(71 FR 44466-01) § 1.482-1T(d)(3)(ii)(C), Example 6, was adopted without changes in
the final regulations.
1938. This example seems influenced by the developer-assister rule example discussed
in sec. 19.2.2.
611
Company X fully developed the intangible by itself. The example therefore,

in principle, does not pertain to the allocation of profits to group entities
that contributed to the development of the IP, but rather to the appropriate
transfer pricing treatment when a sole developer transfers a fully developed
intangible to another group entity. Thus, the arm’s length charge for the
transfer of R&D data should be governed by the transfer pricing methods,
not the IP ownership provisions. This will likely necessitate an application
of the cost-plus method or comparable profits method (CPM) to allocate a
normal market return to Company Y. As the IP was developed solely by
Company X, it alone should be allocated the residual profits. The exam-
ple, however, states that Company X should be allocated an “appropriate
portion” of the residual profits. If this is to be taken as an indication that
a profit split is deemed appropriate, the author would take issue with the
allocation result. He does not see any basis in the transfer pricing methods
for allocating residual profits to Company Y, as Company X alone rendered
all intangible development inputs.
It is the author’s view that the example should be revised to make it clear
that the application of the economic substance exception to determine IP
ownership should be reserved for cases in which more than one group en-
tity has contributed to the development of the IP. If one group entity alone
has developed the IP and then later transfers it to another group entity, the
profit allocation should be governed by the transfer pricing methods, not
the IP ownership provisions.
21.3.3. Imputation of contingent payment terms: A lead-in
The transfer pricing methods normally require concurrent compensation in

each income period. This point of departure has been relaxed in order to
make room for controlled pricing structures akin to those commonly ap-
plied among unrelated enterprises in business sectors in which significant
up-front high-risk investments are required, often long before the poten-
tial super-profit rewards materialize (e.g. in the pharmaceutical, extractive
and software sectors). The current US regulations recognize the use of
controlled contingent payment terms, as long as the terms are consistent
with the economic substance of the controlled transaction.1939 When the
1939. Treas. Regs. § 1.482-9(h)(i)(2)(ii). The regulations contain a comparability pro-

vision dealing with market penetration strategies; see Treas. Regs. § 1.482-1(d)(4)(1);
and the comments in sec. 6.6.5.4.4. of this book. Temporarily increased expenses or
lower prices than charged in comparable uncontrolled transactions (CUTs) in the same
market will not lead to a reassessment, provided that an uncontrolled taxpayer would
612
acceptance of such terms was proposed in 2003,1940 it was deemed to pro-

vide sought-after flexibility for the pricing of high-risk R&D that stretches
across multiple income years.1941
The recipient of a service will not pay the service provider on a concur-
rent basis in contingent payment arrangements. Payment is only due if and
when the specified contingency occurs. The US regulations only accept
controlled contingent payment terms when the payment basis reflects the
recipient’s benefit from the service and the risks borne by the renderer.1942
The author interprets this to mean that non-payment on a concurrent basis
for services rendered in the context of IP development will normally only
be acceptable under the US regulations if the service provider is allocated
a portion of the subsequent residual profits.1943
The flip side of the recognition of taxpayer contingent payment terms is that
the US Internal Revenue Service (IRS) may impute such terms if necessary
to reflect the economic substance of the controlled transaction.1944 This is
particularly relevant when a US group entity incurs intangible development
risks through R&D (or incremental marketing efforts) without receiving
have engaged in a comparable strategy under the same circumstances. The market pen-
etration provision can be seen as a parallel to the recognition of contingent payment
arrangements, in the sense that both rules tolerate the absence of concurrent compensa-
tion in anticipation of future benefits. Both sets of rules may be particularly relevant in
the context of intra-group-developed marketing IP, for which significant expenses must
be incurred in order to establish a trademark in a new market.
1940. The 2003 proposed regulations (68 FR 53448-01) set, as a condition for recog-
nizing controlled contingent payment terms, that “it is reasonable to conclude that such
payment would be made by uncontrolled taxpayers that engaged in similar transactions
under similar circumstances”, at § 1.482-9(i)(1), indicating that a CUT that demon-
strated third-party use of a similar contingent payment term was necessary. This led
to questions from commentators, resulting in an omission of the language in the 2006
temporary regulations (71 FR 44466-01). It is therefore clear that there is no require-
ment for a CUT in order to trigger recognition of a controlled contingent payment struc-
ture. Additional requirements for the recognition of contingent payment arrangements
are set out in the regulations (including written contract and specified contingency); see
Treas. Regs. § 1.482-9(h)(i)(2)(i).
1941. See Terr (2004), at p. 565.
1942. Treas. Regs. § 1.482-9(h)(i)(2)(i)(C).
1943. See the analysis of the examples illustrating the imputation authority in sec.
21.3.5.
1944. Treas. Regs. § 1.482-9(h)(i)(3). See also the discussion in sec. 21.3.5. Terr
(2004), at p. 566 finds that the power of the US Internal Revenue Service to impute
contractual terms may arguably exist independently of the US Internal Revenue Code
(IRC) sec. 482 regulations to the extent authorized by common law doctrines, such as
substance over form and economic substance.
613
concurrent compensation.1945 In these cases, there will be an asymmetry

between the burdens and benefits associated with the intangible develop-
ment. The US subsidiary incurs costs and risks, while the upside in the
form of potential subsequent residual profits are allocated to the foreign
entity. The imputation provision will aim to correct this.
In sections 21.3.4.-21.3.6., the author will analyse the examples drawn up

by the US regulations on the contingent payment terms imputation provi-
sion before concluding remarks are made in section 21.3.7.
21.3.4. Imputation of contingent payment terms base example:

Successful contract R&D arrangement with contingent
profit split payment structure
The base example pertains to a contract R&D arrangement, under which

ompany X shall perform R&D for Company Y.1946 Any patents or other
rights resulting from the R&D will be owned by Company Y alone. Fur-
ther, Company X will receive payment only if the research leads to com-
mercial sales of derivative products. If so, Company Y is obliged to pay
Company X a percentage of its gross worldwide sales of the derivative
products for a period of 7 years. At the time at which the agreement is
entered into, the possibility that the R&D will result in new products is
highly uncertain. Further, the potential markets for any new products are
unknown and cannot be reliably estimated. R&D is carried out by Com-
pany X in years 1-4, resulting in a compound. Company Y registers patents
for the compound in several jurisdictions in year 4. Products based on the
compound are sold by Company Y in year 6 and continue through year 9.
As agreed, Company Y pays a percentage of its gross sales to Company X.
The first question pertains to the recognition of the contingent payment

terms. The actual behaviour of the controlled parties is found to be consist-
ent with the contractual risk allocation.1947 Further, Company X had the
financial capacity to bear the risk that the R&D could be unsuccessful, and
that it therefore would not receive any compensation. Also, Company X
exercised managerial and operational control over its own R&D activities,
which made a contractual allocation of the R&D risks to Company X rea-
1945. See the analysis of the economic substance imputation authority applied to mar-
keting IP in sec. 23.3.
1946. Treas. Regs. § 1.482-9(h)(i)(5), Example 1.
1947. See the analysis of risk as a comparability factor in sec. 6.6.5.5.
614
sonable. On this basis, the example deems the contingent payment terms to
be consistent with the economic substance of the arrangement.
The second question is whether the remuneration paid by Company Y in

years 6-9 was at arm’s length. The example takes into account that Compa-
ny X bore the risk that it possibly would not receive any compensation for
its services. Further, the realistic alternative available to Company X was
to undertake the R&D activities on its own behalf and license out the rights
to exploit any successfully developed intangibles.1948 The relevance of this
alternative is accentuated by the demonstrated willingness of Company X
to incur, and its ability to bear, the R&D expenses and risks. On this basis,
the example finds that the IRS is entitled to consider royalties charged for
comparable intangibles.
The example is ambiguous on the further specifics of the profit alloca-

tion. Given the focus on the realistic alternatives of the US entity, however,
the author deems it likely that the IRS, in this situation, would demand a
contingent payment that effectively would put the US entity in a position
equal to that which would have been the result if it had performed R&D on
its own behalf and subsequently licensed out its own intangible. In other
words, the lion’s share (or all) of the residual profits generated by the patent
would likely be allocated to the US subsidiary (analogous to an application
of the CPM, and the income method in the context of CSAs). Company Y
will then only receive a normal market return for its contributions to the
value chain, if any.1949

Unsuccessful contract R&D arrangement with
contingent profit split payment structure
The factual pattern of the second example is the same as in the base ex-
ample in section 21.3.4., apart from the twist that the R&D proves un-
1948. The “realistic alternatives available” concept is central in current transfer pricing
methodology. See the comments in the context of unspecified methods in sec. 12.2., the
new 2017 OECD guidance on valuation in sec. 13.5. and the income method for pricing
buy-ins in CSAs under the US cost-sharing regulations in sec. 14.2.8.3., with further
references. The “realistic alternatives” valuation principle was codified in the last sen-
tence of IRC sec. 482 in the December 2017 tax reform; see the discussion in sec. 1.7.
1949. The 2017 OECD TPG take essentially the same stance, as they finds that “reim-
bursement of costs plus a modest mark-up” will not sufficiently remunerate the R&D
entity under a controlled contract R&D agreement; see OECD TPG, para. 6.79.
615
successful.1950 Pursuant to the controlled agreement, Company Y did not

provide Company X with any compensation for the year 1-4 R&D efforts.
The example deems the contingent payment term to be consistent with the
economic substance of the actual conduct of the parties, and the result-
ing non-payment in the examined years is deemed arm’s length. In other
words, it is accepted that the US subsidiary bears all R&D costs and risks
without being allocated any profits. The regulations ensure symmetry. If
successful, the US entity enjoys the residual profits. If unsuccessful, it may
deduct its R&D expenses without being allocated any income.

Successful contract R&D arrangement with cost-plus-
based contingent payment structure
The facts in this last example are the same as in the base example in section
21.3.4., apart from the twist that the contingent payment arrangement does
not provide Company X with a percentage of the worldwide revenues of the
derivative product, but instead reimburses its R&D costs with the addition
of a normal return mark-up.1951 In year 6, Company Y makes the single
payment to Company X as required under the contract.
The cost-plus reimbursement is not deemed to reflect the recipient’s benefit

and the renderer’s risk. It is therefore not consistent with the economic
substance of the transaction. This triggers the IRS’s authority to impute
contingent payment terms. It is found that one possible basis of payment
that will reflect the recipient’s benefits and the renderer’s risks is a charge
equal to a percentage of the commercial sales of the derivative products.
The example takes into account the alternatives realistically available to

the parties for the purpose of determining an appropriate royalty rate.
As in the base example, the realistic alternative available to Company X
was to undertake the R&D activities on its own behalf and license out
the rights to exploit any successfully developed intangibles. As in section
21.3.4., the author deems it likely that the IRS, in similar factual patterns,
would impute contingent payment terms that allocated the residual prof-
1950. Treas. Regs. § 1.482-9(h)(i)(5), Example 2. This example was introduced in the
2006 temporary regulations (71 FR 44466-01), in response to requests from commen-
tators for guidance on the scenario in which the R&D efforts are unsuccessful.
1951. Treas. Regs. § 1.482-9(h)(i)(5), Example 3. The mark-up is within the range of
mark-ups on costs of independent contract researchers that were compensated under
concurrent remuneration terms.
616
its to the US subsidiary, analogous to an application of the CPM and the

income method.1952
21.3.7. Concluding remarks on the application of

the economic substance exception to impute
contingent payment terms
The US guidance on contingent payment terms should be seen as a clear

rejection of contingent cost-plus-based payment structures. In essence, this
is the same message as conveyed by the athletic gear example for market-
ing IP (discussed in section 23.3.3.), which rejects a solution in which the
US subsidiary is reimbursed incremental marketing costs with a mark-up
after the marketing has proved to yield residual profits.
On the flip side, however, the examples discussed in sections 21.3.4.-21.3.6.

provide, rather surprisingly, support for the argument that concurrent cost-
plus-based compensation of a US R&D entity that contributes to the devel-
opment of IP is sufficient, regardless of the extent of the contribution. The
R&D entity will then not be entitled to any of the residual profits. It will
produce a steady stream of net income equal to the cost plus mark-up on
the R&D costs incurred throughout the development phase, entailing that
there are no effective tax deductions for the development costs in the Unit-
ed States. In this sense, the United States has no stake in the subsequent
residual profits, as no risk has been assumed, from a fiscal point of view.
The US contingent payment term imputation authority presents multina-

tionals with a dilemma: if concurrent cost-plus remuneration is provided,
the multinational will not be able to receive effective US tax deductions for
the R&D expenses, as the US entity is reimbursed for its expenses with an
arm’s length mark-up. The entity will therefore report a normal return prof-
it throughout the R&D project, and the subsequent residual profits from the
exploitation of the developed intangible are allocated to the foreign owner-
ship entity. If, however, concurrent cost-plus remuneration is not provided,
the multinational will be able to effectively deduct the R&D expenses in
the United States, but is faced with the risk that the entire residual profits
from subsequent exploitation of the developed intangible will be taxed in
the United States.
1952. See also the comments on the athletic gear example in sec. 23.3.3., pertaining to
the determination of ownership of co-developed marketing intangibles, pursuant to the
economic substance exception.
617
On this basis, the author assumes that when the R&D risk is relatively low,
combined with a relatively high likelihood that the R&D results will gen-
erate residual profits, a multinational will opt to remunerate the US R&D
service provider on a concurrent cost-plus basis. One could argue that the
United States, in this situation, is allocated a level of income that does not
sufficiently reflect the value of the intangible development contributions
from the US entity. In essence, the cost-plus allocation will depart from
the paradigm that intangible profits should be allocated to the jurisdiction
where the value was created.
Nevertheless, R&D projects with such risk profiles tend to be based on

pre-existing intangibles connected to commercially established products.
The contribution of such pre-existing intangibles to the R&D process (in
practice, typically through CSAs) should be taxed pursuant to the pricing
methods, as the author will revert to in section 21.4.1953 This will ensure
that intangible value is taxed where it was created, in spite of the cost-plus-
based remuneration of the ongoing “second-generation” R&D efforts.
If the R&D is truly “blue sky” research in the sense that it does not build
upon any pre-existing intangibles or established products, concurrent cost-
plus remuneration of a US R&D provider will ensure that only a normal
return is allocated to the United States, while the residual profits, if any, are
allocated to the foreign funding entity. In these cases, at least, intangible
value is not allocated to the jurisdiction where it is created. The author is
somewhat surprised that the United States accepts this. Perhaps this should
be seen as an indication that such cases are not very practical and that con-
tract R&D arrangements in which the US entity is the R&D provider are
normally based on pre-existing intangibles and organized through CSAs.
Further, in cases in which the R&D risk is relatively high, combined with
significant uncertainty as to whether the R&D will provide successful re-
sults and generate residual profits, the author assumes that it will be more
relevant for a multinational to not remunerate the US R&D service pro-
vider on a concurrent basis in order to effectively deduct the R&D costs in
the United States. Again, such R&D will normally be organized through
CSAs. When it is not, however, the imputation authority for contingent pay-
ment arrangements is relevant. The potential subsequent profit allocation to
the US entity will be the same as if it were deemed to own the developed
intangible. It will attract the entire residual profits. Thus, non-concurrent
1953. The realistic alternatives available to the controlled parties will underlie the ap-
plication of the transfer pricing methods; see supra n. 1948.
618
The US stance on contract R&D arrangements in light of the 2015
provisions on the arm’s length standard and best-method rule
payment of a US R&D group entity is a risky tax-planning strategy, seen

from the point of view of a multinational.
At the same time, the United States will have a significant stake in the
R&D development in these cases, due to concurrent tax deductions for
R&D expenses. Any income resulting from the research should then be
allocated to the United States in order to ensure symmetry. Also, it would
simply not make sense for the US entity, on the basis of its realistic alter-
natives, to contribute unique inputs to the intangible’s development and in
return receive no concurrent remuneration or future residual profits. The
rule, as it stands now, is therefore, in the author’s view, logical and must be
assumed to correspond to what third parties would have agreed in similar
circumstances, i.e. that significant risks should be followed by an opportu-
nity to earn significant profits.
21.4. The US stance on contract R&D arrangements

in light of the 2015 provisions on the arm’s length
standard and best-method rule
The result expressed in section 21.3.7. that concurrent cost-plus-based com-
pensation of a US R&D entity that has contributed to the development of
an intangible is seemingly sufficient, regardless of the extent of the IP de-
velopment contribution, is unexpected. It also marks a pronounced devia-
tion from the approach of the current OECD Transfer Pricing Guidelines
(OECD TPG), which allocate the residual profits to the group entity that
performs the important R&D functions.1954 From this, it could possibly be
inferred that the migration of intangible ownership pursuant to popular
contract R&D schemes in which the R&D provider, located in a high-
tax jurisdiction (here, the United States), is remunerated on a concurrent
cost-plus basis, while a foreign IP holding or funding entity, located in a
beneficial tax environment with respect to intangible income, is assigned
ownership to any intangibles resulting from the R&D, and thus the residual
profits, is acceptable under US law, while it is not under the OECD TPG.
The US economic substance exception provisions (including the contin-

gent payment terms imputation authority) must, however, be read in light
of the transfer pricing methods. Contract R&D projects will normally
build upon pre-existing intangibles, know-how and the experience of the
1954. See the analysis of the current OECD approach for determining the ownership of
manufacturing IP in ch. 22.
619
research team providing the R&D services. In these cases, a proper arm’s
length charge for the contribution of the pre-existing unique inputs to the
intangible development process by the R&D provider will ensure that the
intangible value is allocated to the jurisdiction in which it was created. If
the R&D arrangement qualifies as a CSA, the income method will be the
go-to method for pricing the pre-existing intangibles.1955
Alternatively, if the arrangement does not qualify as a CSA, the pricing

will likely be based on an unspecified method, in which the alternatives
realistically available to the controlled parties are the key factors for deter-
mining the arm’s length charge, along with an analogical application of the
income method.1956 Thus, where contingent payment terms are imputed, it
is likely that the minimum amount of royalties allocated to the US entity
will be equal to the amount of operating profits that it would have received
under its best realistically available alternative. In most cases, this will be
carrying out the R&D on its own behalf, as the owner, and then licens-
ing out the fully developed intangibles to other group entities at a royalty
rate that extracts the residual profits from local entities’ exploitation of the
completed intangible. In other words, a multinational’s risk connected to
not remunerating a US R&D service provider on a concurrent basis will be
significant.
New provisions on the application of the arm’s length standard and the
best-method rule under section 482 with other tax code provisions were
introduced into the regulations in September 2015.1957 These are relevant
to the tax treatment of controlled contract R&D arrangements and require
that the entire controlled arrangement must be considered in order to en-
sure arm’s length compensation of all intangible value transferred intra-
group.1958
The new provisions include an example that pertains to US Company P,

which has developed product X through R&D carried out by its own R&D
team. Company P enters into a contract R&D arrangement with a foreign
subsidiary, under which it shall provide R&D services to create a new line
of products that build upon the product X platform.1959 Any IP developed

1956. See supra n. 1948.
1957. T.D. 9738. See the comments in the context of (i) the relationship between IRC
secs. 367 and 482 in sec. 3.5.8.; (ii) the best-method rule in sec. 6.4.; and (iii) aggre-
gated valuation in sec. 6.7.2.
1958. Treas. Regs. § 1.482-1T(a)(i)(A).
1959. Treas. Regs. § 1.482-1T(f)(2)(i)(E), Example 10.
620
under the arrangement is to be owned by the foreign subsidiary. Company

P has not transferred the product X platform or its R&D team, but is pro-
viding value associated with the use of the platform and know-how through
its rendering of the R&D services. The example requires that this value
must be compensated under IRC section 482 through an aggregated valu-
ation that reflects the interrelated value of the services and the embedded
value of the platform and know-how.1960
The example is ambiguous with respect to the profit allocation conse-

quences of the required arm’s length charge. The author nevertheless finds
that it likely requires a significant portion of the residual profits from the
IP developed under the contract R&D agreement to be allocated to the US
R&D entity. The end allocation result should thus be similar to that yielded
by the income method in the context of CSA buy-ins. Given the ambiguity
of the new 2015 provisions, there will likely be diverging allocation asser-
tions from multinationals and the IRS.
A noteworthy aspect of the example is that the US R&D entity was reim-
bursed for its R&D costs (on a concurrent basis) by the foreign subsidiary.
In other words, the US entity did not incur any of the financial risks associ-
ated with the R&D, and the United States had no stake in the development,
as there were no concurrent US tax deductions for the R&D expenses. Even
still, an arm’s length charge for the interrelated value of the services and
the platform and know-how is imposed.
As there are clear similarities between a simple reimbursement of R&D

costs without any mark-up and a cost-plus remuneration for R&D services,
the new 2015 provisions introduce significant uncertainty as to whether
concurrent cost-plus remuneration for R&D expenses throughout the de-
velopment phase still should be considered a possible safe harbour from the
US entity being allocated any residual profits generated through the sub-
sequent exploitation of intangibles developed through the contract R&D
agreement.
1960. The example also contains an extension, in which the facts are otherwise the
same, but Company P assigns all or a portion of the R&D team and the rights to the
product X platform to the foreign subsidiary. The taxpayer assertion is that the trans-
ferred platform rights must be compensated, but that the transferred know-how is not
compensable, as it is purportedly not encompassed by the intangibles definition in IRC
sec. 936(h)(3)(B). The example requires compensation for all value transferred to the
foreign subsidiary, apparently including the value of the know-how, based on an im-
puted agreement.
621
The author’s impression is that there might be a conflict between the al-
location results dictated by the economic substance provisions on imputed
contingent payment terms and the new 2015 provisions. It may be argued
that the former provisions are clearer than the latter, and that the possible
concurrent cost-plus safe harbour therefore should prevail until the eco-
nomic substance provisions are altered to reflect otherwise. This would,
however, in the author’s view, contradict the fundamental requirement un-
der the arm’s length principle that there should be parity in the taxation of
related and unrelated enterprises. There cannot be any serious doubt that
third parties must pay full value for unique development inputs such as
know-how, pre-existing intangibles, goodwill and workforce in place. In
order to realize this fundamental principle, the new 2015 clarification that
there should be an arm’s length charge for all intangible value provided in
controlled transactions must be decisive.
Thus, in cases in which the contract R&D builds upon pre-existing re-
sources, the arm’s length charge should, as indicated above, be based on
an aggregated valuation approach, akin to an analogical application of the
income method. This will ensure that the residual profits are allocated to
the R&D entity.
When the contract R&D is pure “blue sky” research that does not build
upon any specific pre-existing resources, the R&D will still, in itself, be
the primary value driver behind the creation of unique intangibles.1961 The
arm’s length charge for the R&D services should reflect the best realistic
alternatives available to the R&D entity. This will normally entail that the
residual profits should be allocated to the R&D entity, based on the posi-
tion that it alternatively could have developed the IP itself outside of the
contract R&D arrangement and reaped the entire residual profits through
licensing out the fully developed intangible to other group entities.
Further, regardless of whether the R&D is based on pre-existing resources

or is pure “blue sky” research, the foreign group entity that funds the R&D
should, in principle, be allocated an arm’s length return. There is currently,
however, no legal basis in the US regulations to impose such an allocation
for funding contributions. The funding entity is simply cut off from profit
allocation. Also on this point, the US regulations provide a markedly dif-
ferent outcome than the current OECD TPG, which afford a group entity
1961. See the discussion of IP value drivers in sec. 21.2.
622
that funds R&D efforts a risk-adjusted rate of return on the funding con-
tribution.1962
In conclusion, a US R&D entity should, under the current regulations, be

allocated the entire residual profits from the exploitation of IP developed
through controlled contract R&D arrangements. The foreign funding en-
tity, however, will not be entitled to any remuneration for its R&D funding
contribution.
623
Chapter 22

Developed Manufacturing IP under the OECD TPG
22.1. Introduction
The question in this chapter is how the intangible property (IP) ownership
rules of the current OECD Transfer Pricing Guidelines (OECD TPG) al-
locate profits generated through the exploitation of an internally developed
manufacturing IP among the group entities that contributed to the devel-
opment.1963
The OECD rules governing this problem have been significantly revised in
the 2017 OECD TPG. The predecessor rules contained in the 2010 OECD
TPG were underdeveloped and allowed controlled pricing structures that
resulted in BEPS. Tightening those rules was a focus point in the latest
revision of the OECD TPG. The core aim of the revised rules is that the
residual profits generated by unique and valuable manufacturing IP shall
be allocated to the jurisdictions in which the real intangible value creation
took place.1964 As will be shown in this chapter, the new OECD rules make
up a relatively complex profit allocation system that divides the IP profits
among the group entities that contributed to the development of the IP. The
key IP development contributions that must be remunerated are functions,
R&D financing (funding) and pre-existing IP.
1963. With respect to terminology, it should be noted that the new 2017 OECD own-
ership guidance applies the comprehensive wording of “development, enhancement,
maintenance, protection and exploitation” (the so-called “DEMPE functions”) to de-
scribe intangible development contributions; see OECD TPG, para. 6.50. The terminol-
ogy seems unnecessarily broad. For instance, the word “development” overlaps with
“enhancement”, and “maintenance” overlaps with “protection”. For the sake of simplic-
ity, and because the guidance is geared towards the research and development (R&D)
of manufacturing intangibles, the author will only use the word “development”. Unless
otherwise stated, this also encompasses the enhancement, maintenance and protection
of intangibles.
1964. As illustrated by the critical comments made in Brauner (2014a), at p. 99, it will
likely not always be straightforward to determine in which jurisdictions the significant
intangible property (IP) value creation in fact takes place.
625
Chapter 22 - OECD Distribution among Group Entities of Residual Profits
from Exploitation of Internally Developed Manufacturing IP
The author discusses the allocation of profits for functions in section

22.3., funding in section 22.4. and pre-existing IP in section 22.5. First,
however, he provides a lead-in to the profit allocation issue at hand in sec-
tion 22.2.

internally developed manufacturing IP
For a general discussion of the value drivers behind the creation of manu-
facturing IP, the author refers to section 21.2., which forms the background
for the following comments.
The goal of the OECD TPG on intangible ownership is to ensure that re-
sidual profits are allocated to the jurisdiction(s) in which intangible value
was created. The rules thus fundamentally seek to allocate profits based on
causality. As research and development (R&D), capital and pre-existing
intangibles are the main forces behind intangible value creation,1965 it may
logically be inferred that the residual profits should be split among the ju-
risdictions in which the multinational:
– carried out ongoing R&D (often high-tax jurisdictions);
– drew R&D funding from (often low-tax jurisdictions); and
– created unique pre-existing intangibles (often high-tax jurisdictions).
As will be analysed extensively in the rest of this chapter, however, the new
2017 ownership guidance does not accept an outright split of the profits
among the IP development contributions in this manner. Even though there
is unquestionable causality between funding and intangible value creation,
this development input is not afforded residual profits, but a separately de-
termined return. This stands in contrast to the contributions of both ongo-
ing R&D and pre-existing intangibles, which are allocated residual profits.
The reason for this discrimination is likely twofold.
First, capital is a more generic development input than “high-level”

R&D.1966 A genuinely disruptive technology in early-phase R&D with

1966. See, however, Musselli et al. (2017), at p. 337, where it is argued that the OECD,
through the 2017 OECD TPG, places too much weight on functions (and too little on
capital) for profit allocation purposes, with the result that profit allocation will not be
aligned with the underlying value creation. The author does not agree (for the reasons
stated in his analysis in secs. 22.3.2.-22.3.3.).
626
A lead-in to the profit allocation problem for internally developed
manufacturing IP
blockbuster potential will likely not struggle to raise funding, but it will
be difficult to find such investment opportunities, even if the capital is
available.1967 Second, if IP development contributions in the form of
funding were to attract residual profits, it would essentially be up to the
multinational to choose whether or not to tax those profits, as it is free
to locate its capital in any jurisdiction. The balancing of these factors
with the acknowledgement that there indeed is causality between funding
contributions and intangible value creation has proven severely difficult
for the OECD, as the author will revert to in his discussion on R&D
funding.1968
Further, on a more concrete level, the new 2017 OECD ownership guid-
ance may be seen as a response to controlled transactions designed by
multinationals to shift intangible profits away from high-tax “R&D juris-
dictions” to low-tax “funding jurisdictions”.1969 In this respect, two main
culprits have emerged: (i) contract R&D arrangements that shift residual
profits from ongoing R&D; and (ii) cost-sharing arrangements (CSAs)
that shift profits from pre-existing intangibles as well as ongoing R&D.
The hallmark of contract R&D arrangements is that the group entity that
contributes R&D efforts is remunerated with a normal market return
for its ongoing R&D, normally on a cost-plus basis, while the residual
profits are assigned to the group funding entity (which pays the cost-plus
consideration). CSAs work differently, but yield similar results. For in-
stance, say that valuable and unique pre-existing IP (e.g. version 1 of a
piece of software) created in a high-tax R&D jurisdiction is contributed
to a CSA for the purpose of further development. While version 1 was
1967. See also Brauner (2010), at p. 17 with respect to the buy-in issue under the US
cost-sharing arrangement (CSA) regulations.
1968. See sec. 22.4.
1969. See, e.g. Navarro (2017), at p. 230. See also Musselli et al. (2017), at p. 331,
where it is argued that the new OECD approach to IP ownership (which focuses on im-
portant functions) places too much weight on labour relative to capital, and thus favours
highly industrialized (and high-tax) jurisdictions with large pools of educated work-
forces over developing (and low-tax) jurisdictions with respect to tax base allocation. In
this direction, see also Rocha (2017), at p. 243. See also Milewska (2017), in particular
at p. 54, for an interesting perspective from the Mexican side, where it is noted that em-
ployees in developing countries (while often numerous) generally perform only routine
functions and follow strategic decisions made by the (foreign) parent company. The
new OECD rules will indeed allocate the residual profits from intra-group-developed
manufacturing IP only to the jurisdictions in which important development functions
were carried out. However, it should, in the author’s view, also be taken into account
that such functions may very well also be carried out in developing (and low-tax) juris-
dictions. See also Fichtner et al. (2016), at pp. 27-28; and the discussions in secs. 11.2.
and 26.6.
627
100% owned by the R&D entity resident in the high-tax jurisdiction,

version 2 will be mainly owned by a foreign group entity that provides
the bulk of the R&D funding. Thus, through the CSA, most of the re-
sidual profits from version 2 (and subsequent versions) of the IP have
been shifted from the high-tax R&D jurisdiction to the low-tax funding
jurisdiction.
Thus, the implicit assertions underlying the profit allocation results of

such controlled R&D arrangements and CSAs is that intangible value is
mainly created through funding, not R&D. This conflicts with the position
taken in both the current US and OECD transfer pricing regimes. While
both the United States and OECD end up with the same profit allocation
result, the regimes have selected somewhat different methodological ap-
proaches to attain this result. The OECD approach is to deal with contract
R&D agreements through substance requirements, while the new 2015 US
regulations apply a pricing approach to address the same problem.1970 With
respect to CSAs, both the US and the OECD regimes apply a “realistic
alternatives”-based pricing approach to value contributions of pre-exist-
ing intangibles to IP development (so-called “buy-in pricing”), although
the US methodology is more developed and sophisticated than its OECD
counterparty.

Functions
The question in this section is of how operating profits derived from the
exploitation of internally developed manufacturing IP shall be allocated
among the group entities that contributed to the development of the IP the
most immobile and important development inputs of all: R&D. It is no ex-
aggeration to claim that this issue is one of the most problematic in today’s
international transfer pricing. It is also one of the most significant issues,
as it directly pertains to which jurisdiction may claim entitlement to tax
residual profits from unique and valuable IP.
The discussion is organized in two parts. The author will first tie some
comments to the new OECD guidance on functions in section 22.3.2.
628
Profit allocation for IP development contributions: Functions
Then, he will discuss how this guidance can be applied to controlled con-
tract R&D arrangements in section 22.3.3.
22.3.2. The “important functions” doctrine
The new OECD guidance on IP ownership requires that all group entities
that contribute certain key functions to the development of IP are compen-
sated with an arm’s length amount of operating profits.1971 The profit allo-
cation to such entities should reflect the relative value of the functions they
contributed to the IP development.1972 The guidance singles out a group
of functions that “usually make a significant contribution to intangible
value”, including control over R&D strategy, design, budgets, protection
and quality control.1973 These strategic R&D control decisions should com-
mand relatively more profits than other (and more “generic”) functional IP
development contributions. The author will refer to this as the “important
functions” doctrine. This point is also reflected in the revised intra-group
services chapter of the OECD TPG, which makes it clear that R&D ser-
vices cannot be regarded as low-value services.1974
As there no comparable uncontrolled transactions (CUTs) will exist upon

which to benchmark the allocation of profits to these important R&D
functions, the use of profit splits may become necessary.1975 The new guid-
ance specifically states that a group entity that performs important R&D
functions should not be remunerated as the tested party under a one-sided
method.1976 As the one-sided methods only afford the tested party a normal
market return, it must be concluded that the new guidance requires residual
1971. OECD TPG, para. 6.50. See Andrus et al. (2017), at p. 92 on the DEMPE func-
tions. See also Navarro (2017), at pp. 232-233, where critical remarks on the OECD
DEMPE approach are presented, and the conclusion is drawn that the OECD approach
is “not in line with the arm’s length principle”. The author strongly disagrees with this
reasoning (which seems to elevate the relevance of legal ownership in transfer pricing
to unrealistic heights).
1973. OECD TPG, para. 6.56. See the discussion of the DEMPE concept in Monsenego
(2014), at p. 17 (based on the 2013 OECD intangibles discussion draft). For critical
comments, see Musselli et al. (2017), at p. 340, where the argument is that the deter-
mination of whether a function qualifies as a DEMPE function will entail subjective
assessments, possibly resulting in disagreements among taxpayers and jurisdictions. In
this direction, see also Heggmair (2017), at p. 265.
629
profits to be allocated to group entities contributing important functions to

the development of the IP.
Thus, the important-functions doctrine can, in essence, be regarded as a

peremptory profit split requirement.1977 The doctrine instructs that some
specific development inputs, shall, in all cases, be deemed the main drivers
of IP value creation, and thus capable of attracting residual profits. His-
torically, it is unusual for the OECD TPG to provide such detailed profit
allocation instructions.
The logic behind the important-functions doctrine is convincing. Rou-

tine R&D, such as clinical studies and trial support, is widely outsourced
among unrelated parties on a cost-plus basis, while non-routine R&D is
not.1978 It would be contrary to third-party behaviour, and thus the arm’s
length principle, to allocate only a normal market return to group entities
that contribute unique R&D functions to the development of IP.1979
22.3.3. Outsourcing: Contract R&D arrangements
In this section, the author discusses how the important-functions doctrine

will govern the allocation of profits in controlled contract R&D arrange-
ments. He begins his analysis by commenting on the historical OECD
guidance for contract R&D arrangements in section 22.3.3.2. before dis-
cussing how the current important-functions doctrine can be applied to
contract R&D arrangements in section 22.3.3.3. Lastly, some comments
are made on the use of geographically dispersed employees in section
22.3.3.4.
1977. In this general direction, see also Wilkie (2016), at p. 87. For critical comments,
see Musselli et al. (2017), at pp. 339-340, where it is argued that the split of residual
profits based on important functions will entail highly subjective (and thus problem-
atic) assessments. The author does not share this concern. Subjective assessments are
nothing new in transfer pricing. The important point here is that the residual profits
from unique and valuable IP are allocated only to jurisdictions from which significant
IP value creation contributions emanated from (as opposed to the generally low-tax
R&D funding jurisdiction, which Musselli argues should still be able to claim the re-
sidual profits).
1978. A multinational will likely never outsource unique, core-value-creating R&D;
see the discussion on foreign direct investment in sec. 2.3.
1979. In this direction, see also Brauner (2010), at p. 17 with respect to the buy-in issue
under the US CSA regulations.
630
22.3.3.2. The relationship between the 2015 OECD important-

functions doctrine and the historical 2009 business
restructuring and intra-group services guidance
In 2016, the OECD reached consensus on new wording on the treatment of

cross-border business restructurings in the OECD TPG, replacing the 2009
predecessor text. The new 2016 text was designed to align with the mate-
rial profit allocation positions taken in the new OECD guidance on intan-
gibles. Thus, there are no inconsistencies between the current OECD texts
on business restructurings and intangibles. In this section, however, the
author will discuss the relationship between the current text on intangibles
and the historical 2009 text on business restructurings. He does so because
the discussion provides valuable insight into how the important-functions
doctrine has altered the profit allocation positions historically taken by the
OECD with respect to contract R&D arrangements, and thereby also in-
sight into the practical significance of the important-functions doctrine.
It is no secret that the OECD TPG, until just recently, have been lenient
on the widespread use by multinationals of controlled contract R&D ar-
rangements to shift profits from internally developed manufacturing IP
out of high-tax jurisdictions.1980 Acknowledgement of this historical pre-
BEPS backdrop is necessary in order to gauge the significance of the 2015
important-functions doctrine. The 2009 business restructuring guidance
was written before the financial crisis led to increased focus on declining
public finances, which again led to a political will among OECD member
countries to address such profit-shifting practices.
The 2009 text took the position that a group entity that only funded R&D
could be allocated residual profits, as long as it incurred the financial risks
connected to the IP development.1981 It was not required that the funding
entity performed or controlled core R&D activities. This was clearly illus-
1980. In this direction, see also Schön (2014), at p. 4. See also Musselli et al. (2017),
at p. 337, where it is argued that the OECD should have retained the pre-BEPS solution
that a group entity that solely funds IP development could end up being allocated the
residual profits from later exploitation of the IP (see also Musselli (2006); and infra n.
1998). The author does not share this view. He suspects that, in the long run, it could
have been damaging to the international consensus on the arm’s length principle as the
legal foundation for allocating taxing rights to business profits if the OECD would have
continued to accept controlled profit allocations based on legal formalities akin to the
pre-BEPS contract R&D agreements, which no doubt have resulted in the extraction of
residual profits from the jurisdictions in which the intangible values were created (and
the injection of those profits into tax havens).
1981. See also Brauner (2016), at p. 102; and Brauner (2015), at p. 75.
631
trated in an example pertaining to a contract R&D arrangement in which

the funding entity compensated the R&D entity on a concurrent cost-plus
basis and was allocated the residual profits from the developed IP.1982 This
was accepted, as long as the funding entity made relevant decisions in or-
der to control its “risk of failure of the research”.
Thus, the view of the 2009 guidance was that the incurrence of financial
risk alone was enough to justify allocating residual profits to a funding
entity. In order to accept that the funding entity incurred financial risks,
however, it had to make the following decisions:
– “hiring” and “firing” the R&D entity;
– determining the type of research to be carried out and the objectives of
it;
– determining the R&D budget;
– setting out reporting obligations for the R&D entity; and
– assessing the outcome of the R&D activity.
These are high-level decisions. They are, however, of an administrative

character and may likely be performed by non-technical personnel with
business or legal backgrounds. This interpretation is supported by the em-
phasis of the 2009 business restructuring guidance that the core opera-
tional R&D risk was “distinct from the failure risk borne” by the funding
entity, which justifies allocating residual profits to it.1983
The same reasoning was applied in the previous intra-group services guid-
ance on contract R&D arrangements. The OECD TPG here stated, in un-
ambiguous wording, that when the R&D entity “has discretion to work
within broadly defined categories … involving frontier research”, which
can be a “critical factor in the performance of the group”, an application
of the “cost plus method may be appropriate”.1984 Of course, this pricing
method will treat the R&D entity as the tested party and allocate only a
normal market return to it, while the residual profits can then be extracted
from the R&D jurisdiction and allocated to a foreign funding entity (often
in a low-tax jurisdiction).
Again, the justification offered by the OECD TPG for this profit alloca-
tion solution was that the R&D entity was “insulated from financial risk”,
as its expenses were reimbursed regardless of whether the research was
1982. 2010 OECD TPG, para. 9.26.

1983. Id.
1984. 2010 OECD TPG, para. 7.41.
632
successful or not.1985 The implication of this position was that the residual
profits were allocated to the funding jurisdiction, not the R&D jurisdiction.
Intangible income would then not be allocated to the jurisdiction where
intangible value was created.1986
This pre-BEPS reasoning of the historical business restructuring and intra-

group service guidance conflicts with the 2017 important-functions doc-
trine. The former guidance required, in order to accept an allocation of re-
sidual profits to a funding entity, that the entity incurred the financial risks
associated with the R&D project. Little or no involvement was required by
the funding entity with respect to the “core R&D” activities performed by
the R&D entity.
In contrast, the important-functions doctrine requires that the funding en-

tity perform some important R&D functions in order to claim entitlement
to any residual profits. The question of how much “R&D substance” the
important-functions doctrine requires from a foreign funding entity before
it may be allocated residual profits is the topic for discussion in section
22.3.3.3.
In conclusion, it is clear that the previous generation of OECD guidance on

business restructurings and intra-group services allowed group entities that
performed core R&D functions – and thus, in reality, were responsible for
IP value creation – to be compensated with only a normal market return on
a concurrent basis throughout the IP development phase under a one-sided
pricing method (typically cost-plus) without being entitled to any of the
later exploitation “fruits” of the IP that they had contributed so significant-
ly to the development of.1987 This OECD position has now been discarded.
The new guidance explicitly prohibits an R&D entity being treated as the
tested party and only allocated a normal market return under a one-sided
1985. The guidance also added that the funding entity, as the owner of the developed
IP, would assume the commercial exploitation risks. This argument is clearly inappro-
priate. Any risks connected to subsequent exploitation should have no bearing on how
group entities that contributed to the IP development (i.e. the pre-exploitation phase)
should be remunerated for their efforts.
1986. In this direction, see also Brauner (2016), at p. 121, where it is stated that “no
research company would enter into an agreement under which it shares its crown jewel
intangibles with another company that brings nothing to the table but cash”. See also
supra n. 1206; and Durst (2012b), who argues that R&D agreements in which the R&D
provider relinquishes all rights to residual profits (from unique and valuable IP that it
creates) likely do not exist among third parties.
1987. In this direction, see also Schön (2014), at p. 4.
633
method.1988 Group entities that contribute important functions to the de-

velopment of IP must now be allocated residual profits from subsequent
exploitation of the IP. Thus, the important-functions doctrine has taken the
opposite profit allocation position to the one historically embraced by the
OECD, and has thus significantly altered how IP profits must be allocated
intra-group.1989
22.3.3.3. The important-functions doctrine and contract R&D

arrangements
If the important-functions doctrine entailed that the R&D entity’s own em-
ployees must carry out R&D functions physically on site in the entity’s
residence jurisdiction in order for it to be allocated residual profits, it would
put a brutal stop to contract R&D schemes designed to shift profits. Only a
1988. OECD TPG, para. 6.58. See also Pankiv (2016), at p. 467, where cost-plus remu-
neration of an R&D entity is rejected. However, Navarro (2017), at pp. 240-241 (and
also p. 286) seems to blatantly disregard the position of the OECD TPG on this point.
The author argues – based on the circular and unrealistic premise that the taxpayer
can demonstrate that the cost-plus method (or TNMM) will result in an arm’s length
profit allocation – that the legal owner of the IP may compensate other group entities
that perform DEMPE functions on a cost-plus basis, with the apparent result that the
legal owner (i.e. the funding entity) ends up with the residual profits. The reasoning
offered for this solution is that the OECD’s 2017 DEMPE approach “impl[ies] the dis-
regard of the ownership (i.e. an implicit transactional adjustment) not in accordance
with the arm’s length principle”. The author strongly disagrees. Navarro’s reasoning (i)
disregards the manifest OECD position on the issue (2017 OECD TPG, para. 6.58); (ii)
seems to be based on an unfounded conception of the relevance of legal ownership in
transfer pricing; and (iii) uses an ill-conceived “best method”-like line of argumenta-
tion to support the use of one-sided pricing methodology to remunerate group enti-
ties performing DEMPE functions. With respect to this last part, Navarro seems to be
under the impression that the important-functions doctrine (i.e. the key element of the
2017 OECD IP ownership provisions) represents an application of the specified profit
split method (which it does not), and thus that the one-sided methods could be preferred
if better suited to the circumstances. The author’s take on this is that both the reason-
ing and the conclusion of Navarro on this point represent an unfounded interpretation
of the OECD TPG and should thus be disregarded. It is clear that group entities that
contribute DEMPE functions shall, under the 2017 OECD TPG, attract residual profits,
and thus cannot be remunerated under one-sided pricing methodologies.
1989. See also, in this direction, Brauner (2016), at p. 110, where it is noted that the
current OECD TPG have “reduced the importance of (financial) risk taking in transfer
pricing”. See also Musselli et al. (2017), at p. 332, where this BEPS revision of the
OECD TPG is described as follows: “This is a Copernican revolution … the OECD
has progressed from a model where intangibles are owned by funding capital necessary
for intangibles development and assuming the risk of a negative research outcome to
a model where intangibles are owned by companies performing important functions
related to the same intangibles development, enhancement, maintenance, protection or
exploitation (DEMPE).”
634
“true” R&D entity, which both performs and controls its own R&D, would
then qualify for residual profits. While this, at an early stage of the BEPS
Project, may have been the OECD’s ambition,1990 it is clear that the impor-
tant-functions doctrine does not go this far.1991
The final consensus text now expressly states that the group entity claiming
entitlement to residual profits does not have to “physically [perform] all of
the functions related to the development … through its own personnel… to
retain or be attributed a portion of the return”.1992
The question is to which extent R&D may be performed in one jurisdiction

while the residual profits are allocated to a group entity resident in another
jurisdiction, typically a funding (cash-box) entity.1993 The guidance on this
important issue is ambiguous and convoluted, as may be expected for an
issue this controversial.1994 This question was a key part of the intangibles
(and the larger BEPS) project. It was debated extensively and was among
the last points on which to reach consensus. The criterion is that that the
funding entity must control the R&D functions outsourced to other group
entities in order to attract residual profits.1995 Thus, a group funding entity
in jurisdiction X may, in principle, be allocated the residual profits from IP
developed by a group R&D entity in jurisdiction Y, if the R&D can be said
to have been controlled by the financing entity. If the financing entity did
1990. For a debate comment on this in the relatively early phase of the BEPS revision
of the intangibles guidance, see Durst (2012b), at p. 1124; and Durst (2012d).
1991. The wording contained in the 2012 OECD Discussion draft Revision of the spe-
cial considerations for intangibles in Chapter VI of the OECD Transfer Pricing Guide-
lines and Related Provisions (2012D) more directly indicated a requirement to have
competent “white coat” employees on the ground locally. It read: “It is expected, how-
ever, that where functions are in alignment with claims to intangible related returns in
contracts and registrations, the entity claiming entitlement to intangible related returns
will physically perform, through its own employees, the important functions related to
the development, enhancement, maintenance and protection of the intangibles” (see
2012D, para. 40). For comments on the 2012 OECD intangibles draft, see Wittendorff
(2012d); and Helderman et al. (2013).
1993. On outsourcing, see also Swaneveld et al. (2004).
1994. In this direction, see also Brauner (2016), at p. 110, where it is noted: “Even
though the cash-box is targeted by these rules, they state that the taxpayer does not have
to perform the functions of controlling the risk itself; paying for them suffices. This is
quite vague and would likely result in many interpretation conflicts and litigation but
perhaps would eliminate the most egregious cases where one party is a true cash-box or
an empty shell solely owning the intangible.”
1995. This criterion has been criticized for being open to tax planning (placement of
R&D control functions in “cash boxes” to attract residual profits); see, e.g. Ballentine
(2016). See also Osborn et al. (2017), under sec. I.C, and the practical example of a tax
planning design in sec. IV.A-C.
635
not control the R&D, the residual profits will be allocated to jurisdiction Y,
where intangible value was created.
First, it is clear that if the funding entity retains all important control func-
tions over the outsourced R&D, it will be entitled to all residual profits.
This interpretation is supported by the first “Shuyona” example, which
pertains to a multinational with two R&D centres.1996 One is operated by
the parent in country X and is responsible for all R&D within the group. It
designs R&D programmes, develops and controls budgets, decides where
R&D is to be carried out and monitors progress on R&D projects. The
second centre is subordinated to the first, is operated by a subsidiary in
country Y and carries out specific projects assigned by the parent R&D
centre. It needs approval for modifications to the R&D programme and for
budget increases, and reports its progress on a monthly basis to the parent.
The parent funds the subsidiary’s R&D through a concurrent service pro-
vider remuneration, while the residual profits are allocated to the parent.
The example accepts this allocation, as the parent performs the important
functions.
Second, it is clear that if the funding entity does not perform any of the
important R&D control functions, it will not be entitled to any residual
profits. This interpretation is supported by the second Shuyona example,
which modifies the factual pattern of the first with the twist that the group
now sells two lines of products.1997 R&D for product lines A and B is car-
ried out and funded autonomously by the parent and the subsidiary, re-
spectively. In particular, the subsidiary R&D centre now develops its own
programmes, establishes its own budgets, determines which projects to go
for, hires its own staff and does not report to the parent R&D centre. The
example concludes that the residual profits from the product line B intan-
gibles should be allocated to the subsidiary, as it performed the important
1996. OECD TPG, annex to ch. VI, Example 14. The language contained in the 2013
OECD Revised discussion draft on transfer pricing aspects of intangibles (2013 RDD)
was not as clear as the final 2017 version. It entitled the subsidiary to remuneration
“reflecting the anticipated contribution to intangible value”. This could, in the author’s
view, be interpreted as requiring the allocation of (some) residual profits to the subsidi-
ary. This was likely an unintended ambiguity, as the rest of the 2013 example indicated
that a concurrent normal market return allocation to the subsidiary would suffice.
1997. OECD TPG, annex to ch. VI, Example 15. Some additional wording was intro-
duced in the final version of this example, emphasizing that it would be “inappropriate
to treat Company S as the tested party in an R&D service agreement”, as the subsidiary
performs all of the important functions. The author, of course, agrees with this. If the
subsidiary was remunerated as the tested party under a one-sided method, this would
entail that the residual profits were allocated to the parent, contrary to the result of the
important-functions doctrine.
636
R&D functions for these intangibles. The parent did not control any of
these functions.
These examples address “polar” situations. In the first example, the R&D
entity is both controlled and funded by another group entity. In the second,
it acts completely autonomously in both respects. The important-functions
doctrine leads to obvious one-sided allocations of residual profits in these
scenarios: there is no profit split, as all residual profits are allocated to
the group entity that contributed the important R&D functions. These two
Shuyona examples just reiterate the OECD’s position that important R&D
functions attract residual profits. The examples do not convey useful guid-
ance for more nuanced factual patterns.
Cases that lie between these polar situations will likely trigger challeng-
ing assessments. What if the entity claiming entitlement to residual profits
outsources some important R&D functions while retaining others? For in-
stance, what if strategic R&D decisions and quality control for the develop-
ment of a pharmaceutical product’s IP are outsourced, while control over
budgets and some design aspects of the research are retained?
In the obvious absence of CUTs, the author sees no other alternative than
to apply a profit split. In principle, the split should be based on an assess-
ment of the relative value contribution from each type of important R&D
function contributed to the IP development. It will likely be unrealistic to
identify clear causal relationships between the specific types of important
R&D functions and the creation of intangible value. This may leave no
other alternative than to split the residual profits equally among the entities
performing important R&D functions.
Does this entail that the important-functions doctrine is more restrictive

than the pre-BEPS allocation guidance contained in the 2010 OECD TPG?
At least one thing is clear. The “black-or-white” allocation patterns de-
scribed in the 2009 business restructuring and intra-group services texts
discussed in section 22.3.3.2. are no longer acceptable. As mentioned, in
typical contract R&D arrangements, a mere funding entity could, under
the 2010 OECD TPG, be allocated all residual profits on the basis of finan-
cial risk alone, while the entity performing R&D could be remunerated on
a separate, cost-plus basis. The important-functions doctrine flips the table
completely, allocating all residual profits to the R&D entity.
The tax-planning contract R&D arrangements of the future will likely

place some important R&D functions in the low-tax funding entity while
637
the remaining R&D functions are performed by the high-tax R&D entity,
resulting in a profit split. In these cases, the important-functions doctrine
will at least reject the allocation of all residual profits to the funding entity.
Tax authorities in R&D jurisdictions should carry out a thorough func-

tional analysis of the separation of functions in R&D projects when a for-
eign funding entity asserts entitlement to a portion of the residual profits. It
must be clarified whether the foreign entity actually performs the functions
it asserts to perform and whether the functions can be deemed important
R&D functions in the context of the specific project at hand. If they can,
the significance of those functions relative to the functions performed in
the R&D jurisdiction should be determined. In order to realize the purpose
behind the important-functions doctrine, relatively more residual profits
should, in the author’s view, be allocated to “core” R&D functions than to
more auxiliary functions, such as budgeting.
To sum up, the important functions doctrine should be seen, in the context
of international transfer pricing, as a considerable leap towards aligning
the allocation of intangible profits with value creation.1998 As opposed to
the pre-BEPS (2010) OECD TPG, financial risk alone is no longer suf-
ficient to attract residual profits. The result will likely be a smoother prof-
it allocation, with the residual profits allocable only to the jurisdictions
where actual R&D functions are carried out. It remains to be seen how
multinationals will adapt their controlled R&D structures to the important-
functions doctrine. There is no reason to entertain illusions in this respect.
The author’s guess is that they will turn their focus towards combining
elements of important R&D functions with funding in a low-tax entity.
If some R&D functions (e.g. control over design aspects and budgets) are
placed in a low-tax funding entity, this will attract a split of the residual
profits. Due to the absence of clear causality, it may prove difficult for
1998. For an opposing view, see Musselli et al. (2017), at p. 331 and pp. 337-338, where
the 2017 OECD approach based on the allocation of residual profits to important func-
tions is highly criticized. Musselli argues that the OECD should have upheld the pre-
BEPS position that residual profits could be allocated to a group entity that only funds
the R&D efforts, possibly combined with stricter ex post valuation adjustment rules,
as well as anti-“cherry-picking” rules. This argument seems largely to be based on the
belief that third parties actually enter into agreements in which an R&D enterprise that
is in possession of resources capable of generating superprofits (IP residual) willingly
would surrender this profit potential in return for only a normal market profit (concur-
rent cost-plus remuneration during the development phase). The author does not share
this belief. His clear impression is that such third-party “comparables” in fact do not ex-
ist. He finds Musselli’s argument (which claims support in loose references to economic
theory and empirical research) to carry little persuasive power.
638
source-jurisdiction tax authorities to challenge the specific profit split ap-

plied. If one adds to this that the funding entity is entitled to a risk-adjusted
return on its R&D funding,1999 a multinational may, in practice, come very
close to achieving full extraction of the residual profits from the jurisdic-
tion where the bulk of the important R&D functions are carried out. This
will bring the allocation of intangible profits in international transfer pric-
ing back to square one, with the same result as that yielded by the pre-
BEPS guidance.
22.3.3.4. Performance of important R&D functions through

geographically dispersed employees
The question in this section is whether a funding entity may be allocated

residual profits if it performs important R&D control functions through
employees that live and work in a country different from its residence ju-
risdiction. For instance, assume that a low-tax funding entity enters into a
contract R&D agreement with a high-tax R&D entity, pursuant to which it
is allocated the residual profits from the developed intangibles in exchange
for the payment of a concurrent cost-plus remuneration to the R&D entity.
All R&D is carried out in the R&D jurisdiction. The funding entity, with
no technical personnel in its residence jurisdiction, employs key senior
members of the research team to control the R&D, while they continue to
live and work in the R&D jurisdiction. Viewed in isolation, this structure
should satisfy the important-functions doctrine, as the funding entity per-
forms the important R&D control functions.
Nevertheless, as the performance of the R&D control functions will form

an essential and significant part of the business activities of the funding
entity, they may trigger a permanent establishment (PE) in the R&D ju-
risdiction.2000 In determining the allocation of income between the head
office in the low-tax jurisdiction and the PE in the high-tax jurisdiction,
functions, assets and risks shall be attributed to the PE pursuant to the
“significant people functions” doctrine of the authorized OECD approach
(AOA).2001 The important-functions doctrine described in the new OECD
TPG on IP ownership should, in the author’s view, normally qualify as
1999. See the discussion on R&D funding profit allocation in sec. 22.4.
2000. OECD Model Tax Convention (OECD MTC), at art. 5.
2001. See sec. 25.2. for a discussion of the concept of “significant people functions” in
relation to art. 7 of the OECD MTC. See also the OECD Commentary on Article 7, at
paras. 20 and 21; and the OECD 2010 Report, at para. 16.
639
significant people functions in the context of intangible development.2002

Thus, it is likely that the economic ownership of the developed intangible
will be attributed to the PE, thereby attracting the residual profits from the
developed intangible to the source state where the employees carry out
their work and intangible value is created.2003 The author reverts to the al-
location of profits to a PE in chapter 25.
22.4. Profit allocation to IP development contributions:

Funding
The question in this section is the extent to which IP development funding

contributions should be allocated operating profits generated through the
subsequent exploitation of the developed IP under the new 2017 OECD
TPG. This is one of the most important – and problematic – issues in the
international transfer pricing of intangibles. It was one of the key issues
in the 2015 revision of the TPG on intangibles, and among the last to be
reached consensus on.
22.4.2. A lead-in to the issue
The remuneration of IP development funding can be regarded as the “flip

side” of determining which group entity should be entitled to the residual
profits. IP development may span over several years, require significant
funding and entail high risks. Because of this, the return allocable for the
funding may be considerable. This return will come out of the operating
profits from the subsequent exploitation of the developed IP. Therefore,
the larger the funding return, the less will be left as residual profits for the
“owner” of the IP that contributed the important R&D functions.
2002. This is not to say that the opposite (i.e. that significant people functions normally
will be equal to the important functions) is true; see the discussion in sec. 25.2.
2003. This will have several consequences. First, the residual profits will, of course,
likely be taxed at a much higher rate than in the low-tax jurisdiction. Second, as the per-
manent establishment (PE) is assigned ownership to the developed intangible for profit
allocation purposes, a subsequent migration of it from the PE jurisdiction will likely
trigger a charge. Third, if the funding entity provided the R&D entity with concurrent
remuneration, this may have generated a taxable loss in its residence state throughout
the development phase, which may prove difficult to utilize, as the funding entity is not
allocated the subsequent residual profits.
640
Profit allocation to IP development contributions: Funding
This highlights a central aspect pertaining to the relationship between the

important-functions doctrine and the guidance on IP development fund-
ing remuneration. When the important-functions doctrine strips a foreign
funding entity of residual profits, the entity goes from being in an “equity”
position in the IP development project to a “creditor” position. In other
words, it goes from being entitled to a negatively defined amount (the re-
sidual profits) to a positively defined amount (the arm’s length return for
the funding provided).
Because it is not very simple for a multinational to relocate its talented re-
searchers to other jurisdictions, the important-functions doctrine may turn
out to be successful in hampering a great deal of common profit-shifting
schemes, typically contract R&D agreements and CSAs. The doctrine
may, however, entice multinationals to extract IP profits from the source
jurisdiction under the “label” of funding remuneration.
It was therefore crucial for the OECD to impose legal restrictions on the
allocation of profits for funding. Otherwise, a funding entity might be able
to extract a return so high that it in reality could be treated as the “owner”
of the developed intangible for profit allocation purposes. That would be
detrimental to the new OECD approach to IP ownership, which is centred
around the important-functions doctrine.
The OECD has, in the current OECD TPG, taken a two-tiered approach to
the remuneration of IP development funding. The main rule is that if the
funding entity controls the financial risk associated with its funding con-
tribution, it will be entitled to a risk-adjusted rate of return on the invested
capital.2004 It can be noted that there is an inconsistency between this new
OECD position and the historical position taken by the OECD in the 2009
business restructuring guidance. The 2009 guidance requires, in order to
give effect to a contractual allocation of risk, that the entity allocated the
risk also, in substance controls, it.2005 It defines control as “the capacity to
make decisions to take on the risk (decision to put the capital at risk) and
decisions on whether and how to manage the risk, internally or using an
external provider”.2006 (Emphasis added)
However, it is unclear as to what extent this is actually a requirement.

The guidance, on several points, states that the alignment of control in
2004. See Andrus et al. (2017), at p. 94 on this point.

2005. 2010 OECD TPG, paras. 9.22-9.28.
2006. 2010 OECD TPG, para. 9.23; see also 2010 OECD TPG, para. 1.49.
641
substance and contractual allocation of risk is simply “a relevant factor”,

and a “relevant, although not determinative” factor.2007 Thus, when the
2009 guidance spoke of risk, it focused on financial risk.2008 The author
draws support for this interpretation from an example in paragraph 9.26
of the 2010 OECD TPG, pertaining to a straightforward contract R&D
agreement. The example accepts a significantly lesser degree of func-
tions being performed by a funding entity than required under the new
2017 OECD TPG in order to assign the residual profits to it. The ratio
nale behind the 2009 stance is that the funding entity is deemed to bear
the “failure risk”, i.e. the financial risk. The example indicates that the
mere selection by the funding entity of a party to perform the R&D, the
decision of what type of research is to be carried out, as well as a re-
porting obligation for the R&D entity in combination with its funding
will constitute sufficient control to allocate the financial risk, and thereby
the residual profits, to the funding entity. Thus, there is a crucial differ-
ence between the concept of control in the 2009 business restructuring
guidance and the new intangibles guidance, as control over financial risk
alone could entitle a funding entity to the residual profits under the 2009
text.2009 In contrast, financial risk will only entitle the funding entity to
a risk-adjusted rate of anticipated return on its capital invested under the
new guidance, but not more. An opposite conclusion would contradict the
important-functions doctrine.
Further, the current OECD TPG limit this remuneration to a funding entity
that does not control the financial risk to a risk-free rate of return on the
invested capital.2010 The new two-tiered OECD approach to IP development
funding remuneration was revealed in the draft texts leading up to the final
2017 consensus text.2011 There was, however, no indication in the drafts that
the OECD was willing to allocate any profits to a funding entity that did
not even control the financial risk associated with its funding. Under the
2014 draft, such an entity was completely cut off from funding remunera-
tion, while it rather surprisingly is entitled to a risk-free rate of return in the
current consensus text.
The difference between a “risk-adjusted” and “risk-free” rate of return on

the R&D funding capital may be significant. How the risk-adjusted rate
2007. 2010 OECD TPG, para. 9.22; and the illustration in 2010 OECD TPG, para. 9.33.
2008. See also, in this direction, Monsenego (2014), at p. 13. See also the discussion of
significant people functions in the context of art. 7 of the OECD MTC in sec. 25.2.
2009. On the contractual shifting of risks, see Schön (2014).
2010. See Andrus et al. (2017), at p. 93 on this.
2011. See 2013 RDD, at para. 84; and 2014D, at para. 6.61.
642
shall be determined under the current OECD TPG is not entirely clear, as
the author will revert to in the discussion in sections 22.4.4.-22.4.11. A risk-
free rate is easier to determine. At the time of writing, this rate is low, with
US treasury bonds of 3 months’, 5 years’ and 10 years’ maturity offering
yields of 0.06%, 1.51% and 2.14%, respectively.2012
In essence, this entails that a group entity that is not in control of the finan-
cial risk associated with its R&D funding, and is therefore only entitled to
a risk-free return, will incur a “loss” on its funding, as the risk-free return
will be below the return that it could have received on an alternative R&D
investment of similar risk. Nonetheless, this “loss” will not be as dramatic
as it would have been had the OECD gone through with its initial plan of
entirely cutting off funding entities that do not control the financial risks
associated with their R&D funding.
22.4.3. Control over financial risk
A group funding entity must control the financial risk connected to its IP
development funding contribution in order to qualify for profit allocation
equal to a risk-adjusted rate of return.2013 If it is not in control, it will only
attract a risk-free rate of return.2014
2012. See http://www.bloomberg.com/markets/rates-bonds/government-bonds/us (ac-

cessed 4 Sept. 2015).
2013. OECD TPG, para. 6.61 (see also paras. 1.98 and 1.65). See also Andrus et al.
(2017), at p. 94 on the control threshold.
2014. OECD TPG, para. 1.103. See also the example in OECD TPG, para. 1.85, where
the funding entity does not have the capacity to make decisions to take on or decline the
financing opportunity. It is therefore not deemed to be in control of the financial risk,
and is thus only allocated a risk-free rate of return on its funding; see OECD TPG, para.
1.103. See also OECD TPG, annex to ch. VI, Example 16, where company T does not
control the financial risk and is only entitled to a risk-free return on its R&D funding
(see OECD TPG, annex to ch. VI, para. 58). See also OECD TPG, annex to ch. VIII,
Example 5, where company A does not control the funding risk, and is therefore not
allocated any operating profits. Even though not entirely clear, the example does not
seem willing to allocate a risk-free return to the funding entity. The author finds that
comprehensively strange. Such a solution would depart significantly from the positions
taken in the guidance on risk contained in the OECD TPG, which is applicable also
to CSAs (see OECD TPG, paras. 8.9 and 6.59), and would represent an unjustified
discrimination of IP development funding, depending on the contractual mechanism
through which it is rendered, as funding through contract R&D would be allocated a
risk-free return, while funding through CSAs would not. The author therefore finds that
the example must be interpreted as allocating a risk-free return to a funding entity that
is not in control of the financial risk.
643
At least in theory, financial risks are distinguishable from operational R&D

risks.2015 The latter is the risk that the R&D will not be successfully com-
pleted and commercialized, while financial risk is risk that the funder will
not recover the principal investment amount together with a risk-adjusted
rate of return. Intangible development funding, however, is project-specific
financing, where the financial risk is derived from the operational R&D
risks. A return on investment is contingent on the project being successful
so that it will generate operating profits sufficient to pay the risk-adjusted
rate of return to the funding entity. The divide between operational and
financial risk for project-specific R&D funding is therefore rather artificial.
Nevertheless, the approach of the current OECD TPG to IP ownership is
built on this distinction.
In light of this, it comes as no surprise that the guidance on control in the

context of IP development funding is relatively ambiguous. After all, the
issue of allocating operating profits to a funding entity will typically arise
because the important functions doctrine has stripped the entity of all re-
sidual profits since it does not control the important R&D functions. This
entails that the entity will likely have very little influence (if any) on the
substantial R&D operating decisions and the operational risks. Given the
close relationship between operational R&D risks and financial risks, the
author finds it somewhat illogical to require that the funding entity be in
control of financial risks when the important R&D functions are performed
by a different group entity that also generates the operational R&D risks.
It is unclear precisely what type of control is required over the financial

risks. “Control” in general is defined in the OECD TPG as the “the capac-
ity to make decisions to take on the risk and decisions on whether and
how to respond to the risk”.2016 Thus, the group entity purporting to control
the relevant risk must be able to assess potential risk outcomes and their
effects on business.2017 The funding entity must also have competent and
experienced people in place that are able to assess the creditworthiness of
the group entity receiving funds, the R&D risks, whether it is likely that
additional funding is required, etc.2018 The extent of required control ac-
tivities increases with the amount of R&D financing and the R&D risk.2019

2016. OECD TPG, para. 1.65. See Monsenego (2014), at p. 15, on financial capacity
(under the 2010 OECD TPG); and Bullen (2010), at pp. 496-506, on the control of risk
under the historical 1995/2010 OECD TPG text.
644
The author’s interpretation of the control criterion’s wording in the current

OECD TPG is that it requires the funding entity to be capable of:
– assessing the commercial prospects of the R&D project in order to
decide whether and how much to invest;
– assessing the relative profitability of the R&D investment against other
realistic investment opportunities; and
– monitoring the progress of the R&D project and revising its invest-
ment analysis accordingly, e.g. to halt funding if the prospects of the
project change.2020
Nevertheless, there is no getting around the fact that there is tension be-
tween this general OECD guidance on the control criterion and some spe-
cific examples included in the OECD TPG on IP development funding.
The examples indicate a lower threshold for establishing control over fi-
nancial risks than follows from the above interpretation of the general con-
trol guidance.2021 The examples fall into two categories: those that do not
accept that the funding entity is in control of the financial risk and those
that do. The first group of examples describes the funding entity as doing
nothing apart from channelling funds, with no technical personnel capable
of supervising R&D.2022
Further, the examples that accept that the funding entity is in control of
the financial risk offer little in the way of intuition for their conclusions.
One example indicates that the funding entity must be capable of “analys-
2020. This issue is distinguishable from whether the funding entity must, at least to
some extent, be in control of the R&D entity’s use of the funding in order to be assigned
the financial risk. An interesting point here is that the 2013 RDD allowed the allocation
of operating profits for intangible development funding, even if the funding entity did
not have “any control over the use of the contributed funds or the conduct of the funded
activity”; see 2013 RDD, at para. 84. The quoted wording was omitted in the 2014 draft
text and in the final 2017 text. The author finds it doubtful as to whether third-party
investors would be willing to contribute funds to an R&D project for which they had
no control over the spending. However, if the R&D team is talented and has a proven
track record and the project is likely to be highly profitable, that may not be the case.
Also, the R&D activity will, in itself, drive the use of the contributed funds, and the
new guidance clearly does not demand that the funding entity be capable of exercising
any control over R&D functions. The author therefore finds that control over the use of
contributed funds should not be required. Had the entity controlled the important R&D
functions, it would have been entitled to residual profits, not just a risk-adjusted return
on its funding.
2021. The author discusses these examples more thoroughly, including their profit al-
location consequences, in sec. 22.4.6.
2022. See the first example in OECD TPG, paras. 1.85 and 1.103; the second example
in OECD TPG, annex to ch. VI, para. 58; and the third example in OECD TPG, annex
to ch. VIII, Example 5. See also OECD TPG, annex to ch. VI, Example 16 (Shuyona).
645
ing” the intangible at stake and making some profit estimates.2023 Another
example simply states that the funding entity is in control without elabo-
rating.2024 In both of the examples that accept that the funding entity is in
control of the financial risk, however, the funding entity only contributes
funding to the IP development, while all R&D and pre-existing intangi-
bles are contributed by other group entities. There is no indication that the
funding entity is capable of performing any technical assessments of the
potential of the R&D project.
In conclusion, the examples apparently deem only pure “post-box” enti-

ties – which are comprehensively stripped of all relevant decision-making
capacity – to not be in control of the financial risks associated with their
IP development funding. The examples provide a solid basis for asserting
that as long as there is some capability in the funding entity with respect
to making a financially informed investment decision, the threshold for
establishing control over financial risk will likely be passed. In the author’s
view, this is where the line generally must be drawn under the current
OECD TPG.
This interpretation is compatible with the fact that a funding entity will
not perform the important R&D functions, and will therefore not control
the operational R&D risks. The required investment decisions may likely
be carried out by experienced, non-technical personnel with managerial,
business and legal backgrounds with insight into project financing within
the relevant business sector. It will probably not be demanding for a mul-
tinational to ensure that such functions are in place in a foreign (low-tax)
funding entity. It will therefore likely be rare that the profit allocation to
a group funding entity must be limited to a risk-free return on the R&D
funding amount. Normally, the funding entity will be entitled to a risk-
adjusted rate of return.
22.4.4. The point of departure for determining the risk-

adjusted rate of return
The guidance on how the rate of return allocable to a group entity that
provides IP development financing (and is in control of the associated fi-
nancial risks) shall be determined is comprehensively ambiguous.2025
2023. OECD TPG, annex to ch. VI, Example 6.

2024. OECD TPG, annex to ch. VIII, Example 4.
2025. The core provisions encompass OECD TPG, paras. 6.59-6.62.
646
Nevertheless, two main directions can be made out. First, the remunera-
tion must mirror that of “similar funding arrangements among independent
entities”.2026 This is problematic for the “usual” reason, i.e. the absence of
CUTs. The author will elaborate on this in section 22.4.5.
Second, the remuneration is restricted to a “risk-adjusted” rate of return.2027

This is a separately determined arm’s length return to the funding contri-
bution, and stands in contrast to the residual profits typically allocated to
the funding entity in the controlled pricing subject to review. As discussed
in section 22.4.3., due to the low threshold for constituting control over
financial risk, the funding entity will normally be entitled to a risk-adjust-
ed rate of return. This makes the determination of this rate an important
transfer pricing issue.
Ironically, the current OECD TPG on IP ownership – a main point of

which is to restrict the allocation of profits to foreign, low-tax funding enti-
ties – provides precious little guidance on how to arrive at an appropriate
risk-adjusted rate of return. The only indication provided is that the rate
may be “based on the cost of capital or the return of a realistic alterna-
tive investment with comparable economic characteristics”,2028 and that it
is “important to consider the financing options realistically available to the
party receiving the funds”.2029
It is clear that these parameters are not exhaustive. The author will discuss
how the risk-adjusted rate of return shall be determined in sections 22.4.6.-
22.4.11.
22.4.5. Uncontrolled transaction analogy for determining

the risk-adjusted rate of return: Venture capital
financing
The basic point of departure for allocating profits for IP development

funding under the current OECD TPG is, as mentioned in section 22.4.4.,
that the remuneration must mirror that of “similar funding arrangements
among independent entities”.2030 A problem is that there will normally be

2029. Id.
647
no CUTs to the typical R&D contract agreements or CSAs used by mul-

tinationals to extract residual profits from source jurisdictions. There are
several reasons for this.
First, an unrelated party that possesses a valuable and unique resource will
normally be in a bargaining position that makes it unwilling to surren-
der its “equity” position in the potential future profits to a party that only
brings funding to the table. It will simply have better realistic alternatives
available, such as self-financing or external loan financing. In very early-
stage R&D for particularly innovative products, when the risk is extreme
and considerable amounts of capital are required to further develop and
commercialize the project, however, that bargaining position may be se-
verely challenged.
Nevertheless, even in such circumstances, the entrepreneur will rarely give

up all residual profits to an investor. There will normally be an equity split
so that the entrepreneur owns part of the company, with the rest being
owned by investors. Also, the farther the entrepreneur gets along the IP
development cycle, the more bargaining power he will command, and the
less residual profits he will be willing to part with.2031
Conversely, the hallmark of controlled R&D agreements and CSAs is that

the group entity in the best bargaining position, i.e. that which contributes
R&D and unique pre-existing intangibles, relinquishes its residual profits
to a foreign funding entity that “only” brings cash to the table.
Second, early-stage, high-risk investment opportunities in specific projects

are not publicly available to just any investor, in particular, not to passive
financial investors. Commercial R&D is normally carried out by multina-
tionals. A new, early-stage project may be carried out as part of a larger
R&D portfolio. An unrelated financial investor who buys shares in a listed
multinational will implicitly buy into its entire R&D project portfolio, con-
sisting of a range of projects, with some at the initial idea phase and others
far developed, and some with large profit potentials and others that will
fail. The investor will likely have severely restricted information on the
R&D conducted by the company, including future strategies to be pursued.
Thus, the investor must accept the whole R&D portfolio as such. It cannot
“cherry-pick” only the R&D projects that it finds most alluring.
2031. On the use of the bargaining power concept in transfer pricing, see Blessing
(2010b). See also supra n. 1125.
648
Conversely, related parties normally enter into R&D contract arrangements

or CSAs only for specific projects. Given the perfect flow of information
between the controlled parties, this triggers an obvious cherry-picking
risk, which lacks a direct parallel among unrelated parties.
To the extent that third parties even have access to investments in specific,
early-phase, promising, high-risk R&D projects, this will normally per-
tain to start-up companies – often consisting of a few talented individuals
with technical backgrounds – that seek funding. Investment opportunities
in these projects will normally only be available and relevant for a limited
circle of investors, entirely different from the financial investors that trade
in noted financial instruments. Such early-stage investors are known as
venture capitalists, i.e. active investors that provide capital for extremely
risky, pre-commercialized projects.2032 In exchange, they require a signifi-
cant return.
Such investments are organized through contractual structures vastly dif-

ferent from the R&D contract schemes and CSAs that frequently appear in
related-party transfer pricing structures. Nevertheless, venture capital in-
vestment practices could provide a sound point of departure for determin-
ing the risk-adjusted return allocable to a controlled funding contribution
to early-phase R&D. The author will therefore provide an overview of the
relevant aspects of venture capital (VC) investments.2033
2032. Private equity capital may also contribute, but is normally reserved for invest-
ments that are carried out with the purpose of restructuring an existing company, with
existing products and cash flows, to optimise its financial performance. These invest-
ments typically come at a later stage in the life cycle of a project or company than
venture capital (VC) investments, which typically concern projects that have not yet
resulted in any concrete marketable products or associated revenues, or are even not yet
incorporated. Private equity investments are typically triggered by the identification of
underutilized assets and optimization potential, as opposed to VC investments, which
may be triggered by the early-phase potential of certain individuals, teams, innovations
or projects.
2033. Very few early-stage projects attract funding from VC funds; see National Ven-
ture Capital Association Yearbook 2013 (NVCA, YB 2013), at p. 27. In 2012, soft-
ware projects received 31% of VC funding in the United States, life sciences projects
received 26% and renewable technology received 10.5%. The few that are picked are
illiquid investments. Considerable amounts of capital may be “locked in” for a long de-
velopment period before an exit opportunity is available. There is also a high probabil-
ity of failure for single projects, without any option for recovery of the invested capital.
For instance, in the United States, only approximately 16% of VC projects become
listed companies, and 33% are acquired. The two main exit opportunities available to
VC funds are public offerings and acquisitions.
649
A VC fund is typically organized as a limited partnership.2034 The owner-

ship structure of a VC fund consists of a general partner (GP) and several
limited partners (LPs). The relationship between a GP and an LP with
respect to remuneration in a VC project provides a useful analogy for the
relationship between a group entity that contributes important R&D func-
tions to IP development and one that merely contributes funding.
The GP is a professional VC enterprise responsible for organizing the le-

gal structure of the VC fund, marketing the project towards institutional
investors2035 and picking the projects to be included in the fund’s project
portfolio. A GP is actively involved with the day-to-day management of the
selected start-up companies and exercises a significant amount of control
over their development activities. While the GP contributes a significant
amount of human capital (in the form of functions) to a VC fund, it nor-
mally only provides a relatively modest portion of the invested funds.2036
Conversely, an LP is a passive, risk-willing investor. An LP has no active
involvement in the projects of the VC fund, but will provide the lion’s share
of the funding.2037
The author will illustrate this with an example.2038 Let us say that a VC
fund, in total, raises capital of 1,000, to be gradually invested into promis-
ing ideas for software development over a period of 5 years. The GP invests
1% of the capital, while the remaining 99% comes from LPs. The VC fund
successfully develops a few of the selected start-up projects and sells them
for a total of 1500.
Under the VC fund agreement, the GP is entitled to an annual management

fee of 1% of the invested capital, in addition to a carried interest (carry) of
20% of the profits realized when exiting an individual project.2039 Under
2034. NVCA, YB 2013, at p. 8. A limited partnership is a form of partnership similar

to a general partnership, except that in addition to at least one general partner with un-
limited liability for the limited partnership’s obligations, there are one or more limited
partners with liability limited to their capital contributions. Limited partnerships are
recognized, for instance, under the company law of Germany, Japan, New Zealand, the
United Kingdom and the United States.
2035. Such investors include pension funds, insurance companies, foundations and
high-net-worth individuals.
2036. In the United States, this portion is generally around 1% of the invested capital,
while it is larger in, for instance, Norway, with portions ranging from 1-10% of the
invested capital.
2037. Often as much as 99% of the capital invested.
2038. The example is inspired by Breslin (2013), at p. 4.
2039. In this example, profits are simply calculated as the difference between the exit
price for all sold companies minus the total invested capital on all projects, regardless
650
these assumptions, the GP is entitled to a management fee of 50,2040 in ad-

dition to the 100 in carry,2041 for a total of 150.2042 The remaining 350 will
be distributed to the LPs, who will then realize a return on their investment
of 35%.2043
If one calculates the return on investment for the GP, it will be an immense
1500%. It is, however, a mistake to interpret this as a return on the financ-
ing contribution of the GP.2044 In fact, it is clear that the GP and the LPs
realize the same 35% return on their funding. The extraordinary excess
return of the GP is remuneration for its functions, in particular, for its tal-
ent in selecting promising start-up projects and for leading those projects
towards commercial maturity.2045
If one were to apply this typical VC structure for allocating profits among
functions and funding analogously to a transfer pricing scenario in which the
important R&D functions are performed by one group entity while another
provides funding, the result would be that the funding entity should receive
only a financial return on its funding contribution. That return should be set
relatively high, however, due to the risks connected with early-phase R&D
investment. Venture capitalists in 2011 generally required a return, often
referred to as the “target rate of return”, on their investments of 30-70%.2046
This rate certainly should not be persuasive for the remuneration of fund-
ing entities for transfer pricing purposes in general. Nonetheless, it should
of success.
2040. 1,000 × 0.01 × 5 = 50
2041. 500 × 0.2 = 100
2042. The management fee is calculated as 1,000 × 1% × 5 = 50. The vested interest is
calculated as (1500 − 1000) × 20% = 100.
2043. The return on investment to limited partners (LPs) is calculated as the distrib-
uted profits divided by their total investment, which, in this case, is (350 ÷ 1000) × 100
= 35.35%. LPs will also be entitled to an annual interest on the capital invested (hurdle
rate), for instance, 8%, which the author disregards for the purpose of this example.
2044. See also Breslin (2013), p. 5.
2045. The classification of carried interests in VC and private equity has also proven
problematic in tax contexts apart from international transfer pricing. For instance, the
Norwegian Supreme Court, on 12 November 2015, ruled in favour of the taxpayer in
a case pertaining to the domestic income tax treatment of carried interests (Rt. 2015
s. 1260). The question was whether the carried interest was to be regarded as income
on capital (in which case it would be tax-exempt under the Norwegian participation
method) or as income from employment (which, in addition to a marginal tax rate for
the personal taxpayer of 47.8%, would trigger employer social security contributions of
14.1% of the carried interest for the VC fund).
2046. Damodaran (2012), p. 647. The target rate of return is the internal rate of return
required by venture capitalists.
651
be recognized that the general rate of return achieved by financial investors

in VC funds represents the return to investors that (i) do not exercise direct
control over the R&D that they invest in; and (ii) incur high levels of risks.
Likely, this will be the closest one can get to a CUT for benchmarking
the operating profits allocable for intangible development funding of early-
phase, high-risk R&D.
It is the author’s view that the VC financial investor returns are, in general,
helpful for determining the profit allocable to group entities that fund genu-
inely high-risk, early-phase R&D projects. This will ensure that the fund-
ing remuneration mirrors that of “similar funding arrangements among
independent entities”.2047 The VC target rate of return, however, of course
does not qualify as a CUT under the specified CUT method of the OECD
TPG, and there are obvious comparability concerns connected with the use
of this rate as a reference point. It is therefore crucial that the profit alloca-
tion assessment also takes into account the specific economic characteris-
tics relevant to the controlled transaction.
22.4.6. Do the OECD TPG examples contribute to

allocable to the funding entity?
The OECD has, in several of the examples contained in the current OECD
TPG, addressed the allocation of profits to a group entity that funds R&D.
Some of them provide unusually concrete indications of the level of remu-
neration envisioned by the OECD as an arm’s length return. This clarifica-
tion was a long time coming. The 2012, 2013 and (albeit to a lesser extent)
2014 drafts of the new intangibles chapter were rather ambiguous on this
point. While there are still “blank spots” in the final 2017 consensus text,
there is now at least some guidance pertaining to the relatively low-risk
R&D scenarios addressed (i.e. the funding of second-generation R&D).
The first example pertains to company B, which performs R&D through

experienced staff and owns valuable pre-existing IP that is used in the de-
velopment.2048 It enters into a contract R&D agreement with related compa-
ny A, pursuant to which it shall perform and control all activities related to

2048. OECD TPG, annex to ch. VI, Example 6. The example was not included in either
the 2012D or 2013 RDD texts, but surfaced for the first time in the 2014D, where it was
bracketed, indicating that it was not a consensus text.
652
the development, maintenance and exploitation of the intangible. Company

A funds the development with 100 annually in years 1-5 and is assigned le-
gal ownership of any intangibles developed. It is estimated that worldwide
exploitation of the developed intangibles will generate operating profits of
550 annually in years 6-15. The controlled profit allocation assigns 200 of
the annual profits to company B and 350 to company A.
Company A is deemed to be in control of the financial risks associated with

its funding contribution, and is therefore allocated a risk-adjusted rate of
return. With respect to the amount, the example just assumes that the fund-
ing entity should receive 110 annually. Thus, of the global annual profits of
550, 110 is allocated to company A, with the residual profits of 440 allo-
cable to company B. This equals an internal rate of return on company A’s
funding investment of approximately 11%,2049 which apparently is the rate
that the OECD is willing to allocate for funding contributions in projects
akin to the relatively low-risk R&D described in the example.
The second example pertains to companies A and B, which enter into a

CSA to develop an intangible.2050 Company B provides a pre-existing in-
tangible, which forms the basis for the development, as well as the R&D
team, while company A funds the R&D with an annual 100 in years 1-5.
Company A is deemed to be in control of the financial risk associated with
its funding contribution to the CSA, and therefore entitled to a return on its
funding contribution.
The exploitation of the cost-shared IP is estimated to generate global prof-

its of 550 per year in years 6-15.2051 The controlled pricing allocates 330 in
operating profits to company A in each year. This pricing is benchmarked
against the best realistic alternative of company A, which is to receive a
risk-adjusted rate of return on its funding capital. This return is determined
2049. The final 2017 version of the example explicitly states that the return equals 11%,
while the draft version in 2014D did not do so (but did, nevertheless, contain the same
information on investment and profit amounts so that it was possible to calculate the
internal rate of return). The fact that the OECD has specified the 11% rate in the final
text adds to the impression that the intention of the example is to pinpoint a concrete
arm’s length remuneration for intangible development funding of relatively low-risk
R&D ventures.
2050. OECD TPG, annex to ch. VIII, Example 4. The funding contribution is, accord-
ing to the author’s calculations, afforded a rate of return of 11.64% in this example.
Thus, the OECD affords the same amount of return to a funding contribution regardless
of whether the funding is contributed through a contract research agreement or a CSA.
See the discussion of the example in the context of CSAs in sec. 14.3.
2051. With 60% of the profits generated in the market of Company A and 40% in the
market of Company B.
653
to be 110 per year in years 6-15, equal to an internal rate of return of 12%.
Thus, in order for the pricing to be at arm’s length, company A must sur-
render (through balancing payments) 220 per year to company B so that the
profits allocable to company A in each year is 110, equal to its best realistic
alternative.2052
Both of the above examples indicate a risk-adjusted rate of return of ap-

proximately 11%. A profit allocation that results in this level of return now
seems to be the OECD-endorsed arm’s length profit allocation for similar
cases.2053 This is undeniably interesting, both because it offers concrete
guidance and because it seems relatively restrictive. 11% is a far cry from
the VC rate of return of 30-70% suggested in the discussion in section
22.4.5. The explanation for this is likely that the described R&D projects
entail relatively low risks. All projects are based on pre-existing IP, devel-
oped by experienced staff, and are expected to be highly profitable.
In the author’s view, this rate of return should be viewed as a minimum rate
that should only be appropriate in cases in which there is no genuine un-
certainty connected to the R&D project. The rate is clearly inappropriate
for analogical application in the context of high-risk R&D projects, typi-
cally when the R&D is not based on pre-existing and valuable intangibles.
It would therefore be comprehensively inappropriate for tax authorities in
R&D jurisdictions to use this as a go-to rate for remunerating a foreign
2052. The OECD TPG include a third example in the annex to ch. VI, Example 17. One
would assume that the funding entity in this example, Company S, should be deemed as
not being in control of the financial risk. The example is, however, ambiguous. It simply
states that the entity is entitled to a “financing return”, the level of which depends on
whether it is in control of the funding risk. The author cannot comprehend why the
example does not dismiss the possibility that the funding entity controls the financial
risk. Clearly, it should have done so, as there is no indication that the funding entity is
capable of controlling this risk, and should therefore only be entitled to a risk-free rate
of return. This ambiguity renders the example impotent. Irrespective of the extent of
the funding remuneration, the residual profits are allocated to Company A. For com-
ments on the example, see Musselli et al. (2017), at p. 333.
2053. It should be noted that footnote 1 of the new CSA guidance (see OECD TPG,
annex to ch. VIII, para. 20) does not explain how the 110 in annual income from the
funding investment in years 6-15 is derived. It is just assumed that this level of profit
represents an arm’s length level of return. It is further stated that this result is shown
just to demonstrate the principles of the example and is not meant to offer any “guid-
ance as to the level of arm’s length returns to participants in [cost contribution agree-
ments]”. While the author recognizes this reservation, he finds it difficult to see that the
rate of return used in the example should have no relevance in cases pertaining to “pure
cash-box” entities. After all, the rate of return selected is used to illustrate an arm’s
length return.
654
funding entity without first establishing that the funded R&D project in-
deed is a low-risk venture.
22.4.7. Do the US CSA regulations contribute to

The OECD’s funding guidance applies generally, irrespective of whether

the IP development is carried out through an R&D contract agreement,
a CSA or a different legal vehicle.2054 Comparatively, the US cost-shar-
ing regulations determine the allocation of profits only when the IP de-
velopment is organized through a qualifying CSA. Nevertheless, within
this scope of application, the US CSA rules also face the issue of how
to remunerate (normally foreign) R&D funding contributions. The author
therefore finds it interesting to investigate whether the US solution may
provide any arguments for the interpretation of the OECD’s guidance on
the determination of the risk-adjusted return.
A typical scenario under the CSA regulations is where a US group entity

performs R&D, owns valuable, pre-existing IP and enters into a CSA with
a foreign group funding entity to develop new versions of the pre-existing
IP, pursuant to which the lion’s share of the residual profits from the ex-
ploitation of IP developed under the CSA are assigned to the foreign entity.
The CSA regulations strip the foreign funding entity of the residual profits
by way of determining a buy-in payment. This is a deemed payment from
the foreign entity, as consideration for the pre-existing IP and other unique
resources (e.g. know-how) contributed by the US entity to the CSA.
In these scenarios, where only one of the parties to the controlled agree-
ment contributes unique inputs to the IP development process, the income
method will likely be applied to determine the buy-in amount.2055 That
amount will, based on the best alternatives realistically available to the US
entity, be set so high that the US entity is taxed for all residual profits from
the IP developed under the CSA.2056

2056. Normally, it will be a realistic option for the US entity to continue its ongoing
R&D and develop the intangible by itself by way of self-funding the project, thereby
reaping all residual profits from the developed intangible itself. Thus, the income meth-
655
An important result is that this methodology, when applied in the generic

profit-shifting scenario described, will allocate only a comparable prof-
its method (CPM)-like normal return to the funding entity. The residual
profits are allocated to the US parent.2057 In other words, the US CSA regu-
lations do not allocate any operating profits for IP development funding
contributions.2058
The result is surprising in light of the fact that the 2005 proposed US
CSA regulations offered relatively generous compensation for IP devel-
opment funding. Under the CPM application of the income method in
the 2005 proposed regulations, funding contributions were compensated
with a portion of the residual profits through a so-called “cost contribution
adjustment”.2059 The residual profit split method also allocated a portion
of the residual profits to IP development costs through the so-called “cost
contribution share” of the residual profits.2060
od requires that, if the total present value for the US entity from participating in the
CSA is less than the total present value it alternatively could have derived from the re-
alistic option of self-funding and licensing out the developed intangible, the difference
in value must be allocated to the US entity through a deemed buy-in amount.
2057. Treas. Regs. § 1.482-7(g)(4)(iii)(B).
2058. This is illustrated in an example on the income method pertaining to a US parent
that performs R&D through an experienced team and has developed version 1 of an
established piece of software. It enters into a CSA with a foreign subsidiary to develop
future versions of the software. The CSA assigns exclusive exploitation rights to the
developed intangible for the United States and for the worldwide market to the parent
and the subsidiary, respectively. On the basis that the subsidiary does not provide any
unique contributions to the intangible development, it is allocated only a normal market
return on its routine functions pursuant to the CPM, set to 14% of its revenues. It is not
allocated any operating profits that are consideration for its intangible development
2059. See 2005 Prop. Treas. Regs. (70 FR 51116-01) § 1.482-7(g)(4)(iv)(B)(4), if the
calculation were to be performed on the basis of sales); or § 1.482-7(g)(iv)(C)(3), if the
calculation were to be performed on the basis of profits. Funding returns would also be
allocated to external routine contributions; see § 1.482-7(g)(4)(iv)(D)(v). This was il-
lustrated through an example in § 1.482-7(g)(4)(iv)(D), pertaining to a US pharmaceu-
tical parent that entered into a CSA with a foreign funding entity to further develop a
partially researched vaccine. The development costs were 100, of which the subsidiary
funded 50. The return allocated for that investment was 50 (in other words, a 100% re-
turn on the funding). The example does not provide information on how many income
periods there were under the CSA investment, but this is not relevant, as the return is
stated on a present value basis.
2060. 2005 Prop. Treas. Reg. (70 FR 51116-01) § 1.482-7(g)(7)(iii)(C)(2). An example
illustrated this, pertaining to a CSA in which both the US and foreign entities contrib-
uted unique pre-existing intangibles to the R&D; see § 1.482-7(g)(7)(v). The intangible
development costs were 10 billion, of which the subsidiary funded 6. The return allo-
cated to that funding was 6 (in other words, a 100% return).
656
After the 2005 proposed CSA regulations, the US approach changed. The
2007 CIP, while indicating that CSA funding should be remunerated, was
markedly ambiguous on how the return should be determined.2061 With-
out offering any explanation, the 2009 temporary US CSA regulations did
not allow any return on IP development funding contributions. One would
expect such a significant change from the proposed regulations to be com-
mented on.2062
The effects of the US position are particularly pronounced in generic cases

in which the foreign subsidiary contributes solely funding, which is priced
under the income method. The R&D funding contribution is simply ig-
nored, with the subsidiary only being allocated a normal market return
on its routine functions. The effects of the US stance will not be so pro-
nounced under the profit split method (PSM), as no single entity here will
be left completely stripped of residual profits.
In conclusion, the US CSA regulations simply do not remunerate IP devel-

opment funding. This is an aggressive – and likely effective – approach. As
the US and OECD regimes differ significantly on this issue,2063 it is clear
that the US CSA regulations provide no argument for the interpretation of
the OECD’s guidance on the determination of the risk-adjusted return. The
US solution, however, is most noteworthy from a de lege ferenda perspec-
tive. The author will revert to this in his concluding comments in section
22.4.11.
2061. See the discussion of the 2007 IRS Coordinated Issue Paper (CIP) in sec. 14.2.3.
2062. Precisely why the final 2011 CSA regulations (76 FR 80082-01) ended up not
remunerating funding is unclear. The preamble states, with respect to the residual profit
split method, that “market returns are not assigned to cost contributions because, under
this method, resources, capabilities, and rights that benefit the development of cost
shared intangibles (and thus make such development more valuable its costs) are com-
pensated as platform contributions”; see TD 9568, Explanation of provisions, E, 4.
In other words, the argument is that because operating profits are allocated to unique
development contributions (R&D, pre-existing intangibles, etc.), funding should not
receive any. The author struggles to see the logic. The fact that one particular type of
intangible development contribution is remunerated is obviously not an argument for
refusing to remunerate another type. It is unfortunate that this was not challenged. Per-
haps the complex nature of the CSA regulations allowed for this issue to “fly under the
radar” in the process towards the final regulations.
2063. Thus, double taxation may occur when the US enters into a treaty based on the
OECD model and the foreign jurisdiction claims entitlement to tax an arm’s length
funding remuneration.
657
22.4.8. Does the cost of capital contribute to

The current OECD TPG state that the risk-adjusted rate of return may be
determined by reference to the “cost of capital”.2064 The author will briefly
comment on this.
The determination of a return on risky R&D investments is not a mechani-

cal exercise. There is no financial method available that will yield an ob-
jectively “correct” result. When unrelated parties (e.g. VC funds) value
investments in early-phase, high-risk, single R&D projects for the purpose
of assessing whether they should invest or not, they normally do not base
their valuations on discounted cash-flow methodology, but rather on the
likelihood of the project being successfully developed and commercial-
ized.
Often, the so-called “venture capital method” is used,2065 which consists of

five steps, as follows:
(1) the operating result of the project for a future year is estimated. The
future year will typically be the year in which the venture capitalists
expect that they can sell their participation in the project through an
initial public offering or mergers and acquisitions;
(2) on the basis of the market pricing of comparable companies in the
same line of business, an earnings multiple is extracted; 2066
(3) the exit value is estimated by multiplying the estimated future earnings
by the market multiple; 2067
(4) the exit value is discounted to the present value, using the “target rate
of return” as the discount rate; and
(5) the net present value of the project is found by subtracting the IP devel-
opment investments from the present value of the exit value.

2065. Damodaran (2012), p. 646. See also Metrick (2010), at p. 178.
2066. An earnings multiple is extracted by dividing the market price of a project or
company by its earnings. E.g. if a company has a price-to-earnings ratio of 10, this
implies that it has a market value (equity) ten times its annual earnings.
2067. For instance, a VC fund invests in an early-phase, high-risk software develop-
ment project that is expected to be successfully exited in an initial public offering in 5
years. If the expected earnings from the project are estimated to be 4 in year 5 and the
market multiple extracted from comparable software companies is 25, the estimated
exit value will be 100.
658
The target rate of return is normally set at a significantly higher rate than
a conventional cost-of-equity discount rate, or a weighted average cost of
capital (WACC) for that matter. Leading financial literature suggests 30-
70%, depending on the stage of development.2068 In contrast, a traditional
risk-and-return model will rarely produce discount rates above 20%. The
target rate of return is determined through historical experience and plain
guesswork.2069
In conclusion, the cost of capital does not offer any firm benchmarks for
remunerating IP development funding contributions. The methodology,
however, does correspond to (and support) the above analogy (in section
22.4.5.) to VC financing, as well as the fact that a return of 30-70% may be
appropriate for high-risk projects.
22.4.9. Do the financing alternatives realistically available

contribute to the determination of the risk-adjusted
rate of return allocable to the funding entity?
The current OECD TPG state that the “financing options realistically
available to the party receiving the funds”, as well as the return that the
funder could get from a “realistic alternative investment with comparable
economic characteristics” should be taken into account when determining
the risk-adjusted rate of return allocable to IP development funding.2070 The
wording is ambiguous on several points.
First, the logic behind taking into account the perspectives of both parties
in this particular context is not immediately apparent. The whole point
behind the approach of the new OECD guidance on IP ownership – in
which the R&D entity performing the important development functions is
allocated the residual profits and the funding entity a risk-adjusted rate of
return – is to restrict the extraction of intangible profits from the jurisdic-
2068. Damodaran (2012), p. 647.

2069. Id. The target rates of return are high due to a combination of risk factors: (i)
stand-alone, early-phase projects are more exposed to macroeconomic risks than the
rest of the market; (ii) venture capitalists are often focused on particular sectors, such
as biotechnology and software, and thus are not diversified between business sectors,
and therefore demand a premium return for project-specific risks; (iii) the target rate
of return takes into account that the project may fail; And (iv) target rates of return for
venture capitalists may also be influenced by the fact that a high rate of return will pro-
vide a better bargaining position when entering into negotiations with an entrepreneur
for an equity stake in the project.
659
tion where IP value creation takes place. Thus, the problem is seen from
the point of view of the borrowing entity, i.e. the R&D entity resident in the
high-tax jurisdiction.
Why should the realistically available alternative return of the funding

entity matter in this assessment? One would assume that in third-party
contexts, the price that a borrowing entity will be willing to pay should not
be directly influenced by the lending entity’s investment options. Neverthe-
less, there is no getting around the fact that the guidance requires the alter-
native return of the funding entity to be taken into account in determining
the risk-adjusted rate of return allocable to it.
It will be important to clarify the risk of the R&D project being funded.
It will, in the case of contract R&D agreements and CSAs, often be that
the R&D pertains to the further development of established and successful
products with relatively low risk (second-generation R&D). In these cases,
it will be natural to use the “11% examples” (see section 22.4.6.) as the
point of departure for determining the realistically available return allo-
cable to the funding entity. In more impractical, high-risk R&D scenarios,
it must be assessed whether the funding entity actually has an alternative
investment available with a similar risk profile. That may very well not be
the case. If so, one should, in the author’s view, use the 11% “low-risk”
return, as that may provide the best estimate of the return of a realistically
available R&D investment.
Further, for the reasons stated above, it is the author’s view that the de-
termination of the risk-adjusted rate of return on IP development funding
should mainly be framed from the point of view of the borrowing entity.
This aligns with the way the realistic alternatives principle typically is ap-
plied in other transfer pricing contexts, such as IP valuation. In these other
applications, the pronounced focus lies on the point of view of the entity
performing R&D and contributing unique functions, as this party normally
will have a superior bargaining position vis-à-vis an entity that only con-
tributes funding. The interests of the funding entity are normally well ca-
tered to, as long as the end allocation result ensures that it is not put worse
off by entering into the funding transaction.2071
Second, the wording quoted at the beginning of this section does not indicate
whether the realistically available financing alternatives of the borrowing en-
2071. The author refers in particular to the discussion of the Pervichnyi example in sec.
13.5.
660
tity should be based solely on the borrowing ability embodied in the specific
R&D project being funded, or whether also the income, projects, functions
and assets connected to other business activities of the borrowing entity can
be taken into account. The realistic options available for procuring funding
will vary depending on this premise. If the assessment is restricted to solely
taking the commercial prospects of the specific IP development project into
consideration, disregarding all other income and assets of the borrowing enti-
ty, that would likely lead to the conclusion either that funding simply was un-
available for a project that was particularly risky or that the cost of financing
would be set relatively high, typically on par with a VC target rate of return.
This could lead to a scenario in which the funding entity would, more or
less, always be allocated a relatively high return on its funding contribu-
tion, regardless of whether the borrowing entity is highly liquid and solvent
and easily could have commanded a relatively low market interest rate on
third-party loans on its own. Such a result would be contrary to the arm’s
length principle, as there would not be parity in the prices applied between
related and unrelated parties.
The author therefore concludes that the determination of the return allocable
to a funding entity should be based on the realistic options available to the
borrower entity, also taking into consideration the business income and assets
of that entity that are not connected to the specific R&D project being funded.
Third, and as an extension of the above, the OECD guidance on this point
does not provide any direction as to whether the realistic funding alter-
natives available should encompass those that are contingent on the bor-
rowing entity’s affiliation with its group, i.e. whether funding alternatives
that rely on the business income and assets owned and controlled by other
group entities are relevant.
An inherent feature of the arm’s length principle is, of course, the separate
entity approach. Each entity within a group should, for transfer pricing pur-
poses, be treated as if it were not a member of the group. This paradigm,
however, has been relaxed in some contexts. For instance, the OECD TPG
take the position that a controlled service transaction should not be deemed
to have occurred – with the result that compensation is not required – when
a group entity obtains “incidental benefits attributable solely to its being part
of a larger concern, and not to any specific activity being performed”.2072
2072. See OECD TPG, para. 7.13, which distinguishes passive association that is not
compensable from active promotion that is. See also Treas. Regs. § 1.482-9(l)(3)(v),
661
Further, synergistic benefits that arise purely as a result of group member-

ship without the deliberate concerted action of group members or the per-
formance of any functions are not required to be separately compensated
or specifically allocated among members of the group.2073
This rule is illustrated in an example pertaining to a subsidiary that is able

to procure external financing on better terms than it would have been able
to on its own, purely as a result of its consistently AAA credit-rated par-
ent.2074 The subsidiary borrows an amount from a third-party lender on the
beneficial terms. It then borrows an identical amount from a group entity.
The interest rate on the latter loan is deemed to be at arm’s length because
it is the same rate charged by an unrelated party in a comparable loan
transaction, as well as because the synergistic benefit of the low interest
rate is a result of the “group membership alone and not from any … con-
certed action” of the group members.2075
In this, the author finds a solid basis for asserting that the borrowing ca-
pacity indicated by the business and assets of other group entities should
be taken into account. If this would not be the case, there would be incon-
sistency between the provisions for pricing funding agreements in general
and those pertaining to funding in the specific context of IP development.
Synergetic benefits would then be persuasive for the pricing in the former
category of transactions, but not the latter. The author finds no good reason
to accept such inconsistency.
This imposes a severe limitation on the returns allocable for funding con-
tributions. For instance, it may be that a borrowing entity alone, and based
solely on the commercial merits of the specific R&D project being funded,
would only be able to procure financing on VC terms, for instance, with an
which states that a taxpayer is not considered to obtain a benefit when the benefit results
from passive association with a group. See also § 1.482-9(l)(5), Examples 15-19.
2073. OECD TPG, para. 1.158.
2074. OECD TPG, para. 1.164.
2075. OECD TPG, para. 1.166. The example is unclear in the sense that it does not
specifically state whether the low interest rate would be deemed a compensable benefit
for the subsidiary if there was no CUT. E.g. if the subsidiary had beneficial third-party
financing readily available but no transaction was actually entered into with the third
party, would the interest rate on a controlled loan arrangement that corresponded to the
interest rate available on third-party financing be considered a compensable benefit if it
were clear that the interest rate was lower than the subsidiary would have been able to
procure had it not been a member of the group? The general OECD TPG position that
benefits that arise solely due to group affiliation are not compensable clearly indicates
a negative conclusion. However, due to the specific inclusion of the comparable uncon-
trolled loan agreement in the example, that conclusion cannot be regarded as certain.
662
annual rate of 70%. If, however, the entity (taking its group affiliation into
consideration), would be able to borrow for 10% from unrelated parties,
the latter rate should be decisive.
22.4.10. Does the financial risk assumed contribute to

The 2014 OECD TPG intangibles chapter draft text mentioned that “the
financial risk assumed by the funding entity” should be taken into account
to determine the risk-adjusted rate of return allocable to IP development
funding.2076 Even though this parameter is not positively listed in the final
consensus text of the current OECD TPG, the author finds it clear that it
may nevertheless be taken into account, as the list is not exhaustive.
It is not immediately clear to the author what the 2014 draft meant by
“financial risk”. The OECD TPG describe several types of risks that, in
general, are important in a functional analysis of controlled IP transfers,
including those pertaining to R&D, infringement, product obsolescence
and liability.2077 The key risk here must be that the R&D efforts prove un-
successful, with the consequence that the investment is lost, along with the
potential returns. This will be the relevant risk for the funding entity and is
the same risk faced by unrelated VC investors.
The return attributable to the funding entity must be commensurate with

the degree of this risk.2078 The author will tie some reflections to this. Con-
sistent with his stance above, he will use the target rate of return of 30-70%
for unrelated passive financial investors in VC funds as a point of reference
for the upper limit for funding remuneration, relevant only for relatively
high-risk investments. Alternatively, if the funding should be seen as a
relatively low-risk investment, this could indicate that the funding entity
should be allocated a return significantly lower than a typical VC financial
investor return, e.g. the 11-12% return indicated by the new OECD exam-
ples.2079 It goes without saying that this assessment should be carried out
with a sceptical frame of mind, as there is an obvious danger that R&D
2076. 2014D, para. 6.61.

2078. It will not, however, affect the deemed obligation of the borrower entity to repay
the principal amount.
663
jurisdictions, on the one side, and funding jurisdictions and multinationals,

on the other, will default to low and high-risk assertions, respectively.
Although not exhaustive, the following factors should be relevant for the
determination of whether particular R&D funding should deemed as a
relatively high or low-risk investment: (i) the funding entity’s control over
early-phase, high-risk R&D; (ii) the inherent risk of the particular R&D
project; and (iii) whether the funding is a single project investment or part
of a portfolio.
With regard to factor (i), a pure funding entity will not be in control of
the important R&D functions.2080 This is, to a certain degree, parallel to
limited partners in VC funds that are not involved in the active day-to-day
management or control of the backed projects. Nevertheless, the general
partner of a VC fund has a custodial responsibility towards the limited
partners with respect to choosing the right projects and supervising their
progression. LPs in VC funds are therefore indirectly in some control over
the important decisions pertaining to the further development of each sup-
ported project.
More importantly, the limited partners in VC funds decide whether they

want to contribute more funds to the development of the supported projects
on a concurrent basis as the need for additional cash flow arises. If they
deem the risk to be too high in relation to the possible benefits, they will
simply halt funding. Investments in early-phase, high-risk IP development
would likely never be undertaken by third parties without a severe degree
of control over how the investments were spent, combined with concurrent
monitoring of the progress made and results achieved.
The author would therefore assert that the mere existence of a pure funding
entity is an argument for restricting its funding remuneration to a relatively
low-risk return. The logic behind this is that this entity has no true basis
for exercising control over its investment. No rational third-party investor
would likely be willing to enter into such an agreement unless it pertained
to a relatively low-risk investment. Therefore, the existence of group fund-
2080. It will, in principle, depend on the controlled agreement between the funding
entity and the entity incurring intangible development costs as to whether the funding
entity has the authority to exercise control over the development process. If the funding
entity has such authority, however, it will likely also perform some of the important
development functions and thus be entitled to a portion of the residual profits. For the
purpose of this discussion, the author therefore assume that the funding entity has no
such authority.
664
ing entities with no control or influence over the R&D processes they are
funding should logically indicate that the underlying R&D risk is in fact
relatively low.
With regard to factor (ii), an opinion on the inherent R&D risk must be
formed in order to uncover the derivative financial risk. The 2014 draft
accentuated the importance of the timing of the funding for the level of fi-
nancial risk.2081 How far the research has been developed (and in particular,
how close it is to commercialization), combined with its profit potential,
will be important for the chance of attracting third-party funding. Even
VC funds tend to avoid investing in research without immediately apparent
practical applications, regardless of how promising it may seem.
If the controlled R&D is genuine, high-risk, early-stage, “blue sky” re-

search, this will justify deeming the funding a relatively high-risk invest-
ment. Typical transfer pricing structures in the form of contract R&D and
CSAs, however, will often not pertain to such “made from scratch” pro-
jects. The multinational’s experience, know-how and pre-existing IP form
the basis for a continuum of gradually developing knowledge, where suc-
ceeding products often are improvements on the previous ones, such as
the case is for Apple’s line of iPhone mobile phones. Over time, gradual
improvements may collectively alter the core product, but that is not likely
to happen from version to version. Xilinx, Veritas and BMC all pertained to
factual patterns in which the cost-shared IP development processes, in real-
ity, were nothing other than straightforward continuations of established,
US-based R&D for existing and profitable products.2082
In such cases, the R&D risk will likely not be comparable to that of early-
phase, high-risk IP development, and these instances should therefore be
classified as relatively low-risk investments. The return attributable to the
2081. 2014D, at para. 6.60. This text was added first in the 2014 draft version of the
guidance.
2082. In Xilinx Inc. and Subsidiaries v. CIR, 125 T.C. No. 4 (Tax Ct., 2005), affirmed
by 598 F.3d 1191 (9th Cir., 2010), recommendation regarding acquiescence AOD-2010-
03 (IRS AOD, 2010), and acq. in result, 2010-33 I.R.B. 240 (IRS ACQ, 2010), pre-ex-
isting IP in the form of rights to integrated circuits and development software systems
were transferred to a CSA with an Irish manufacturing and distribution subsidiary for
the European market. In Veritas Software Corporation & Subsidiaries v. CIR, 133 T.C.
No. 14 (U.S. Tax Ct. 2009), IRS nonacquiescence in AOD-2010-05, a US parent con-
tributed pre-existing IP in the form of existing storage management software to a CSA
with an Irish distribution subsidiary. Finally, in BMC Software Inc. v. CIR, 141 T.C. No.
5 (Tax Ct., 2013), reversed by 780 F.3d 669 (5th Cir., 2015), a US parent entered into
a CSA with a European holding subsidiary with respect to existing software solutions
owned by the parent.
665
funding of further R&D for mature “milk cow” products should be signifi-
cantly lower than the general return to financial VC investors.
It might also be that a contract R&D arrangement or CSA is entered into for
which there is no existing commercialized product based on pre-existing
IP, but for which the know-how and unique skills of an experienced R&D
team in place with a good track record will be an obvious value driver. It is
the author’s view that investments into such projects also should be seen as
relatively low-risk investments. In such cases, the return attributable to the
funding should be lower than the general return to financial VC investors.
With regard to factor (iii), it must be ascertained whether the funding per-
tains to a portfolio of “diversified” R&D projects or to a single project. VC
investors invest in funds that contain a range of start-up projects at differ-
ent stages of development, pertaining to different products with different
growth potential and risk profiles. In this respect, a diversification benefit
is achieved. The investment risk connected to a single project is mitigated,
to some extent at least, by the inclusion of other projects in the investment
portfolio. Conversely, contract R&D arrangements and CSAs are normally
entered into for specific projects.2083
As a point of departure, one would logically expect an investor to demand

a higher return on an investment in a specific early-phase, high-risk project
than for a portfolio of such projects, due to the lack of diversification and
the resulting increase in risk. This logic, however, is founded on the prem-
ise that there is asymmetrical information between the transacting parties.
The author is sceptical of applying such reasoning for the purpose of al-
locating income among related parties. Due to the perfectly symmetrical
information between the parties to a controlled arrangement, there will
be an obvious risk of cherry-picking high-profit-potential R&D projects
for structuring through contract R&D arrangements and CSAs in order to
extract residual profits from the source jurisdiction.
The question should, in the author’s view, therefore be reversed. Does the
fact that the controlled arrangement includes only a specific project not
indicate that its funding risk is lower than the typical VC-financed, ear-
ly-phase, high-risk projects? The economic substance of such controlled
schemes may be that only the most promising research is selected for the
purpose of profit shifting. If such an observation can be made and it is sup-
2083. See, e.g. Treas. Regs. § 1.482-7(d)(1)(i) on the definition and scope of the intan-
gible development activity covered by the CSA.
666
ported by the analysis of the inherent R&D risks, the funding should likely
be seen as a relatively low-risk investment and be remunerated accordingly.
The return attributable to the funding entity in such cases should be lower
than the general return to financial VC investors.
22.4.11. Concluding comments on IP development funding
In this section, the author will summarize his interpretation of the OECD’s
guidance on the allocation of profits for IP development funding and tie
some de lege ferenda reflections to the rule.
The allocation of financial risk to a funding entity will be respected if

it controls the risk. The low threshold for establishing control will likely
result in only foreign “post-box” funding entities being cut off from a risk-
adjusted rate of return. These will nevertheless receive a risk-free rate of
return. In the vast majority of cases, group-funding entities will be entitled
to a risk-adjusted rate of return.
The extent of this return depends on a two-fold assessment: the return

must (i) mirror that of “similar funding arrangements among independent
entities”; 2084 and (ii) be restricted to a “risk-adjusted return”.2085
The author has argued that the remuneration required by passive third-
party financial investors in VC structures, ranging from 30-70% of the
invested amount, could form a meaningful point of departure for deter-
mining the operating profits allocable to the funding of early-phase, genu-
inely high-risk R&D. Contrary to such relatively high-risk investments,
the projects selected by multinationals for development through contract
R&D schemes and CSAs often pertain to the further development of IP
connected to established and successful products, with the R&D based on
pre-existing IP and carried out by the same experienced and proven R&D
team that developed the first-generation IP. The funding of such projects
should be seen as relatively low-risk investments. The examples included
in the current OECD TPG suggest that such investments could be remuner-
ated with a rate of approximately 11% of the invested amount.
The question then becomes how to distinguish between relatively high and
low-risk funding endeavours. The author argues that the degree of financial

667
risk is driven by (i) the funding entity’s control over the R&D risk; (ii) the
inherent risk of the particular R&D project; and (iii) whether the funded
project is part of a portfolio of R&D projects or should be assessed on a
stand-alone basis. The author finds that factors (i) and (iii), in the context of
typical contract R&D arrangements or CSAs, should lead to the conclusion
that IP development funding is a relatively low-risk investment. The same
goes for factor (ii), except for in cases in which it is clear that the funded
R&D project in fact pertains to genuinely high-risk, early stage, “blue sky”
research. Further, the author has argued that the realistic alternatives avail-
able, in particular to the R&D entity being funded, often should entail that
the funding is seen as a relatively low-risk investment.
Thus, the author’s interpretation of the current OECD TPG entails that
IP development funding should normally be seen as a relatively low-risk
investment and remunerated accordingly. While the specific level of profit
allocation must be assessed concretely, it is the author’s view that the 11%
rate indicated by the OECD examples should be guiding. Due in particular
to the restriction imposed on the profit allocation by the realistic alterna-
tives of the controlled parties, the author is not convinced that it will be
practical to remunerate funding on the premise that it is a relatively high-
risk investment. Should it, in specific and narrow circumstances, be appro-
priate to do so, the third-party VC remuneration level of 30-70% ought to
be guiding in the absence of true CUTs.
In the end, however, the wording contained in the current OECD TPG on
the risk-adjusted rate of return allocable to a funding entity is ambiguous,
leaving room for divergent interpretations.2086 It has likely been difficult
to achieve consensus on more specific directions.2087 Conflicting views on
the allocation of profits to IP development funding could lead to double
taxation.
R&D jurisdictions will likely default to low-risk characterizations of fund-

ing, while the funding jurisdiction and the multinationals will default to
high-risk characterizations. Even if there is agreement on the risk charac-
2086. See also Andrus et al. (2017), at p. 104, where it is recognized that the current
IP development funding rules are too unclear and that further work to clarify the rules
will be necessary. In this direction, see also Storck et al. (2016), at p. 217.
2087. An indication of this is that the generic “11%” example was introduced late in
the BEPS process. It first emerged, in bracketed form, in 2014D. Substantially the same
example was used in the Public Discussion Draft, BEPS Action 8: Revisions to Chapter
VIII of the Transfer Pricing Guidelines on Cost Contribution Arrangements (OECD
2015) [hereinafter 2015 OECD CSA draft], also in bracketed form.
668
terization, it will likely be problematic to agree on an appropriate level of

remuneration. For instance, the VC third-party target rate of return repre-
sents an extremely large interval, stretching from 30% to 70%. Further, is
the 11% rate indicated by the OECD in the 2015 examples truly appropri-
ate for the generic scenarios? Unfortunately, no objectively correct answer
exists.2088
Having said this, the author finds that the negative sides of the guidance
on the remuneration of IP development funding contained in the current
OECD TPG must not be exaggerated.
First, ambiguity and inherent imprecision with respect to the amount of op-
erating profits allocable is not specific to the guidance on IP development
funding. It is a recurring theme in all transfer pricing, fundamentally due
to the lack of CUTs for the type of transactions carried out by multination-
als. Realistically, this is just a problem one must live with.
Second, the OECD approach to funding is sound in theory. It requires that

funding, in principle, should be treated equally as all other IP development
contributions, in the sense that such inputs should also qualify for an arm’s
length allocation of profits.
Third, the OECD approach attempts to remunerate IP development fund-

ing separately, based on a tailored analysis of the profit potential and risks
in each R&D project. This will likely yield a more reliable result than if
the funding entity were remunerated through, for instance, a profit split.2089
Fourth, the inherent feature that the funding return must be calculated on
the basis of a nominal investment amount will, to some extent, mitigate
the negative effects of the ambiguity pertaining to the determination of an
arm’s length, risk-adjusted rate. For instance, assume that 100 is invested in
2088. This problem is largely analogous to domestic law attempts to apply the general
arm’s length standard to benchmark controlled interest rates on the internal debts of
multinationals, which has resulted in widespread adoption of domestic mechanical thin
capitalization rules.
2089. It would be difficult to assess the relative value of pre-existing intangibles and
R&D against funding in order to perform a profit split. Such a comparison would be
akin to comparing apples and oranges. Draft versions of Examples 16 and 17 indicated
that a profit split approach should also be adopted to remunerate a funding entity (see
2013D, Example 13, para. 274; 2013D, Example 14, para. 278; 2014D, Example 17,
para. 60; and 2014D, Example 18, para. 64), but this was fortunately scrapped in the
final 2017 consensus text.
669
a risky R&D project over 5 years and that the arm’s length return is deter-
mined to be 30%. The funding entity will not be entitled to any more after
it has been allocated 30 on top of the invested amount. This is important,
as the value of unique IP will often not be correlated with the investment
costs necessary to create it. Thus, the residual profits may be significantly
larger than the return allocable to funding and may be generated long after
the funding remuneration has been completed. This in itself entails that the
funding guidance offers some resistance against profit shifting.
The big question, however, is whether the current OECD approach to IP

development funding remuneration offers a sustainable solution for avoid-
ing BEPS. The core ambition of the 2017 OECD approach to IP ownership
is to align the allocation of intangible profits with intangible value creation.
This is operationalized through the important-functions doctrine, assign-
ing residual profits from internally developed IP to the R&D jurisdiction.
The route taken by the OECD on funding may prove detrimental to the
realization of this ambition.
Risk-willing capital is necessary to create intangible value. While they are

not as scarce or unique development inputs as, for instance, the patent for
a blockbuster drug or the Apple logo, risk-willing capital must be remuner-
ated with a clearly higher rate of return than the normal market return al-
locable to pronouncedly generic inputs, such as administrative or fiduciary
functions. This is the crux of the current OECD approach to IP ownership.
In the author’s view, there is no doubt that the primary purpose behind
multinationals’ use of foreign funding entities, regardless of whether the
controlled structure is organized as a contract R&D arrangement or a
CSA, is to extract operating profits from taxation at source. The author
doubts that it will be onerous for a multinational to set up a foreign fund-
ing entity with enough substance to be allocated financial risk and to
channel IP development funding through it.2090 Thus, even if the impor-
tant functions doctrine allocates the residual profits to the R&D jurisdic-
tion, much of these profits can be extracted by way of a risk-adjusted
return to the foreign funding entity, significantly impairing the effect of
the doctrine.
2090. See also Osborn et al. (2017), under sec. IV.A-C, for a practical example of a
tax-planning package designed to “fill” up a cash-box entity with enough substance
not only to attract a risk-adjusted rate of return for its funding of R&D, but also some
residual profits from the IP developed, by way of ensuring that the cash box controls
financial risks and some R&D risks.
670
Given that the goal is to allocate IP profits to the jurisdiction in which they
were created, it could be argued that (i) the shifting of profits out of an
R&D jurisdiction to a funding jurisdiction by way of funding remuneration
is unworthy of protection; and (ii) the best would be to disallow it entirely,
as the United States has done in the context of CSAs.
The basic idea that all IP development inputs should receive appropriate
compensation, however, lies at the very heart of the arm’s length princi-
ple. This is reflected in the design of the OECD’s profit allocation rules. It
would contradict this system to single out one particular type of IP devel-
opment input for which remuneration is refused. To do so would be to take
a big step away from a coherent and logical profit allocation system.
The reason why the United States, with all of its experience with the trans-
fer pricing of IP, has chosen to cut off funding remuneration in the context
of CSAs is probably twofold. First, reductions of the domestic tax base
through what is, realistically, outright profit shifting by multinationals,
simply will not be tolerated. Second, the United States likely views unique
development inputs (in the form of pre-existing IP, R&D results, talented
R&D teams in place, etc.) as the true intangible value drivers. Regardless
of what motivates it, it is clear that the US position conflicts with the basic
notion that all IP development inputs should be remunerated at an arm’s
length level.
In this light, the OECD’s solution represents a compromise. In cases in

which the economic substance of the controlled transaction does not sup-
port the assignment of financial risk to the funding entity, the OECD does
not (contrary to the US regulations) cut off funding remuneration entirely,
but restricts it to a meagre risk-free rate of return. Outside of these narrow
“post-box entity” cases, however, a risk-adjusted return must be provided,
leaving the door open for profit shifting. The result is a rule far less restric-
tive than its US counterpart.
Nevertheless, the author finds the OECD compromise to be well founded.

After all, the OECD has managed to stand its ground and uphold the in-
tegrity of the arm’s length transfer pricing system by affirming that all IP
development inputs should receive an arm’s length return. Only time will
tell as to whether it is possible in practice to uphold such a principled stance
while at the same time realizing the allocation intention of the important-
functions doctrine. With that said, the author finds it likely that the OECD
position may, in practice, prove too lenient, allowing for BEPS, and that it
will need to be revised in the future. If so, what then?
671
As the author sees it, there are two main alternative routes available. First,
the OECD could, as the US CSA regulations have done, cut off the funding
entity entirely, regardless of whether the entity incurs financial risk. That
would, in the author’s view, breach the arm’s length principle, resulting in
an incoherent transfer pricing system. However, it would be an effective
weapon against BEPS.
Second, the OECD could – as some suspected they would do already in

the latest revision round – mechanically “cap off” the funding remunera-
tion at a specific and restrictive level (for instance, at 3, 5 or 7%) of the
funded amount, regardless of the risk of the R&D efforts. The author fa-
vours this alternative. Such a rule would be rather analogous to domestic
thin capitalization rules for intra-group financing transactions, which have
proven effective. It would strike a reason able balance between the need
to respect the arm’s length transfer pricing system by remunerating all IP
development inputs and ensuring that the extent of IP profit extraction by
multinationals from R&D jurisdictions is contained at sustainable levels.
Not to mention, the problems associated with applying the current ambigu-
ous OECD funding guidance would become a thing of the past. The main
problem with this type of allocation rule would be its lack of flexibility.
All R&D projects would be remunerated with the same rate of return, irre-
spective of inherent risk. That would not align well with how third parties
price analogous investments, and therefore would not reflect a “true” arm’s
length rate. In the author’s view, however, this would be a relatively modest
price to pay for a more effective profit allocation system.

Pre-existing unique IP
In sections 22.4.10.-22.4.11., the author discussed the provisions of the cur-

rent OECD TPG with respect to the remuneration of IP development con-
tributions in the form of important functions and funding, which govern
the allocation of residual profits and a risk-adjusted return, respectively.
In cases in which the IP development is truly “blue sky” research, in the
sense that it is not based on any existing R&D results, know-how, etc., the
focus will be on these two development inputs. Nonetheless, R&D carried
out by multinationals will often not start from scratch, but will build upon
and enhance current R&D results (embodied in patents, know-how, etc.).
672
Profit allocation for IP development contributions: Pre-existing unique IP
Thus, a third development input is introduced: pre-existing unique IP. The

question then becomes of how much of the profits from the subsequently
developed IP shall be allocated to this IP development contribution, taking
into account that the two other development inputs (important functions
and funding) also shall be ensured their share of the profits (residual profits
and a risk-adjusted rate of return, respectively).
22.5.2. The point of departure for pricing the contribution

of the pre-existing IP
The wording contained in the OECD TPG on the contribution of assets to

IP development is geared towards funding.2091 Contributions of other types
of assets are hardly discussed. Nevertheless, it is, in principle, clear that
all assets contributed to IP development should receive an “appropriate
compensation”.2092 The wording indicates that the transfer pricing methods
should be used. This will likely result in the application of unspecified
versions of either the transactional net margin method or the PSM, as the
specified methods cannot be applied directly in the context of IP develop-
ment. The latter methods are designed to allocate operating profits among
the inputs to the intangible value chain (as discussed in part 2 of the book),
not among the contributions to the development of a specific value chain
input (in this case, unique IP).2093
For instance, it could be that the pre-existing IP and ongoing R&D are
contributed by one group entity, while another provides funding. The latter
entity could then be treated as the tested party and allocated a return on
2091. OECD TPG, paras. 6.59-6.61.

2092. OECD TPG, para. 6.59. As indicated, the author’s discussion is limited to the
contribution of unique pre-existing IP. Routine assets (laboratory gear, plant and equip-
ment, generic software applications, etc.) should be remunerated with a normal return
pursuant to the one-sided methods, likely the cost-plus method. In other words, these
assets, as opposed to unique pre-existing IP, will not be allocated residual profits, sim-
plifying the remuneration of them to some extent.
2093. The TNMM allocates a normal market return to the tested party based on the
operating margins realized by comparable enterprises. Because it will not be possible
to find such margins that include IP funding remuneration, an unspecified version of
the method must be used (this was also the problem with applying the CPM directly
to allocate income to the funding entity in a CSA prior to the introduction of the US
income method; see the discussion of the 2007 CIP in sec. 14.2.3.). Its aim will be to
allocate the risk-adjusted rate of return for the funding contribution, in addition to the
normal TNMM margin as compensation for routine functions, to the tested party. The
same will apply for the first step of the residual profit split method. See the analysis of
the TNMM and OECD PSM in ch. 8 and ch. 9, respectively.
673
its funding contribution. The residual profits are allocable to the former
entity alone. The challenge is to determine a reliable funding remuneration.
Alternatively, the pre-existing IP could be contributed by entity 1, ongoing
R&D by entity 2 and funding by entity 3. In this case, the PSM could be
applied, resulting in the allocation of a “normal return” to entity 3 for its
funding and a split of the residual profits between entities 1 and 2. Both
steps will likely be problematic to carry out reliably.
A practical scenario is that the R&D is organized through a CSA. The

CSA wording in the current OECD TPG supplements the intangibles
and services chapters.2094 It provides no detailed guidance on the pric-
ing of pre-existing IP contributed to a CSA. Instead, the basic position
is that the pricing should be “based on [the pre-existing IP’s] value …
in order to be consistent with the arm’s length principle”.2095 Further, the
fact that the controlled R&D process is organized through a CSA does
“not affect the appropriate valuation of the separate contributions of the
parties”.2096 This provides a firm basis for asserting that the pricing of
pre-existing IP contributed to the development of a new intangible should
be the same regardless of whether the R&D is organized through a CSA
or other structures, typically an R&D contract research agreement. One
will ultimately end up in the intangibles chapter of the OECD TPG in
any case.2097
Both the important-functions doctrine and the guidance on funding remu-

neration look towards current R&D and financing. Conversely, the transfer
pricing methods applied to price the contribution of pre-existing IP will
allocate profits to the group entities that created the pre-existing intangible.
The author will tie some comments to the coordination of these rules in
the following sections. He will discuss the allocation of profits for when
the group entity that performs the current important R&D functions is the
same as or different from the entity that contributes the pre-existing intan-
gible in sections 22.5.3. and 22.5.4., respectively.

2097. The absence of detailed instructions on how to price the contribution of a pre-
existing intangible to a CSA stands in stark contrast to the US CSA regulations. While
this principled OECD stance was likely necessary in order to achieve consensus, the
author suspects that it may prove difficult to apply the ordinary transfer pricing meth-
ods uniformly in the context of intricate CSAs. He refers to his discussion of CSAs in
ch. 16 for further comments.
674
22.5.3. The pre-existing IP is contributed by the same group

entity that carries out the ongoing R&D functions
In this scenario, it will be easy to identify which jurisdiction should be

allocated the residual profits, as both unique development inputs (pre-
existing IP and ongoing R&D) are contributed by the same group entity.
Still, it must be distinguished between the residual profits allocable to the
pre-existing IP and the ongoing R&D because the timing of their profit al-
location will differ, and it must be ensured that the R&D jurisdiction is not
allocated the residual profits allocable for the pre-existing IP twice. These
problems can be illustrated through two examples outlined in the OECD
TPG.
The first example pertains to a parent (Shuyona), which performs impor-

tant R&D functions and controls the R&D carried out by its foreign subsid-
iary S in jurisdiction Y.2098 The parent transfers valuable, self-developed IP
to foreign subsidiary T in jurisdiction Z for an arm’s length fee and enters
into a contract R&D arrangement under which subsidiary T funds ongoing
R&D and is assigned all residual profits from the developed IP. The parent
and subsidiary S continue their R&D activities as before the transfer.
The 2014 draft version of the example found that (i) the parent was entitled
to compensation for its important R&D functions; (ii) subsidiary S was en-
titled to compensation for R&D (as a service provider); and (iii) subsidiary
T was entitled to compensation for its investment in the acquired intangi-
bles and for funding on-going R&D. The example concluded:
[I]t may be extremely difficult or impossible to identify comparable transac-
tions with such a structure and use of profit split methods, valuation tech-
niques, or other methods may be necessary to identify the appropriate level of
compensation to Shuyona for its functions, assets and risks.
Thus, the 2014 draft version provided no clear direction for the profit allo-
cation. The quoted text was omitted in the final 2017 consensus version of
the example (included in the current OECD TPG), which focuses solely on
2098. OECD TPG, annex to ch. VI, Example 16. The text in the 2013 and 2014 draft
versions of the example is identical. The example was bracketed in the 2014 version,
indicating that there was no consensus on the text. The example was also included in
the 2014 discussion draft on the use of profit splits in the context of value chains. The
final 2015 version of the example focuses on the remuneration of the intangible devel-
opment funding contributions. The author discusses this example in sec. 22.3.3.3. on
the important-functions doctrine and contract R&D.
675
the remuneration of the IP development funding contribution.2099 The ques-

tion is how the contribution of the pre-existing IP shall be remunerated.
The factual pattern is significantly complicated by the transfer of the pre-

existing IP. Without it, the analysis would be relatively straightforward. All
pre-existing IP was created by the parent, which also performs the ongoing
important R&D functions. Thus, the residual profits should be allocable
to it alone under the rationale that the intangible value was created in its
jurisdiction. Subsidiary S should be allocated a normal return, likely based
on the cost-plus method, while subsidiary T should receive a risk-adjusted
return for its funding contribution, provided that it is in control of the finan-
cial risk.2100 This is the result of a direct application of the OECD guidance
on important functions and funding. The end allocation result cannot be
altered by the fact that there is a transfer of the pre-existing IP. After all,
the substance of the controlled transaction remains the same: all important
R&D functions, past and present, are carried out by the parent.
The transfer, however, entails that the parent, through the price paid by
the subsidiary, is allocated the estimated value of the transferred IP at the
time of the transfer. Thus, it should not also be allocated the residual profits
that later are realized through the exploitation of the transferred IP based
on the argument that it performs the concurrent R&D functions. To do so
would mean that it would be allocated the same value twice. That would be
nonsensical and contrary to the arm’s length principle.2101
Further, the example states that the transferred IP rights include the right to
use the pre-existing IP for the purpose of further research. This is distinct
from the exploitation rights and must be priced separately.2102 Let us say
that the transferred IP forms the basis for R&D of a new version. Further,
40% of the estimated value from the exploitation of the new intangible
is due to the predecessor IP. The 40% will be the arm’s length price for
the transfer of the pre-existing IP. When these 40% later materialize, they
must be allocated to subsidiary T, or otherwise the parent would get double
pay for the transfer of the same rights.
2099. See the discussion in sec. 22.4.3. on control over financial risk.
2100. The final 2017 version of the example concludes that subsidiary T is not in con-
trol of the financial risk and therefore should be allocated a risk-free return only.
2101. See also Ballentine (2016), where the conclusion is reached that “T cannot pay
Parent a price equal to the present value of all future cash flows and then also share
those same cash flows with Parent. The result is a negative expected present value to T,
which it would not accept at arm’s length”.
2102. In practice, the exploitation rights and the right to use the intangible for the pur-
pose of further R&D would most likely be priced together.
676
The second example pertains to a similar scheme.2103 A parent carries out

R&D and transfers patents and other IP connected to an existing product
to a foreign cash-box entity without technical personnel for an arm’s length
fee. A contract R&D agreement is then entered into, under which the par-
ent is to carry out R&D as before. R&D funding comes from the subsidi-
ary, which is also assigned the residual profits from the exploitation of the
developed IP. In the author’s view, there cannot be any doubts as to the end
allocation result. Only the parent performs important R&D functions. The
R&D jurisdiction should therefore be entitled to the residual profits, while
a risk-adjusted return on the funding contribution is allocated to the “fund-
ing jurisdiction”.2104
The allocation of profits to the transferred pre-existing IP is complicated

by the fact that they are only partially developed and not yet commercial-
ized. The example emphasizes that the fee relates “strictly to the existing
intangibles and does not include compensation for future R&D services”
of the parent.2105 Nevertheless, this is, in principle, the same type of as-
sessment as in the first example. The point is to single out the transfer of
the pre-existing IP and remunerate this separately from the ongoing R&D.
Nevertheless, when the transferred IP is only partially developed, this line
becomes blurry.
22.5.4. The pre-existing IP is contributed by a group entity

different from that which carries out the ongoing
R&D functions
In this scenario, the residual profits will be split between two jurisdic-
tions, as the unique development inputs (the pre-existing intangible and
2103. OECD TPG, annex to ch. VI, Example 17. The text in the 2013 and 2014 draft
versions of the example is identical. The example was bracketed in the 2014 version,
indicating that there was no consensus on the text. The example was also included in
the 2014 discussion draft on the use of profit splits in the context of value chains. The
final 2017 version of the example remains largely the same as the draft versions, but
focuses on the remuneration of IP development funding, and the reference to a possible
profit split solution is now removed. For comments on Example 17, see Musselli et al.
(2017), at p. 333.
2104. Also here, “double counting” must be avoided. Thus, when the estimated value
paid for later materializes through the exploitation of the developed IP, it must be allo-
cated to the funding jurisdiction so that the same value is not taxable twice in the R&D
jurisdiction.
2105. This clarification was introduced in the version of the example contained in the
2014 draft, which was otherwise similar to the 2013 version.
677
the R&D) are contributed by different group entities.2106 Thus, in addition

to the timing and double-counting factors discussed in section 22.5.3., the
allocation takes on an additional dimension. The difficult-to-assess line
between intangible value caused by the pre-existing intangibles and the
ongoing R&D is decisive for the amount allocable to the R&D jurisdic-
tions. The jurisdiction in which the pre-existing IP was created will likely
argue that most of the value of the intangible developed through the CSA
or contract R&D arrangement comes from the pre-existing IP. The other
(R&D) jurisdiction will likely claim that ongoing R&D is the main value
driver. The assessment will be pronouncedly difficult if the pre-existing IP
is only partially developed.2107
For instance, assume that a US entity, through its own R&D and funding,
has developed an innovative and successful software program. A Norwe-
gian group entity, operating autonomously, overtakes the software develop-
ment functions, building extensively on the original version. The US entity
is not involved in further R&D. The structure is set up so that the US entity
enters into a CSA with an Irish subsidiary and the Norwegian company.
All three entities contribute ongoing R&D funding in proportion to the
benefits they expect to derive. The residual profits from the new version
are allocated accordingly. In addition to its part of the funding, the US en-
tity contributes the pre-existing software code to the CSA. Let us assume
that 90% of the value of the new version is attributable to the pre-existing
software.
How should the residual profits from the new version be allocated pursuant
to the current OECD TPG? A proper allocation, in the author’s view, may
only be attainable through a layered analysis.
First, it must be ascertained as to which group entities are entitled to the

residual profits from the pre-existing IP. This assessment should – in par-
allel to an application of the PSM – take into consideration all functions,
assets and risks contributed to the development of the pre-existing IP. The
US entity alone developed the original software in the example. Had other
2106. A premise for this discussion is that the group entities are resident in different
jurisdictions.
2107. With respect to the profit split assessment under the specified OECD PSM, the
author argues that in cases in which more than one unique input has caused the intan-
gible value and these inputs were necessary components of the value creation, the best
solution might be to divide the residual profits equally among them in the absence of
reliable evidence on causality indicating that a different allocation would be more ap-
propriate; see the discussion of the residual profit split under the OECD PSM in ch. 9.
The author does not rule out that the same approach could be useful in this context.
678
group entities resident in different jurisdictions also contributed unique

development inputs, additional complexity would have been added to the
assessment (as they would have then also been entitled to compensation).
Second, it must be estimated how much of the value of the developed intan-
gible is caused by the pre-existing IP versus the ongoing R&D. This is an
essential determination, as it divides the estimated net present value of the
residual profits from the developed intangible into two parts: one allocable
to the jurisdiction where the pre-existing IP was created, and the other
allocable to the jurisdiction where the ongoing R&D is carried out. This
valuation will likely be challenging, and ultimately subjective. Key prem-
ises include assumptions concerning the extent to which the new intangible
builds upon the pre-existing IP, the useful life of the pre-existing IP, as well
as the life of the new intangible.2108 Again, the above example is simplified,
as the value of the original software is assumed to account for 90% of the
value of the new version.
The third step is to ascertain which group entities are entitled to the re-
maining residual profits from the developed intangible due to the ongoing
R&D. In the example, the Norwegian entity alone performs R&D. The al-
location is then straightforward: the entire 10% should be allocated to the
Norwegian entity. Had other group entities also performed important R&D
functions, as the case may be in practice, the assessment would become
more complicated. The Irish entity is allocated a risk-adjusted return on its
The above should apply also to other controlled R&D structures in which
pre-existing IP is contributed to IP development, including contract R&D
agreements.
22.5.5. Concluding remarks on the contribution of

pre-existing IP to intangible development processes
When pre-existing IP is contributed to the development of an intangible, it

is paramount that the guidance on important functions and funding is not
applied in isolation based solely on the current stage of development, but
in conjunction with an application of the transfer pricing methods to price
the contribution of the pre-existing IP. Only then will the profit alloca-
2108. See the analysis of CSA buy-in pricing in ch. 14, in particular, secs. 14.2.8.-
14.2.8.7.
679
tion reflect the intangible value creation made by different group entities
throughout all development phases. For instance, in the example in section
22.5.4., had the Norwegian tax authorities asserted that the entire residual
profits from the new software developed should be allocated to Norway,
based on the argument that the Norwegian entity performs the important
ongoing R&D functions, this clearly should be rejected. Such an alloca-
tion would fail to reflect that only 10% of the intangible value creation
was carried out in Norway. It is therefore somewhat ironic that the current
OECD TPG on IP ownership is so preoccupied with ensuring that ongoing
R&D and funding contributions are appropriately remunerated that sight is
nearly lost of the fact that a proper arm’s length profit allocation may also
require compensation to group entities that contribute pre-existing IP to
the development process.
680
Chapter 23

Developed Marketing IP under the US Regulations
23.1. Introduction
A marketing intangible developed within a group may generate residual
profits when it is exploited in a marked jurisdiction. The typical set-up is
that a distribution subsidiary in the market jurisdiction licenses the intan-
gible property (IP) from a group entity resident in a foreign (often low-tax)
jurisdiction and builds the local value of the IP through marketing efforts.
The question in this chapter is how the current US regulations allocate
residual profits that are generated through exploitation of such marketing
IP in the United States among the group entities that contributed to the
development of the intangible.
As discussed in chapter 20, the point of departure is that the group entity
that holds legal title to an item of IP is entitled to the residual profits.
The US regulations, however, are not willing to give such pricing effect
to legal ownership unless there is enough economic substance underly-
ing it. Section 23.3., is dedicated to a discussion of this issue. If there
is enough economic substance underlying the foreign legal ownership of
the internally developed marketing IP, the legal ownership will be given
pricing effect, and the residual profits will thus be allocated to the foreign
group entity that holds legal title to the IP. The question then turns to how
other group entities, in particular the local US distribution entity that has
developed and exploited the IP, shall be remunerated for its efforts. This
issue is discussed in section 23.4. Concluding comments are provided in
section 23.5.
First, however, the author will present a high-level introduction to the profit
allocation problem for internally developed marketing IP in section 23.2.
681
Exploitation of Internally Developed Marketing IP

internally developed marketing IP
A multinational is generally free to register which group entity it desires to
hold legal title to particular marketing IP it controls.2109 The focus of IP law is
on protecting IP rights against third parties, not on how legal title is allocated
among the legal entities within the same economic group. This freedom,
combined with the historical point of departure under international tax law
that the group entity that holds legal title to an item of IP should be entitled to
the residual profits, provides multinationals with an incentive to register legal
ownership of their unique and valuable marketing intangibles to group enti-
ties resident in jurisdictions that offer beneficial tax treatment of IP profits.
This incentive effect is amplified by the fact that a multinational will nor-
mally develop and exploit its valuable marketing intangibles in every ju-
risdiction where it sells its products and services, as marketing IP must be
promoted locally in order to have value there. In contrast, the development
of manufacturing intangibles will normally involve vastly fewer jurisdic-
tions. R&D, and the funding of it, will typically source from just a cou-
ple jurisdictions (where the R&D and financing centres of the group are
located).2110 Thus, as opposed to the case for marketing intangibles, almost
all jurisdictions in which a manufacturing intangible is exploited will have
no factual basis to assert that they should be entitled to a portion of the
residual profits (because the intangible was not developed there).
Marketing activities are intangible development contributions. Where such

contributions are rendered by a local distribution entity, fundamental profit
allocation issues are triggered. On the one side, the US Internal Revenue
Code (IRC) section 482 regulations (and the new 2017 OECD Transfer
Pricing Guidelines (OECD TPG) on intangibles) in principle aim to en-
sure that intangible profits are allocated among jurisdictions in a manner
that reflects where the intangible value creation takes place. Applied in the
2109. There may, however, in some jurisdictions, be intangible property (IP) law re-
strictions requiring that a local legal entity is registered as title holder to trademarks
and the like. The factual pattern in DHL Corp. & Subsidiaries v. CIR, T.C. Memo.
1998-461 (Tax Ct., 1998), affirmed in part, reversed in part by 285 F.3d 1210 (9th Cir.,
2002), provides an illustration of how global trademark registrations are carried out by
multinationals in practice. See the discussion in sec. 19.2.5.3.
2110. See the discussions on the value drivers in IP development with respect to manu-
facturing IP in sec. 21.2. for the US regulations and sec. 22.2. for the OECD Transfer
Pricing Guidelines (OECD TPG). On the risk aspects of developing manufacturing
IP through research and development (R&D) versus marketing IP through marketing
expenditures, see Roberge (2013), at p. 220.
682
A lead-in to the profit allocation problem for internally developed
marketing IP
context of locally developed marketing intangibles, this basic reasoning

provides an argument for allocating the residual profits to the source state.
Source state tax authorities would likely advocate this position.
On the other hand, it is clear that unrelated enterprises do enter into distri-
bution agreements in which one party licenses a marketing intangible and
must contribute to the local value of the intangible in the form of marketing
expenditures without being allocated any residual profits.2111 Multination-
als tend to advocate that controlled profit allocations should correspond to
such third-party behaviour. In other words, the local market jurisdiction
should allow the group distribution entity to deduct its marketing (IP de-
velopment) costs without being allocated any of the residual profits that are
subsequently generated by the IP in the local market jurisdiction. The chal-
lenge faced by the US and OECD is to design allocation rules that strike a
reasonable balance between these two axioms.
A related issue is how incremental operating profits due to specific market

conditions should be allocated between the local group entity that performs
marketing and sales, on the one side, and a foreign group entity that holds
legal title to the marketing intangible, on the other side. While, in princi-
ple, it is clear that profits caused by the local exploitation of foreign-owned
marketing IP and local market characteristics are two different notions en-
tirely (that also follow different profit allocation norms),2112 the two types
of profits tend to be connected, as both stem from the marketing and sales
part of the value chain.2113
The allocation of residual profits from a locally developed marketing intan-

gible is not a problem particular to developing countries. The GSK case, in-
volving the United States as the market jurisdiction and the United Kingdom
as the residence jurisdiction of the legal owner of the marketing intangible,
illustrates the relevance of the issue among developed countries.2114 Ulti-
2111. It is, however, unclear how profitable such third-party distributors indeed are. It
has been asserted in the literature that such distributors make low, stable returns; see,
e.g. Allen et al. (2006), at sec. 5.2.
2112. See the discussion of the new OECD allocation rules for profits from local mar-
ket characteristics in ch. 10 and on the corresponding US rules (on comparability with
respect to economic conditions) in sec. 6.6.5.4.
2113. Separation of residual profits generated through local exploitation of a foreign-
owned marketing intangible from incremental profits due to specific market charac-
teristics may be challenging, as experienced by tax authorities in developing countries
when multinationals extract all local marketing (residual) profits based on the assertion
that they are caused solely by locally exploiting a foreign-owned marketing intangible.
2114. See the discussion of GlaxoSmithKline Holdings v. CIR (2006) in sec. 19.2.5.2.
683
mately, however, the author’s impression is that the problem is pronounced

when the market jurisdiction is a developing country, due to the existence
of local market characteristics that generate incremental operating profits.
23.3. When are the US regulations unwilling to give

pricing effect to foreign legal ownership of
marketing IP due to a lack of economic substance?
In sections 23.3.2.-23.3.8., the author will discuss the application of the US
“economic substance” exception for internally developed marketing intangi-
bles. The point of departure is that the US regulations will respect controlled
pricing agreements that allocate residual profits from local US exploitation
of marketing IP to a foreign group entity that holds legal title to the IP as
long as there is sufficient economic substance supporting the legal owner-
ship. The question then becomes: what is the threshold for applying the eco-
nomic substance exception in order to alter the controlled profit allocation so
that the local distribution entity is allocated a portion of the residual profits?
While the US regulations provide no general guidance for this problem,

two extensive examples are drawn up that illustrate the application of the
economic substance exception. These examples will be discussed in sec-
tions 23.3.2.-23.3.4. Thereafter, the author will discuss three different as-
pects of the threshold for the economic substance exception: (i) concur-
rent compensation throughout the development phase as a safe harbour
(discussed in section 23.3.5.); (ii) the “incremental marketing” criterion
(discussed in section 23.3.6.); and (iii) that the economic substance excep-
tion should only be relevant for incoherent pricing structures (discussed in
section 23.3.7.). Lastly, in section 23.3.8., the author will briefly compare
the current economic substance exception to its 1994 predecessor.
23.3.2. The wristwatch example: US distributor incurs

value of foreign-owned trademark
The first example on the application of the economic substance exception

for internally developed marketing IP addresses the allocation of a US dis-
tribution subsidiary’s operating profits generated through US exploitation of
684
When are the US regulations unwilling to give pricing effect to foreign legal
ownership of marketing IP due to a lack of economic substance?
a marketing intangible for wristwatches, where legal title to the intangible

is held by a foreign manufacturing parent.2115 The US subsidiary purchases
the watches from the parent at a fixed, arm’s length price per watch. The
parent and subsidiary together undertake (without separate compensation)
carrying out US marketing activities to establish the trademark (compa-
rable uncontrolled transactions (CUTs) contain similar marketing obliga-
tions). In years 1-6, the subsidiary also undertakes incremental marketing,
for which it is not compensated. In year 7, the sale of watches begins gen-
erating residual profits. The controlled pricing allocates the residual profits
to the foreign parent, on the basis that it holds legal title to the trademark.
The mismatch between the subsidiary’s incremental marketing activities

in years 1-6 and the subsequent allocation of residual profits to the parent
in year 7 is deemed inconsistent with the economic substance of the actual
conduct of the controlled parties. The logic is that an unrelated party would
not have carried out incremental marketing to develop the value of market-
ing IP that it did not own unless it received contemporaneous compensation
or otherwise had reasonable anticipation of receiving a future benefit.2116
The example therefore rejects the controlled profit allocation that allocated
all residual profits to the foreign title-holding entity, and instead prescribes
a split of the residual profits. This new allocation is achieved through the
imputation of an agreement that, consistent with the economic substance
of the behaviour of the controlled parties, affords the US subsidiary “an
appropriate portion of the premium return”.2117
23.3.3. The athletic gear base example: US subsidiary incurs

value of foreign-owned trademark
The second example addresses the allocation of operating profits that a US

subsidiary has earned through exploiting exclusive US rights to athletic
2115. Treas. Regs. § 1.482-1(d)(3)(ii)(C), Example 3.

2116. Id., at (iii).
2117. The current regulations offer separate guidance for the imputation of contin-
gent compensation structures. See the discussion in sec. 21.3.3. The imputation of any
specific type of agreement is not required, as long as the prescribed profit allocation
is achieved. For example, the imputed agreement may be (i) a service agreement that
affords the subsidiary contingent payment for its incremental efforts in years 1-6; (ii) a
long-term, exclusive distribution agreement that allows the subsidiary to benefit from
its incremental marketing activities; or (iii) that the parent compensates the subsidiary
for terminating its imputed long-term distribution agreement.
685
gear manufacturing and marketing IP that it licensed from its foreign par-
ent (which holds legal title to the IP).2118 Both the subsidiary and parent are
obligated to undertake specific US marketing activities without separate
compensation (similar obligations are applied in CUTs). The subsidiary,
however, also performs incremental marketing activities without compen-
sation. In year 7, US sales begin generating residual profits. A separate ser-
vices agreement is then entered into, pursuant to which the parent agrees to
compensate the subsidiary on a cost basis for its incremental marketing ex-
penditures incurred in years 1-6 and forward. The royalty rate is increased
so that all residual profits from US sales are allocated to the parent.
This example also rejects the controlled profit allocation that allocates all
residual profits to the foreign entity,2119 and prescribes instead a split of the
residual profits, thereby allocating a portion of the residual profits to the
US subsidiary.2120
23.3.4. The athletic gear example: The 2006 extension
An extension of the athletic gear example was introduced in 2006 to clarify

that the economic substance exception should not be applied simply if the
US Internal Revenue Service (IRS) disagrees with the taxpayer’s transfer
pricing,2121 as “normal” mispricing should be corrected by way of the trans-
fer pricing methods. Only if the mispricing is “substantial” should applica-
tion of the economic substance exception be considered.
The factual pattern in the extension is the same as in the base example
(discussed in section 23.3.3.) with respect to years 1-6, apart from the twist
that in year 1, the parent agrees to compensate the subsidiary quarterly
for its incremental marketing activities on a cost-plus basis, regardless of
whether the marketing proves successful. For year 4, it is determined that
the mark-up on costs is outside the arm’s length range. A reassessment is

2119. Id., at (iii).
2120. Langbein (2005), at p. 1086 comments on the 2003 proposed regulations (68 FR
53448-01), stating that he finds the allocation of residual profits to the subsidiary to rest
on a “nebulous and discretionary identification of a return on ‘services’” rather than a
more concrete, cost-based allocation for IP. As above in the first example (Treas. Regs.
§ 1.482-1(d)(3)(ii)(C), Example 3, as discussed in sec. 23.3.2.), there is no require-
ment as to the imputation of any specific type of agreement, as long as the resulting
allocation of income is as envisioned by the example. See the discussion on contingent
compensation structures in sec. 21.3.3.
686
made, requiring compensation based on the median of the interquartile

range of the arm’s length mark-ups charged by uncontrolled parties.
The example emphasizes that the fact that the mark-up in year 4 was out-
side the arm’s length range does not indicate that the economic substance
exception should be triggered, as the subsidiary did not bear the risks asso-
ciated with the incremental marketing activities. Thus, the transfer pricing
methods – and not the economic substance exception – should be applied
to correct the controlled pricing. The example is, however, ambiguous with
respect to whether a mark-up significantly outside the arm’s length range
could trigger the economic substance exception.2122 The author will revert
to this issue in the discussion in section 23.3.7.
23.3.5. Concurrent remuneration during the IP development

phase as a safe harbour from the economic substance
exception
The current US regulations, through the wristwatch and athletic gear exam-
ples discussed in sections 23.3.2.-23.3.4., seek, more clearly than their 1994
predecessor, to induce vigilant transfer pricing behaviour when a US group
entity contributes to the domestic value of foreign-owned marketing IP.
The economic substance exception turns on the lack of contemporaneous
compensation for incremental marketing expenses. While the regulations
do not elaborate on what is meant by concurrent compensation, the author’s
interpretation is that the language refers to an arm’s length compensation
for the incremental marketing contributions, typically determined under
the cost-plus method or the comparable profits method (CPM).
Provided that the US subsidiary is allocated such remuneration throughout

the development phase, there should be no legal basis to apply the eco-
nomic substance exception (to split any US residual profits generated in the
exploitation phase). Concurrent normal market return compensation to the
United States will thus provide the multinational with a safe harbour from
the economic substance exception. This conclusion corresponds to the au-
thor’s observations on the economic substance exception applied to inter-
nally developed manufacturing IP (see the discussion in section 21.3.7.).
Multinationals are given the choice of either (i) reimbursing excessive mar-
keting expenditures with a steady cost-plus margin income to the US entity
2122. It is stated that the mark-up being outside the arm’s length range does not “with-
out more” trigger the economic substance imputation authority.
687
on a concurrent basis throughout the IP development phase; or (ii) risking

a residual profit split further down the line when the IP is exploited.
The likely problem with this safe harbour, seen from the perspective of
a multinational, is that it will effectively result in the marketing costs not
being deductible in the United States (as the costs are reimbursed by the
foreign parent with a mark-up). This may be problematic when significant
marketing expenses are necessary to establish a new product and there is a
risk that the marketing may prove unsuccessful, typically because there is
fierce competition in the marketplace.
23.3.6. A premise for triggering the economic substance

exception: “Incremental” marketing expenditures
The economic substance exception is triggered in both the wristwatch and

athletic gear examples (discussed in sections 23.3.2. and 23.3.3., respec-
tively) because the US subsidiary incurs “incremental” marketing expen-
ditures without compensation during the IP development phase (before the
product begins generating residual profits).2123 The effect of triggering the
economic substance exception could be serious: the entire residual profits
from US exploitation of the marketing intangible may be split between
the US subsidiary and the foreign group entity that holds legal title to the
marketing IP.
Unfortunately, the regulations provide no guidance as to what constitutes

incremental marketing. Those who commented on the 2003 proposed reg-
ulations requested clarification of the term.2124 The IRS responded, rather
2123. The term “incremental marketing” was likely adopted as a response to the criti-
cism of the Ninth Circuit in DHL Corp. & Subsidiaries v. CIR, T.C. Memo. 1998-461
(Tax Ct., 1998), affirmed in part, reversed in part by 285 F.3d 1210 (9th Cir., 2002), per-
taining to the determination of intangible ownership based on the incurrence of relative
marketing expenditures under the 1968 regulations (33 Fed. Reg. 5848). See the discus-
sion of DHL in sec. 19.2.5.3. See also Levy (2002); and Terr (2004), at p. 564. Lang-
bein (2005), at p. 1085 notes that the new terminology weakens the link to intangible
development expenses, without being clear on the implications of that. The incremental
marketing concept is also significant for determining an arm’s length remuneration for
intangible development contributions rendered by other group entities than the entity
assigned ownership. See the discussion under sec. 23.4.3. Terr (2006), at p. 881 also
finds that the examples contain insufficient guidance on the question of whether and
to what extent a contribution to intangible development shall be remunerated. See also
Wittendorff (2010a), at p. 637 on incremental marketing expenses.
2124. See the preamble to the 2006 regulations (71 FR 44466-01), at sec. B.2. See also
Terr (2004), at p. 564.
688
obtusely, in the preamble to the 2006 temporary service regulations that in-
cremental marketing activities refer to “activities … quantitatively greater
(in terms of volume, expense, etc.)” than activities undertaken by third par-
ties in comparable transactions.2125 The incremental marketing threshold
therefore remains ambiguous.
Whether the threshold is surpassed can, in the author’s view (in line with
the statement in the 2006 preamble), only be assessed in a meaningful
manner if there is a third-party “reference level” of marketing expendi-
tures, incurred under similar terms and conditions, that the controlled level
of marketing expenses can be compared to.2126
Apparently, it is common in third-party distribution and franchising agree-

ments that the local franchisee is obliged to perform a certain level of
marketing without reimbursement from the franchiser. For valuable inter-
national marketing IP, the franchiser will likely have a strong bargaining
position vis-à-vis local unrelated franchisees that only provide fairly rou-
tine inputs to the value chain. Presumably, the more valuable the licensed
IP is, the more burdens (e.g. marketing expenses) will be placed on the
licensee.2127
2125. The author agrees with Terr (2006), at p. 881, who found little comfort in the
statement of the 2006 preamble that consideration was being given to discounted cash-
flow analysis as a method to analyse non-owner contributions to intangible value, with
respect to avoiding subjective evaluations of contributions.
2126. See also, in particular, Levey et al. (2006), at p. 7 for a strong analysis of this
“bright line test” (i.e. whether the related distributor’s marketing expenses exceed an
arm’s length level, thereby going from being “routine” to “non-routine” marketing ex-
penses and thus creating local marketing IP). Levey criticizes the mere notion of there
being a clear industry standard in the first place against which to benchmark a related
distributor’s marketing expenses, and emphasizes that there are serious problems con-
nected to attaining reliable information at all on third-party distributors’ marketing
expenses in publicly available financial statements. This is due to a lack of disaggre-
gated and properly categorized cost data, as well as accounting timing issues (some
marketing expenses may be booked on the balance sheet for expensing in later periods).
Further, if the related distributor nonetheless is guaranteed a normal return on its mar-
keting expenses, Levey argues that it should not be relevant as to whether its expenses
exceed an arm’s length level. See also Helderman et al. (2013), at p. 365 and p. 368,
with respect to the incremental marketing expenses issue (discussed in relation to the
US Glaxo case).
2127. For example, Apple would likely have no struggle in outsourcing local distribu-
tion of its products for a normal market fee to unrelated distributors. It may be, how-
ever, that third-party outsourcing is not a realistic alternative for the multinational due
to its need to (i) ensure quality control over each step in the value chain of its products
so that its marketing intangibles maintain their value; or (ii) make sure that know-how
and business processes remain secret from competitors. See the discussion of foreign
direct investment in sec. 2.3.
689
Logically, one would nevertheless assume that there are limits as to the
level of marketing expenses that an unrelated distributor would be willing
to incur, even when negotiating with a powerful franchiser for the licence
of an attractive marketing IP. An unrelated licensee would presumably,
in general, reject conditions that could put it worse off than it would have
been under its best realistic alternative, for instance, when the estimated
present value of the marketing efforts is negative.2128 Also, a local third-
party distributor may control unique intangibles (local goodwill, market-
ing know-how, etc.) that may have synergetic effects with the licensed mar-
keting intangible. It seems unlikely that such a distributor (with a relatively
strong bargaining position) would accept incurring incremental marketing
expenditures without assurance that its efforts would be profitable. Further,
it may be that the foreign licenser’s trademark is successfully established
and valuable in foreign jurisdictions, but unknown in the United States. If
so, that might indicate that the franchiser is in a lesser bargaining position
than it would have been had the brand already been established and highly
valuable in the United States.
Given the wide range of possible bargaining constellations between un-

related licensers and licensees, the author does not find it realistic that, in
most cases, it will be possible to identify any “normal” third-party level of
marketing expenditures.
If it is not possible, on the basis of CUTs, to determine whether the con-

trolled level of marketing expenses is “incremental”, the point of departure
should be that a related distribution entity that provides only routine inputs
to the value chain should be able to incur marketing expenses at a level that
is not incompatible with a normal market return operating margin. Thus,
if the US subsidiary, concurrently throughout the IP development phase,
experiences a level of operating profits that corresponds to the amount that
would be allocated to it under the CPM, there should be no basis for an
application of the economic substance exception. Further, if its operating
margin lies outside the arm’s length range, the natural approach would, in
the author’s view, be to perform a transfer pricing adjustment, as opposed
to imputing new pricing terms based on the economic substance excep-
tion. If, however, the incurrence of incremental marketing expenditures re-
sults in an operating margin that lies significantly outside the arm’s length
range, it may be appropriate to resort to the economic substance exception.
2128. It is, however, conceivable that a licensee would be willing to incur short-term
losses if there are prospects of higher future returns as a result of the licence arrange-
ment.
690
Due to the blatant lack of guidance on the “incremental” marketing crite-

rion, the author suspects that, in practice, it will prove challenging to deter-
mine whether there indeed is a legal basis on which to apply the economic
substance exception.2129 Future revisions of the US regulations should seek
to clarify the incremental marketing expenses threshold.
23.3.7. The economic substance exception should be relevant

only for incoherent pricing structures
The wristwatch and athletic gear examples illustrate that the economic
substance exception is reserved for rather extreme scenarios. Comments
on both the 2003 proposed and the 2006 temporary service regulations
expressed a concern that the economic substance exception could be used
to impute new pricing terms if the IRS took issue with the taxpayer’s trans-
fer pricing.2130 Even though the author certainly agrees that the threshold
for applying the economic substance exception is ambiguous, he does not
share this concern.
The essence of the factual patterns of the wristwatch and athletic gear ex-
amples is that a US group entity takes the entrepreneurial role in an intra-
group marketing IP development process and assumes the associated risks
without receiving a commensurate benefit in return. It is difficult to see
such arrangements as anything other than schemes constructed for the pur-
pose of achieving concurrent tax deductions for IP development expenses
(in a high-tax jurisdiction), combined with the shifting of residual profits
to the foreign group entity that holds legal title to the intangible (often resi-
dent in a low-tax jurisdiction).
Take the wristwatch example (discussed in section 23.3.2.), for instance, in

which the US subsidiary incurs incremental marketing costs over a 6-year
period. Even if the marketing efforts prove successful (in the sense that the
local brand value is increased), the subsidiary has no prospect of reaping
any of the residual profits. It is not even reimbursed for its costs. To accept
this pricing structure would be contrary to the realistic options available to
the controlled parties. The subsidiary would simply be significantly better
off by not entering into the licensing structure.
2129. See also Vincent (2006), p. 28:10.

2130. See the preamble to the 2006 proposed regulations (71 FR 44466-01), sec. C;
and the preamble to the 2009 final regulations (74 FR 38830-01), sec. C. See also the
discussion of the 2006 extension of the second example illustrating the economic sub-
stance provision in sec. 23.3.4.
691
Further, in the athletic gear example (discussed in section 23.3.3.), the US

subsidiary is reimbursed for its incremental marketing costs plus a normal
return mark-up only after the IP development process proves successful.
All residual profits are allocated to the foreign parent. While less extreme
than the wristwatch example, it must be assumed that a rational third-party
enterprise would not enter into such an agreement, as the risk-adjusted re-
turn of the incremental marketing investment would likely lie significantly
below a normal market return. After all, there is a pronounced risk at the
outset that the subsidiary will not even be able to recover its incremental
marketing expenses.
In the author’s view, the wristwatch and athletic gear examples establish a
high threshold for applying the economic substance exception. These fac-
tual patterns do not pertain to marginal transfer pricing misstatements, but
to incoherent pricing structures.
The author is therefore surprised by the introduction of the 2006 extension

of the athletic gear example (discussed in section 23.3.4.), as it may be
interpreted as placing the threshold for applying the economic substance
exception up for discussion. There, the US subsidiary is reimbursed for its
incremental marketing costs on a concurrent basis (with the addition of a
mark-up) regardless of whether the intangible development proves success-
ful, and therefore assumes no risk. Comparatively, in the wristwatch and
athletic gear base examples, the US entity assumes the entire development
risk. The problem in the extension is simply that the mark-up lies outside
the arm’s length range, for which an adjustment is made.2131 The extension
states that a mispriced mark-up “does not, without more, support imputa-
tion of additional contractual terms”.2132 Thus, even though the extension
does make it clear that the economic substance exception is inapplicable, it
manages, by way of the wording “not, without more”, to introduce uncer-
tainty with respect to how far from qualifying for imputed pricing terms
the factual pattern really lies.
The problem with the 2006 extension is that it can now be questioned
whether the economic substance exception truly requires as relatively ex-
treme scenarios as indicated by the wristwatch and athletic gear examples,
or whether a fairly “normal” mispricing will be sufficient to trigger the
imputation authority.2133
2131. See the discussion of the arm’s length range in sec. 6.5.
2132. Treas. Regs. § 1.482-1(d)(3)(ii)(C), Example 5, at (iv).
2133. The answer to this question may have significant profit allocation implications. If
treated as a transfer pricing issue, the effect will be to adjust the normal market return
692
The original wording proposed in 2006 for the extension contained an ad-
ditional reservation:
[H]ad the compensation paid to USSub been significantly outside the arm’s
length range, that might lead the Commissioner to examine further whether,
despite the contractual terms that require cost-plus reimbursement of USSub,
the economic substance of the transaction was not consistent with [foreign
patents] bearing the risk associated with promotional activities in the United
States market. (Emphasis added) 2134
This wording was removed in the final 2009 version of the service regula-
tions, as the US Treasury and the IRS agreed that it did not convey the
intended meaning.2135
To the author, the removal does not seem to entail much reality, as at the
same time, it was made clear that the imputation authority would be rel-
evant “where the economic substance of the transaction is consistent with
such terms”.2136 Also, the current text contains the “without more” reserva-
tion. In the author’s view, the 2006 extension failed its mission to clarify
the boundaries of the economic substance exception. Instead, the exten-
sion, as it currently reads, seems to indicate almost an overlap of the scope
of application for the economic substance exception and the ordinary trans-
fer pricing methods.
The author finds that there are compelling reasons to interpret the eco-
nomic substance exception in accordance with the ordinary meaning of
the wording “inconsistent with the economic substance of the underlying
transactions” of the economic substance provision in the 2009 Treas. Regs.
§ 1.482-4(f)(3)(i)(A), and to rely significantly on the wristwatch and ath-
letic gear base examples. These sources of law indicate a high threshold for
allocable to the routine inputs provided by the US subsidiary (e.g. adjustment of the
cost plus mark-up). However, if the imputation authority is used to establish pricing
terms comparable to the ones used in the wristwatch and athletic gear examples (see
secs. 23.3.2.-23.3.4.), the US entity could be allocated residual profits, which, of course,
can be significantly greater than the normal return resulting from an application of a
one-sided pricing method. The author would object to such an allocation. Even if the
IRS should impute new pricing terms on the basis of the economic substance exception,
the income allocation results achieved thereby must lie within the boundaries set by the
transfer pricing methods. The pricing solution must be at arm’s length. This is a severe
restriction on the imputation authority. See also Terr (2006), at p. 884, who notes that
“if the shortfall is significant”, the US Internal Revenue Service (IRS) could impute
terms that supported an adjustment of that shortfall.
2134. 2006 Temp. Treas. Regs. (71 FR 44466-01) § 1.482-1-T(d)(3)(ii)(C), Example 5.
2135. See the preamble to the final 2009 regulations (74 FR 38830-01), at sec. C.
2136. Id.
693
applying the economic substance exception. Further, the inherent lack of

clarity in the 2006 extension, the initial intention behind it to clarify that
the economic substance exception is inapplicable if the problem simply
is that the IRS disagrees with the taxpayer’s pricing, as well as the ob-
vious problems associated with using similar thresholds for applying the
economic substance exception and the transfer pricing methods provide
further support for the author’s interpretation.
The economic substance exception should be reserved for cases in which

the US subsidiary incurs the intangible development risks and it is not pos-
sible to determine a proper allocation of operating profits without the im-
putation of additional terms. This is in line with the wristwatch and athletic
gear base examples, where there are no arrangements in place to remuner-
ate the US subsidiary, either on a concurrent cost-plus basis or a contingent
residual profit basis. In such cases, the transfer pricing methods cannot
readily be applied, as there are no pricing terms to adjust. Application of
the economic substance exception should therefore be justified.
Where there is a pricing mechanism in place, however, as is the case in the

2006 extension, income adjustments should be based on ordinary pricing
adjustments. The basic reason for this is that the US entity does not bear the
risks of incremental marketing in these latter cases. The question is simply
as to which extent the US entity should profit. That is, of course, a transfer
pricing matter, not an issue of rejecting a controlled contractual risk alloca-
tion on the basis that it does not align with the economic substance of the
controlled transaction (and replacing it with new imputed pricing terms).
23.3.8. The economic substance exception is balanced

relative to its 1994 predecessor
It has been debated as to whether the current economic substance excep-

tion is more far-reaching than its 1994 predecessor.2137 As illustrated by the
1994 economic substance example,2138 the authority was already signifi-
cant. It has therefore been speculated as to whether the new guidance was
introduced to replace the previous “excess marketing expenditures” ter-
minology with “incremental marketing activities” due to criticism levied
by the Ninth Circuit in DHL,2139 or whether the new economic substance
2137. See Terr (2004), at p. 567.

2139. See supra n. 2123.
694
exception guidance could allow the IRS to circumvent the statute of limi-
tations.2140
The author’s impression is that the coordination of the economic substance

exception with the imputation authority for contingent payment terms (see
the discussion in section 21.3.3.) necessitated more elaborate guidance and
that the combined authority likely is more far-reaching than the 1994 rule.
The current economic substance exception offers more aggressive and le-
nient solutions than its 1994 predecessor, in the sense that it – in specific
contexts – allocates more or less operating profits to the United States.
On the aggressive side, the current regulations require allocation of residual

profits to a US distribution entity, whereas the 1994 regulations did not. For
instance, in cases of incremental marketing expenditures incurred under
non-exclusive licences, the second cheese example of the 1994 regulations
allocated the entire residual profits to the foreign group entity (that held le-
gal title to the marketing IP) under the rationale that the US subsidiary was
an “assister”. The US entity was only afforded a cost-plus remuneration.2141
The wristwatch example discussed in section 23.3.2. also pertains to a non-

exclusive licence in which the US subsidiary incurs incremental market-
ing costs, but it splits the residual profits between the subsidiary and the
foreign parent that holds legal title to the marketing IP. Had the factual
pattern of the wristwatch example been assessed under the 1994 regula-
tions, the result would, in the author’s view, likely have been to allocate the
entire residual profits to the foreign legal owner of the IP. Thus, the current
regulations allocate more income to the United States, relative to the 1994
regulations, when incremental marketing efforts are carried out under non-
exclusive licence agreements.
The athletic gear example (discussed in sec. 23.3.3.) provides a different re-
sult than the “multiple owners rule” of the third cheese example in the 1994
regulations. Pursuant to the latter rule, the US entity was allocated all re-
sidual profits from the US exploitation of the intangible. This differs from
the allocation pattern professed by the athletic gear example, in which the
profits from the US exploitation of the trademark are split between the US
entity and the foreign parent. Thus, in the context of exclusive licence ar-
rangements, the current regulations allocate less profits to the United States
than the 1994 regulations.
2140. See Terr (2004), at p. 567.

2141. See the discussion of the 1994 “cheese” examples in sec. 19.4.6.
695
The author finds that the wristwatch and athletic gear examples together
offer a coherent and balanced alternative to the 1994 cheese examples. As
discussed in the analysis of the 1994 regulations,2142 the author does not
find any convincing reason as to why the treatment of the factual patterns
in the second and third cheese examples should differ as drastically as
they did. The 1994 threshold for allocating residual profits to the US entity
seemed somewhat arbitrary. The current regulations instead use a profit
split approach in cases in which the US entity, under the 1994 regulations,
was not entitled to any of the residual profits, as well as in cases in which
it was entitled to all of it. Thus, the rough edges of the 1994 regulations are
now smoothed, offering an approach that, on average, likely will be more
acceptable, both from a US and treaty partner perspective.
23.4. Remuneration of a US distribution entity when

the economic substance exception is not triggered
The premise for the following discussion is that the controlled assignment
of legal title to the marketing IP (that is developed and exploited in the
United States) to a foreign group entity is respected and given transfer
pricing effect under the legal ownership rule, as it is supported by suffi-
cient economic substance. In other words, the economic substance excep-
tion (discussed in under section 23.3.) has not been triggered. The residual
profits from US exploitation of the intangible should therefore be allocated
to the foreign group entity that holds legal title to the IP.
The question then turns to how the US distribution subsidiary should be

remunerated for its IP development contributions that increased the US
value of the licensed marketing intangible. This will, in principle, depend
on an application of the transfer pricing methods.2143
The US regulations, through guidance provided in two base examples, split

the treatment of this issue into two main categories. The first is where the US
distribution entity has incurred an arm’s length level of marketing expenses,
and the second is where it has incurred an incremental level of market-
ing expenses. The author will discuss this guidance in sections 23.4.2. and
23.4.3., respectively. Thereafter, in section 23.4.4., he will tie some com-
2142. See the discussion of the “multiple owners” exception in sec. 19.4.6.
2143. Treas. Regs. § 1.482-4(f)(4)(i).
696
ments to the selection of an appropriate pricing methodology to remunerate

the “assister” entity that contributed to the development of the marketing IP.
23.4.2. Arm’s length marketing expenditures are incurred
The question is of how much profit shall be allocated to a US group entity

that has developed the US market value of a trademark owned by a foreign
group entity by incurring an arm’s length level of marketing expenditures.
In the first example, a foreign manufacturing parent is registered as the own-

er of a trademark for wristwatches.2144 Its US subsidiary is given the right to
resell watches in the US market under an exclusive 5-year agreement. Both
entities are obligated to carry out specified types and levels of US marketing.
The subsidiary purchases watches from the parent at a fixed price per watch,
which entails (i) compensation for the exclusive US resale rights; (ii) the right
to use the trademark in the United States; and (iii) the marketing activities
of both parties. The question is whether the wristwatch transfer price should
be adjusted in order to provide the subsidiary with an arm’s length return.
The example is ambiguous with respect to the required profit allocation.

On a stand-alone basis, it may possibly be interpreted to invite the applica-
tion of the profit split method (PSM), based on the observations that both
entities perform marketing and that the CUT method likely will be inappli-
cable due to comparability concerns.2145 Such arguments are not convinc-
2144. Treas. Regs. § 1.482-4(f)(4)(ii), Example 2.

2145. The original version of the example, as it read in the 2003 proposed regulations
(68 FR 53448-01), stated that if it were not possible to identify comparable uncon-
trolled transactions (CUTs) that incorporated a similar range of integrated elements
and if there were non-routine contributions from both the parent and the subsidiary,
“the most reliable measure of the arm’s length price for the wristwatches may be the
residual profit split method”, pursuant to which “an appropriate allocation of the com-
bined operating profits from the sale of wristwatches and related activities” could be
carried out. See also 2003 Prop. Treas. Regs. § 1.482-9(g)(1), which stated that “the
profit split method is ordinarily used in controlled services transactions involving high-
value services or transactions that are highly integrated and that cannot be readily
evaluated on a separate basis”. The sentence was removed in the 2006 temporary regu-
lations [71 FR 44466-01], “thus eliminating any implication that the profit split method
is a ‘default’ method for controlled services transactions that have value significantly
in excess of cost”; see the 2006 preamble, sec. A.6. The final 2009 regulations [74 FR
38830-01] version of § 1.482-9(g) was modified in order to make it consistent with
the § 1.482-7 cost-sharing regulations, clarifying that the residual profit split method
(RPSM) is inapplicable to arrangements in which only one of the controlled parties
make “significant non-routine contributions”. The reference was criticized by commen-
tators, who interpreted it to imply a preference for the PSM in all cases pertaining to
697
ing. First, the US subsidiary does not seem to contribute any unique inputs
to the value chain. The PSM should therefore be inapplicable (likely, the
CPM should be applied to remunerate the US entity). Second, the subsidi-
ary does not incur any incremental marketing expenditures. Third-party
distributors will presumably be willing to provide the same level of mar-
keting for a normal market return. It is therefore difficult to see why the US
subsidiary should demand any part of the residual profits. The author’s in-
terpretation is supported by the new examples on the best-method rule.2146
The second example on this issue pertains to a scenario in which a foreign

group entity holds legal title to the AA trademark for athletic gear and
licenses exclusive US make-sell rights to a local subsidiary in return for
an annual royalty.2147 Both entities are obliged to perform specified types
and levels of US marketing without compensation, apart from that embed-
ded in the royalty rate. The question is whether an allocation is warranted
with respect to the royalty rate in order to provide the subsidiary with an
arm’s length return. In this example, it is also unlikely that a CUT is avail-
able, rendering the CUT method inapplicable. The US entity does not pro-
vide any non-routine contributions to the value of the trademark. It should
therefore likely be compensated based on either the cost-plus method or the
CPM.2148 The author’s interpretation is supported by the new examples on
the best-method rule.2149
In conclusion, it is the author’s view that the above examples do not provide
any legal basis for allocating residual profits to a US distribution subsidiary
that does not provide any non-routine IP development inputs and incurs
only an arm’s length level of marketing expenditures to build the local
value of marketing IP owned by a foreign group entity. The residual profit
intangible development contributions; see the 2006 preamble, at sec. A.6. The 2006
temporary regulations therefore removed the reference. This may seem to be a consid-
erable revision. It was, however, offset by the inclusion of strict comparability require-
ments, rendering the CUT method inapplicable in most cases. Thus, even though the
pronounced preference for the RPSM is removed in the final 2009 regulations, it will,
in practice, be the go-to method in cases in which both parties contribute unique inputs
to the intangible’s development.
2146. The author refers to the discussion of the cost-plus method and comparable prof-
its method (CPM) in sec. 23.4.4.
2147. Treas. Regs. § 1.482-4(f)(4), Example 3.
2148. Terr (2006), at p. 881 seems to suggest that Example 12 on the application of the
RPSM in § 1.482-8(b) of the current regulations (for a discussion of the example, see
sec. 23.4.4.) could advise on the allocation of income in this scenario. For the reasons
stated in this section, the author disagrees.
2149. The author refers to the discussion of the cost-plus method and CPM in sec.
23.4.4.
698
should, in these cases, be allocated to the foreign group entity that holds
legal title to the IP. The US distribution subsidiary shall, in these cases, be
treated as an “assister”.
23.4.3. Above-arm’s length marketing expenditures

are incurred
The question then turns to the compensation of a US distribution subsidi-

ary that contributes incremental marketing efforts to increase the US value
of marketing IP that it licenses from a foreign group entity that holds legal
title to the IP.2150 The guidance on this issue distinguishes between sce-
narios in which (i) the US subsidiary is deemed owner of a licence; (ii) the
US subsidiary is remunerated for its marketing efforts under a separate
service agreement; and (iii) the foreign group entity that holds legal title to
the marketing IP is remunerated under a separate service agreement. The
discussion will follow this structure.
23.4.3.2. Scenario 1: The US subsidiary is deemed owner of a licence
The question in this first scenario is that of how much of the operating
profits earned by a US subsidiary from exploitation of foreign-owned mar-
keting IP shall be retained by it when it contributed to developing the lo-
cal value of the IP by incurring an above-arm’s length level of marketing
expenditures when it is deemed owner of the licence.
The US regulations provide guidance for this problem in the form of an ex-
tension of the athletic gear base example (discussed in section 23.4.2.).2151
The factual pattern for the first year is the same as in the base example.
In year 2, however, the subsidiary undertakes marketing efforts above the
level required in the licence agreement, but the agreement is of sufficient
duration that it is reasonable to expect that the subsidiary could obtain
benefits from its incremental marketing activities by way of increased US
sales.
2150. It has been asserted in the literature that this problem is highly practical, as most
group distribution entities incur more marketing expenses than third-party distributors;
see Allen et al. (2006), at sec. 5.3. For an analysis of the corresponding problem under
the (2010 version of the) OECD TPG, see, in particular, Roberge (2013).
699
The example concludes that if it is reasonable to anticipate that the incre-

mental marketing activity will increase the value of only the US rights
to the trademark (and not add to the value of the trademark in other
jurisdictions),2152 it will not be necessary to compensate the US entity for
the incremental marketing. The entire residual profits from the US exploi-
tation of the marketing IP may then be allocated to the foreign group entity
that holds legal title to it.
The author’s impression is that the example, at least at the outset, forms
a solid basis for treating the US subsidiary as an assister that will only be
entitled to a normal market return for its contribution to the development of
the local value of the marketing IP (and thus not any of the residual profits
subsequently generated by the IP). It is, however, doubtful as to whether
this conclusion can be upheld, as the example deems the subsidiary the
owner of “intangible property (that is, USSub’s license to use the AA trade-
mark for a specified term)”. (Emphasis added) 2153
Thus, even if the foreign parent (that holds legal title to the IP) is seen as
the owner of the trademark as such, the subsidiary is deemed owner of the
licensed US exploitation rights. The “discrete owner rule” (see the discus-
sion in section 20.2.3.), although ambiguous, will likely allocate a portion
of the US residual profits to the subsidiary based solely on the fact that it is
the legal owner of the licence, even if the intangible as such is owned by a
foreign group entity.2154 The question is if the US subsidiary, on top of this,
should be allocated a separate remuneration for its incremental marketing
costs.
As mentioned, the position taken in the example is that if the incremental

marketing costs only increase the value of the US rights to the foreign-
owned marketing IP, no additional profits should be allocated to the US
subsidiary (with respect to its contribution to the development of the local
value of the IP). The rationale behind this is likely that the subsidiary will,
under the terms of the licence agreement, reap its portion of the residual
profits from increased US sales due to the additional marketing.
If the incremental marketing costs also increase the value of the foreign
rights to the marketing IP, however, it will be necessary to compensate
2152. The 2003 proposed regulations (68 FR 53448-01) referred to these rights as
“USSub’s intangible”.
2153. See Treas. Regs. § 1.482-4(f)(4), Example 4, at (ii).
2154. The author refers to the analysis of the “discrete owners rule” in sec. 20.2.3.,
which forms the backdrop for the discussion in this section.
700
the US subsidiary. The author assumes that the rationale simply is that the
United States will not tolerate deductions for incremental marketing costs
that the subsidiary will not benefit from.2155 No guidance is provided as to
the extent of the required additional remuneration of the US subsidiary
or on how to determine whether the incremental marketing costs actually
increased the value of the foreign rights to the marketing IP. It is unclear
whether the subsidiary should be entitled to additional residual profits.
The result is similar to that of the multiple owner rule of the third cheese
example of the 1994 regulations,2156 in which a US subsidiary that incurred
incremental marketing expenditures was entitled to the residual profits
from the US rights to a trademark legally owned by a foreign entity. The
fact that a purpose of the current regulations was to move away from the
criticized solutions of the 1994 cheese examples further enhances the lack
of clarity surrounding this example.
There are, however, more fundamental problems at issue here. First, had
the example instead chosen to focus on applying the ordinary transfer
pricing methods to determine an arm’s length compensation for the incre-
mental marketing services provided by the subsidiary, the author finds it
unlikely that any additional residual profits would have been allocated to
the United States. The reason for this is that even though the subsidiary
certainly assumes financial risk in funding the marketing, the marketing
services themselves should likely be regarded as routine services, entitled
only to a normal rate of return (typically cost-plus based).
Second, and most importantly, it could be argued that the US subsidiary’s

incurrence of incremental marketing expenditures itself may indicate that
there is not sufficient economic substance supporting the foreign legal
ownership of the US rights to the marketing intangible. A rational solu-
tion would then be to apply the economic substance exception (discussed
in section 23.3.) and treat the US subsidiary as the economic owner of the
US intangible as opposed to the licence, and assert that the entire residual
profits from US exploitation of the marketing IP should be allocated to
the US subsidiary. Instead, the regulations have ended up with an unclear
add-on to the already unclear “discrete owner rule” for cases in which a US
2155. See also the discussion in sec. 20.2.3. on the legal ownership rule with respect
to a licence. Terr (2004), at p. 564 seems to be of the opinion that the example does not
take a stance with respect to whether the US licensee should be compensated for its
contribution. For the reasons stated in this section, the author does not agree.
2156. See the discussion of the “cheese” example in sec. 19.4.6.
701
subsidiary incurs incremental marketing costs that also increase the value
of the foreign rights to the relevant marketing intangible.
Thus, the conclusion is that the US subsidiary in these cases will receive
(i) remuneration pursuant to its licence agreement, likely entailing a small
portion of the residual profits (pursuant to the discrete owner rule); and (ii)
a return for increasing the value of the foreign rights to the marketing IP,
likely entailing a normal return fee (on the basis of the cost-plus method
or the CPM).
Whether this intricate mix of profit allocation to a US subsidiary based on

a combination of the discrete owner rule and the (one-sided) transfer pric-
ing methods applied to remunerate intangible development contributions
at all is practicable remains to be seen. The author’s impression is that the
ambiguity associated with the profit allocation consequences of deeming
a US licensee the legal owner of the licence adds unnecessary uncertainty
to cases that could have been dealt with in a satisfying manner through the
economic substance exception (allocating all residual profits from the US
exploitation of the foreign-owned marketing IP to the US subsidiary) and
the (one-sided) transfer pricing methods (compensating the US subsidiary
for increasing the value of foreign rights to the marketing IP).
23.4.3.3. Scenario 2: The US subsidiary is compensated under

a separate service agreement
The question in this second scenario is how much of the operating profits
earned by a US subsidiary from exploitation of a foreign-owned marketing
IP shall be retained by the subsidiary when it contributed to developing the
local value of the IP by incurring an above-arm’s length level of marketing
expenditures when the marketing efforts of the subsidiary are compensated
through a separate agreement.
The guidance provided for this problem here is also in the form of an ex-
tension of the athletic gear base example (discussed in section 23.4.2.).2157
The factual pattern for the first year is the same as in the base example, but
in year 2, a separate service agreement is entered into, pursuant to which
the subsidiary shall carry out incremental US marketing activities. It is
made clear that the subsidiary will not be entitled to any of the residual
702
Remuneration of a US distribution entity when the economic substance
exception is not triggered
profits from US sales, as it does not provide any non-routine contributions

to the value of the foreign-owned trademark.
Comparatively, the previous regulatory profit allocation solution for this

scenario (under the developer-assister (DA) rule and the 1994 third cheese
example) used the incurrence of US incremental marketing costs as the
key benchmark for determining whether the US entity should be entitled
to residual profits, based on the position that it was the deemed owner of
the US rights. Nevertheless, as the incremental marketing services in this
context are subject to separate compensation, their remuneration should
naturally be seen as a transfer pricing matter as opposed to questioning
whether there is sufficient economic substance supporting the assignment
of legal ownership to the foreign group entity.
The fundamental reason for this is that when incremental marketing ser-
vices are separately compensable, the US subsidiary will ultimately be re-
imbursed for its costs, with the addition of a profit margin. Thus, the incur-
rence of incremental marketing costs by the US subsidiary no longer rep-
resents a form of behaviour that is contrary to the practices of comparable
unrelated distributors. The author therefore fully agrees with the example
in that the US subsidiary should not attract residual profits in this case, as
the services provided seem to be of a routine nature. The US entity will be
entitled to a normal market return (typically cost-plus or CPM-based).2158
23.4.3.4. Scenario 3: The foreign legal owner is compensated

under a separate service agreement
The question in this third scenario is of how much of the operating profits
earned by a US subsidiary from the exploitation of foreign-owned marketing
IP shall be retained by the subsidiary when it contributed to developing the
local value of the IP by incurring an arm’s length level of marketing expendi-
tures when the incremental marketing efforts of the foreign group entity (that
holds legal title to the IP) are compensated through a separate agreement.
Also here, guidance is provided in the form of an extension of the athletic

gear base example (discussed in section 23.4.2.).2159 The factual pattern
2158. Also here, the proposed 2003 regulations (68 FR 53448-01) contained a prefer-
ence for the RPSM, which is removed in the current 2009 regulations (74 FR 38830-01)
and replaced with a reference to the best-method rule and the new examples. See the
discussion on the cost-plus method and CPM in sec. 23.4.4.
703
for the first year is the same as in the base example, but in year 2, the
parent and subsidiary now enter into a separate service agreement that ob-
ligates the parent to perform incremental marketing activities in the form
of advertising the AA trademark in international sporting events (e.g. the
Olympics).
The separate service agreement obligates the subsidiary to compensate the

parent for the benefit that it could anticipate to reap as a result of the par-
ent’s marketing efforts (an increase in the US value of the marketing IP).
The question is with regard to the extent of the payment.
The example sees the subsidiary as the owner of the US rights to the
trademark under the discrete owners rule (see the discussion in section
20.2.3.). While the profit allocation consequences of that rule are, in gen-
eral, ambiguous,2160 it seems clear that this particular example takes the
position that the US subsidiary should be allocated a portion of the US
residual profits.
The author is not convinced by this reasoning. The point of departure is

that the foreign parent owns the marketing IP, as the economic substance
exception has not been triggered. Pricing effect should thus be given to the
foreign legal ownership of the IP, and the foreign parent should therefore
be allocated the residual profits from the US exploitation of the IP. As the
US subsidiary only contributes routine inputs to the value chain, it should
be compensated based on a one-sided pricing method.
However, the example does not adhere to this logic due to the discrete
owners rule. Instead, the point of departure is turned “upside-down”, as
the basic assumption is that the US subsidiary owns the US rights and
thus should reap the residual profits yielded through their exploitation. The
question then becomes whether the US subsidiary must surrender some of
those residual profits because the foreign parent incurs incremental mar-
keting expenses.
The answer to this question will depend on the transfer pricing mechanism
applied in the separate service agreements. Normally, marketing should
be viewed as a routine contribution that alternatively could be outsourced
to third parties for a normal market fee, and should be remunerated under
the one-sided methods (typically on a cost-plus basis). This will lead to the
2160. See the analysis of the remuneration of a US licensee that incurs incremental
marketing expenditures without separate compensation in sec. 23.4.3.2.
704
Remuneration of a US distribution entity when the economic substance
exception is not triggered
conclusion that a foreign group entity that owns the US rights to a market-
ing intangible will not be allocated the residual profits from the rights, even
when it incurs incremental marketing expenditures to increase their value.
Instead, the foreign owner will simply be compensated under a one-sided
method.
The author is not convinced that this result reflects an allocation of operat-
ing profits that corresponds to what would have resulted from third-party
behaviour under similar terms and conditions. Ideally, the profit allocation
guidance contained in the US regulations for this scenario should be re-
vised so that the foreign group entity that holds legal title to the marketing
IP will be entitled to the residual profits from the US exploitation. After all,
the economic substance exception is not triggered here. Legal title should
therefore be respected for profit allocation purposes. At the very least, a
profit split should be mandatory so that the foreign group entity gets a por-
tion of the US residual profits, even if it only contributes routine marketing
efforts. Such a result cannot, however, be based on the transfer pricing
methods, as they will only allocate a normal market return to routine value
chain inputs.2161 To achieve this result, the guidance on the economic sub-
stance exception must be revised.
23.4.4. What is the appropriate transfer pricing methodology

to remunerate an assister?
Here, the author will tie some comments to the position of the US regu-
lations on appropriate transfer pricing methodology to remunerate a US
group entity that contributes to the US value of marketing IP that it licenses
from a foreign group entity that holds legal title to the IP.
As discussed in sections 23.4.2. and 23.4.3., the US entity will, in most

cases, be entitled only to a normal market return for its marketing IP de-
velopment contributions. Exceptions to this are only made when the US
subsidiary is deemed to own the licence of the US rights to the foreign-
owned marketing IP under the discrete owners rule, or in the rare cases
in which a profit split is deemed appropriate. It is in this light that the US
service regulations introduced three new examples in 2009 that illustrate
the application of the best-method rule to select transfer pricing method-
ologies to deal with these scenarios. The core focus of the examples is on
2161. Only in the presumably rare cases in which the foreign entity performs particu-
larly valuable marketing (e.g. advertising in the Olympics) may a profit split be relevant
705
applying the one-sided methods in the service regulations to provide the

US subsidiary with a normal market return, thereby in effect allocating the
residual profits from US exploitation of the foreign-owned marketing IP to
the foreign group entity that holds legal title to it.
The first of the new examples pertains to the application of the cost-plus
method for services when the US subsidiary is remunerated for its market-
ing efforts through a separate agreement.2162 The example discusses the
relationship between the cost-plus method and the CPM, and finds that if
sufficiently disaggregated third-party accounting data had been available,
the CPM would likely be the most reliable method. The cost-of-services-
plus method is found to be the most reliable method only because the sub-
sidiary is in possession of CUTs.2163
The second of the new examples pertains to the application of the CPM
for services, in which the US subsidiary is remunerated for its marketing
efforts through a separate agreement that applies (also here, the cost-plus
method).2164 The example finds the CUT and cost-plus methods inapplica-
ble due to the lack of CUTs. As it is possible to find third-party advertising
companies that perform similar marketing activities, their accounting data
can be used as CUTs under the CPM to benchmark the operating margin
of the US subsidiary with respect to its marketing efforts.2165 The CPM is
therefore deemed the most reliable method.2166
The last example illustrates an application of the PSM, where the question
is of how much of the residual profits from the exploitation of foreign-
under the transfer pricing methods, then based on the rationale that such efforts are
non-routine IP development contributions.
2162. Treas. Regs. § 1.482-8(b), Example 10.
2163. See the cost-of-services-plus method in Treas. Regs. § 1-482-9(e). While the
author does not analyse the general cost-plus method as such in depth in this book, he
does comment on the relationship between the gross and net profit methods in sec. 6.3.,
in particular with respect to the classification of costs.
2164. Treas. Regs. § 1.482-8(b), Example 11.
2165. The accounting treatment by the unrelated marketing companies of material
items (e.g. the classification of costs of goods sold and administrative expenses) could
differ from the treatment applied by the subsidiary. The example deems such incon-
sistencies less important when using the ratio of operating profits to total service costs
under the CPM for services. The author agrees, as the classification of expenses will
not influence this profit measure (all costs are included in the operating profits and in
total service costs). See the analysis of operating profits in sec. 6.2., in particular, the
discussion on the classification of expenses in sec. 6.2.4., with further references.
2166. See the CPM for services in Treas. Regs. § 1.482-9(f). See also the general CPM
in Treas. Regs. § 1.482-5, and the discussion in ch. 8.
706
Concluding remarks
owned marketing IP can be kept by the local US distribution entity when

both it and the foreign IP owner entity contribute unique marketing IP de-
velopment inputs.2167 The presence of unique contributions renders the CUT
method, cost-plus method and CPM inapplicable. The PSM is deemed to
yield the most reliable result,2168 entailing a residual profit split pursuant to
the relative value of the unique IP development contributions.2169
In conclusion, a local US group distribution entity that contributed to the

value of foreign-owned marketing IP that it exploits will rarely be entitled
to any of the residual profits generated by the IP, and will thus generally be
remunerated through the one-sided methods (typically cost-plus or CPM).
The PSM will only be appropriate if the local entity contributes something
relatively unique to the marketing IP development (e.g. advertisements in
the Olympics, promotional contracts with famous athletes or highly benefi-
cial distribution arrangements).
23.5. Concluding remarks
Multinationals can generally extract residual profits generated in the Unit-

ed States from the local exploitation of marketing IP to the foreign group
entity that holds legal title to the IP. As discussed in section 23.2., the intan-
gible value behind marketing IP has to be created locally in each marked
jurisdiction in order to have value there. Thus, a US distribution subsidiary
that carries out marketing will contribute to creating the intangible value
that later manifests in US residual profits. In this light, it would only seem
fair if the US subsidiary would be entitled to a portion of the residual prof-
its that it contributed to the creation of.
Nevertheless, if the owner of valuable marketing IP can license it to an un-

related distribution company without surrendering any part of the residual
profits from the exploitation of the IP, it will not be meaningful to put a
related US distribution subsidiary in a better position, as the arm’s length
principle fundamentally aims for parity in taxation between related and
unrelated parties. From the discussions in this chapter, it can be observed
that this axiom is respected under US law to a large extent, as the threshold
2167. Treas. Regs. § 1.482-8(b), Example 12. See the RPSM for services in Treas.
Regs. § 1.482-9(g); and the analysis of the general US PSM in ch. 9.
2168. See the RPSM for services in Treas. Regs. § 1.482-9(g).
2169. The relative values may be difficult to measure. See the discussion of the use of
an unspecified residual profit split under the US PSM in sec. 9.3.2., where this is devel-
oped further.
707
for not giving pricing effect to foreign legal ownership of marketing IP

indeed is set high. The US economic substance exception applied in the
context of internally developed marketing intangibles will only look past
legal ownership if the US subsidiary carries out incremental marketing
efforts without being entitled to concurrent (normal return) compensation
(typically cost-plus-based) or to a portion of the potential future residual
profits. Thus, the economic substance exception will only be available for
application in the relatively extreme scenarios in which the US subsidiary
deducts high marketing costs without there being any link to potential fu-
ture income.
The current economic substance exception was meant to distinguish more

clearly between the rules on intangible ownership and transfer pricing.2170
That ambition has, in the author’s opinion, not been realized (quite the
contrary). The intangible ownership rules are now heavily influenced by
the transfer pricing methods. This is illustrated clearly by the fact that the
income allocation effect of applying the economic substance exception to
determine IP ownership is consistently highlighted in the wristwatch and
athletic gear examples (see sections 23.3.2. and 23.3.3.). The focus is on
whether the US entity receives “an appropriate portion of the premium re-
turn”. How this is achieved seems secondary, as long as the allocation result
itself is acceptable.2171 Also, the lack of guidance for determining the “ap-
propriate” profit spilt under the economic substance exception will, in prac-
tice, make it necessary to resort to analogies to the transfer pricing meth-
ods, in particular, to the PSM.2172 This further reduces the degree to which
it is meaningful to distinguish between the rules for determining ownership
of internally developed intangibles and the transfer pricing methods.
As opposed to the 1968 DA rule, the current economic substance excep-

tion will likely not identify only one group entity that should be allocated
all residual profits from the co-developed intangible, but rather allocate
the residual profits among the group entities that were involved in the in-
tangible value creation, which, in typical scenarios, will be the US distri-
2170. It can be noted that the economic substance exception is systemically a part
of the general provisions on transfer pricing comparability, and therefore inherently
part of the transfer pricing provisions. The comparability provisions of the US regula-
tions are placed in Treas. Regs. § 1.482-1(d), with the economic substance exception in
§ 1.482-1(d)(3)(ii)(B). See also the legal ownership rule in § 1.482-4(f)(3)(A); and the
discussion of comparability in the context of transfer pricing in sec. 6.6.
2171. This is underlined by the fact that the examples leave the door open for the IRS
to “impute one or more agreements … consistent with the economic substance”. See
Treas. Regs. § 1.482-1(d)(3)(ii)(C), Examples 3, 4 and 6.
2172. Treas. Regs. § 1.482-6.
708
Concluding remarks
bution subsidiary and the foreign group entity that holds legal title to the
marketing IP. This is, in the author’s view, a significant shift, as it rejects
the traditional DA-rule construction that a single group entity (the “devel-
oper”) is entitled to the residual profits, while the other entities involved in
the development (the “assisters”) are remunerated on a separate (normal
return) basis.
The fact that the current US regulations do not distinguish clearly between
rules on intangible ownership and transfer pricing should be seen as a mere
consequence of the fact that both categories of rules fundamentally seek to
allocate operating profits among the controlled parties in an arm’s length
manner. A sharp divide between the categories would be rather artificial
in any case.
Further, as discussed in section 23.4., if the economic substance exception

is not triggered – as it normally will not be due to its high threshold for
application – the foreign legal owner of the marketing IP will be entitled
to the residual profits. The question then turns to how the US subsidiary
should be remunerated for its IP development contributions.
There will be no legal basis under US transfer pricing law to allocate a

portion of the residual profits to the US distribution subsidiary when it only
incurs an arm’s length level of marketing expenditures (as discussed in
section 23.4.2.). The real question is whether the situation will be different
if it incurs marketing expenses above an arm’s length level. As discussed
in section 23.4.3., when a US subsidiary develops a foreign-owned market-
ing intangible by incurring incremental marketing expenditures that are
not compensated separately, the picture becomes blurry.2173 The subsidiary
will likely be entitled to a portion of the residual profits under the discrete
owner rule. In addition, it will be allocated an arm’s length return for its in-
cremental marketing efforts, likely under the cost-plus method or the CPM.
As seen from the above, the current US regulations, in some instances,

afford the US subsidiary a portion of the residual profits. Thus, assisters
may now be allocated a portion of the residual profits under the current
regulations.2174 This departs from the “black and white” allocation pattern
of the 1968 DA rule, which allocated the residual profits to the developer,
2173. See the discussion in sec. 23.4.3.2., as well as the analysis of the discrete owner
rule in sec. 20.2.3.
2174. For this reason, the author does not agree with Terr (2004), at p. 564, who argues
that the changes from the 1968 regulations (33 Fed. Reg. 5848) to the proposed 2003
regulations (68 FR 53448-01) regarding intangible ownership were changes more of
709
while the assisters were left with a normal market return, regardless of the
extent to which their development inputs contributed to the creation of the
intangible value. The 1968 DA rule method for compensating assisters is
thus finally abandoned, at least in principle.2175
Nevertheless, the cases in which a US distribution subsidiary may be al-

located residual profits are few. Normally, the foreign legal owner of the
marketing IP will be able to extract the entire residual profits. Thus, the
purpose behind the new 2009 provisions on intangible ownership and re-
muneration of IP development contributions to move away from the “black
and white” allocation patterns under the DA rule seems, in practice, to have
been realized only to a very limited extent. The author’s impression is that
it is still very much decisive for the allocation of residual profits from in-
ternally developed marketing IP as to which group entity is assigned legal
title to it.
A multinational will presumably be inclined to assign legal title to US-

developed marketing IP to foreign group entities (in low-tax jurisdictions)
and focus on establishing sufficient economic substance to substantiate that
this assignment should be given pricing effect in the United States. If there
is a high degree of uncertainty as to whether significant marketing efforts
will succeed, the multinational will likely prefer to deduct the marketing
expenses at the level of the US subsidiary to receive income tax deductions
in a high-tax jurisdiction (i.e. no concurrent cost plus remuneration during
the IP development phase). The current regulations deliver a clear message
for such cases: foreign legal ownership of marketing IP will not be given
pricing effect unless the US subsidiary is allocated a portion of the future
residual profits.
terminology and emphasis than of substance. The author does, however, sincerely agree
with Terr’s expectation that the new regulations do not appear more likely than the 1968
and 1994 approaches (59 FR 34971-01) to avoid the challenging facts-and-circumstan
ces-based assessment akin to that which was assessed in DHL Corp. & Subsidiaries v.
CIR, T.C. Memo. 1998-461 (Tax Ct., 1998), affirmed in part, reversed in part by 285
F.3d 1210 (9th Cir., 2002).
2175. See also Langbein (2005), at p. 1087 on the corresponding rule of the 2003 pro-
posed regulations (68 FR 53448-01).
710
Chapter 24
Distribution among Group Entities of Residual

Developed Marketing IP under the OECD TPG
24.1. Introduction
A marketing intangible developed within a group may generate residual
profits when it is exploited in a market jurisdiction. The typical set-up is
that a distribution subsidiary in the market jurisdiction licenses the intan-
gible property (IP) from a group entity resident in a foreign (often low-tax)
jurisdiction and builds the local value of the IP through marketing efforts.
The question in this chapter is of how the 2017 OECD Transfer Pricing
Guidelines (OECD TPG) allocate residual profits from such marketing IP
among the group entities that contributed to the development of the intan-
gible.
For a general overview of the profit allocation problem for internally devel-
oped marketing intangibles, the author refers to the discussion in chapter
23, under section 23.2.
As discussed in chapter 20, the point of departure is that the group entity
that holds legal title to an item of IP is entitled to the residual profit. The
2017 OECD TPG go far in accepting this in most situations with respect
to marketing IP.2176 Nevertheless, three scenarios are singled out in which
the remuneration of the local distribution subsidiary is up for discussion.2177
These scenarios represent a sliding scale with respect to the level of risks
incurred by the subsidiary in rendering development contributions to the
local value of the foreign-owned marketing IP:
(1) first scenario: the subsidiary’s marketing expenses are reimbursed on a
cost-plus basis. It therefore incurs no marketing risks. This is analysed
in section 24.3.;
2176. OECD Transfer Pricing Guidelines (OECD TPG), para. 6.77. On the legal own-
ership of marketing intangible property (IP), see, in particular, Roberge (2013), at
p. 216. See also Musselli et al. (2008a).
2177. These scenarios are described in the OECD TPG, paras. 6.77 and 6.78. On the
economic ownership of marketing IP, see Roberge (2013), at p. 217.
711
from Exploitation of Internally Developed Marketing IP
(2) second scenario: the subsidiary sustains an arm’s length level of mar-
keting costs, and therefore incurs relatively low marketing risks. This
is discussed in section 24.4.; and
(3) third scenario: the subsidiary sustains marketing costs above an arm’s
length level, and therefore incurs relatively high marketing risks. This
is analysed in section 24.5.
Concluding comments on the profit allocation positions taken in the 2015

OECD TPG are provided in section 24.6.
First, however, the author will, in section 24.2., clarify as to which kind of
marketing profits the discussions in this chapter will pertain to.
24.2. The 2017 OECD TPG require differentiation of

market-based super profits
The 2017 OECD TPG on the ownership of internally developed marketing
IP focuses on how residual profits generated through the local exploita-
tion of a foreign-owned marketing intangible (e.g. trademarks and trade
names) should be allocated between the local distribution subsidiary and
the foreign group entity that holds legal title to the marketing IP. Such
marketing IP is normally the primary source for marketing-related super
profits. Nevertheless, such profits may also encompass profits from local
marketing IP that the subsidiary owns (e.g. goodwill and know-how), as
well as profits from local market characteristics. Assume, for the sake of
argument, a theoretical scenario in which a large multinational distributes
the latest version of its globally best-selling smartphone in China through a
local subsidiary. Residual profits from local sales will likely be due to both
the unique manufacturing IP and marketing IP associated with the device.
Let us disregard the profits that are due to the manufacturing intangibles
and focus on the profits that are marketing-related. These latter profits may
stem from three distinguishable elements:
(1) foreign-owned marketing IP (trademarks and trade names). Such prof-
its will, under the 2017 OECD TPG, normally be allocated to the for-
eign group entity that holds legal title to the IP. As will be discussed in
sections 24.3.-24.6., the OECD has struggled to reach clear consensus
with respect to when the local distribution subsidiary shall be allocat-
ed a portion of the residual profits that are due to the marketing IP; 2178
712
The 2017 OECD TPG require differentiation of market-based super profits
(2) unique intangibles owned by the local distribution entity (e.g. goodwill
and know-how). For instance, the local subsidiary may have developed
unique and valuable customer relationships over a longer period, its
employees may be highly qualified and possess special know-how and
its business may have reputational value (goodwill). The source juris-
diction alone should be entitled to tax profits from such intangibles, as
they are owned there; and
(3) location-specific advantages (e.g. local purchasing power, product
preferences, growing market and good infrastructure). Such profits
will normally be extracted from the source state under the 2017 OECD
TPG.2179
The profits from these three elements may be allocable to different group
entities, as the profit allocations are governed by different transfer pric-
ing norms.2180 To ensure a true arm’s length allocation of marketing-based
super profits, it is therefore crucial that a thorough functional analysis is
performed in order to identify the extent to which the marketing-related
super profits of the local subsidiary stem from these three elements. If it
is automatically assumed that the entire marketing profits are due to the
foreign-owned marketing IP, the result may be that the allocation of operat-
ing profits among the contracting jurisdictions are distorted.
Thus, it must be clarified as to whether the marketing-related super prof-

its of the local subsidiary, in addition to the profits generated through the
exploitation of the foreign-owned IP, also stem from a unique intangible
owned by the subsidiary (e.g. goodwill and know-how). The profits from
such locally owned IP must be allocated to the source state. The wide-
spread use of the transactional net margin method (TNMM) to remunerate
source jurisdictions may be interpreted as an indicator that multinationals
and source-state tax authorities have not identified unique local marketing
intangibles in many cases. The author finds that puzzling. The emphasis
of the 2017 OECD IP guidance on the profit split method (PSM) could, at
least partially, be seen as a reaction to the likely unrealistic assumption that
local distribution entities do not own any unique marketing intangibles.2181
Lastly, it must be clarified as to whether parts of the marketing-based su-
2179. The author refers to the analysis of the allocation of incremental operating profits
due to location-specific advantages in sec. 10.7.
2180. See the discussion of the new OECD allocation rules for profits from local mar-
ket characteristics in ch. 10, and on the corresponding US rules (on comparability with
respect to economic conditions) in sec. 6.6.5.4.
2181. See the comments in ch. 11 on the transfer pricing of intangibles under the
OECD TPG in the post-BEPS era.
713
per profits stem from local market characteristics. If so, the allocation of
the latter profits must be carefully considered when choosing comparables
under the appropriate transfer pricing method.
There should be great potential, both for multinationals and source-state

tax authorities, for achieving profit allocations that more reliably ensure
that intangible value is allocated to the jurisdiction where it was created
through differentiation of the local entity’s marked-based super profits into
the above-mentioned three categories.
The discussions in sections 24.3.-24.6. below will address only the alloca-
tion of market-based super profits that stem from one of the three elements
mentioned above, namely from foreign-owned marked IP (e.g. trademarks
and trade names). The reason for this is that the IP ownership rules of the
2017 OECD TPG with respect to marketing IP focus only on this. For other
unique marketing IP that the local subsidiary has developed by itself (e.g.
goodwill), there will be no allocation discussion. The profits from such
IP shall be taxed in the source state, as the local subsidiary owns it. With
respect to the allocation of profits from local market characteristics, the
author refers to the analysis in chapter 10.

is reimbursed on a cost-plus basis
The question in this first scenario is whether the 2017 OECD TPG, based
on economic substance considerations, will deny controlled profit alloca-
tions that extract the residual profits that a local distribution subsidiary has
generated through the exploitation of marketing IP (that it licenses) to the
foreign group entity that holds legal title to the IP in cases in which the
subsidiary, throughout the development phase, is reimbursed by the foreign
group entity for its marketing costs to build the local value of the IP.2182
The historical 1995 OECD TPG position was that the subsidiary would
only be entitled to a normal market return on its distribution and marketing
services, as opposed to a stake in the subsequent residual profits.2183 The
2010 consensus text is identical.2184 Seemingly in line with the historical
position, the new 2017 guidance states that the local subsidiary will only

2183. 1995 OECD TPG, para. 6.37.
2184. 2010 OECD TPG, para. 6.37.
714
Scenario 1: The local group distribution entity is reimbursed on
a cost-plus basis
be entitled to a normal market return “where a distributor acts merely as

an agent, being reimbursed for its promotional expenditures and being di-
rected and controlled in its activities by the owner of the trademarks and
other marketing intangibles”.2185 The rationale behind this position is that
the subsidiary should not be entitled to any additional profits above a nor-
mal market return when it does not “assume the risks … with the further
development of the … intangibles”.2186
The application of this rule is illustrated in an example (the “Primair” ex-

ample, which is clearly inspired by the US regulations; see the discussion
in section 23.4.2.) pertaining to a newly established subsidiary in juris-
diction Y, which markets and sells watches purchased from its parent in
jurisdiction X under the trademark and trade name R.2187 The marketing
intangibles are owned by the parent and are widely known and valuable in
country X, but not in country Y. An exclusive, long-term, royalty-free mar-
keting arrangement is entered into, pursuant to which the subsidiary shall
build the local value of the licensed marketing IP. The purchase price per
watch paid to the parent is at arm’s length (apparently determined on the
basis of comparable uncontrolled transactions (CUTs)) and provides the
subsidiary with sufficient remuneration for its distribution activities. The
question is of how the subsidiary should be remunerated for its marketing
development contributions to the local value of the R watch intangibles.
The example highlights that the foreign parent “develops the overall mar-
keting plan based largely on its experience in other countries, it develops
and approves the marketing budgets, and it makes final decisions regarding
advertising designs, product positioning and core advertising messages”.2188
The subsidiary executes the global marketing strategy locally under the di-
rection of its parent and provides feedback on the local effectiveness of the
marketing. In doing so, it incurs marketing expenses, which are reimbursed
through a concurrent cost-plus remuneration, with the mark-up determined
on the basis of the mark-ups experienced by independent advertising agents
that are deemed comparable.2189 The example concludes that the foreign

2186. Id.
2187. OECD TPG, appendix to ch. VI, Example 8, at para. 22. The wording in the final
version of the example is identical to that used in the 2013 and 2014 drafts. The example
contains elaborate and specific premises and is likely inspired by the US “wristwatch”
examples; see the analysis in secs. 23.3.2., 23.3.6. and 23.4.2.
2188. OECD TPG, appendix to ch. VI, Example 8 (Primair), at para. 22.
2189. Normally, the author would object to such comparables; see the discussion of
the US Treas. Regs. § 1.482-8(b), Example 11 on the comparable profits method in sec.
23.4.4. In the US example, marketing was only an auxiliary part of the subsidiary’s
715
parent, because it owns the licensed marketing IP, “is entitled to retain any
income derived from exploiting the R trademark and trade name in the
country Y market that exceeds the arm’s length compensation”.2190 Thus,
the controlled cost-plus remuneration of the subsidiary’s marketing efforts
is accepted, with the result that the residual profits are extracted from the
market jurisdiction and allocated to the foreign parent.
In order to avoid any residual profits being allocated to the distribution sub-
sidiary, it must – in addition to being reimbursed for its marketing expenses
with an arm’s length mark-up – be “directed and controlled in its activities”
by the foreign group entity that holds legal title to the licensed market-
ing IP so that it acts “merely as an agent”.2191 This requirement is clearly
based on the new OECD comparability guidance on risk and economic
substance,2192 and forms a parallel to the “important functions” doctrine for
internally developed manufacturing IP.2193
In terms of context, multinationals tend to centralize high-level marketing

decisions, ensuring that local sales entities pursue a uniform global mar-
keting pattern. That was the case in GSK, and was one of the main taxpayer
arguments for allocating the residual profits from the US exploitation of
the trade name Zantac in the United Kingdom.2194 Thus, a local subsidiary
will rarely perform all important marketing functions, but may neverthe-
less have a significant amount of freedom with respect to how it chooses
to approach local marketing within the global framework of the group.
This may entail that some important functions are performed locally in the
source state.2195
business. The author finds it inappropriate to compare its operating margins on the mar-
keting activity to the margins realized by specialized third-party marketing enterprises
in the absence of suitable comparability adjustments. In this example, however, country
Y’s subsidiary is newly established, with distribution and marketing of the watches
sold under the R trademark as its only activities. The marketing carried out by the
subsidiary should therefore likely not be deemed auxiliary. Experienced and special-
ized third-party marketing enterprises would likely realize higher operating margins
on their promotional activities than a newly established company that only markets one
type of product. Comparability adjustments should therefore be implemented to make
the third-party margins more reliable, and the interquartile range should be used.
2190. OECD TPG, appendix to ch. VI, Example 8 (Primair), at para. 25.
2192. See the analysis on comparability and risk in sec. 6.6.5.5.
2193. See the analysis of the important-functions doctrine in sec. 22.3.2.
2194. The group entity that held legal ownership to the trade name was resident in the
United Kingdom. See the discussion of the GSK case in sec. 19.2.5.2.
2195. For instance, it has been asserted that the Canadian subsidiary in the Glaxo case
(see the discussion in sec. 6.7.4.) had a large degree of autonomy with respect to local
marketing efforts; see Roberge (2013), at p. 229.
716
Scenario 1: The local group distribution entity is reimbursed on
a cost-plus basis
The wording “directed and controlled” and “merely as an agent” lends sup-
port to the interpretation that the control of the foreign owner must be rela-
tively intense and encompass all local activities that pertain to the licensed
marketing IP in order for the foreign group entity to retain entitlement to
the entire residual profits. Clearly, all high-level decisions pertaining to the
local marketing efforts must be governed by the foreign owner (e.g. mar-
keting strategy design and budget determination).
The question is whether important local day-to-day operating decisions

(e.g. localization of global marketing material and advertising media pur-
chases) must also be controlled by the foreign owner. The 2017 OECD
TPG do not provide any specific direction for this question. The answer
must therefore be based on ordinary transfer pricing considerations.
First, if the subsidiary makes substantial local marketing decisions, it

could arguably retain sufficient autonomy to not qualify as “merely an
agent”, even if its decisions are within the boundaries of the global market-
ing strategy of the group.
Second, marketing activities performed in the source state should be seen

as IP development contributions that enhance the local value of the licensed
marketing IP. It could be argued, on the basis of causality, that because the
development activities that create the local value are carried out by the
subsidiary, it should be entitled to reap some of the residual profits that may
result from its contributions.
Third, it could be argued that the subsidiary incurs the IP development

risks. The day-to-day operational decisions it makes influence whether its
marketing efforts are successful or not, and thus whether the local value of
the licensed marketing IP increases. It could be questioned as to whether
it would align with the economic substance of the actual behaviour of the
controlled parties if the intangible development risks (and the residual
profits) were allocated to the foreign group entity that holds legal title to
the IP in such circumstances.
Let us say, for the sake of argument, that a distribution subsidiary makes
significant day-to-day decisions pertaining to local marketing strategies
and design that fit within the multinational’s global marketing programme,
and, within the upper limits established by the foreign owner, it is entitled
to determine the amount of marketing expenses to be incurred. The ex-
penses are reimbursed on an arm’s length, cost-plus basis. If one were to
assess this factual pattern based on the interpretation of the current guid-
717
ance suggested above, the residual profits would likely be split between
the local subsidiary and the foreign group entity that holds legal title to the
marketing IP, as the former would not qualify as a “mere agent”. Would
such an allocation yield an arm’s length result?
The allocation result would then be that a group entity that does not incur
the financial risks associated with the IP development is allocated a portion
of the residual profits. This, in itself, should not be seen as a controversial
position, as the same result may follow from an application of the important-
functions doctrine to internally developed manufacturing IP.2196 Further, it
could be argued that third-party licensers of unique and valuable marketing
intangibles do generally not accept terms that allocate residual profits to a
local licensee. The licenser would likely have better realistic alternatives
available, because only he contributes unique inputs to the value chain.2197
The basic reason for this is that the services provided by the local distributor
would likely be relatively generic, and could be outsourced to third parties
for a normal market return fee. Thus, the argument is that such a licenser
would generally be in a superior bargaining position vis-à-vis a local dis-
tributor that only brings routine inputs to the table. The licenser supplies the
necessary rights to the unique marketing IP, the capital necessary to build
its local value and the high-level instructions for how the marketing should
be carried out, and therefore also bears all risks connected to the IP devel-
opment. It seems unrealistic to assume that an unrelated licenser in this case
would not demand all residual profits from the developed intangible.
The author finds this decisive. An arm’s length profit allocation must align
with the realistic alternatives available to the controlled parties. This result
will normally not be attained if a local distribution subsidiary, which is reim-
bursed for its marketing expenses with the addition of an arm’s length mark-
up margin, is allocated a portion of the residual profits, even if it does make
some significant marketing decisions on a relatively independent basis.2198
Also, the new “directed and controlled” add-on (to the wording carried
over from the 1995/2010 OECD TPG) is, as touched upon above in this
section, influenced by the new 2017 guidance on risk and economic sub-
stance, as well as the important-functions doctrine for manufacturing IP.
2196. For instance, the residual profits are allocated to the group entity that performs
research and development (R&D), even though the intangible development was funded
by a foreign “cash-box” entity or IP holding company.
2197. See the discussion of foreign direct investment (FDI) in sec. 2.3.
2198. See also Barbera (2003), at p. 71 on royalty structures based on the reimburse-
ment of marketing expenses.
718
Scenario 2: The local group distribution entity bears an arm’s length level
of marketing costs
Understandably, the OECD also wanted to apply the important-functions

doctrine to the development of marketing IP. An obvious problem with this
is that the distinction between risk and intangible development funding is
more meaningful in the context of manufacturing intangibles. The value
of unique manufacturing IP is driven by research and development (R&D)
carried out by talented researchers. Such talent represents scarce resources,
and is therefore highly valuable. This has no direct parallel in the context
of marketing IP, as the increase in local marketing value will often rely
directly on the amount of, for example, local media exposure, as well as the
quality of the underlying product being promoted. The important-functions
doctrine is not fit for the internal development of marketing intangibles due
to the absence of unique development contributions on the side of the local
distribution subsidiary.
In conclusion, it is the author’s view that the wording of the current 2017
guidance should be interpreted narrowly. The local subsidiary should be
deemed to act “merely as an agent” even if it is free to decide on local
marketing matters on a day-to-day basis, as long as its actions align with
the global marketing framework indicated by the foreign owner of the mar-
keting IP.2199 A more liberal interpretation would result in an allocation of
operating profits among jurisdictions that would be contrary to third-party
behaviour and the realistic alternatives available to the controlled parties.
Thus, there will be no legal basis under the 2017 OECD TPG to allocate
any residual profits to a local distribution subsidiary that is reimbursed for
its arm’s length IP development contributions (in the form of marketing
expenses) on a concurrent cost-plus basis throughout the IP development
phase. The material content of the OECD TPG here is in line with the US
Internal Revenue Code (IRC) section 482 regulations, which provide the
same result for this scenario (see the discussion in section 23.3.5.).

bears an arm’s length level of marketing costs
24.4.1. The subsidiary must earn a normal market return

throughout the IP development phase
The question in this second scenario is whether the 2017 OECD TPG,
based on economic substance considerations, will deny controlled profit al-
2199. Such a framework should include the overall marketing strategy and country-by-
country marketing budgets.
719
locations that extract the residual profits that a local distribution subsidiary
has generated through the exploitation of marketing IP (that it licenses) to
the foreign group entity that holds legal title to the IP in cases in which the
subsidiary is not reimbursed for its arm’s length marketing costs incurred
to build the local value of the marketing IP under a long-term distribution
agreement.2200
The point of departure for this problem under the 1995 (and identical 2010)
OECD TPG consensus text was that “the ability of a party that is not the
legal owner of a marketing intangible to obtain the future benefits of mar-
keting activities that increase the value of that intangible will depend prin-
cipally on the substance of the rights of that party”.2201
Thus, if there were no basis in the controlled pricing agreement between

the foreign group entity that held legal title to the IP and the local distribu-
tion entity that the latter should be allocated a portion of the residual prof-
its, there would be no legal basis for such allocation under the 1995/2010
OECD TPG. Nevertheless, the 1995/2010 OECD TPG recognized that the
local subsidiary should be allocated a normal market return for its IP de-
velopment contributions in the form of increased local sales when the con-
trolled distribution agreement provided for long-term and exclusive distri-
bution rights for the trademarked product. The distributor’s (normal return)
share of benefits should then be determined with reference to the operating
profits earned by independent distributors in comparable circumstances
(typically TNMM-based).2202
While the new 2017 OECD TPG clearly retain this point of departure,2203 it
is also added that the “distributor’s efforts may have enhanced the value of
its own intangibles, namely its distribution rights”.2204
This addition seems to draw heavily on the US discrete owner rule (see the
discussion in section 20.2.3.). Unfortunately, it also suffers from similar
ambiguities with respect to the extent that the local distribution subsidi-

2201. 1995 and 2010 OECD TPG, para. 6.38.
2202. Id.
2203. OECD TPG, para. 6.78. It is added that the residual profits would be the result
of the subsidiary’s “functions performed, assets used, and risks assumed currently or
in the future”. The 2013 OECD Revised discussion draft on transfer pricing aspects
of intangibles (2013 RDD) also emphasized the costs incurred by the local subsidiary
(see paras. 95 and 96), but this wording was taken out in the final consensus text. This
particular revision is, in the author’s view, inconsequential.
720
of marketing costs
ary should be entitled to residual profits from its local exploitation of the
licensed marketing IP.
The 2017 OECD TPG include an example on the application of this “US
discrete owner”-influenced rule, which uses the factual pattern in the ex-
ample discussed in section 24.3. as its base (the Primair example; see the
discussion in section 24.3.), and adds the twist that the subsidiary now
develops, funds and executes its own market plan.2205 The foreign group
entity that holds legal title to the marketing IP exercises a lesser degree of
control over the subsidiary’s marketing efforts. The differences between
the controlled transfer pricing structures applied in the (Primair) base ex-
ample and this extension example relevant to the current problem are as
follows:
– In the base example, the subsidiary is compensated in two ways: (i)

it receives separate cost-plus remuneration for its marketing activities
on a concurrent basis throughout the IP development phase; and (ii)
the purchase price for the R watches is set to a level that enables the
subsidiary to realize an arm’s length operating margin on its sales ac-
tivities.
– In the extension example relevant to the current problem, the subsidi-

ary is remunerated only through a lowered purchase price for the R
watches, which must be low enough to provide it with an operating
margin large enough to yield an arm’s length compensation for both its
marketing and sales activities. In other words, there is no concurrent
compensation to the subsidiary with respect to its IP development con-
tributions throughout the IP development phase. The subsidiary bears
its (arm’s length) marketing expenses itself.
The question is how the profit margin allocated to the subsidiary through
the lowered purchase price for the watches shall be determined. There is
no doubt that the subsidiary should be allocated “a greater anticipated total
profit” than in the base (Primair) example (in which it was reimbursed
for its marketing expenses) because of its additional functions and risks,
because it now bears its own marketing expenses.2206 The question is of
how much more profit (than cost plus) shall be allocated to the subsidiary.

2206. OECD TPG, annex to ch. VI, Example 9, para. 26, second bullet point. The
wording used in the 2013 RDD, annex, Example 6, was “earning a greater total profit”.
The author does not attach any particular consequence to the revised wording in the
final text.
721
Since the subsidiary bears the marketing costs, it experiences high operat-
ing expenditures and modest margins in years 1-3.2207 CUTs indicate that
the functions performed, the level of marketing expenditures incurred and
the operating profits realized by the subsidiary in these years are similar
to those of independent distributors bearing the same types of risks and
costs in the initial years of comparable long-term marketing and distribu-
tion agreements for similarly unknown products.
Unfortunately, the example offers little guidance on the specifics of how

the subsidiary should be remunerated. It essentially just refers to its own
premise that the allocation of profits between the subsidiary and the foreign
owner entity corresponds to the allocation indicated by the CUTs.
An underlying point, however, seems to be that the exclusive, long-term

distribution agreement should offer the subsidiary an “opportunity … to
benefit (or suffer a loss)” from its marketing efforts, akin to what would be
the case for comparable third-party distributors.2208 The author interprets
this wording to entail that the subsidiary should be entitled to a normal
market return profit margin for its marketing and distribution activities
(typically determined on the basis of the TNMM). This, however, is where
the guidance becomes problematic.
Logically, one would assume that if the subsidiary were to benefit from its
marketing activities based on the view that it has “enhanced the value of its
own intangible, namely its distribution rights”,2209 one would stop here and
say that the subsidiary should be entitled to keep its net profits, or at least a
healthy portion of it. Provided that the subsidiary’s purchase price for the
R watches is at arm’s length, this would, in effect, be the same as saying
that the subsidiary is entitled to the residual profits generated by the local
exploitation of the marketing intangible.
The basic problem with this reasoning is that the example at the same time
instructs that the profits allocable to the subsidiary should be determined
by reference to the profits made by comparable unrelated enterprises. Of
course, third-party enterprises that carry out marketing and sales for prod-
ucts trademarked by unique intangibles that they do not own must pay roy-
alties to the licensers. How much of the licensee’s operating profits are
allocated to the unrelated licensers will depend on the bargaining posi-
2207. OECD TPG, annex to ch. VI, Example 9, at para. 27.

2208. Id., at para. 28.
722
of marketing costs
tions of the parties, in particular on whether the licensee brings any unique
inputs to the table. Normally, when the licensed intangible is unique and
valuable and the licensee only provides routine marketing and distribution
functions, it should be presumed that the licensee would likely not retain
any residual profits.2210
This results in the basic observation that the operating profits of unrelated
distributors (as reflected in the operating profits in the financial statement)
are the profits they retain after deductions have been made for royalties
paid to the licensers. Thus, even though the new guidance states that the
subsidiary should have the “opportunity … to benefit (or suffer a loss)”
because it “enhanced the value of its own intangible”, its profits will be
benchmarked, likely under the TNMM, against the post-royalty profits of
unrelated comparable distributors. In this light, the author fails to see that
there is any substance to the impression that one might immediately get
from the wording of the new 2017 OECD guidance that the subsidiary
should be entitled to some of the residual profits because it acts relatively
independently and incurs risks.
Further, as the example stands, it should be regarded as fairly certain that

a cost-plus-based remuneration rendered after the IP has been developed
will not be sufficient when the subsidiary bears its own marketing expendi-
tures. Such cost-plus remuneration will not compensate the subsidiary suf-
ficiently for the risks it incurs through its marketing expenditures through-
out the IP development phase.
In conclusion, the 2017 OECD TPG should be read as requiring concur-

rent allocation of a normal market return (likely TNMM-based) through-
out the IP development phase to a distribution subsidiary that incurs an
arm’s length level of marketing expenses in order to uphold the controlled
pricing that allocates the residual profits to the foreign group entity that
holds legal title to the marketing IP. Reimbursement of the subsidiary’s
marketing expenses on a cost-plus basis after the IP development is com-
pleted and the marketing IP begins generating residual profits should
not be regarded as sufficient. The material content of the profit alloca-
tion solution offered by the OECD for this scenario is the same as that
drawn up by the US IRC section 482 regulations (see the discussion in
section 23.4.).
2210. See the discussion on FDI in sec. 2.3.
723
24.4.2. The first twist: The duration of the distribution

agreement is shortened; can residual profits then
be allocated to the subsidiary?
The 2017 OECD TPG provide two twists to the exclusive long-term distri-
bution (Primair) example discussed in section 24.4.1.
The first extension alters the factual pattern so that the distribution agree-
ment is a short-term, royalty-free agreement, which spans over 3 years with
no option to renew.2211 The distribution subsidiary is now unable to benefit
from the arm’s length marketing costs it incurs. The example assumes that
third parties enter into short-term distribution agreements only when “they
stand to earn a reward commensurate with the functions performed, the as-
sets used, and the risks assumed within the time period of the contract”.2212
Further, the short-term nature of the distribution agreement makes it “un-

reasonable to expect” that the subsidiary may obtain “appropriate benefits
under the contract within the limited term of the agreement”, as it incurs
substantially higher risks – without compensation – than would have been
the case in a long-term agreement.2213 On this basis, the OECD TPG con-
clude that the subsidiary is entitled to compensation for its risk contribution
to the value of the marketing IP.
However, the example is ambiguous with respect to the extent of this re-
muneration. It states that the compensation could take the form of either
“direct compensation … for the anticipated value created through the mar-
keting expenditures and … functions”, or a “reduction in the price paid …
for R watches during Years 1 through 3”.2214
While the second alternative merely pertains to form, the first alternative
does seem to provide a hint as to the level of the required profit allocation.
It links the compensation to the anticipated increase in value of the market-
ing IP. The author interprets this wording as requiring that the subsidiary,
2211. OECD TPG, annex to ch. VI, Example 11. On short and long-term distribution
agreements and the effects for transfer pricing purposes, see Roberge (2013), at p. 217,
where doubt is expressed with respect to the common assumption that most intra-group
distribution agreements are long-term. See also Becker (2008), at p. 461, who argues
that not updating the royalty rate in long-term distribution agreements creates risk for
the distributor.
2213. Id., at para. 37.
2214. Id., at para. 38.
724
of marketing costs
over the 3-year term of the agreement, shall be allocated operating profits
equal to the net present value of the expected residual profits from the local
exploitation of the trademark in future income years. This may trigger the
allocation of substantial amounts of operating profits to the subsidiary. It
will likely be necessary to resort to the 2017 guidance on valuation in order
to estimate the net present value.2215
This result is somewhat surprising. After all, the foreign parent is the own-
er of the valuable marketing IP, and therefore provides a unique value chain
input. The subsidiary certainly provides marketing functions and expenses
and assumes associated risks, but these inputs should likely be deemed
as rather generic. From this perspective, it is difficult to argue that the al-
location of the residual profits to the subsidiary would be aligned with the
realistic alternatives of the controlled parties.
Nevertheless, even if the licenser has a superior bargaining position, it must

at least provide the licensee with an arm’s length normal market return on
a concurrent basis. That has not been done. The example therefore bears
resemblance to the imputation authority under the economic substance ex-
ception of the US regulations, where the lack of concurrent normal return
compensation may trigger a claim for the subsequent residual profits (see
the discussion in section 23.3.).2216 Ultimately, the example seems to treat
the subsidiary as the owner of the local rights to the marketing intangible,
and thus entitled to the residual profits from their exploitation.
In conclusion, it can be observed that the OECD is unwilling to give pric-

ing effect to foreign legal ownership of internally developed marketing IP
in cases in which the local distribution subsidiary incurs (what is at the out-
set) an arm’s length level of marketing expenditures when the distribution
agreement is of such limited duration that the subsidiary will be unable to
reap any reasonable benefit from its efforts. In these cases, the OECD guid-
ance seems to provide a legal basis for allocating the residual profits from
the local exploitation of the marketing IP to the subsidiary.
2215. See the analysis of the new guidance on valuation in ch. 13.
2216. The fact that the example should be seen as an application of the economic sub-
stance doctrine is further strengthened by the premise that “evidence derived from
comparable independent enterprises” shows that they do not enter into short-term
agreements with no hope of benefitting from their investments; see OECD TPG, annex
to ch. VI, Example 11, para. 36. See also the analysis of the US imputation authority
under the economic substance exception in the context of manufacturing and marketing
intangibles in secs. 21.3.3. and 23.3., respectively.
725
24.4.3. The second twist: A royalty payment is introduced;

can residual profits then be allocated to
the subsidiary?
The second extension of the exclusive, long-term distribution (Primair) ex-

ample (discussed in section 24.4.1.) introduces a royalty payment at the
end of the third year when the brand has been successfully established in
the local market.2217 Without this royalty payment, the subsidiary would
have been allocated the residual profits due to the combination of higher
margins and an unaltered purchase price for the R watches sold from the
foreign parent. The example makes two assumptions. First, the marketing
activities and expenses of the subsidiary from year 4 are consistent with
those of comparable unrelated distributors. Second, there is no evidence
that such distributors would have agreed to pay royalties in similar circum-
stances.
This factual pattern is relatively similar to that discussed under the eco-
nomic substance exception in the US regulations,2218 pertaining to a US
distribution subsidiary that incurs marketing costs during the first 6 years
of a long-term distribution agreement. Once the product is successfully
established, all residual profits are allocated to the foreign legal owner. The
US regulations take the position that this controlled allocation of operating
profits contradicts the economic substance of the transaction. It is found
unlikely that an unrelated distributor would enter into a long-term market-
ing and distribution agreement unless it received concurrent remuneration
or had an expectation of deriving some future benefit. Thus, the foreign
legal owner may only extract the subsequent residual profits if it remuner-
ates the subsidiary for its marketing on a concurrent basis.
A significant difference between the factual pattern discussed in the US

regulations and that of the current OECD example is that the local dis-
tribution subsidiary incurs incremental marketing costs in the former,
while it does not in the latter. Thus, the US guidance pertains to a more
aggressive transfer pricing structure in which the local subsidiary is put
even worse off.
2217. OECD TPG, annex to ch. VI, Example 12. The renegotiated agreement is an-
other long-term agreement, similar in all respects to the original agreement, apart from
the royalty.
2218. See Treas. Regs. § 1.482-1(d)(3)(ii)(C), Example 3; and the discussion in sec.
23.3.2.
726
of marketing costs
The position taken by the OECD example, in surprisingly clear language, is

that a transfer pricing adjustment disallowing the royalty payments should
be carried out. The basis for this conclusion is two-fold.
First, a royalty would generally not be expected when “a marketing and

distribution entity obtains no rights for transfer pricing purposes in trade-
marks and similar intangibles other than the right to use such intangibles in
distributing a branded product supplied by the entity entitled to the income
derived from exploiting such intangibles”.2219
The author is far from convinced by this reasoning. In his view, there
seems to be no principal basis as to why a royalty payment could not be ap-
propriate in such circumstances. After all, the purpose of the royalty would
be to allocate the residual profits to the foreign group entity that holds legal
title to the marketing IP, and the OECD does, in the quoted wording, ad-
mit that the owner is “entitled to the income derived from exploiting such
intangibles”.2220 Whether the income is extracted from the market jurisdic-
tion through (i) the sales price for the distributed products from the owner
of the intangible to the local distributor; or (ii) through a royalty should be
beside the point.
Second, and considerably more persuasive, is the argument that the operat-
ing margins of the subsidiary in year 4 and on are “consistently lower than
those of independent enterprises with comparable functions performed,
assets used and risks assumed during the corresponding years of simi-
lar long-term marketing and distribution agreements”.2221 In other words,
there is a mismatch between the operating margins that the transfer pricing
methods (e.g. the TNMM) indicate should be allocated to the subsidiary
and its reported income.
Thus, given that the guidance on temporary pricing strategies cannot be

applied to justify the subsidiary’s low returns, there must be an adjust-
ment.2222 However, the example disallows the royalties entirely. A more ap-
propriate result would, in the author’s view, be to adjust the subsidiary’s in-
come as indicated by the applicable transfer pricing method so that it falls
within the arm’s length range of the operating margins of the comparable
2220. Id.
2221. Id.
2222. Temporary pricing strategies are discussed in sec. 6.6.5.4.4. The guidance, in
the author’s view, is not relevant, as the trademark is established and generates residual
profits in year 4 onwards.
727
unrelated enterprises. It should be perfectly clear that this does not neces-
sarily require the disallowance of the entire royalty amount. The royalty
payments could be partially upheld as long as the subsidiary’s operating
margins align with the results of the transfer pricing method.
Even though it is categorically stated in the OECD example that the entire
royalty payments should be disallowed, the author finds that his interpreta-
tion is reconcilable with the message provided by the example. After all,
the point is that the royalty payments should be disallowed because the
subsidiary’s profit margins are “consistently lower than those of independ-
ent enterprises”.2223 Thus, if the controlled margins are adjusted so that
they align with those of the unrelated distributors, there should be no jus-
tification in the OECD transfer pricing methodology for disallowing the
exceeding royalties, if any. The author therefore finds it likely that this
extension example envisions the TNMM-based profit allocation of the base
example to be continued from year 4 and on, entailing that the residual
profits remaining (after upwards TNMM adjustment of the subsidiary’s
profit margin) are extracted from the market jurisdiction.
In light of the fact that the extension pertains to a distribution subsidiary

that incurs an arm’s length level of marketing expenditures, the OECD’s
stance is more restrictive towards allocating residual marketing profits to a
foreign group owner than the US regulations (see the analysis of the paral-
lel problem under the US regulations in section 23.4.2.). Nevertheless, the
author finds it unlikely that a conflict may arise between the two sets of
rules. As the US regulations are concerned with the allocation of residual
profits to a US distribution subsidiary, it will only support the view that the
OECD seems to go further in the same direction.
In conclusion, the author finds no legal basis in the “second twist” OECD
(Primair) example on which to allocate any portion of the residual profits
from the local exploitation of foreign-owned marketing IP to the local sub-
sidiary. The residual marketing profits should be allocated to the foreign
owner of the IP. Only a normal market return shall be allocated to the sub-
sidiary under the one-sided methods (likely the TNMM). This may neces-
sitate additional (normal return) profit allocation to the subsidiary, but this
allocation should not include a portion of the residual profits.
2223. OECD TPG, annex to ch. VI, Example 12, at para. 41. The 2013 draft of the
example (2013 RDD, annex, Example 9) emphasized this even further, as it, in addition
to the current wording, stated that “Company S’s profit margins are consistently lower
than the profit margins of independent enterprises during the corresponding years of
similar long-term marketing and distribution agreements because of the royalty”.
728
Scenario 3: The local group distribution entity bears an above-arm’s length
level of marketing costs

bears an above-arm’s length level of marketing costs
The question in this third scenario is whether the 2017 OECD TPG, based
on economic substance considerations, will deny controlled profit allo-
cations that extract from the market jurisdiction the residual profits that
a local distribution subsidiary has generated through the exploitation of
marketing IP (that it licenses) to the foreign group entity that holds legal
title to the IP in cases in which the subsidiary incurs marketing expen-
ditures exceeding those that an unrelated comparable distributor would
have agreed to.2224 This is a classic transfer pricing problem that has been
the subject of repeated OECD negotiations and is analogous to the issue
discussed in the context of the US IRC section 482 regulations in section
23.4.3.2225 The analysis of this problem is divided into two sections. The
author discusses the threshold set out in the 2017 OECD TPG for apply-
ing economic substance considerations to the controlled pricing in section
24.5.2. Thereafter, in section 24.5.3., he discusses the material content of
the profit allocation that should be applied in cases in which the economic
substance threshold has been breached.
24.5.2. The OECD TPG threshold for additional profit

allocation to the source-state distribution subsidiary
In this third scenario, the local subsidiary incurs marketing expenditures

and risks that go beyond what normal low-risk distributors (LRDs) would
be willing to assume. Such high levels of marketing may be necessary
to establish an unknown product in a competitive market. As a result of
concurrent income tax deductions for such marketing costs (combined
2224. OECD TPG, para. 6.78. It has been asserted in the literature that this problem
is highly practical, as most group distribution entities incur more marketing expenses
than third-party distributors; see Allen et al. (2006), at sec. 5.3. On this issue, see thor-
ough reflections (based on the 2010 version of the OECD TPG) in Roberge (2013).
2225. This problem pertains to the so-called “bright line test”, where the question is
whether it can be accepted that the local group entity performs marketing on a service-
provider basis (and is thus only entitled to a normal market return), or if the entity
should be deemed to own the local marketing value it has created (and thus be entitled
to residual profits). The “service provider” classification is normally contingent on the
local entity only incurring a “routine” level of marketing expenses, while the “owner”
classification is contingent on a “non-routine” level of marketing expenses. See, e.g.
Musselli et al. (2008a), at p. 263.
729
with the allocation of income premised on an LRD characterization), the

distribution entity may fail to report a net taxable position. This scenario
has proven to be particularly relevant when multinationals distribute their
products in developing countries. India has a substantial body of case law
pertaining to local distribution entities incurring significant advertising ex-
penditures coupled with low returns, resulting in tax losses.2226 China has
had similar experiences.2227
Such taxpayer loss return positions may be difficult to reconcile with the
classification of the local subsidiary as an LRD. Herein lies the heart of the
matter. An LRD is expected to earn a modest (but steady) market return on
the low-risk, routine functions it performs. When it does not, the question
arises as to whether the economic substance of the transaction is such that
a low-risk characterization of the distribution entity should not be accepted
for pricing purposes, i.e. whether the entity should be remunerated with a
(higher) normal market return pursuant to the one-sided methods, or even
be entitled to a split of the residual profits.
The point of departure in the 1995 and 2010 OECD TPG was, as men-
tioned in sections 24.3.-24.4., that residual profits yielded through the lo-
cal exploitation of foreign-owned marketing IP should be allocated to the
foreign group entity that holds legal title to the IP (i.e. extraction of residual
profits from the source), while the source-state group distribution entity
should be remunerated on a separate basis to provide it with a normal mar-
2226. See, e.g. Maruti Suzuki India Ltd v. CIT, W.P.(C) 6876/2008 (Delhi High Court,
2010); Aztec Software & Technology Services Limited v. CIT, 2007 107 ITD 141 Bang
(Income Tax Appellate Tribunal – Bangalore, 2007); Rolls Royce Plc v. Dy. Director
Of Income-Tax, (2008) 113 TTJ Delhi 446 (Income Tax Appellate Tribunal – Delhi,
2007); Sony India (P.) Ltd. v. CBDT, (2006) 206 CTR Del 157 (Delhi High Court,
2006); Amadeus Global Travel Distribution v. CIT, 113 TTJ Delhi 767 (Income Tax
Appellate Tribunal – Delhi, 2007); UCB India (P) Ltd. v. CIT, 30 SOT 95 (Income Tax
Appellate Tribunal – Mumbai, 2009); Global Vantedge Pvt. Ltd. v. CIT, ITA Nos. 116
& 323/Del/2011 (Income Tax Appellate Tribunal – Delhi, 2010); Intel Asia Electronics
Inc. v. Asstt Director Of Income Tax, ITA No. 131/Bang/2010 (Income Tax Appellate
Tribunal – Bangalore, 2009); Amadeus India Pvt Ltd v. CIT, ITA 938/2011 (Delhi High
Court, 2011); and Symantec Software Solutions Pvt Ltd v. CIT, ITA No. 7894 (Income
Tax Appellate Tribunal – Mumbai, 2010). With respect to India’s position on the alloca-
tion of operating profits generated through marketing intangibles, see the UN Transfer
Pricing Manual (UN TPM), paras. 10.4.8.5, 10.4.8.7 and 10.4.8.12-10.4.8.18. See also
Levey et al. (2011). For insightful comments on the Maruti case, see, in particular,
Casley et al. (2011). For further information on India’s approach to the transfer pricing
of IP, see Lagarden (2014), at p. 344.
2227. China will likely assert the application of a profit split approach in such cases;
see UN TPM, paras. 10.3.5.10-10.3.5.11. See also Bell (2012) on this issue; and Lagar-
den (2014), at p. 345 on China and IP transfer pricing.
730
ket return for its development contributions.2228 Additional operating profits

could, however, be allocated to the source state if the local distribution en-
tity incurred (without being reimbursed) “extraordinary marketing expen-
ditures beyond what an independent distributor with similar rights might
incur for the benefit of its own distribution activities”.2229 In these cases, the
local subsidiary “might obtain an additional return from the owner of the
trademark, perhaps through a decrease in the purchase price of the product
or a reduction in royalty rate”.2230
This wording was, in the author’s view, ambiguous both with respect to the
“extraordinary marketing expenditures” criterion for triggering additional
source-state compensation2231 as well as the profit allocation consequences
indicated by an “additional return”. The absence of clarity was likely due to
difficulties in reaching consensus on sharper criteria in 1995.2232
The 2017 OECD TPG retain the point of departure of the 1995 and 2010
text, but add some modifications.2233 There now seem to be two cumulative
criteria for triggering this particular allocation rule.
First, the 1995 reference to “extraordinary marketing expenditures” is re-

placed with the following: “[A] distributor may perform functions, use as-
sets or assume risks that exceed those an independent distributor with simi-
lar rights might incur or perform.”2234 Thus, while the 1995 text focused
solely on the level of marketing expenditures incurred by the subsidiary,
the new wording seems to demand a broader perspective, encompassing
the entire spectrum of the local subsidiary’s IP development contribu-
tions. This does not, however, alter the fact that the value of marketing
IP generally is correlated with the level of marketing expenditures. Also,
the more marketing costs the subsidiary incurs, the more likely it will be
2228. 1995 and 2010 OECD TPG, para. 6.37.

2229. In contrast, the wording of the 1994 cheese examples was more reserved than
that of its 1995 OECD TPG counterparty. The second cheese example referred to “sig-
nificantly larger [expenses] than the expenses incurred by independent distributors un-
der similar circumstances”; see 1994 Treas. Regs. (59 FR 34971-01) § 1.482-4(f)(3)(iv),
Example 3; and the discussion in sec. 19.4.6.
2230. 1995 and 2010 OECD TPG, para. 6.37.
2231. See the discussion of the parallel criterion for “incremental” marketing expendi-
tures under the US regulations in sec. 23.3.6.
2232. The author takes it that the 1995 OECD TPG text was inspired by the 1994 US
cheese examples, which distinguished between cases in which the local distribution
subsidiary incurred an arm’s length level of marketing expenditures, and the expendi-
tures were incremental. See the discussion of the 1994 cheese examples in sec. 19.4.6.
2234. Id.
731
that additional functions, assets and risks are utilized in the process. The
new 2017 OECD guidance could therefore be seen as clarifying the 1995
predecessor text (with respect to the required marketing expenses) rather
than conveying entirely new concepts. The level of marketing expenditures
will, in the author’s view, remain the key factor for determining whether
a distribution subsidiary’s marketing IP development contributions exceed
those of comparable third-party distributors. Nevertheless, the new word-
ing emphasizes, to a stronger degree than the 1995 OECD TPG, that the
risks incurred by the subsidiary should factor into the assessment of the
amount of operating profits allocable to the source state.2235
It will likely not be straightforward to establish whether the marketing ex-

penditures incurred by the subsidiary “exceed those an independent dis-
tributor” would incur.2236 The author refers to the discussion on “incremen-
tal marketing expenditures” with respect to the US regulations in section
23.3.6.
Second, the new 2017 OECD guidance adds that the subsidiary’s incre-
mental development inputs must create “value beyond that created by other
similarly situated marketers/distributors”.2237 The interpretation of this
new wording is not straightforward. It seems to require a determination of
whether the value yielded by the subsidiary’s incremental development in-
puts goes beyond the (likely theoretical and supposedly comparable) value
created by similarly situated third-party distributors. The author suspects
that this comparison may be challenging to carry out in practice. Given
that the subsidiary has incurred a level of marketing costs that goes beyond
what unrelated distributors would be willing to bear, it may be questioned
as to what the resulting value should be compared with. Should it be bench-
marked against the intangible value created by an unrelated distributor that
provides only a “normal level” of development contributions? That would
be tantamount to comparing apples and oranges, as the value created by
the larger cost base of the controlled subsidiary then would be compared
to the value created by the smaller cost base of the otherwise comparable
third-party distributor. Indeed, the author would be surprised if the former
2235. The author reverts to the significance of risk below in this section.
2236. See also Vincent (2006), at 28:10; and, in particular, Levey et al. (2006), at p. 7
(see supra n. 2126). See, however, Roberge (2013), at p. 230, where it is suggested (with
respect to the pharmaceutical industry) that the level of marketing assistance provided
to the distribution subsidiary by the owner of the marketing IP (e.g. in the form of free
goods or marketing samples) could be of aid in determining whether the subsidiary
carried out incremental marketing efforts.
732
value did not go “beyond” the latter. Of course, even if there should be
differences in terms of absolute value, the relative return might still be the
same. In summary, such a comparison seems comprehensively inappropri-
ate. Alternatively, should it be compared to the value created by unrelated
distributors that also provide “incremental” development contributions, in
particular, marketing costs? That would be problematic, as third parties, by
definition, will not incur such a level of costs.2238
The author finds neither of the two interpretation alternatives satisfying.

The solution to this conundrum is, in the author’s view, to see the wording
of “functions … assets [and] risks that exceed” and “create value beyond”
as one collective criterion rather than two separate criteria. The point of the
2017 OECD guidance seems to be that the local subsidiary here contrib-
utes intangible development inputs to an extent that is incompatible with a
normal market return remuneration for low-risk routine value chain con-
tributions. The basic justification for a low-risk distribution classification is
ignored where the local subsidiary bears excessive marketing costs during
the marketing IP development phase, combined with the allocation of the
entire residual profits to the foreign group entity that holds legal title to the
marketing IP when the marketing efforts start to pay off in later income
periods.
The author finds it unclear whether the “create value beyond” wording
should be deemed to have any independent significance. Let us, for the
sake of argument, suppose that the subsidiary incurs incremental market-
ing costs that result in an unsuccessful marketing campaign. Should this
failure to “create value beyond that created by other similarly situated mar-
keters/distributors” entail that the subsidiary, in spite of its above-arm’s
length level of marketing costs during the IP development phase, is not
entitled to any additional income on top of the profits it earns as a low-risk
distributor? The answer to this must clearly be negative. The failure to cre-
ate value is due to the additional risks associated with the incremental mar-
2238. Let us, for the sake of argument, say that, in theory, it would be possible to iden-
tify unrelated comparable distributors that provide “incremental” development contri-
butions. The question is then whether the value created by the “incremental” contribu-
tions of the local distribution subsidiary goes beyond that created by the third-party
distributors. This question does not, in the author’s view, adhere to logic. One would
assume that related and unrelated parties derive the same amount of utility from each
additional unit of marketing expenditures. Also, it does not seem reasonable to demand
that the value created by a related distributor should go “beyond” the value created by
a comparable unrelated distributor when these entities incur the same amount of mar-
keting expenditures. The subsidiary should be remunerated in the same way as these
“incremental” third-party risk-takers.
733
keting expenditures materializing. At arm’s length, it cannot be expected

that third parties would be willing to incur such incremental risks without
receiving a commensurate reward.
The author therefore does not see the “create value beyond” wording as an
additional threshold. The decisive criterion under the current OECD TPG
must – as was the case under the 1995 text – be that if the local subsidiary
has incurred incremental marketing expenditures, it should not be compen-
sated as an LRD.
On this basis, the author concludes that the 2017 OECD TPG have not,
in substance, altered the 1995/2010 OECD TPG threshold for triggering
additional profit allocation to a distribution subsidiary in cases in which
it incurs incremental marketing expenditures. The essence still remains
whether the subsidiary has incurred extraordinary marketing expenditures
(absent of concurrent normal return compensation throughout the market-
ing IP development phase) without being entitled to any of the potential
future residual profits. If so, there is a legal basis in the OECD TPG to
set aside the controlled profit allocation (which assigns all of the residual
profits to the foreign group entity that holds legal title to the marketing IP)
and allocate more profits to the local distribution subsidiary. The extent to
which the subsidiary in this scenario should receive additional profits will
be discussed in section 24.5.3.
24.5.3. The material content of the profit allocation
The premise for the following discussion is that the local subsidiary has in-
curred above-arm’s length marketing expenditures to build the local value
of marketing IP that it licenses from a foreign group entity that holds legal
title to the IP. In such cases, the OECD TPG allow additional profits to
be allocated to the subsidiary (see the discussion in section 24.5.2.). The
question in the following is of how much more profit can be allocated to it.
The 1995 and 2010 OECD texts stated that the subsidiary, in cases in which
it incurred incremental marketing expenditures, “might obtain an addition-
al return from the owner of the trademark, perhaps through a decrease in
the purchase price of the product or a reduction in royalty rate”.2239 The
author finds it unclear whether this wording entitled the subsidiary to a
normal market return remuneration on a separate basis (e.g. cost-plus or
2239. 1995 and 2010 OECD TPG, para. 6.38.
734
TNMM) or to a portion of the residual profits (which would otherwise be

allocable to the foreign IP owner).
The guidance in the current OECD TPG entails a similar degree of

ambiguity,2240 but presents several suggestions as to the manner in which
additional compensation may be allocated to the subsidiary. These are (i) a
“decrease in the purchase price of the product”; (ii) a “reduction in royalty
rate”; or (iii) a “share of the profits associated with the enhanced value of
the trademark or other marketing intangibles”.
The first two alternatives merely pertain to the form of the additional re-
muneration. They do not convey clear positions on the material content of
the profit allocation. The third alternative, however, does provide material
direction. It indicates that the subsidiary could be allocated a portion of the
residual profits. This lends support to the interpretation that the preceding
two alternatives require only a normal market return, as it would be un-
necessary to include them in the guidance if their meaning were simply
to convey that the subsidiary could be entitled to a portion of the residual
profits (which is explicitly indicated by the third alternative).
Thus, the new guidance basically says that the subsidiary can be compen-
sated for its incremental IP development contributions either through a
separate normal market return (typically cost-plus or TNMM) or through
a portion of the local residual profits. As the wording reads, the three profit
allocation alternatives are equal in priority.
A rule that encompasses such a wide selection of profit allocation patterns

of equal standing would be disconcerting. The patterns offer the subsidiary
significantly varying degrees of operating profits, thus creating the poten-
tial for irreconcilable profit allocations. A multinational (and the residence
jurisdiction of the foreign group entity that holds legal title to the market-
ing IP) would likely always opt for the first or second alternative so that a
minimum amount of profits could be allocated to the source state, while
the source-state tax authorities would likely, in practice, always opt for the
residual profits alternative to allocate a maximum amount of profit there.
2240. The 2017 guidance states that “an independent distributor … would typically
require additional remuneration from the owner of the trademark … in order to com-
pensate … for its functions, assets, risks, and anticipated value creation” before refer-
ring to the new examples for further guidance; see OECD TPG, para. 6.78. Again, it is
unclear as to whether this additional remuneration should be a normal market return for
the incremental development contributions or a portion of the residual profits.
735
The conflict would be difficult to resolve, as the rule accommodates both

positions.
Thus, the general guidance contained in the 2017 OECD TPG does not
determine a specific profit allocation method that should be applied to re-
munerate a local distribution subsidiary that incurs an above-arm’s length
level of marketing expenditures, but leaves the door open for both a normal
market return and a portion of the residual profits from the local exploita-
tion of the marketing IP. The question in the following is whether an exten-
sion of the Primair example included in the 2017 OECD guidance, which
pertains to this scenario (above-arm’s length marketing expenditures), pro-
vides further direction.
The example pertains to a distribution subsidiary that, under a long-term,

exclusive distribution agreement, carries out marketing to build the local
value of foreign-owned marketing IP.2241 In doing so, it incurs marketing
expenses in year 1-5 that “far exceed” those of comparable third-party dis-
tributors.2242 It is expected that the marketing will result in higher local
prices and margins for the R watches. The subsidiary expects to benefit
from its promotional activities in the form of increased profits from sales.
The marketing expenses in years 1-5 (the IP development phase), however,
drive the subsidiary’s profits in this period significantly below those real-
ized by comparable third-party distributors, leading to the observation that
the subsidiary has not been “adequately compensated by the margins it
earns on the resale of R watches”.2243
The example concludes that the subsidiary should be remunerated consist-

ently with what independent enterprises would have earned in comparable
transactions. Aligned with the general guidance discussed above, the ex-
ample lists three main profit allocation options (seemingly of equal legal
rank): (i) adjustment of the purchase price for R watches through the re-
sale price method or the TNMM; (ii) application of the residual profit split
method; 2244 or (iii) direct compensation for the subsidiary’s incremental
marketing expenditures on a cost-plus basis.

2242. Id., at para. 31.
2243. Id., at para. 33.
2244. Its application will consist of two steps (see the analysis of the OECD profit split
method in ch. 9): (i) a normal market return is allocated to the parties for their routine
contributions to the development; and (ii) the residual profits are split pursuant to the
relative values of the parties’ intangible development contributions.
736
The first alternative is to adjust (downwards) the subsidiary’s purchase price

for the R watches by applying the resale price method or the TNMM.2245
The methods, consistent as they are, will extract the entire residual prof-
its from the market jurisdiction to the foreign group entity that owns the
marketing IP. In other words, they will “stop” the lowering of the purchase
price at a level that affords the subsidiary only a normal market return on
its marketing and distribution activities (corresponding to the profit level
earned by comparable third-party distributors).2246 It will not lower the
price beyond this point. An illustrative example of the use of the TNMM
to determine a purchase price is the taxpayer’s transfer pricing in the Nor-
wegian case Wingcard (the author refers to the analysis of this ruling in
section 15.6.).
Further, the resale price method and TNMM will compensate the sub-
sidiary as if it only provided low-risk distribution activities.2247 That prem-
ise will not be fitting when it incurs incremental marketing expenses and
thereby, in reality, has an entrepreneurial role in the development of the lo-
cal value of the foreign-owned marketing IP. The subsidiary will also have
financed the incremental marketing efforts on a concurrent basis without
reimbursement or a promise of a part of the potential future residual prof-
its. The author therefore cannot fathom why a low-risk characterization
2245. OECD TPG, annex to ch. VI, Example 10, para. 33, first bullet point.
2246. Also, an application of the resale price method and transactional net margin
method (TNMM) may raise significant comparability concerns here. The third-party
comparables must be assumed to have incurred only a “normal” level of marketing
expenses, while those of the tested party “far exceed” this level. The profit data will
therefore not be immediately comparable. It may be difficult to carry out proper com-
parability adjustments if aggregated financial statements are the only accounting data
available on the third-party comparables, which almost always will be the case. See
the discussions on aggregated financial accounting data in secs. 6.2.4. and 6.2.5., with
further references. See also the discussion on comparability in sec. 6.6.
2247. This result (if accepted) is confounding. The resale price method and the TNMM
provide the subsidiary with an arm’s length return only if it incurred an arm’s length
level of marketing expenditures and risks, which precisely is not the case here. Perhaps
it could be argued that the application of some “custom-made” version of the resale
price method or the TNMM, pursuant to which a normal return was calculated on the
incremental marketing cost base of the subsidiary, could remunerate it sufficiently. In
effect, this would be equal to applying the cost-plus method. For the reasons stated
below in this section (comments on the third profit allocation alternative described by
the 2017 OECD example), however, the author does not see this as providing an arm’s
length result. Also, such “custom-made” versions must be regarded as unspecified
transfer pricing methods; see the discussion of unspecified transfer pricing methods
under the OECD TPG in sec. 12.3. The author finds it unlikely that the 2015 guidance
envisioned the application of such methods. Had it done so, it likely would not have
specifically indicated that the resale price method and the TNMM could be applied.
737
of the subsidiary should be acceptable, as this would be contrary to the

economic substance of the controlled transaction.2248
The author would also argue that an application of the resale price method
or the TNMM to adjust the purchase price would be contrary to the options
realistically available to the subsidiary.2249 In the author’s view, it cannot
be presumed that an unrelated distributor would be willing to enter into an
arrangement pursuant to which it was obliged to incur marketing expen-
ditures significantly above the market level (with the additional risks that
such costs entail), with the best-case scenario being that it could possibly
earn a profit margin equal to that which it could have earned had it not
incurred such incremental marketing costs. Such an investment will likely
have a negative net present value. The subsidiary would be better off by not
entering into it.
In conclusion, with respect to this profit allocation alternative, given the

extent of the subsidiary’s IP development contribution, the extraction of the
residual profits from the source by the foreign owner (which is the impli-
cation of the example’s reference to the resale price method and TNMM)
seems to violate the stated purpose of the new 2017 OECD IP guidance
stating that intangible profits should be allocated to the jurisdiction in
which the intangible value was created.2250 This profit allocation alterna-
tive should therefore not be used.
The second profit allocation alternative suggested by the 2017 OECD ex-
ample is the PSM.2251 While it may be relatively straightforward to iden-
tify which controlled parties contributed the most valuable development
inputs in the context of internally developed manufacturing IP (typically
R&D and unique pre-existing intangibles), it may prove more problematic
to identify similarly unique contributions in the context marketing intan-
gibles, as the values of the latter IP tend to be mainly cost-driven. The
relative significance of functions is therefore diminished.2252 A distribution
2248. See the discussion of the 2017 OECD TPG on risk in sec. 6.6.5.5.
2249. OECD TPG, para. 6.139. See the discussions of the realistic options available in
the context of unspecified methods in sec. 12.3., and in the context of valuation in sec.
13.5.
2251. OECD TPG, annex to chapter VI, Example 10, para. 33, second bullet point.
2252. Multinationals will likely argue otherwise. The taxpayer in GSK argued that
more income should be allocated to the foreign legal owner of the trademark even
though the US distribution subsidiary performed local marketing. See the discussion
of GlaxoSmithKline Holdings (Americas), Inc. v. CIR, 117 TC 1 (Tax Ct. Docket No.
5750-04, 2004) in sec. 20.2.4.2. The reason for this is that purportedly key marketing
738
subsidiary that incurs incremental marketing expenditures to build the lo-

cal value of the IP should be allocated a significant portion of the residual
profits under the PSM, aligned with the stated purpose of the 2017 OECD
guidance on intangibles to allocate profits in accordance with where the
intangible value was created.
This PSM approach should, in the author’s view, be the preferred allocation
method, as it will take into account both parties’ contributions of functions
to (and, in particular, risks borne in respect of) the intangible’s develop-
ment.2253
The third profit allocation alternative described by the 2017 OECD exam-
ple is to reimburse the subsidiary’s incremental marketing costs with the
addition of a mark-up (in other words, to apply the cost-plus method).2254
The guidance pertaining (also) to this alternative is somewhat ambiguous.
One possible interpretation is that the foreign group entity that holds legal
title to the IP – at the end of year 5 – is entitled to extract all future residual
profits from the exploitation of the locally developed marketing intangible
simply by, at this point in time, reimbursing the subsidiary on a cost-plus
basis. The subsidiary has, in this case, assumed all risks associated with
the incremental costs, while the foreign legal owner of the IP is allocated
the fruits thereof. The latter group entity will have the benefit of know-
ing whether the marketing investments were successful before deciding
whether or not to compensate. If the investments were unsuccessful, there
will be no future residual profits, and the legal owner will not be interested
in reimbursing the local subsidiary. It is a win-win situation for the foreign
owner and a lose-lose situation for the subsidiary. During the first 5 years
of developing the local trademark, the subsidiary has no chance of receiv-
ing concurrent arm’s length remuneration, nor is it promised a portion of
the potential future residual profits. In fact, the best conceivable outcome
functions were performed centrally (e.g. the designing of a global marketing strategy
and preparation of budgets). That argument did not seem to have much impact on the
final settlement with the US Internal Revenue Service (IRS) and, in general, is not
very convincing. Marketing content will normally not be created by the employees of
the legal owner of the marketing intangible, but will rather be outsourced by the group
to specialized third-party advertising firms. Also, high-level marketing decisions will
often result from relatively generic upper management functions, as opposed to the
specialized R&D functions necessary for the creation of manufacturing intangibles.
Reliance on relatively generic functions for profit allocation purposes might entail that
the legal owner easily could extract profits from the source state.
2253. In this direction, see also Roberge (2013), at p. 236, with regard to the result
under the 2010 OECD TPG.
2254. OECD TPG, annex to ch. VI, Example 10, para. 33, third bullet point.
739
for the subsidiary is to be reimbursed for its incremental marketing expens-

es at the end of the development phase (of course, only if the promotion
efforts prove successful). Such an investment will likely have a negative net
present value. The subsidiary would be better off by not entering into it.
This cost-plus alternative should therefore not be used, as it will not adhere
to the pricing boundaries dictated by the best realistic alternatives of the
controlled parties. From a comparative viewpoint, it can be noted that the
economic substance provisions of the 2009 US IRC section 482 regulations
also reject such controlled pricing structures.2255
To summarize the above discussion in a question, which alternatives for

allocating additional profits to a distribution subsidiary that has incurred
above-arm’s length marketing expenses does the 2017 OECD guidance of-
fer? Apparently, all bets are off. Both the general guidance and the spe-
cific example indicate that the entire spectrum of specified transfer pricing
methods (apart from the CUT method) are acceptable alternatives for al-
locating profits to the source jurisdiction. The author attributes this un-
fortunate lack of direction to the likely inability of the OECD to reach
consensus on an appropriate allocation pattern.
As argued above, the wording of the 2017 example cannot be read in isola-
tion from the rest of the OECD TPG on intangibles. In particular, the allo-
cation patterns suggested by the 2017 example based on applying the resale
price method, cost-plus method and TNMM would, in the author’s view,
conflict not only with the overarching purpose of the new guidance on the
allocation of residual profits from unique and valuable IP that residual prof-
its should be allocated to the jurisdictions in which the intangible value was
created, but also with the economic substance of the controlled transaction,
which would result in pricing that is contrary to the realistic alternatives
of the subsidiary. These approaches should therefore be rejected. The new
guidance should be interpreted so as to allow only the profit split method to
remunerate a subsidiary that has incurred incremental marketing expenses.
This is consistent with the economic substance of the controlled transac-
tion, which is that the subsidiary alone incurs all risks connected to the IP’s
development (and performs all necessary functions).2256

2256. An extension is provided for the example in OECD TPG, annex to ch. VI, Exam-
ple 13, pursuant to which, at the beginning of year 4, (i) the foreign owner outsources
the manufacturing of the watches to a third party, from which the subsidiary purchases
them directly at arm’s length prices; and (ii) the distribution agreement is renegotiated
into a new, long-term agreement with terms similar to the initial agreement, apart from
obliging the subsidiary to pay a royalty for its right to process, market and distribute the
740
24.5.4. The profit allocation reservation
The example discussed in section 24.5.3. contains a potentially potent res-

ervation, which may incapacitate the authority to allocate additional in-
come to the source-state subsidiary all together. The example states that
if the subsidiary “could expect” to obtain an arm’s length return on the
incremental marketing costs during the remaining term of the distribution
agreement then “a different outcome could be appropriate”.2257 The author
interprets this to entail that if it can be established, at the end of year 5,
that the subsidiary has the potential to earn an arm’s length return on its
investment for the next 5 years (the distribution agreement had a duration
of 10 years), the source-state tax authorities will not be entitled to reassess.
No guidance is provided on how to determine whether the distribution sub-

sidiary could expect to receive an arm’s length return for its incremental
marketing efforts during the remaining term of the distribution agreement.
Nevertheless, as concluded in section 24.5.3., it would be contrary to the
economic substance of the controlled transaction and the realistic alterna-
tives available to the subsidiary to remunerate it on the basis of the one-
sided methods.
Thus, an arm’s length result will only be possible to attain in this situation
if the subsidiary, from and including year 5, is allocated a portion of the re-
sidual profits generated through the local exploitation of the foreign-owned
marketing intangible that it licenses. This portion should likely be the li-
on’s share of the profits. If the controlled agreement does not give such re-
muneration to the subsidiary, it should not be deemed at arm’s length. The
controlled allocation of profits to the subsidiary should then be reassessed.
Thus, in conclusion, it can be observed that the reservation will only bar
source-state tax authorities from reassessing the income of the distribution
subsidiary when the controlled pricing structure is based on a profit split.
watches. For years 1-3, the adjustment should be the same as in the main example. The
author found there that a profit split approach should be adopted. The guidance states
that the same allocation should be performed in the following years, but also allows
compensation of the subsidiary for the renegotiated agreement pursuant to the business
restructuring guidance in ch. IX of the OECD TPG. Such an adjustment cannot, how-
ever, possibly come in addition to the other adjustments suggested above in this section.
Otherwise, the subsidiary would, in effect, be compensated twice for the same income
item. See the discussion in sec. 22.5. of the somewhat parallel problem of allocating
residual profits from pre-existing intangibles by way of substance views and transfer
pricing charges.
2257. OECD TPG, annex to ch. VI, Example 10, para. 34.
741
When the pricing is based on a one-sided pricing method (e.g. cost-plus

or TNMM), the reservation will not be relevant, and the source-state tax
authorities may reassess.
24.6. Concluding comments
As shown in this chapter, the OECD has traditionally taken the position
that the group entity that holds legal title to internally developed marketing
IP shall be allocated the residual profits that it generates.2258 This approach
is highly susceptible to tax planning, as a multinational is generally free
to designate whichever group entity it may choose to hold legal title to IP
that is controlled by the group. This “restrictive” OECD profit allocation
approach has only afforded local distribution entities (and their residence
jurisdictions) a portion of the marketing-based super profits in very narrow
circumstances, if at all.2259
It was therefore interesting to see whether the 2017 BEPS revision of the
OECD TPG on intangible ownership induced the OECD member coun-
tries to agree on a clear and progressive stance on this issue. Based on the
analysis in this chapter, the author concludes that this did not turn out to be
the case. Precisely why is an open question. As the author sees it, the BEPS
revision targeted abusive profit-shifting structures pertaining to manufac-
turing intangibles. The more “ordinary” problem of allocating marketing-
based super profits took a backseat position in the OECD debates leading
up to the 2017 consensus text.
As the current OECD TPG now read, it will be easy for a multinational
to set up a structure with legal ownership of the marketing IP in a foreign
group entity (resident in a low-tax jurisdiction), license the IP to a local
distribution subsidiary (resident in a high-tax jurisdiction) and extract the
residual profits by way of royalty payments. As long as the multinational
ensures that the distribution subsidiary does not incur excessive market-
ing expenses (see the discussion in section 24.5.2.) during the marketing
IP development phase, the source-state tax authorities will have no legal
basis in the OECD TPG on which to downwardly adjust outbound royalty
payments that extract the entire marketing-based super profits from the
2258. See also the comments on the historical OECD position in sec. 19.5.
2259. In such circumstances, however, where the local group entity is allocated re-
sidual marketing profits, it will be deemed to own a local value of the marketing IP in
question. If the local entity is later stripped down, this may trigger exit compensation
for the transfer of the local marketing IP; see Musselli et al. (2008a), at p. 262.
742
Concluding comments
source jurisdiction. The “excessive marketing expenses exception” to profit

allocation to the (foreign) legal owner of the marketing IP is now an old
approach to transfer pricing that calls back to the 1979 OECD Report and
the first “cheese” example of the 1992 proposed US regulations that sought
to modify the 1968 developer-assister rule,2260 and should now be regarded
as highly overdue for rethinking and revision.
In respect of allowing profit allocation of marketing-based super profits, to

hinge on legal formalities in this manner is contrary to a clear aim behind
the OECD BEPS Project that intangible value should be taxed where it is
created. As the current OECD TPG now stand, the foreign group entity that
holds legal title to the marketing IP may extract the entire residual profits
from the marketing jurisdiction without providing much in the form of
functions, assets and risks, apart from legal ownership, to the value chain.
It may be questioned as to whether such profit allocation indeed should be
regarded as being at arm’s length. The OECD should, in the author’s view,
seek to revise the guidance on intangible ownership of marketing IP so that
it will ensure a more balanced profit allocation that takes into account the
functions, assets and risks provided by both the licenser and the licensee in
marketing IP structures.
2260. See the comments on the IP ownership approach of the 1979 OECD Report in
sec. 19.5., and on the proposed 1992 first “cheese” example in sec. 19.3.
743
Chapter 25
The Allocation of Residual Profits from Unique

and Valuable IP to Permanent Establishments
25.1. Introduction
This chapter will address two questions. The first is how intangible prop-
erty (IP) ownership shall be assigned under article 7 of the OECD Model
Tax Convention (OECD MTC) 2261 – more specifically, whether the head
office or the permanent establishment (PE) shall be assigned economic
ownership of an item of IP, and thereby also the residual profits generated
by it. As this issue is relevant both for internally developed and externally
acquired manufacturing and marketing IP, the author will analyse it sepa-
rately for each type of IP. The discussion of this problem should be read
in light of the introduction to the “significant people functions” concept in
chapter 17. Here, the author will focus on interpreting the concept for the
particular purpose of assigning ownership of IP among the head office and
the PE of an enterprise. The second question that will be discussed in this
chapter is of whether the PE threshold under article 7 of the OECD MTC
may have a distortive effect on the allocation of business profits compared
to allocations that are carried out under article 9.

developed manufacturing IP
The question is whether the head office or the PE shall be assigned eco-
nomic ownership of internally developed manufacturing IP, and thereby
also the residual profits generated by it.
The performance of research and development (R&D) functions and their

funding may be rendered by different parts of the enterprise, typically with
R&D being performed by a research centre PE that is funded by the head
office. The question then becomes whether the assignment of economic
IP ownership requires that the relevant part of the enterprise performs the
2261. The transfer pricing issues that arise when a permanent establishment (PE) con-
tributes to an intangible property (IP) value chain are analysed in ch. 17.
745
Chapter 25 - The Allocation of Residual Profits from Unique and Valuable IP
to Permanent Establishments
important R&D functions or whether it is enough that it funds the R&D

efforts.2262
The 2010 OECD PE Report (the 2010 Report) applies the significant-
people-functions concept to assign economic ownership of IP, pursuant to
which the key functions are “those which require active decision-making
with regard to the taking on and management of … risk … associated with
the development of the intangible”.2263
The Report indicates that such functions will include the active manage-
ment of individual R&D programmes, the testing of specifications, the
determination of the research framework, review and evaluation of data,
the setting of milestones for specific R&D projects and the determination
of whether a project should be continued or abandoned.2264 Such R&D
functions lie close to the important functions described in the 2017 OECD
Transfer Pricing Guidelines (OECD TPG) on IP ownership.2265
The 2010 Report, however, also makes clear that the performance by one
part of the enterprise of “most or all of the functions by which a trade intan-
gible, e.g. a complex software operation, has been created” does not neces-
sarily entail that it should be deemed economic owner of the developed
IP.2266 It is also stated in the 2010 Report that “under the authorised OECD
approach the ‘developer’ of the assets would have to bear such losses”.2267
The term “developer” in this context should be read as a reference to es-
tablished transfer pricing terminology based on the long-standing 1968 US
developer-assister (DA) rule, under which the entity that bears R&D costs
shall be entitled to the residual profits.2268 Thus, the 2010 Report seemingly
accepts dealings that, in substance, are similar to contract R&D arrange-
2262. It is, however, clear that it does not suffice for economic ownership that one par-
ticular part of an enterprise has the necessary capital available to assume the research
and development (R&D) risks (but has not actually assumed the costs); see the OECD
2010 Report, at para. 91. With respect to the cost side of the IP development, this could
be the case if the R&D project fails (and does not generate income), and “ownership”
of the project is assigned to the PE, resulting in a loss position (for the PE); see Schön
(2014), at p. 9; and Schön (2008), where it is argued that the source state will not be
obliged under international tax law to provide a deduction for the loss, and that this
OECD MTC art. 7 treatment differs from what would be the case under art. 9 (if the
R&D investment were carried out through a subsidiary in the source state instead).
2263. 2010 Report, para. 85.
2264. 2010 Report, paras. 87 and 88.
2266. 2010 Report, para. 83.
2267. 2010 Report, para. 89.
2268. See the analysis of the historical 1968 US developer-assister rule in sec. 19.2.
746
Assigning economic ownership to internally developed marketing IP
ments in which the R&D entity is remunerated on a cost-plus basis while

the residual profits are allocated to the funding entity.2269
The author finds that the 2010 Report must be interpreted as accepting the
assignment of economic ownership to the part of the enterprise that renders
R&D funding, and thus incurs the financial risks associated with the R&D
efforts, even if this part of the enterprise does not perform any of the R&D
functions connected to the developed IP.
This result is incompatible with the allocation pattern dictated by the 2017
OECD TPG on important functions, which requires that residual profits
are allocated to group entities that contribute important R&D functions.
Remuneration of R&D funding is restricted under the 2017 OECD TPG to
a separately determined risk-adjusted return.2270 As the 2010 Report must
be applied by taking into account the OECD TPG as modified from time
to time,2271 the important-functions doctrine must be applied analogically
to assign economic ownership.
The result will be that all residual profits from the developed IP must be allo-
cated exclusively to the part of the enterprise that performs the important R&D
functions.2272 The part of the enterprise that only renders a “funding contribu-
tion” will merely be entitled to a risk-adjusted return.2273 If important R&D
functions are performed by both the head office and the PE, the residual profits
from the developed IP must be split between the residence and source states.
The allocation of residual profits under article 7 of the OECD MTC will then
– as is the case under article 9 – reflect where IP value creation takes place.

developed marketing IP
nomic ownership of internally developed marketing IP, and thereby also
the residual profits generated by it.
2269. 2010 Report, paras. 84 and 85.

2270. See the analysis of the remuneration of IP development funding contributions in
sec. 22.4.
2271. 2010 Report, preface, at para. 10.
2272. This follows from an analogical application of the 2017 OECD TPG instruction
that a group entity that contributes important R&D functions cannot be remunerated as
the tested party under a one-sided method; see OECD TPG, para. 6.58.
2273. OECD TPG, para. 6.62; and the analysis in sec. 22.4. of this book.
747
The 2010 Report is ambiguous with respect to which specific significant

people functions shall be determinative for the IP ownership assessment.
The 2010 Report merely reiterates the point of departure that the relevant
functions are those that are “associated with the initial assumption and
subsequent management of risks of the marketing intangibles” and indi-
cates that this may include the creation of and control over branding strate-
gies, trademark and trade name protection and the maintenance of estab-
lished intangibles.2274
These are typical head office functions, as the marketing efforts of multi-
nationals tend to be coordinated through centralized, high-level manage-
ment. The specific application and localization of the marketing strategy
may, however, be determined in the source state where the marketing ef-
forts are actually carried out. For example, the decision to invest in a mar-
keting campaign may be taken at the head office, while the more specific
marketing content is determined in the source state.
Even though the 2010 Report, to some extent, recognizes the “creative”
day-to-day marketing functions performed at the PE level, the author finds
that the functions that directly affect the financial risk connected to the
marketing IP development (investment decisions, design of the marketing
strategy, IP protection, etc.) must prevail, as such functions lie at the core of
the significant-people-functions concept. If these functions are performed
at the head office, economic ownership of the local marketing IP should be
allocated to the residence state.2275
A transfer pricing problem closely related to the ownership issue arises

where the head office is assigned IP ownership, but the PE incurs incre-
mental marketing expenditures to develop the local value of the marketing
IP in the source state.2276 In this context, the PE is deemed to license ex-
ploitation rights to the marketing IP through a dealing with the head office.
The question is whether the PE in such cases should be allocated a portion
of the residual profits generated through its local exploitation of the IP. This
2274. 2010 Report, para. 97. See the comments in Levey et al. (2006), at p. 4, on the
key entrepreneurial risk-taking functions criterion for determining the ownership of
marketing IP.
2275. Even if the head office is assigned economic ownership of the marketing IP
developed in the source state, it cannot be taken for granted that all source-state mar-
keting profits are due to this IP. It may be that the PE, by itself, developed other local
marketing IP (e.g. goodwill or know-how) that is separate from the marketing IP owned
by the head office. The residual profits from such other IP may be allocable exclusively
to the PE.
2276. With respect to the cost side of the IP development, see supra n. 2262.
748
Assigning economic ownership of acquired manufacturing IP
is the article 7 parallel to the allocation of profits to a local distribution

subsidiary that incurs incremental marketing expenditures under article 9
(see the discussion in section 24.5.). There is irony to this scenario, as the
basis for allocating economic ownership to the residence jurisdiction is that
the “financial risk” functions are performed there, while the incremental
marketing expenses that cause the financial risk are deducted in the source
jurisdiction. This does not, however, entail that there is a lack of reality in
the head office’s economic ownership. The financial risk functions may
very well in substance be performed in the residence state, while the costs
themselves are deducted at source.
A PE will be wholly comparable to a distribution subsidiary in the paral-

lel situation under article 9 with respect to the functions performed, assets
used and risks incurred to develop the marketing IP. To ensure neutrality
in the international tax law treatment of business profits, a PE should be
entitled to a portion of the residual profits from local marketing IP to which
it has contributed incremental marketing expenditures to develop its value.
An unrelated distributor would simply not be willing to incur incremental
marketing expenditures without a stake in the potential super profit returns.
A different result would thus conflict with the arm’s length principle.

manufacturing IP
nomic ownership of acquired manufacturing IP, and thereby also the re-
sidual profits generated by it. This depends on which part of the enterprise
performs the significant people functions connected to the acquired intan-
gible, i.e. the active decision-making functions relating to the taking on and
management of financial risks.2277 The 2010 Report indicates that relevant
functions include the evaluation of whether the IP should be purchased,
the performance of follow-up R&D, as well as the management of risks
pertaining to deployment of the IP.2278
If this ambiguous guidance should be understood as assigning economic

ownership of the acquired manufacturing IP to the part of the enterprise
that “funds” the IP (through the investment decision) as opposed to the
part of the enterprise that carries out R&D functions relevant to the IP
2277. 2010 Report, at para. 92.

2278. 2010 Report, at para. 94.
749
(further development, maintenance R&D, etc.), the 2010 Report, on this

point, will arguably be incompatible with the important-functions doctrine
of the 2017 OECD TPG. As the 2010 Report requires that the OECD TPG
be applied by analogy to allocate profits among the head office and PE,
there is no question that ownership of the acquired IP must be assigned
to the part of the enterprise that performs the important R&D functions.
Only then will the allocation pattern dictated by the 2017 OECD TPG be
respected. If important R&D functions are performed by both the head
office and the PE, the residual profits must be split between the residence
and source states.
Application of the 2017 important-functions doctrine is unproblematic

when the enterprise actually carries out R&D. The people responsible for
ongoing R&D will then typically also be the ones that were qualified to as-
sess whether the acquired IP would be useful for the business and whether
it would be more rational to acquire it than to develop a similar intangible
internally.2279 It may, however, also be that the enterprise does not carry
out any R&D, or at least not any R&D relevant to the acquired intangible.
For example, this could be the case if competing software is acquired for
the purpose of letting it “wither on the vine” without further development.
Ownership of the acquired manufacturing IP should then be allocated to
the jurisdiction where the people performing the financial risk decisions
(decisions on whether to acquire the IP, to use equity or debt financing,
whether the investment should be hedged or insured, etc.) are located. This
will tie the profit allocation directly to the closest proxy that there is to
intangible development value creation in this scenario, namely the decision
to purchase the IP.

marketing IP
nomic ownership of acquired marketing IP, and thereby also the residual
profits generated by it. The financial risks will vary depending on the value
of the acquired marketing IP, which again will depend on whether the IP is
new and untested or established and well known. If it is a new trademark,
the enterprise will need to invest in marketing subsequent to the acquisi-
tion to develop the value of the IP. The profit allocation must then adhere
to the rules discussed in section 25.3. for internally developed marketing
2279. 2010 Report, at para. 93.
750
The cliff effect under article 7 of the OECD MTC and the important-
functions doctrine
IP. If the trademark is established, however, these rules will not be useful,
as the value of the acquired marketing IP already has been developed by
the seller. The following discussion pertains solely to this latter scenario.
The 2010 Report does not indicate which particular significant people
functions are relevant for the “taking on and management of risks” with
respect to acquired marketing IP.2280 The 2010 Report’s guidance on the
corresponding issue for acquired manufacturing IP (discussed in section
25.4.) may aid in the interpretation. The relevant risk-assuming function
will, in particular, be the decision to purchase the IP.
Thus, economic ownership of acquired established marketing IP should be

assigned to the part of the enterprise that performs the financial risk func-
tions related to the acquisition, typically the purchase decision.2281
25.6. The cliff effect under article 7 of the OECD MTC

and the important-functions doctrine
In this section, the author will discuss the potential impact of the PE
threshold of article 7 on the allocation of residual profits from unique and
valuable IP under the 2017 important-functions doctrine, using recent case
law as an illustration in the analysis.
In a range of recent cases on controlled distribution structures involving

products based on foreign-owned IP, source-state tax authorities have
2280. 2010 Report, at para. 94.

2281. The author finds it fair to ask whether it is useful to base the assignment of eco-
nomic ownership on the financial-risk-centred significant-people-functions concept. As
ownership here largely hinges on a one-time event, the multinational could arrange for
this single decision to be taken in the jurisdiction of its choice. The ownership assign-
ment has added importance in this context, as the “incremental marketing expenses
exception” will not be triggered, because the acquired marketing IP is established. The
PE will thus likely only need to incur a “normal” arm’s length level of marketing ex-
penses to maintain the already established value of the acquired marketing intangible,
in contrast to when the value of new and unknown marketing IP must be built from
scratch, which will likely require intensive marketing efforts in the initial years. Once
the ownership issue has been solved, it is unlikely that there will be any basis on which
to assign profits to the PE. Ultimately, however, there may be no better allocation pa-
rameter available than the financial risk functions relied upon by the significant-people-
functions doctrine.
751
claimed taxing jurisdiction over residual profits based on PE assertions.

The lack of a thorough profit allocation analysis is a common factor in
these rulings. Typically, the PE assertion was rejected by the courts, leav-
ing the allocation issue untried.2282 If it was found that there was a PE,
the source-state allocation assertion was accepted or adjusted without a
proper analysis. Some cases were settled before or during the proceed-
ings.2283 This case law therefore does not contain much useful guidance on
the profit allocation issue. As the factual patterns pertain to popular distri-
bution structures, however, it may be fruitful to gain insight into how the
2017 important-functions doctrine would allocate profits to the source state
(if there is a PE), and thereby how much profit is at stake in these common
scenarios. This is particularly relevant in the current post-BEPS climate, as
the OECD has now adjusted the PE threshold of article 7 so that there no
longer is a “commissionaire safe harbour” (as discussed in section 17.6.).
If there is no PE, the source state will be cut off from these profits (the so-
called “cliff effect” of article 7). Even though it will no longer be possible
to avoid PE status through commissionaire structures, the PE threshold
remains relevant for profit allocation cut-off purposes in a host of other
situations. The article 7 cliff effect is interesting because, as the author
argued in sections 17.1. and 17.3., the material content of the allocation
patterns under articles 7 and 9 is similar. Contrary to article 7, however,
article 9 does not contain a threshold that must be breached before profit
allocation kicks in. The analysis aims to provide an impression of the pos-
sible distortion effect on the international allocation of operating profits
caused by this threshold.2284
2282. The author’s impression is that the combination of a restrictive PE threshold in

OECD MTC, art. 5 with the ambiguous pre-2010 OECD MTC, art. 7 profit allocation
guidance may have enticed multinationals to favour a fight regarding the PE threshold
over the allocation issue. In this sense, art. 5 may have served as an “all or nothing”
allocation proposal. If the multinational succeeded in convincing source-state courts
that there was no PE, no additional profits would be allocable to the source state. Con-
versely, if a PE were found, it would likely prove challenging to convince the courts
to significantly alter the tax authority’s profit allocation assertion due to the ambiguity
of the pre-2010 allocation rules, typically resulting in an allocation of the reassessed
income to the source state, or at least substantial parts of it.
2283. E.g. the General Motors dispute between the United States and Canada, in which
the United States did not agree on the allocation of profits to a Canadian PE of General
Motors, and thus refused to provide credit relief under the 1980 Canada- United States
treaty. The case was settled. This dispute is apparently one reason as to why the United
States is unwilling to interpret art. 7 of its existing treaties on the basis of the 2010
authorized OECD approach.
2284. The analysis is limited to the 2010 version of the OECD MTC, art. 7. The com-
ments should not be perceived as positions on how the allocation issues should have
been resolved in these specific cases under the pre-2010 version of art. 7.
752
functions doctrine
25.6.2.
Philip Morris (Italy, 2001)
In the 2001 Phillip Morris case, the Italian Supreme Court reviewed a reas-
sessment based on the assertion that the Phillip Morris group had a PE in
Italy.2285 The profit allocation issue was not addressed.2286 The Italian tax
authorities allocated intangible business profits from local cigarette sales to
the PE. The profits were consideration for make-sell rights to manufactur-
ing and marketing IP associated with the cigarettes, in the form of royalty
payments from the Italian tobacco monopoly.2287
As the author understands the facts of the case, the intangibles were legally
owned by, and licensed from, a German group entity.2288 The group had a
subsidiary in Italy, which made and distributed cigarette filters and con-
trolled whether the monopoly’s cigarette distribution satisfied the quality
criteria of the licence agreement with the German entity. Further, its em-
ployees participated in licensing negotiations between the German entity
and the monopoly.
How should the operating profits be allocated, provided that there is a PE

and the issue is assessed under the important-functions doctrine of the 2017
OECD TPG? Only fragments of the relevant information are offered in the
ruling, but the author will provide some thoughts based on what is known.
The question was not whether the royalty payments were at arm’s length,
as the licence agreement was between the German entity and the unre-
lated Italian tobacco monopoly, but whether they were allocable to the PE
(in other words, whether the economic ownership of the Italian make-sell
rights to the IP should be assigned to Italy).
The author will assume that the important people functions necessary for
the creation of the manufacturing IP were performed by the German entity.
This entity should then be regarded as the economic owner under article 7
and should be allocated the profits from the exploitation of the manufactur-
2285. Ministry of Finance (Tax Office) v. Philip Morris (GmbH), Case No. 7682/05
(Supreme Court of Italy, 2002).
2286. The case was reversed and remitted to the Court of Appeal of Lombardy for
renewed assessment and was ultimately settled. The author suspects that the allocation
issue would have been the crux of the case, had it gone forward.
2287. Under Italian law, the monopoly was obliged to perform tasks such as transporta-
tion, distribution, conservation and diffusion of tobacco products.
2288. See the judgment in IT: Supreme Court, 2002, Case No. 7682/05, Ministry of
Finance (Tax Office) v. Philip Morris (GmbH), at para. 3.7.
753
ing IP in Italy. This should also be the point of departure for the marketing
IP. The only basis for allocating a portion of the marketing royalties to the
PE would be if it incurred incremental marketing costs (see the discussion
in section 25.3.). The facts referred to in the ruling provide no basis for
assuming so. Even if they did, it would likely not justify an allocation of
the entire residual profits.2289 The author’s impression is that none of the
intangible profits are allocable to Italy.
Nevertheless, it is clear that the PE did perform functions and incur risks
for which it should be remunerated. In particular, it performed manufac-
turing, distribution, control and negotiation functions. These are rather
generic services, the performance of which alternatively could have been
outsourced to third parties for a normal fee. Thus, the PE should be remu-
nerated as a routine service provider under a one-sided method.
A point here is that, while not clear from the ruling, the subsidiary was
likely remunerated for a range of routine functions performed in Italy, at
least for the manufacturing and distribution functions. Functions compen-
sated at the level of the subsidiary should not also be compensable for the
PE (see the analysis in section 17.6.4.). It may be that the only element
not compensated at the level of the subsidiary was the royalty payments.
Since these should not be allocated to the PE under the important-functions
doctrine, there will be no cliff effect here, as all routine functions are com-
pensated in the source state under the article 9 compensation of the local
subsidiary.2290
Thus, two observations can be made. First, there is a significant discrepan-

cy between the source state profit allocation assertion, which allocated the
entirety of the royalties to the PE, and the normal market return afforded to
the PE by the important-functions doctrine. Second, there is likely no cliff
effect in this case, as the source state should not lose out on any income
if there is no PE, since the local subsidiary is remunerated for the routine
services performed.
2289. See the discussion in sec. 25.3. on profit allocation for when the PE incurs incre-
mental marketing expenditures.
2290. This was not the case under the 2010 OECD TPG, which allowed separation of
inventory and credit risk from the local distribution subsidiary with effect for the art. 9
pricing, thus resulting in a “cliff effect” if there were no dependent-agent PE under
art. 7 to which the profits associated with these risks could be allocated. The author re-
fers to his analysis of these transfer pricing issues in sec. 17.6.4., and of the new OECD
guidance on risk in sec. 6.6.5.5.
754
functions doctrine
25.6.3.
Rolls Royce (India, 2007)
The 2007 ruling of the Indian Income Tax Appellate Tribunal (ITAT)
in the case of Rolls Royce PLC v. Dy. Director of Income Tax pertained
to whether the UK company Rolls Royce Private Ltd had a PE in India
through its local subsidiary, Rolls Royce India Ltd.2291 The Rolls Royce
group sold aero-engines and spare parts to the Indian Navy and Air Force.
The subsidiary performed local support functions, including business de-
velopment, public relations and technical support, and was remunerated on
a cost-plus basis. The UK parent reimbursed all marketing costs. R&D and
manufacturing were done in the United Kingdom. Employees of the UK
parent frequently visited the local subsidiary’s offices for marketing and
sales purposes.
Based on a PE assertion,2292 the Indian tax authorities allocated 100% of

the local business profits to the PE for 2 of the reassessed years, and 75%
for the remaining years. The ITAT affirmed the assessment on the PE is-
sue, but reduced the profit allocation to 35%.2293 The logic behind the 35%
allocation was that these profits were due to marketing efforts carried out
in India. The remaining 65% were allocated to the United Kingdom.2294

and the issue is assessed under the important-functions doctrine of the
2017 OECD TPG?
First, it is clear that economic ownership of the manufacturing IP should

be assigned to the UK parent, as R&D was carried out in the United King-
dom. The residual profits from sales of aero-engines and spare parts in
India are therefore allocable in their entirety to the United Kingdom.2295
Even though the 65/35 ITAT allocation may seem conservative and in fa-
vour of the United Kingdom, it cannot be ruled out that an even larger
portion of the total profits should be due to manufacturing IP connected
to specialized, high-tech products with narrow customer bases, such as the
aero-engines in this case.
2291. Rolls Royce Plc v. Dy. Director of Income Tax, 125 ITD 136, Delhi, 2007.
2292. See Kapoor (2008) for critical comments on the PE issue.
2293. The ruling does not devote much attention to the allocation issue.
2294. 50% to UK manufacturing and 15% to UK R&D.
2295. It may be questioned as to whether the technical support performed by the PE in
India could be seen as follow-up R&D for the manufacturing intangibles or as technical
know-how separately owned by the PE, which would attract a portion of the residual
profits. The limited information in the ruling provides no basis for assuming so.
755
Second, the ruling does not provide any information on the marketing IP
relevant to the aero-engines and spare parts sold in India or which group
entity owned them. There is, for instance, no mention of whether interna-
tional trademarks or trade names were involved. Due to the worldwide fa-
miliarity of the Rolls Royce name, however, the author would assume that
this was the case. He will further assume that the marketing IP was legally
owned by the UK company. As the marketing costs were reimbursed, the
UK head office should be assigned economic ownership of the marketing
IP. Thus, as the point of departure, there is no legal basis for allocating any
of the residual marketing profits to the Indian PE.2296
The ITAT conversely argued that marketing profits were allocable to India
because marketing was performed there. This is irrelevant under the 2010
Report, as the allocation of residual marketing profits hinges on economic
ownership, not the performance of marketing functions. The functions per-
formed in India seem rather generic, possibly apart from the technical sup-
port functions, and likely possible to outsource to third parties for a normal
market fee. Only a normal market return should therefore be allocable to
India, while the residual profits are allocable in their entirety to the United
Kingdom.
Thus, two observations can be made. First, there is a significant discrep-

ancy between the source-state profit allocation assertion, pursuant to which
the entire Indian operating profits were allocable to the PE, and the normal
market return indicated by the important-functions doctrine. Second, there
realistically does not seem to be a cliff effect in this case, as no additional
profits on top of the article 9 remuneration would be allocable to the source
state if there were a PE.
25.6.4. Zimmer (France, 2010)
In the 2010 Zimmer ruling, the French Supreme Administrative Court re-
jected a reassessment premised on the UK orthopaedic enterprise Zimmer
2296. The PE performed a range of local marketing functions (including business de-
velopment, public relations and conferences). It cannot be ruled out that this activ-
ity created local marketing intangibles (e.g. goodwill, marketing know-how, etc.) that
should be economically owned by the PE alone. If so, the residual marketing profits
should be split between the marketing intangibles owned by the UK head office and
the Indian PE. The author will, however, assume that there were no local marketing
intangibles, as there is no indication of this in the ruling and it was not asserted by the
Indian tax authorities.
756
functions doctrine
Limited, having a PE in France.2297 How should the operating profits be

allocated, provided that there is a PE and the issue is assessed under the
important-functions doctrine of the 2017 OECD TPG?
The profit allocation issue was addressed by neither the first-tier judges
nor the Paris Court of Appeals, which simply accepted the tax author-
ity’s assertion. Key information relevant to the allocation is absent. For
instance, it is not clear as to which group entity owned the manufacturing
and marketing IP used in the locally sold products. As far as the author
can gather, the point of the reassessment was to “reverse” the restructur-
ing so that the profit attributed to the PE plus the controlled allocation of
income to the local subsidiary in total equalled the pre-restructuring op-
erating margin of the local subsidiary.2298 Had there been a PE, it should
likely not have been allocated any residual profits, as the head office
economically owned the manufacturing and marketing intangibles. This
seems to be in line with the controlled pre and post-restructuring alloca-
tion of profits.
2297. Société Zimmer Limited v. Ministre de l’Économie, des Finances et de l’Industrie,

Nos. 304715 and 308525 (Supreme Administrative Court, 2010), reversing No. 05-2361
(Administrative Court of Appeals of Paris, 2007). The case was determined on the PE
issue, as Zimmer Limited was not found to have a dependent-agent PE in France under
art. 4 of the 1968 France-United Kingdom treaty. The Zimmer group restructured its
French orthopaedic distribution operation prior to the reassessment, going from a full-
fledged buy-sell distributor to a limited risk commissionaire structure. Similar restruc-
turings were common in France at the time; see Gelin et al. (2007); and Gelin et al.
(2010). The 2012 Spanish Supreme Court ruling in Roche Vitamins Europe Ltd. v. Ad-
ministración General del Estado, Case No. STS/202/2012, also pertained to a stripped-
down local distribution subsidiary. The Spanish tax authorities asserted that the Swiss
parent had a PE in Spain through the local subsidiary pursuant to the Spain-Switzerland
treaty (1966). The Supreme Court, as in the 2008 ruling of the National Court, ruled in
favour of the tax authorities. The case is somewhat peculiar, in the sense that it is really
an art. 9 case, and the terms of the local manufacturing and marketing agreements were
regarded as being at arm’s length. It is not clear whether the tax authorities asserted that
the entire business profits from the sales in Spain were allocable to the PE. For further
discussions, see Calderón (2012); and Wittendorff (2012e).
2298. Prior to the restructuring, the subsidiary purchased medical device products
from its UK parent for local resale, thus incurring both inventory and credit risk. Its
pre-restructuring profits reflected this, and were likely comparable to the level of profits
achieved by French third-party distributors of medical devices (typically benchmarked
through the transactional net margin method (TNMM)). By contractually stripping
these risks from the subsidiary, through a commissionaire structure, the Zimmer group
was able to allocate less profits to France; see the analysis of dependent-agent PE struc-
tures in sec. 17.6.4. The cost-plus method was likely applied to remunerate the subsidi-
ary.
757
Thus, two observations can be made. First, there is no real discrepancy

between the source-state profit allocation assertion and the allocation indi-
cated by the important-functions doctrine. Second, the question in this case
pertains only to the level of normal return profits allocable to the source
state, in particular, whether the PE should be allocated profits associated
with inventory and credit risk on top of the cost-plus remuneration earned
by the subsidiary. This is no longer relevant under the new OECD guidance
on risk, as the article 9 remuneration of the subsidiary will include com-
pensation for these risks.2299 Thus, there is no cliff effect.
25.6.5. Dell (Norway, 2011)
The 2011 Norwegian Supreme Court ruling in Dell Products v. the State
pertained to a reassessment in which 60% of the business profits from sales
of Dell computers in Norway were allocated to an asserted Norwegian PE
of an Irish entity in the Dell group.2300 The reassessment was upheld in the
first-tier Oslo City Court and the Norwegian Appellant Court.2301 The al-
location issue was argued before the Appellant Court, but the arguments
were not particularly developed, as the focal point was on the PE issue. It
played an even lesser role before the Supreme Court, which rejected the
reassessment on the basis that there was no dependent-agent PE.2302
The Dell group used commissionaire distribution structures in European

markets.2303 As opposed to in Zimmer, however, the Norwegian set-up was
not the result of a restructuring, but, as far as the author can tell, put in
place from the very start. From a total sales revenue in Norway in 2003 of
NOK 1.4 billion, the Norwegian subsidiary reported a taxable income of
approximately NOK 10 million (0.7% of its revenue) under the logic that it
acted as a commissionaire agent for its parent. The legal basis upon which
2299. See the analysis of these transfer pricing issues in sec. 17.6.4., and of the new
OECD guidance on risk in sec. 6.6.5.5.
2300. Ruling by the Norwegian Supreme Court, Utv. 2012, at p. 1, reversing ruling by
Borgarting Appellant Court, Utv. 2011, at p. 807.
2301. Ruling by Oslo City Court, Utv. 2010, at p. 107.
2302. See art. 5, no. 5 of the Ireland-Norway (2000) treaty.
2303. The Norwegian subsidiary, Dell AS, acted as a commission agent for its Irish-
resident parent company, Dell Products. The parent was incorporated in the Nether-
lands and had no employees. The computer products sold in Norway were purchased
by Dell Products from its parent, Dell Products (Europe) BV (also a company resident
in Ireland, but established in the Netherlands, with approximately 4,000 employees,
which manufactured computers in Ireland), which was controlled by the ultimate par-
ent, Dell Computer Corporation, in the United States.
758
functions doctrine
the tax authorities allocated 60% of the Norwegian profits to the asserted
PE is ambiguous.2304

and that the issue is assessed under the 2017 important-functions doctrine?
The international computer hardware industry is competitive, as illustrated

in Sundstrand.2305 Information technology companies must therefore dis-
tinguish themselves through proprietary features. The computers sold in
Norway were likely based on manufacturing and marketing IP owned by
the Dell group, but the ruling provides no details. The following comments
are thus based on factual assumptions and common knowledge.
First, any unique manufacturing IP used in the computers were likely

owned by foreign group entities.2306 Second, it is highly likely that eco-
nomic ownership of the relevant marketing IP, including the Dell trade-
mark, was attributable to a foreign group entity. Thus, the PE must be seen
as licensing the right to use the trademark and other relevant IP in Norway
through a dealing with the relevant entity, for which it must pay a deemed
royalty.
Dell is among the largest international PC vendors, and the Dell trademark
is established worldwide. Thus, the foreign group entity assigned owner-
ship of the marketing IP exploited in Norway would have a dominating
bargaining position vis-à-vis the Norwegian PE, which was a routine ser-
vice provider. This would likely make it feasible for the owner to impose
a royalty high enough to extract the residual profits generated in Norway.
2304. The allocation was based on the pre-2010 art. 7, no. 4 (apportionment method),
as the accounting records of Dell did not provide disaggregated transactional infor-
mation. The Appellant Court, without scrutinizing, agreed in full with the allocation.
Essential factual premises for the allocation issue, including which group entities devel-
oped and owned the manufacturing and marketing intangibles, the level of marketing
expenses incurred in Norway and the remuneration of the Norwegian subsidiary under
the commissionaire agreement are not clarified.
2305. For comments on Sundstrand Corp. and Subsidiaries v. CIR, 96 T.C. No. 12
(Tax Ct., 1991), see the discussion in sec. 5.2.5.3.
2306. The intangibles were most likely owned by Dell Products (Europe) BV, from
which Dell Products BV purchased the computers for resale. The latter company func-
tioned as a distribution hub, taking credit and inventory risk and selling the products in
local market jurisdictions through commissionaire entities. Most of the profits earned
by it were likely extracted through the prices it paid for the computers, leaving it with
a normal return for a distributor that takes on credit and inventory risk. It therefore
seems likely that neither the head office nor the PE owned the relevant manufacturing
intangibles.
759
There could be two possible grounds for allocating residual profits to the
Norwegian PE. The first would be if the PE incurred incremental market-
ing costs. There is no indication that this was the case. Also, the Dell trade-
mark was well known in Norway in the mid-2000s, so it is counterintuitive
that the Norwegian PE should have had to incur incremental marketing
expenditures. Further, the sales in Dell were to large Norwegian businesses
and the Norwegian public sector. These customers are likely not as suscep-
tible to marketing as consumer clients. The second possible ground would
be if the Norwegian PE had developed any unique local marketing IP dis-
tinguishable from the Dell trademark, such as goodwill or know-how.2307
The author does not see any basis in the ruling on which to assume that
there were local marketing intangibles. His impression is that the PE should
not be entitled to any of the residual profits from the marketing IP.
The Norwegian tax authorities argued that the combination of the PE’s ex-
ploitation of the Dell trademark with its performance of sales functions and
incurrence of related risks was the primary value driver, and that the lion’s
share of the profits therefore should be allocated to Norway.2308 This is not
convincing. The fact that the PE exploited the Dell marketing intangible
obviously does not entail that it should be able to reap the residual profits. It
would likely not be difficult for the Dell group to find an unrelated distribu-
tor in Norway that would be willing to distribute Dell computers to large
businesses and the public sector for a normal market fee. The same result
must apply when the distribution is carried out by a related party.
The author would assume that Dell based its remuneration of the Norwegian
subsidiary on the cost-plus method under the argument that it did not incur
inventory or credit risk,2309 resulting in a lesser amount of profits compared
to what would have been allocable to Norway had the remuneration been
based on the TNMM, under the premise that these risks should be included
(in order words, the same set-up as in Zimmer; see section 25.6.4.).
2307. If so, the profits from the marketing intangibles should be split between the for-
eign-owned Dell trademark and the locally developed intangibles economically owned
by the PE. The author does not find it doubtful that the relative significance of the Dell
trademark would be greater than any potential local marketing intangibles developed
by the PE. The lion’s share of the marketing profits should therefore not be allocable to
the PE in any case.
2308. This is similar to the argument of the Indian tax authorities in Rolls Royce; see
the discussion in sec. 25.6.3.
2309. Such pricing structure cannot be upheld under OECD MTC, art. 9, pursuant to
the 2017 OECD TPG on risk. See the analysis in secs. 17.6.4. and 6.6.5.5.
760
functions doctrine
Nevertheless, while the French tax authorities only increased the level of
the normal market return to reflect the inventory and credit risks, the Nor-
wegian reassessment went further by also claiming residual profits. The
correct allocation would, in the author’s view, be to allocate a normal mar-
ket return to Norway that is adjusted for inventory and credit risk, but no
residual profits (in other words, some more profits than allocated by Dell,
but far less than the Norwegian tax authorities claimed).2310
In conclusion, two observations can be made. First, there is significant

discrepancy between the source-state profit allocation assertion and the
allocation indicated by the important-functions doctrine. Only a normal
market return – and no residual profits – should be allocable to Norway
as compensation for the routine distribution functions carried out there.
Second, there should be no cliff effect, as the source state should not lose
out on the incremental normal return premium for inventory and credit
risk if there is no PE, since the local subsidiary should be remunerated for
these risks.2311
25.6.6.
Boston Scientific (Italy, 2012)
At issue in the 2012 Italian Supreme Court ruling in Boston Scientific was
also whether a local distribution subsidiary, acting as a commissionaire
for its Dutch parent for medical device products, could be regarded as a
PE in Italy.2312 The Italian subsidiary, similarly to the Norwegian subsidi-
ary in Dell (see section 25.6.5.), took the return position that the only in-
come allocable to it was the commission fee agreed upon in the distribution
2310. The pre-2010 art. 7, no. 4 apportionment method used as the basis for the 60% re-
assessment allocation is similar to the profit split method applied under art. 9 and under
the 2010 version of art. 7 by analogy; see the analysis of the OECD profit split method
in ch. 9. Based solely on the information available in the ruling, the author struggles to
see why such a significant portion of the profits should be attached to Norway.
2311. See the analysis of these transfer pricing issues in sec. 17.6.4., and of the 2017
OECD TPG on risk in sec. 6.6.5.5.
2312. Ministry of Finance (Tax Office) v. Boston Scientific, Case No. 3769 (Supreme
Court of Italy, 2012). The structure, akin to the ones used in Zimmer and Dell, was
set up as follows: the Dutch company, BSI BV, owned 99% of the Italian subsidiary,
Boston Scientific SpA, with the remaining 1% owned by the US Boston Scientific Corp.
The US entity sold medical device products to the Dutch entity, which had a commis-
sionaire distribution contract with the Italian subsidiary. The Dutch entity functioned
as a hub, having similar distribution contracts with local subsidiaries in several Euro-
pean market jurisdictions.
761
agreement. The PE assertion-based reassessment allocated an additional

EUR 50 million of income to Italy.2313

and the issue is assessed under the 2017 important-functions doctrine?
The ruling is devoid of allocation-relevant information. There is no way

of knowing whether the EUR 50 million reassessment included residual
profits. There is no indication that the PE should be deemed the economic
owner of any unique IP. Perhaps quite on the contrary, the functions per-
formed by it seem routine, including sales, leasing equipment to hospitals
and factoring receivables. For this, it should be allocated a normal return.
The author assumes that the subsidiary was stripped of inventory and cred-
it risks, and thus remunerated under the cost-plus method.2314
In conclusion, there is no basis on which to make an informed assumption

of whether there is a discrepancy between the source-state profit allocation
assertion and the allocation indicated by the important-functions doctrine.
Nevertheless, there should be no cliff effect, as the local subsidiary should
be remunerated for inventory and credit risks regardless of whether there
is a PE.
The cases encompassed by the cursory analysis in this chapter all pertain
to distribution structures for established value chains in which the unique
IP exploited in the source state was owned by foreign group entities. In
none of the cases were the intangibles created in the source state through
the performance of important R&D functions. The residual profits should
thus be extracted from source under the 2017 OECD TPG.
Another notable aspect is that source-state tax authorities have a tendency

to issue ambitious reassessments. In the cases discussed in this chapter,
in Morris, 100% of the royalties were allocated to the PE; in Rolls Royce,
2313. The taxpayer prevailed in the Regional Tax Court, which decided the case solely
on the PE issue. The Supreme Court, while leaning heavily on the Regional Court ar-
guments, rejected the appeal of the Italian tax authorities on procedural grounds. See
Sprague (2012) for further comments.
2314. It is unlikely that this could be upheld as the basis for transfer pricing under
art. 9, pursuant to the 2017 OECD TPG on risk; see the discussions in secs. 6.6.5.5. and
17.6.4.
762
functions doctrine
between 100% and 75% of the profits from the Indian sales were allocated
to the PE; and in Dell, 60% of the profits from sales of computers in Nor-
way were allocated to the PE.2315 While the author finds the source-state
profit allocation in Zimmer to be temperate, the allocations in Morris, Rolls
Royce and Dell seem poorly founded and aggressive (and would have no
legal basis in the 2017 OECD TPG). A contributing factor to this tendency
may have been that the tax authorities focused their efforts on establish-
ing that there was a PE under article 5, thereby accentuating the factual
elements relevant to that assessment, at the cost of a more thorough profit
allocation analysis.2316
If one disregards the unfounded source-state claims for residual profits,

one is left with the core issue in these cases, i.e. whether a profit allocation
based on a “low-risk distributor” characterization of the local subsidiary is
sufficient remuneration to the source state.
Under the 2010 OECD TPG and the 2010 Report, the incremental profits
due to the risks contractually stripped from the local subsidiary (normally
inventory and credit risk) could only be allocated to the source state if there
were a dependent-agent PE, as the stripping was accepted for the purposes
of the article 9 remuneration of the subsidiary. Thus, if there were no PE,
these incremental profits would be extracted from the source.2317 The cliff
effect of the PE threshold in the context of these distribution structures
2315. The actual results were also a far cry from the original reassessments. The re-
manding ruling on the PE issue in Morris likely provided the taxpayer with strong
arguments in the settlement that was reached. Even though the Indian Income Tax
Appellate Tribunal recognized the existence of a PE in Rolls Royce, it modified the al-
location to 35% of the Indian profits. The Norwegian tax authorities lost the case on the
PE issue in Dell, resulting in no allocation of residual profits to Norway.
2316. The PE issue was dominating in all cases. Profit allocation was not discussed
at all in Morris. Only 1.5 pages of a 29-page ruling in Rolls Royce were devoted to
the allocation issue. The Norwegian Appellant Court in Dell touched upon the alloca-
tion issue, but the Supreme Court ruled on the PE issue alone. The author’s point is
illustrated by the reasoning of the Appellant Court in Dell, which stated that “value
creation occurs primarily through sales, and the sale is first and foremost carried out
by [the Norwegian subsidiary]”. There is no legal basis in the OECD TPG to assert
this view on IP value creation. Nevertheless, it is striking how similar the quoted
wording is to the theme for assessment under OECD MTC, art. 5(5), with respect to
the question of whether the Norwegian agent entered into sales in the name of the head
office. If this observation is justified, the Court simply failed to properly recognize the
distinction between the PE threshold under art. 5 and the profit allocation issue under
art. 7.
2317. This problem was likely particularly irritating for source states when the local
subsidiary, prior to a business restructuring, were taxed on the “full profits”, which
included compensation for these risks, as the case was in Zimmer.
763
therefore, in essence, equalled these incremental profits. This cliff effect

should no longer be of much significance, however, as the article 9 remu-
neration of the local subsidiary under the 2017 OECD TPG on risk should
include compensation for inventory and credit risk.2318 Also, the “cliff” in
itself will not be as relevant as before, due to the 2015 OECD removal of
the commissionaire safe harbour.2319 Nevertheless, this lack of relevance
will typically only pertain to distribution structures in which foreign-
owned IP is exploited in the source state through the sale of products in
established value chains.
A more problematic side of article 7 becomes evident when the source

state was involved in intangible value creation. It is clear that the R&D
carried out by multinationals will not always be performed exclusively on
site at its R&D centres. It is practical, for instance, within the software
industry, that research employees travel abroad to work for certain periods.
During these travels, they may create significant intangible value. Assume,
for instance that the core code of a pioneering piece of software is devel-
oped by the employees of a foreign-resident enterprise during a stay in a
source state. The code is refined and further developed in the residence
state when the employees return. It goes on to become a success, generat-
ing residual profits. For the sake of argument, let us say that it is possible
to determine that 60% of the final intangible value of the software was due
to the core code created in the source state. What would the cliff effect of
article 7 be here?
First, if one starts by addressing the allocation issue, it is clear, under the
2017 important-functions doctrine, that significant R&D functions were
carried out in the source state. This intangible development contribution
should therefore, in principle, be assigned the 60% of the profits.
Second, the question is now whether there is a legal basis for the source
state to claim entitlement to the 60% of the profits. That will only be the
case if there is a PE there. If the software engineers were in the source state
only for a temporary visit, for instance, for a month or two, and they moved
around, visiting customers and attending conferences, there would be no
sufficient link between the activities performed by the foreign-resident en-
terprise in the source state and a specific geographical point to establish a
2318. See the analysis of these transfer pricing issues in sec. 17.6.4., and of the 2017
OECD TPG on risk in sec. 6.6.5.5.
764
functions doctrine
PE.2320 There will also be no degree of permanency to the activity,2321 as the

R&D employees are only in the source state for a temporary visit.2322 Thus,
the enterprise will have no taxable presence in the source state, excluding
the intangible profits (that otherwise would have been allocable) from local
taxation. This example is, of course, arranged, but the author doubts that
such scenarios are wholly impractical.
Let us look at the likely implications of this. The PE threshold prevents the
source state from claiming the 60% of the profits. The profits are therefore
allocated to the residence jurisdiction.2323 This entails that the purpose of
the 2017 OECD TPG on IP is not realized because the profits are not al-
located to the jurisdiction where they were created. More precisely, the
allocation rules under article 7 in and of themselves do allocate the income
to the source state, but article 5 does not.
This would not have been the result if the source state R&D alternatively
had been organized through a subsidiary. Article 9, which has no “trigger”
akin to the PE threshold, would then have governed the profit allocation,
and the 60% of the profits would have been taxable at source. There is no
sound justification for assigning intangible profits due to the same develop-
ment contributions so fundamentally differently depending on the legal
form in which the multinational organized its source-state activities.2324
The author is, for this reason, sceptical of the usefulness of the PE thresh-
old in the context of IP development contributions. The historical origin
of the threshold stems from the late 1920s, and its principal structure has
remained largely unaltered since then. Even though the threshold clearly
is susceptible to tax planning, as amply illustrated by the commissionaire
structures discussed in sections 17.6.4. and 25.6.1.-25.6.6. and Action 7 of
2320. This does not, however, necessarily entail that there is no link between the con-
tent of the IP value created and the source state. For instance, it may be that the local
visit enabled the software engineers to tailor the code to the specific needs of a local
customer.
2321. See the OECD Commentary on Article 5 of the OECD MTC, at para. 5.
2322. Id., at para. 6.
2323. The residence state will not be obliged to exclude any profits taxed there through
the provision of treaty relief.
2324. See also Schön (2010a), at p. 260, where an argument is made for the assertion
of taxing rights for any form of cross-border business activities, regardless of formal
thresholds (such as the PE threshold). One may also argue for parity between arts. 7
and 9 on the basis of non-tax-law regulations. See Schön (2007), where he discusses the
question of whether there must be equal treatment of a subsidiary and a PE under EU
law, and concludes that EU law does not affect the allocation of taxing rights between
EU Member States.
765
the BEPS Action Plan on the prevention of artificial avoidance of PE sta-

tus, it cannot entirely be ruled out that it may be useful in some contexts.
Otherwise, one could, in theory, risk source states claiming taxation rights
over profits made by a foreign enterprise that solely exports goods to the
source state. Nevertheless, in a global economy in which significant intan-
gible value (e.g. in the form of software, films, music and technical R&D)
may be created as “one-shot” occurrences over a short period and largely
independent of location, it may prove unwise to cling on to this “black-
and-white” borderline that obviously was not designed to deal with such
scenarios.
While the comments here touch upon the greater debate of whether the PE
threshold represents an equitable trigger for releasing source-state entitle-
ment to tax business profits earned there, it would fall outside the scope of
the book to engage in it. The author merely observes that the PE threshold,
in the context of IP development contributions, may have the potential to
create a cliff effect, one that is of an entirely different scale than the “nor-
mal market return cliffs” at stake in the distribution cases discussed in sec-
tions 25.6.1.-25.6.6. This may distort the international allocation of residual
profits under article 7, in the sense that they will not be taxed where they
were created, in contrast to the operating profits that are allocated under
article 9 according to the 2017 OECD TPG.
766
Chapter 26
Concluding Remarks
26.1. Introduction
Throughout this book, the author analysed the most important profit al-
location problems connected to intangible value chains, both with respect
to how profits shall be allocated among value chain inputs (see the transfer
pricing analysis in part 2 of the book) and among group entities that con-
tribute to the development of a unique intangible (see the intangible prop-
erty (IP) ownership analysis in part 3 of the book). In this final chapter,
the author will tie some closing comments to issues that he finds to be of
particular relevance going forward. These topics are also ripe for further
academic research.
26.2. The methodology for the transfer pricing of IP
The analysis of the historical and current US and OECD transfer pricing
rules reveals that more emphasis is consistently being placed on aligning
profit allocation results with economic substance. Perhaps the best example
of this is the inclination to move away from comparables-based IP pricing
(the comparable uncontrolled transaction method) towards pricing assess-
ments guided by the realistic options available to the controlled parties
and the application of (aggregate) valuation techniques. This is forcefully
illustrated by the 2017 OECD TPG on risk and valuation, the 2009 US
cost-sharing regulations and the 2017 codification in US Internal Revenue
Code (IRC) section 482 of the aggregated valuation principle, based on the
best realistic alternatives of the controlled parties.
The well-established residual profit split method, pursuant to which re-

sidual profits are allocated according to the relative values of the unique
value chain contributions, is also a prominent example of an inherently
substance-based pricing approach. Applications of the comparable prof-
its method and transactional net margin method (TNMM) may also be
in accordance with the realistic options available, provided that the tested
party is a genuine routine input provider (i.e. not in possession of unique
IP). Periodic adjustments and taxpayer-initiated compensating adjustments
ensure that transfer pricing outcomes in each income period are reflective
767
Chapter 26 - Concluding Remarks
of the relative values of the controlled value chain inputs. The emphasis on
economic substance in transfer pricing, in the author’s view, is not merely
appropriate, but it may also be entirely necessary if the existing paradigm
founded on the separate entity approach and arm’s length pricing is to pre-
vail.
Future revisions of the US and OECD transfer pricing rules should, in the
author’s view, take into consideration three aspects in particular.
First, in order to qualify as arm’s length, all profit allocations should, in the
author’s view, be aligned with the realistic alternatives of the controlled
parties. Thus, it should be considered whether it would be appropriate to
further emphasize this in the form of a formal requirement to always cross-
check the results of the transfer pricing methods against the allocation that
follows from the realistic alternatives of the controlled parties. The 2017
US codification in IRC section 482 of the realistic-alternatives pricing
principle will ensure that the United States takes this route.
Second, there should be stricter control with respect to whether the most
appropriate transfer pricing method has been chosen. In particular, it
should be a requirement that the taxpayer provides a convincing justifi-
cation for its application of one-sided methods to remunerate local (pur-
ported) routine input providers, typically contract manufacturers and dis-
tribution subsidiaries. It is crucial that the application of these methods
is founded on a thorough functional analysis that, beyond doubt, clarifies
that the tested party is not in possession of any unique value chain inputs
(goodwill, know-how, etc.).
Third, there should be increased focus on the delineation of the sources of

residual profits. The 2017 OECD TPG on local market characteristics is a
big step in the right direction, but there is, in the author’s opinion, consider-
able room for improvement with respect to the delineation of the sources of
marketing profits, as he will elaborate on in section 26.5.
26.3. Remuneration of IP development funding
Intra-group development of manufacturing IP has historically presented

multinationals with rich opportunities to shift profits out of high-tax re-
search and development (R&D) jurisdictions through cost-sharing arrange-
ments (CSAs) and contract R&D arrangements. Both the United States
and the OECD have recently expensed considerable resources to address
768
Remuneration of IP development funding
this problem. The US approach is geared towards CSAs and buy-in pric-
ing, while the OECD focuses on the performance of concurrent important
R&D functions. Both regimes allocate residual profits to the jurisdiction by
which the unique development inputs are contributed. In the author’s view,
however, there is good reason to question the role that IP development fi-
nancing plays in both regimes.
The remuneration of ongoing R&D contributions under the US regulations

is focused on the relationship between risk and reward, and the message is
that residual profits should follow financial risk.2325 The logic behind this
approach is alluring, as it reflects the economic axiom that higher returns
should follow higher risks. In the world of transfer pricing, however, it is
comprehensively ill-suited, as it will extract the residual profits from the
R&D jurisdiction and move them to the funding jurisdiction. For instance,
if a US R&D entity contributes ongoing research services to a foreign fund-
ing entity for a concurrent cost-plus fee, the subsequent residual profits may
be allocated to the foreign owner alone. This means that a multinational
can choose to which jurisdiction it wants to allocate the residual profits. If
it wants to allocate them to the United States, it will simply not reimburse
the concurrent development costs of the local entity.
This stands in contrast to the US approach for CSA buy-ins, which al-
locates the residual profits to the group entity that contributes the unique
pre-existing IP and other unique development contributions (R&D team
in place, know-how, etc.).2326 No profits are allocated to funding contribu-
tions. The result is an allocation aligned with value creation and the mate-
rial content of the transfer pricing rules. The lack of funding remuneration,
however, ignores the requirement of the arm’s length principle that all de-
velopment contributions should receive compensation.
The author cannot see a convincing rationale as to why there should be

such differing allocation principles depending on whether the contribution
is a stand-alone R&D service (typically rendered through a contract R&D
arrangement) or pre-existing IP contributed to a CSA in which it will serve
2325. See the discussion in sec. 21.3.6. The US position is, in reality, akin to that of
the 2009 OECD business restructuring guidance, pursuant to which the residual profits
could be extracted from the research and development (R&D) jurisdiction, as long as
the foreign funding entity assumed the financial risk, typically by remunerating the
R&D entity on a cost-plus basis; see the discussion in sec. 22.3.3.2.
2326. This result is ensured through the application of the US cost-sharing arrange-
ment (CSA) income or profit split methods; see the analysis in secs. 14.2.8.3. and
14.2.8.6., respectively.
769
as the basis for ongoing R&D. Both types of contributions are non-routine
and drive intangible value.2327
Comparatively, the important-functions doctrine of the OECD applies re-

gardless of how the IP development is organized and funded, and is there-
fore more coherent and neutral than the US approach.2328 While the doc-
trine strives to align profit allocation with value creation and the principles
of the transfer pricing methods, the treatment of IP development financing
provides significant room for profit shifting and dilutes the realization of
its purpose.2329
The allocation of profits to IP development funding under the US and

OECD rules can be summarized in table 26.1.2330, 2331
Table 26.1
United States OECD

“Plain” contract Risk/reward approach: Important-functions
R&D arrangements Funding entity is allocated doctrine: Funding entity is
the residual profits allocated a risk-adjusted
rate of return2330
CSAs CSA buy-in pricing Important-functions
methods: Funding is not doctrine: Funding entity is
allocated a return allocated a risk-adjusted
rate of return2331
2327. The new CSA regulations, and in particular, the income method, represent a sig-
nificant hurdle for multinationals that seek to shift intangible profits out of the United
States by way of CSAs. This may increase the attractiveness of contract R&D struc-
tures in cases in which CSA structures previously would have been adapted. The author
doubts that this strategy will be effective; not all contract R&D arrangements pertain to
“start-from-scratch”, blue-sky R&D efforts. There will likely often be pre-existing in-
tangible property (IP) (R&D results, know-how, etc.) that are contributed by the R&D
entity. If so, they must be priced at arm’s length by applying an aggregated valuation ap-
proach based on the best realistic alternatives of the controlled parties; see US Internal
Revenue Code (IRC) sec. 482, last sentence.
2328. Nevertheless, the OECD CSA rules lack the refinement of the US cost-sharing
regulations. No detailed guidance is provided on the buy-in of pre-existing intangibles,
making it difficult to price them in a uniform manner under the OECD regime, likely
resulting in diverging and equally justifiable valuations, and possibly double taxation.
2330. Provided that the funding entity is deemed to be in control of the financial risk,
otherwise it will only be allocated a risk-free return on its funding contribution.
2331. Id.
770
Allocation of profits for foreign-owned marketing IP
It can be observed that a considerable amount of profits are “leaked” to

funding contributions under both regimes. The transfer pricing reality is
that these profits represent extracted residual profits. It is necessary to re-
strict funding remuneration to a level that ensures that IP profits are allo-
cated pursuant to value creation.
Neither the US nor the OECD solutions are, in the author’s view, balanced.
The “all or nothing” approach of the United States, depending on whether
the development is carried out through a contract R&D arrangement or a
CSA, seems inappropriate. The author fails to see a sensible justification
for allowing the allocation of residual profits to funding contributions in
the contract R&D scenario and denying funding remuneration in the CSA
scenario. This is too much and too little, respectively.2332
A more principled approach would be to allow funding remuneration in

both scenarios, but to cap it at a pre-determined rate of return. This would
considerably decrease the profits allocable for funding in the contract R&D
scenario, and would increase them in the CSA scenario. While the OECD
approach is principled, the author’s view is that the prescribed risk-adjusted
return may easily become too large, and therefore represents a threat to the
ambitions of the important-functions doctrine and also should be capped at
some predetermined level, akin to a thin capitalization rule.2333
This would be a realistic and effective way of balancing the need to address
profit shifting with the need to respect the arm’s length principle by remu-
nerating all IP development contributions while aligning the allocation of
IP profits with value creation.
26.4. Allocation of profits for foreign-owned marketing IP
The point of departure under both the US and OECD regimes is that the
residual profits from the local exploitation of a foreign-owned marketing IP
can be extracted from the source and allocated to the foreign group entity
holding IP ownership (which is often located in a low-tax jurisdiction).
Further, both regimes are open for exceptions to this point of departure in
narrow circumstances, depending on the level of marketing expenses in-
2332. The residual profits should be allocated to the R&D jurisdiction where it was
created in the first scenario, and funding remuneration should not be completely cut off
in the second scenario, as funding is a genuine development contribution that must be
remunerated at arm’s length.
2333. See sec. 22.4.11.
771
curred by the local subsidiary. Essentially, if the expenses are high enough,
there will be a risk that the source state can claim that the local entity
should be deemed owner of the local value of the marketing intangible.
This represents a “safe bet” for multinationals.
As long as the local distribution subsidiary either incurs no (reimbursed)

or only an arm’s length level of marketing expenses, there will be no risk
for the multinational that the residual marketing profits cannot be extracted
from the source. These rules have been left relatively unchanged over the
last decades. It is ironic that the OECD, in its comprehensive 2015 BEPS
revision of the Transfer Pricing Guidelines (OECD TPG) on intangibles,
left the principles in the 1995 OECD TPG for the allocation of residual
profits from internally developed marketing IP largely untouched.
Both the US and OECD approaches are, in essence, risk/reward-based.

The logic is that an unrelated distribution entity would not have agreed to
incur incremental marketing costs without a promise of a stake in the po-
tential residual profits. Again, this economic axiom does not fit well in the
context of transfer pricing, simply because a multinational is free to place
its funding in any jurisdiction, and can therefore easily extract residual
profits if the allocation hinges on funding alone.
The most problematic aspect of the current rules is that the allocation of
marketing profits likely will not align with value creation. The author finds
it fairly obvious that the local value of a global trademark is developed in
the market jurisdiction. That is, in the author’s view, a forceful argument
for allocating at least some of the residual marketing profits there. The
author questions the usefulness of the current US and OECD approaches.
As they stand, it will be fully possible in the lion’s share of cases for a
multinational to develop the local value of its global trademark inside a
market jurisdiction and fully exclude the subsequent profits from the local
exploitation of the intangible from taxation at source.2334
A better solution would, in the author’s opinion, be to apply a profit split

between the ownership jurisdiction and the market jurisdiction. This would
cater to both the fact that marketing value is cost-driven and that IP value
creation takes place in the market jurisdiction.2335
2334. It is mainly just in the presumably narrow circumstances in which incremental

marketing expenditures have been incurred locally that the profits cannot be extracted
from the source state.
2335. To this, it could be objected that a profit split should not be applied if comparable
uncontrolled transactions (CUTs) can be identified in which third-party distributors
772
Is the OECD arm’s length standard heading towards formulary
apportionment?
26.5. The identification of local marketing IP
Residual marketing profits are, in practice, typically presumed to be gen-

erated solely by foreign-owned marketing IP. This assumption may often
be unreasonable. A local distribution subsidiary is present in the market
and develops relationships with its suppliers and customers and exper-
tise through its local sales. This may give rise to the development of local
unique marketing IP (e.g. goodwill and know-how) distinct from the local
value of a foreign-owned trademark. As they stand, the US and OECD
rules seem underdeveloped with respect to the allocation of profits from
marketing IP. There should be a much stronger focus on the identification
and delineation of local marketing IP. Both the US and OECD rules should
be revised to better reflect where intangible marketing value is created.2336
This issue is of great importance, as it has the potential to influence global
transfer pricing practices profoundly. The reason is that if local marketing
IP is identified, the widespread practice of remunerating local distribution
entities under the one-sided methods (in practice, the TNMM) with a nor-
mal market return for routine contributions should no longer be accept
able.2337
26.6. Is the OECD arm’s length standard heading

towards formulary apportionment?
As discussed in chapter 9, the profit split method has gained increased sta-
tus as a methodology for allocating operating profits among group entities,
in particular within the OECD regime, both through an expansion of its
incur a similar level of marketing expenses without being allocated any residual profits,
typically based on the transactional net margin method (TNMM). If such CUTs are
available, the author would agree that there is no justification for allocating a portion of
the residual profits to the market jurisdiction. He suspects, however, that it will be rare
to find genuine CUTs in these cases.
2336. While the identification of local unique marketing intangibles is an ownership
issue, the division of the total residual marketing profits among these and the foreign-
owned trademark is, in principle, a transfer pricing issue. The split should follow the
principles developed under the profit split method, which distributes the profits accord-
ing to the relative values of the unique contributions. The identification of local market-
ing IP must be kept separate from local market characteristics, as incremental profits
from the latter shall not be allocated fully to the market jurisdiction (see the analysis in
ch. 10), as opposed to residual profits due to the former.
2337. The method should not be used when the tested party is in possession of unique
value chain contributions; see sec. 8.3.
773
scope of application as well as by way of incorporation of the methodol-

ogy’s core material content into the OECD IP ownership provisions.
A relevant question is whether this can lead to more evenly distributed op-
erating profits among jurisdictions, avoiding the typical BEPS-driven “cen-
tralized principal model” structures that result in the extraction of residual
profits from source states (and the injection of those profits into low-tax
IP holding companies) and thereby enabling “smoother” profit allocations,
akin to those that would result from formulary apportionment.
If the expanded OECD profit split method is applied to remunerate a group

entity that renders relatively “routine” value chain contributions, the entity
could possibly receive more profits than what would have been the case had
it been remunerated under a one-sided method in the traditional way (e.g.
TNMM-based). More returns to local “routine” entities equals less extrac-
tion of residual profits from source states, and thus less residual profits for
centralized IP holding companies in low-tax jurisdictions. This, however,
will not entail that the arm’s length principle is moving towards formulary
apportionment, but only that the amount of residual profits allocable for
unique IP value chain contributions can possibly become reduced.
A core feature of the arm’s length principle, operationalized through the IP

ownership provisions, is that the residual profits from IP value chains shall
be allocated according to intangible value creation, i.e. to the group entity
that is deemed to own the unique and valuable IP employed in the value
chain. Thus, even if the newly expanded OECD profit split method may
result in a lesser amount of residual profits, the IP ownership provisions
of the 2017 OECD TPG will nevertheless ensure that these residual profits
are distributed only to jurisdictions in which intangible value creation took
place.2338 This allocation result is the opposite of “smooth” international
profit allocation.
Formulary apportionment would, in contrast, spread the amount of residu-

al profits across all jurisdictions in which the multinational does business,
as pre-determined allocation formulas would determine the split.2339 This
2338. The same result is generally attained under US law, with respect to outbound
transactions, through the income method for CSAs and the application of the aggre-
gated valuation principle based on the realistic alternatives of the controlled parties.
2339. A parallel can be drawn here to Schön’s observation on the treatment of individ-
ual group members’ risk profiles under formulary apportionment. See Schön (2014), at
p. 15, where it is stated that “any “formulary” approach to the taxation of multinational
enterprises would indeed disregard the different risk profiles of the involved members
774
Is the OECD arm’s length standard heading towards formulary
apportionment?
would, in an economic sense, overcompensate most jurisdictions, as they

would be allocated more operating profits than the routine value chain in-
puts provided by group entities resident in them were worth. The few juris-
dictions in which IP value was created would then be undercompensated
by not being allocated the entire residual profits.
Thus, the key difference between profit allocation based on the arm’s length
principle and formulary apportionment really lies in how residual profits
are allocated among jurisdictions under the two systems.2340 The arm’s
length standard, by way of the OECD IP ownership provisions, gives tax-
ing rights to residual profits only to the few jurisdictions in which IP value
creation took place, while a formulary apportionment system would give
taxing rights to all jurisdictions in which the multinational does business.
The OECD, in the latest BEPS revision of the OECD TPG, has further de-
veloped and strengthened its IP ownership provisions, precisely to ensure
that residual profits are allocated according to intangible value creation
(i.e. only to jurisdictions that had a role in IP development). Thus, there
are no signs indicating that the arm’s length principle is heading towards
formulary apportionment. Even if the material content of the 2017 revised
OECD IP ownership provisions now draw more on the profit split method-
ology, the residual profits are still just split between the few jurisdictions
that were involved in IP development efforts (R&D, funding, etc.).
The only way in which the OECD arm’s length principle truly could move
towards a formulary apportionment system would be if the IP ownership
provisions were removed, thereby eliminating the requirement that resid-
ual profits shall be allocated to the group entity that is deemed to own
the unique and valuable IP.2341 To do so, however, would remove the heart
of the arm’s length principle. The result would be a formulary apportion-
ment system that would contradict economic reasoning, because it would
of the group”, and further, that “formulary apportionment would basically level out
the individual risk profiles of the involved taxpayers and activities”. See also de lege
ferenda views in Hafkenscheid (2017), at p. 24, where it is argued that art. 9 of the
OECD MTC should be amended so that all group entities are seen as having the same
risk profiles (in order to ensure that profits cannot be shifted within the multinational
enterprise by way of contractual risk transfers).
2340. For de lege ferenda viewpoints on how taxing rights over residual profits can be
allocated in a new tax order, see Schön (2010a), at p. 260, where he discusses the possi-
ble allocation of such profits among four categories of jurisdictions: (i) where the R&D
is performed; (ii) where the (legal) owner of the IP is resident; (iii) where manufactur-
ing takes place; and (iv) where marketing and sales take place.
2341. See Brauner (2010), at p. 18, where it is recognized (in the context of US cost
sharing) that the allocation of residual profits based on the notion of IP ownership has
played a key role in tax-planning structures.
775
be based on the underlying assumption that all value chain inputs – both
routine and non-routine – are worth the same. The system would be unable
to emulate the profit allocation results of transactions among third-party
enterprises with genuinely conflicting interests, the hallmark of which is
that competition will drive the prices for routine value chain contributions
down to a level at which the providers only reap a normal market return.
Whether such an alternative system would provide more equitable inter-
national profit allocation results is, in the end, a political question. It does,
however, seem clear that such an alternative formulary apportionment sys-
tem would fail to realize a key purpose of the current, arm’s length-based
international standard for profit allocation: to provide parity in the tax
treatment of unrelated and related parties.
776
References
Bibliography
ABA (2004) American Bar Association, Section of Taxation, “Comments

on Proposed Regulations on Services, Intangibles, Economic
Substance, and Other Issues”, 23 Jan. 2004, available at Tax
Analysts (www.taxanalysts.com); Doc 2004-1438.
ACC (2003) American Chemistry Council, “Comments on Proposed Sec-
tion 482 Service and Related Regulations”, available at Tax
Ackerman et al. Ackerman, Robert; Chorvat, Elizabeth A., “Modern Finan-
(2002) cial Theory and Transfer Pricing”, 10 George Mason Law
Review (2002), p. 637.
Ahmadov (2011) Ahmadov, Jamil, “The Most Appropriate Method: As the
New OECD Transfer Pricing Standard: Has the Hierarchy
of Methods Been Completely Eliminated?”, 18 International
Transfer Pricing Journal 3 (2011), p. 184.
Ainsworth et al. Ainsworth, Richard T.; Shact, Andrew, “Transfer Pricing,
(2010) Business Restructurings, and Intangibles: Case Studies From
the U.K. and the U.S.”, 60 Tax Notes Int’l (2010), p. 767.
Ainsworth et al. Ainsworth, Richard T.; Shact, Andrew, “Transfer Pricing and
(2011) Business Restructurings: Intangibles, Synergies, and Shel-
ters”, 63 Tax Notes Int’l (2011), p. 341.
Ainsworth et al. Ainsworth, R.; Shact, A., “Transfer Pricing: Case Studies of
(2012) the CUP Method”, 66 Tax Notes Int’l (2012), p. 465.
Aksakal et al. Aksakal, Layla J.; V. Femia, Rocco, “Section 367(d), Intangi-
(2013) bles, and Base Erosion: A Reassessment”, Tax Management
Memorandum, 54 TMM 191, (2013).
Allen (2004) Allen, S.N., “Location savings – A US Perspective”, 11 Inter
national Transfer Pricing Journal 4 (2004).
Allen et al. (2006) Allen, S.; Tomar, R.; Wright, D.R., “Sec. 482 Services Regu-
lations: Implications for Multinationals”, 13 International
Almendral (2013) Almendral, Violeta Ruiz, “Tax Avoidance, the ‘Balanced Al-
location of Taxing Powers’ and the Arm’s Length Standard:
An odd Threesome in Need of Clarification”, Allocating Tax
ing Powers within the European Union: MPI Studies in Tax
Law and Public Finance, Vol. 2 (Richelle, Isabelle; Schön,
Wolfgang; Traversa, Edoardo eds. 2013), p. 131.
777
References
Altshuler et al. Altshuler, Rosanne; Grubert, Harry, “Formulary apportion-

(2010) ment: Is it better than the current system and are there better
alternatives?”, 63 National Tax Journal 4, (2010), p. 1145.
Andresen (2011) Andresen, Helga Marie; Pearson-Wood, Nick; Jørgensen,
Hans-Martin, “ConocoPhilips Case: Implications in Norway
and Beyond”, 17 International Transfer Pricing Journal 6
(2010), p. 461.
Andrew et al. Andrew, M.; Nixon, M., “Exit Charges in Asia Pacific: Cur-
(2012) rent Trends and Approaches in Asia Pacific Tax Authority
Enforcement”, 39 BNA Tax Planning International Review
5 (2012), p. 4.
Andrus (2007) International Fiscal Association, “Transfer pricing and intan-
gibles”, Vol. 92a, Cahiers de droit fiscal international, Chap-
ter on US (2007), p. 629.
Andrus et al. Andrus, Joseph; Feinschreiber, Robert; Kent, Margaret, “De
(2012) scription of the United States transfer pricing provisions”
(2012).
Andrus et al. Andrus, Joe; Oosterhuis, Paul, “Transfer Pricing After BEPS:
(2017) Where Are We and Where Should We Be Going”, Taxes the
Tax Magazine (2017), p. 89.
Angus (2012) Angus, Barbara, “Chipping Away at the Permanent Estab-
lishment Concept”, Tax Management International Journal
(2012), p. 259.
Angus et al. Angus, Barbara et al., “The U.S. International Tax System at
(2010) a Crossroads”, 58 Tax Notes Int’l (2010), p. 45.
Anson et al. Anson, Weston; Suchy, Donna P.; Ahya, Chaitali, “Intellec-
(2005) tual Property Valuation: A Primer for Identifying and Deter-
mining Value”, American Bar Association, Section of Intel-
lectual Property Law (2005).
API (2003) American Petroleum Institute, “Comments on Proposed Trans
fer Pricing Regs”, available at Tax Analysts (www.taxanalysts.
com); Doc 2003-24370.
Ault (2008) Ault, Hugh J., “Reflections on the Role of the OECD in De-
veloping International Tax Norms”, 34 Brooklyn Journal of
International Law 3 (2009), p. 757.
Avi-Yonah (1995) Avi-Yonah, Reuven S., “The rise and fall of arm’s length: A
study in the evolution of US international taxation”, 15 Va.
Tax Rev., (1995), p. 89.
Avi-Yonah Avi-Yonah, Reuven S., “Xilinx and the Arm’s-Length Stand-
(2009a) ard”, 54 Tax Notes Int’l (2009), p. 859.
778
References
Avi-Yonah Avi-Yonah, Reuven S., “The OECD Harmful Tax Competi-

(2009b) tion Report: A Tenth Anniversary Retrospective”, 34 Brook
lyn Journal of International Law 3 (2009), p. 783.
Avi-Yonah (2010) Avi-Yonah, Reuven S., “Between Formulary Apportionment
and the OECD Guidelines: A Proposal for Reconciliation”, 2
World Tax Journal 1 (2010), p. 3.
Avi-Yonah (2012) Avi-Yonah, Reuven S., “Transfer Pricing Disputes in the Unit-
ed States”, in “Resolving Transfer Pricing Disputes: A Global
Analysis” (Baistrocchi, Eduardo; Roxan, Ian eds. 2012).
Avi-Yonah (2015) Avi-Yonah, Reuven S., “Too Big to Tax? Vanguard and the
Arm’s Length Principle”, Law and Economic Research Pa
per Series, Paper No. 15-015 (2015), available at http://ssrn.
com/abstract=2666426.
Avi-Yonah (2017) Avi-Yonah, Reuven S., “Amazon vs. Commissioner: Has Cost
Sharing Outlived Its Usefulness?”, Public Law and Legal The
ory Research Paper Series, Paper No. 551, May 2017, avail-
able at https://papers.ssrn.com/sol3/papers.cfm?abstract_
id=2961235.
Avi-Yonah et al. Avi-Yonah, Reuven S.; Clausing, Kimberly A.; Durst, Mi-
(2009) chael C., “Allocating Business Profits for Tax Purposes: A
Proposal to Adopt a Formulary Profit Split”, 9 Florida Tax
Review (2009), p. 497.
Avi-Yonah et al. Avi-Yonah, Reuven S.; Haiyan, Xu, “Evaluating BEPS”, 10
(2017) Erasmus Law Review 1 (2017).
B&M (2003) Baker & McKenzie on behalf of Software Coalition, “Soft-
ware Coalition Comments on Proposed U.S. Transfer Pricing
Regs”, available at Tax Analysts (www.taxanalysts.com); Doc
2003-24362.
Baistrocchi Baistrocchi, Eduardo, “The Arm’s-Length Standard in the
(2004a) 21st Century: A Proposal for Both Developed and Develop-
ing Countries”, 36 Tax Notes Int’l (2004), p. 241.
Baistrocchi Baistrocchi, Eduardo, “The Transfer Pricing Problem: A
(2006) Global Proposal for Simplification”, 59 Tax Lawyer 4 (2006),
p. 941.
Baker et al. Baker, Philip; Collier, Richard S., “General Report”, in IFA,
(2006) “The attribution of profits to permanent establishments”, Ca
hiers de droit fiscal international, vol. 91b (2006).
Bakker (2009) Bakker, Anuschka (ed.), “Transfer Pricing and Business Re-
structurings: Streamlining all the way” (2009).
Ballentine (2016) Ballentine, J. Gregory, “Ownership, Control, and the Arm’s-
Length Standard”, 82 Tax Notes Int’l (2016), p. 1177.
779
References
Barbera (2003) Barbera, A.J., “Considering Risk in Trademark Royalty Ar-

rangements”, BNA Transfer Pricing Report (2003), p. 71.
Barreiros Barreiros Rosalem, Francine, “The Agent Permanent Estab-
Rosalem (2010) lishment Reconsidered: Application of Arts. 5, 7 and 9 of the
OECD Model Convention”, 17 International Transfer Pri
cing Journal 1 (2010), p. 12.
Bauer et al. (2013) Bauer, Christian J.; Langenmayr, Dominika, “Sorting into
outsourcing: Are profits taxed at a gorilla’s arm’s length?”,
Journal of International Economics 90 (2013), p. 326.
Beck (2008) Beck, Karin E.M., “Business Restructuring in Germany”, 51
Tax Notes Int’l (2008), p. 271.
Becker (2008) Becker, B.C., “Projected and Actual Profit’s Impact on Li-
censees”, 17 Tax Management Transfer Pricing Report 11
(2008), p. 461.
Bell (2007) Bell, Kevin A., “OECD Officials Discuss Tax Treaty Pro-
jects”, 45 Tax Notes Int’l (2007), p. 32.
Bell (2012) Bell, K.A., “China’s Chapter of U.N. Manual Advocates
Reimbursement for Location Savings, Added Profit for Lo-
cal Intangibles”, Tax Management Transfer Pricing Report
(2012), 21 TPR 683.
Bennett et al. Bennett, Mary C.; Dunahoo, Carol, “The Attribution of Prof-
(2005) its to a Permanent Establishment: Issues and Recommenda-
tions”, 33 Intertax 2 (2005), p. 51.
Benshalom Benshalom, Ilan, “Sourcing the Unsourceable: The Cost Shar-
(2007) ing Regulations and the Sourcing of Affiliated Intangible-Re-
lated Transactions”, 26 Virginia Tax Review (2007), p. 635.
Benshalom (2009) Benshalom, Ilan, “Taxing the Financial Income of Multina-
tional Enterprises by Employing a Hybrid Formulary and
Arm’s Length Allocation Method”, 28 Virginia Tax Review
(2008/2009), p. 619.
Benshalom (2013) Benshalom, Ilan, “Rethinking the Source of the Arm’s-
Length Transfer Pricing Problem”, 32 Virginia Tax Review
3 (2013), p. 425.
Bernt and Doublet Bernt, Jan Fridthjof; Doublet, David R., “Vitenskapsfilosofi
(1998) for jurister – en innføring” (1998).
Bhatnagar (2017) Bhatnagar, Vatika, “Appropriate Application of the Berry
Ratio as a Profit Level Indicator”, 24 International Transfer
Pricing Journal 1 (2017), at p. 25.
Biegalski (2010) Biegalski, Adam, “The Arm’s Length Principle: Fiscalism or
Economic Realism? A Few Reflections”, 38 Intertax 3 (2010),
p. 177.
780
References
Bilaney (2017) Bilaney, Sunny Kishore, “Impact of BEPS Project on Cen-

tralized Operating Models”, 24 International Transfer Pri
Birnkrant (2004) Birnkrant, Henry J., “Transfer Pricing for Services Is Broad-
ly Affected by New Prop. Regs., Mostly Not for the Better”,
ResearchGate (2004).
Bischel (1973) Bischel, Jon E., “Tax Allocations Concerning Inter-Company
Pricing Transactions in Foreign Operations: A Reappraisal”,
13 Va. J. Int’l L. (1972/1973), p. 490.
Blair et al. (2005) Blair, Roger D.; Lafontaine, Francine, “The Economics of
Franchising” (2005).
Blessing (2010a) Blessing, Peter H., New York State Bar Association Tax
Section: Report on Section 367(d), in “NYSBA Tax Sec-
tion Submits Report on U.S. Taxation of Some Intangible
Asset Transfers”, 12 Oct. 2010, available at Tax Analysts
(www.taxanalysts.com); Doc 2010-22245.
Blessing (2010b) Blessing, Peter H., “Divergence of Third Party Pricing from
Arm’s Length Results”, in Baker, Philip; Bobbett, Catherine
(eds.), “Tax Polymath – A Life in International Taxation”
(2010), p. 153.
BNA (1999) “IRS Working to Address Sec. 482 Problems with ‘Cheese
Examples’ in Regulations”, BNA Daily Tax Report, 9 Aug.
1999, G-10.
Bodie et al. Bodie, Zvi; Kane, Alex; Marcus, Alan J., Investments, 6th
(2005) ed. (2005).
Boidman (2011) Boidman, Nathan, “Pricing Canada-US Guarantees after GE
Capital: Still Evolving”, Tax Management Transfer Pricing
Report (2011), at p. 1042.
Boidman et al. Boidman, N.; Kandev, M.N., “BEPS: The OECD discovers
(2013) America”, 72 Tax Notes Int’l (2013), p. 1017.
Boos (2003) Boos, Monica, “International Transfer Pricing: The Valua-
tion of Intangible Assets” (2003).
Bowen (2008) Bowen, David N., “Full Value Methods: Has the IRS Final-
ly Hurled the Holy Hand Grenade? A Critical Analysis of
the Scope of §§ 482, 367(d), and 936(h)(3)(B) in Relation to
Goodwill, Going Concern Value, and Workforce in Place”,
37 Tax Mgmt. Int’l J. 3 (2008).
Boykin (1996) Boykin, Richard, “Transfer pricing policy issues: Who is the
developer?”, 12 Tax Notes Int’l (1996), p. 279.
Brauner (2003) Brauner, Yariv, “An International Tax Regime in Crystalliza-
tion”, 56 Tax Law Review (2003), p. 259.
781
References
Brauner (2004) Brauner, Yariv, “A Good Old Habit, or Just an Old One? Pref-
erential Tax Treatment for Reorganizations”, Brigham Young
University Law Review (2004), p. 1.
Brauner (2008) Brauner, Yariv, “Value in the eye of the beholder: The valua-
tion of intangibles for transfer pricing purposes”, 28 Virginia
Tax Review (2008), p. 79.
Brauner (2010) Brauner, Yariv, “Cost Sharing and the Acrobatics of Arm’s
Length Taxation, 38 Intertax 11 (2010), p. 554.
Brauner (2014a) Brauner, Yariv, “What the BEPS?”, 16 Florida Tax Review 2
(2014), p. 55.
Brauner (2014b) Brauner, Yariv, “BEPS: An Interim Evaluation”, 6 World Tax
Journal 1 (2014), p. 10.
Brauner (2015) Brauner, Yariv, “Transfer Pricing in BEPS: First Round –
Business Interests Win (But, Not in Knock-Out), 43 Intertax
1 (2015), p. 72.
Brauner (2016) Brauner, Yariv, “Transfer Pricing Aspects of Intangibles: The
Cost Contribution Arrangement Model”, in “Transfer Pri
cing in a Post-BEPS World” (Lang, Michael; Storck, Alfred;
Petruzzi, Raffaele eds. 2016).
Brauner (2017) Brauner, Yariv, “Assessing BEPS: Origins, standards and
responses – United States branch report”, Cahiers De Droit
Fiscal International, Vol. 102a, Subject 1 (2017).
Brekel et al. Brekel, Ronald van den; Grambusch, Laurette von; Lieven-
(2016) bruck, Matthias; Militor, Vera; “BEPS Action 8: The Death
of Development Cost Contribution Agreements?”, 24 Trans
fer Pricing Report (2016).
Brem et al. (2005) Brem, Markus; Tucha, Thomas Alexander, “On Transfer
Pricing: Conceptual Thoughts on the Nature of the Multi-
national Firm”, Working Paper, Indian Institute of Manage
ment (2005), available at https://papers.ssrn.com/sol3/papers.
cfm?abstract_id=874568.
Breslin (2013) Breslin, Patrick, “An early-stage investor analogy: How re-
lated-party transfers of intangibles contribute to base erosion
and profit shifting”, 22 Tax Management Transfer Pricing
Report 10 (2013).
Bullen (2010) Bullen, Andreas, “Arm’s length transaction structures: Rec-
ognizing and restructuring controlled transactions in trans-
fer pricing” (2011).
Bullen et al. Bullen, Andreas; Anfinsen, Fredrik, “Skatteloven § 13-1
(2013) fjerde ledd: OECDs retningslinjer for internprising som rett-
skildefaktor i norsk rett”, Skatterett (2013).
782
References
Burke (2011) Burke, John J.A., “Re-Thinking First Principles of Transfer

Pricing Rules”, 30 Virginia Tax Review 3 (2011), p. 102.
Burns et al. Burns, Kylie; Hutchinson, Terry, “The impact of ‘empirical
(2009) facts’ on legal scholarship and legal research training”, 43
The law teacher 2 (2009), pp. 153-178.
Calderón (2007) Calderón, Jose, “The OECD Transfer Pricing Guidelines
as a Source of Tax Law: Is Globalization Reaching the Tax
Law?”, 35 Intertax 1 (2007), p. 4.
Calderón (2012) Calderón, J. Manuel, “Subsidiary Constituted PE, Supreme
Court Rules”, 65 Tax Notes Int’l (2012), p. 743.
Casley et al. Casley, Andrew; Kritikides, Artemis, “Transactional Net
(2003) Margin Method, Comparable Profits Method and the Arm’s
Length Principle”, 10 International Transfer Pricing Journal
9 (2003), p. 162.
Casley et al. Casley, Andrew J.; Timbadia, Bhavik, “Marketing Intan-
(2011) gibles: The Latest Controversy”, 18 International Transfer
Pricing Journal 3 (2011), p. 165.
Cauwenbergh Cauwenbergh, Patrick, “Does the Arm’s Length Standard
(1997) Require a Flexible or Rigid Interpretation?”, 4 International
Transfer Pricing Journal 3, (1997), p. 137.
Chakravarty et al. Chakravarty, A.; Ray, S., “Is Business Restructuring and Tax
(2013) Aligned Supply Chain Still Viable?”, 19 Asia-Pacific Tax
Bulletin 6 (2013), p. 415.
Chand et al. Chand, Vikram; Wagh, Sagar, “The Profit Split Method: Sta-
(2014) tus Quo and Outlook in Light of the BEPS Action Plan”, 21
International Transfer Pricing Journal 6 (2014), p. 402.
Chandler et al. Chandler, Clark; Herr, Thomas; Gustavson, Robert, “Prop.
(2003) Regs. On Option Expense for Comparable-Profits Method:
Does the IRS want the Adjustment Applied Regularly?”, 14
Journal of International Taxation 38 (2003).
Chandler et al. Chandler, Clark; Blough, Stephen; Williams, Will, “Recon-
(2010) ciliation of purchase price allocation and transfer pricing”, 18
Tax management transfer pricing report 23 (2010).
Christians et al. Christians, Allison; Shay, Stephen, “Assessing BEPS: Ori-
(2017) gins, standards and responses – General Report”, Cahiers De
Droit Fiscal International, Vol. 102a, Subject 1 (2017).
CIB (2004) Council for International Business, Comments on Proposed
Regulations, available at Tax Analysts (www.taxanalysts.com);
Doc 2004-266.
783
References
Clark (1993) Clark, Robert G., “Transfer Pricing, Section 482, and In-
ternational Tax Conflict: Getting Harmonized Income Al-
location Measures from Multinational Cacophony”, 42 The
American University Law Review (1993), p. 1155.
Clark (1997) Clark, R., “Choosing a Reliable Profit Level Indicator”, 5
Transfer Pricing Report 3 (1997), p. 807.
Collins et al. Collins, Martin J.; Sotos, David J.; Mullaney, Sean W., “Re-
(2013) visiting Certain Fundamentals in Light of the New Wave of
Outbound Reorganization Guidance”, 42 Tax Mgmt. Int’l J.
21 (2013).
Cools (1999) Cools, M., “International Commercial Databases for Trans-
fer Pricing Studies”, 6 International Transfer Pricing Jour
nal 5 (1999), p. 178.
Cottani (2007) Cottani, Giammarco, “Valuation of intangibles for direct tax
and customs purposes: Is convergence the way ahead?”, 14
Cottani (2016) Cottani, Giammarco, “Commissionaire Arrangements/
Low-Risk Distributors and Attribution of Profits to Perma-
nent Establishments”, in “Transfer Pricing in a Post-BEPS
World” (Lang, Michael; Storck, Alfred; Petruzzi, Raffaele
eds. 2016).
Cowan et al. Cowan, M.J.; W. Newberry, Jr., “Reevaluating the Intellec-
(2013) tual Property Holding Company”, 14 Management Account
ing Quarterly 3 (2013), p. 31.
Culbertson (1995) Culbertson, Robert E., “A rose by any other name: Smelling
the flowers at the OECD’s (last) resort”, 11 Tax Notes Int’l,
(1995), p. 370.
Culbertson et al. Culbertson, Robert E.; Durst, Michael C., “Clearing away
(2003) the sand: Retrospective methods and prospective documenta-
tion in transfer pricing today”, 57 Tax Law Review 37 (2003).
Damodaran Damodaran, Aswath, “Damodaran on valuation: Secu-
(2006) rity analysis for investment and corporate finance”, 2nd ed.
(2006).
Damodaran Damodaran, Aswath, “Investment Valuation: Tools and tech-
(2012) niques for determining the value of any asset”, 3rd ed. (2012).
De Simone et al. De Simone, Lisa; Klassen, Kenneth J.; Seidman, Jeri K., “Un-
(2015) profitable Affiliates and Income Shifting Behavior”, Rock
Center for Corporate Governance at Stanford University,
Working Paper No. 157 (2015), available at http://ssrn.com/
abstract=2340693 or http://dx.doi.org/10.2139/ssrn.2340693.
784
References
Deloitte (2000) Deloitte & Touche, “Withdrawal of tax sparing credit for
Japanese tax residents under Singapore-Japan tax treaty”,
Singapore Accountant (2000), p. 46.
Deloitte (2004) Deloitte & Touche, “Comments on Proposed Transfer Pri
cing Regs, Public Comments on Regulations”, available at
Tax Analysts (www.taxanalysts.com); Doc 2004-6535.
Dickenson (2010) Dickenson, Hak K., “An Explanation of the 2009 IRS Cost
Sharing Arrangement Regulations”, 17 International Trans
fer Pricing Journal 1 (2010).
Dix (2010) Dix, Edward B., “From General to Specific: The Arm’s-
Length Standard’s Evolution and Its Relevance in Determin-
ing Costs to Be Shared in Cost-Sharing Agreements”, 64 The
Tax Lawyer 1 (2010), p. 200.
Dolan (1991) Dolan, Kevin, “Cost Sharing: Who Needs It?”, 32 Tax Man
agement Memo (1991), p. 163.
Dolman (2007) Dolman, N., “Life Cycle of Intangible Property and the Im-
plications for Arm’s Length Pricing”, 8 Tax Planning Inter
national Transfer Pricing 6 (2007).
Driscoll (1988) Driscoll, Richard J., “Income Shifting for Intangible Prop-
erty Transferred Abroad”, 14 Int’l Tax J. (1988), p. 165.
Duerr (1972) Duerr, Michael G., “Tax Allocations and International Busi-
ness: Corporate Experience With Section 482 of the Inter-
nal Revenue Code”, Research Report from the Conference
Board, Report no. 555 (1972).
Dujsic et al. Dujsic, M.; Goldberg, S.; Barsalo, B.; Fleming, R., “Digest-
(2008) ing the Glaxo Decision: A Difficult Pill to Swallow for Trans-
fer Pricing Practitioners”, 15 International Transfer Pricing
Journal 5 (2008), p. 203.
Dunning (1977) Dunning, John H., “Trade, Location of Economic Activity
and the MNE: A Search for an Eclectic Approach”, in “The
International Allocation of Economic Activity” (Hesselborn,
Per-Ove; Ohlin, Bertil; Wijkman, Per Magnus eds. 1977).
Dunning et al. Dunning, John H.; Lundan, Sarianna M., “Multinational en-
(2008) terprises and the global economy”, 2nd ed. (2008).
Durst (2010) Durst, Michael C., “It’s Not Just Academic: The OECD
Should Reevaluate Transfer Pricing Laws”, 57 Tax Notes
Int’l (2010), p. 247.
Durst (2011) Durst, Michael C., “OECD Guidelines: Causes and Conse-
quences”, in “Fundamentals of International Transfer Pricing
in Law and Economics”, MPI Studies in Tax Law and Public
Finance 1 (Schön, Wolfgang; Konrad, Kai A. eds. 2011), p. 123.
785
References
Durst (2012a) Durst, Michael C., “Analysis of a Formulary System for

Dividing Income, Part II: Examining Current Formulary
and Arm’s-Length Approaches”, 22 BNA Tax Management
Transfer Pricing Report 5 (2012), p. 270.
Durst (2012b) Durst, Michael C., “A Two-Option Compromise for Intangi-
bles Guidelines”, 66 Tax Notes Int’l (2012), p. 1123.
Durst (2012c) Durst, Michael C., “‘Risk’ and the OECD Discussion Drafts
on Transfer Pricing”, 68 Tax Notes Int’l (2012), p. 285.
Durst (2012d) Durst, Michael C., “The OECD Discussion Draft on Transfer
Pricing for Intangibles, Viewpoints”, Tax Notes Int’l (2012),
p. 450.
Dwarkasing Dwarkasing, Ramon, Associated enterprises: A concept es
(2011) sential for the application of the arm’s length principle and
transfer pricing (2011).
Dworin (1990) Dworin, Lowell, “Transfer Pricing Issues”, 43 National Tax
Journal 3 (1990), p. 285.
Eden (1998) Eden, Lorraine, “Taxing Multinationals: Transfer Pricing
and Corporate Income Taxation in North America” (1998).
Edgar (2005) Edgar, Tim; Holland, David, “Source Taxation and the
OECD Project on Attribution of Profits to Permanent Estab-
lishments”, 37 Tax Notes Int’l (2005), p. 525.
Elliott (2011) Elliott, Amy S., “LB&I May Be Winning Internal Fight Over
Veritas, Practitioners Say”, 133 Tax Notes Int’l (2011), p. 472.
Ernick (2013) Ernick, David, “Base Erosion, Profit Shifting And the Future
of the Corporate Income Tax”, 42 Tax Management Interna
tional Journal (2013), p. 1.
EY (2004) Ernst & Young, “Comments on Proposed Regulations on the
Treatment of Services Under Section 482 and the Allocation
of Income and Deductions from Intangibles”, available at Tax
Faulhaber (2017) Faulhaber, Lilian V., “The Luxembourg Effect: Patent Boxes
and the Limits of International Cooperation”, 101 Minnesota
Law Review (2017), p. 1641.
Fedusiv (2016) Fedusiv, Oleh, “Transactions with Hard-to-Value Intangibles:
Is BEPS Action 8 Based on the Arm’s Length Principle?”, 23
Feinschreiber Feinschreiber, Robert; Kent, Margaret, “A Critical Look at
et al. (2016) the OECD’s Profit-Split Transfer Pricing Revisions”, 83 Tax
Notes Int’l (2016), p. 961.
786
References
Fichtner et al. Fichtner, Jason J.; Michel, Adam N., “The OECD’s Conquest
(2016) of the United States: Understanding the Costs and Conse-
quences of the BEPS Project and Tax Harmonization”, Mer
catus Research (2006), available at: https://www.mercatus.
org/system/files/Fichtner-BEPS-Initiative-v1.pdf.
Finley (2016) Finley, Ryan, “IRS and Taxpayers Still at Odds Over Cost-
Sharing Discount Rates”, 81 Tax Notes Int’l (2016), p. 237.
Fleming et al. Fleming Jr, J. Clifton; Peroni, Robert J.; Shay, Stephen E.,
(2014) “Formulary Apportionment in the U.S. International Income
Tax System: Putting Lipstick on a Pig?”, 36 Michigan Jour
nal of International Law 1 (2014).
Francescucci Francescucci, David L.P., “The Arm’s Length Principle and
(2004a) Group Dynamics – Part 1: The Conceptual Shortcomings”,
11 International Transfer Pricing Journal 6 (2004), p. 55.
Francescucci Francescucci, David L.P., “The Arm’s Length Principle and
(2004b) Group Dynamics – Part 2: Solutions to Conceptual Short-
comings”, 11 International Transfer Pricing Journal 6
(2004), p. 235.
Fris (2003) Fris, Pim, “Dealing with Arm’s Length and Comparability in
the Years 2000”, 10 International Transfer Pricing Journal
6 (2003), p. 194.
Fris et al. (2006) Fris, Pim; Gonnet Sebastian, “A European View on Transfer
Pricing After Glaxo”, 7 Tax Planning International Transfer
Pricing 11 (2006).
Fris et al. (2010) Fris, Pim; Gonnet, Sebastian, “The State of the Art in Com-
parability for Transfer Pricing”, 17 International Transfer
Pricing Journal 2 (2010).
FW (2004) Fenwick & West, “Attorneys Criticize Proposed U.S. Trans-
fer Pricing Regs on Services, Intangibles”, available at Tax
Gazur (1994) Gazur, Wayne M., “The Forgotten Link: ‘Control’ in Sec-
tion 482”, 15 Northwestern Journal of International Law and
Business 1 (1994), p. 1.
Gazzo (2003) Gazzo, Massimiliano, “Permanent Establishment through
Related Corporations: New Case Law in Italy and its Impact
on Multinational Flows”, 57 Bulletin for International Fiscal
Documentation 6 (2003), p. 257.
Gelin et al. (2007) Gelin, Stéphane; Sorel, David, “French Commissionaire:
A PE for Its Foreign Principal”, 47 Tax Notes Int’l (2007),
p. 581.
787
References
Gelin et al. (2010) Gelin, Stéphane; Gutmann, Daniel, “France: Commissio

naire as PE?”, 12 Tax Planning International, European Un
ion Focus 4 (2010), p. 10.
Glaize et al. Glaize, Antoine et al., “Risk Simulation: A Practical Method
(2011) for the Arm’s Length Reallocation of risks in Business Re-
structurings”, 12 International Transfer Pricing Journal 6
(2011), p. 15.
Goede (2012) Goede, Jan de et al., “Interpretation and Application of Ar-
ticle 5 (Permanent Establishment) of the OECD Model Tax
Convention: Response from IBFD research staff”, 66 Bul
letin for International Taxation 6 (2012), p. 313.
Gonnet (2016) Gonnet, Sébastien, “Risks Redefined in Transfer Pricing
Post-BEPS”, in “Transfer Pricing in a Post-BEPS World”
(Lang, Michael; Storck, Alfred; Petruzzi, Raffaele eds.,
2016).
Gouthière (2010) Gouthière, Bruno, “Zimmer: ‘Commissionaire’ Agent Is Not
a Permanent Establishment”, 50 European Taxation 8 (2010),
p. 350.
Gresik (2011) Gresik, Thomas A., “Assessing the Normative Differences
Between Formulary Apportionment and Separate Account-
ing”, in “Fundamentals of International Transfer Pricing in
Law and Economics” MPI Studies in Tax Law and Public Fi
nance 1 (Schön, Wolfgang; Konrad, Kai A. eds. 2011), p. 257.
Gupta (2016) Gupta, Ajay, “The Disappearing Commensurate-With-In-
come Standard”, 82 Tax Notes International (2016), p. 1227.
Haas (1991) Haas, Lawrence J., “Divvying up taxes”, Nat’l Journal
(1991), p. 1106.
Hafkenscheid Hafkenscheid, Rutger, “The BEPS Report on Risk Alloca-
(2017) tion: Not So Functional”, 24 International Transfer Pricing
Journal 1 (2017), p. 19.
Hamaekers (2001) Hamaekers, Hubert, “Arm’s Length – How Long?”, 8 Inter
national Transfer Pricing Journal 2 (2001), p. 30.
Hamaekers (2003) Hamaekers, Hubert, “The Comparable Profits Method and
the Arm’s Length Principle”, 10 International Transfer Pri
Hardy (2006) Hardy, David R., “Assignment of Corporate Opportunities
– The Migration of American Intangibles”, PLI/Tax (2006).
Haugen (2005) Haugen, Lars P., “In Tax Practice the Cost-Plus Method is
Often a TNMM in Disguise: What to do about it?”, 12 Inter
national Transfer Pricing Journal 6 (2005), p. 224.
788
References
Heggmair (2017) Heggmair, Maik, “The New Interpretation of the Arm’s

Length Principle: A Post-BEPS Evaluation”, 24 International
Helderman et al. Helderman, Loek Et al., “A New Era in Determining Arm’s
(2013) Length Compensation for Intangibles? A Comparative Over-
view of Existing and Possible Future Transfer Pricing Prin-
ciples”, 20 International Transfer Pricing Journal 6 (2013).
Hellerstein (1993) Hellerstein, Walter, “Federal Income Taxation of Multina-
tionals: Replacement of Separate Accounting with Formu-
lary Apportionment”, 60 Tax Notes Int’l (1993), p. 1131.
Hellerstein Hellerstein, Walter, “The Case for Formulary Apportion-
(2005a) ment”, 12 International Transfer Pricing Journal 3 (2005),
p. 103.
Hellerstein Hellerstein, Walter, “International Income Allocation in the
(2005b) Twenty-first Century: The Case for Formulary Apportion-
ment”, 12 International Transfer Pricing Journal 3 (2005).
Henderson et al. Henderson, James; Quandt, Richard, Microeconomic Theo
(1980) ry, 3rd ed. (1980).
Higinbotham Higinbotham, Harlow; Guerad, Laurent; Jankowski, Donald,
et al. (1993) “An Economic Perspective on the Best Method Rule: Inter-
company Transfer Pricing under New Section 482 Regula-
tions”, 6 Tax Notes Int’l (1993), p. 731.
Higinbotham Higinbotham, Harlow; Levey, Marc M., “When Arm’s
et al. (1998) Length Isn’t Really Arm’s Length: Issues in Application of
the Arm’s Length Standard”, 26 Intertax 8/9 (1998), p. 235.
Hill (2008) Hill, Dale C., “The Tax Court of Canada on Glaxo: Results
and Implications”, 17 Transfer Pricing Report 6 (2008), p. 269.
Hill et al. (2007) Hill, Dale C.; Di Libero, Santino; Hejazi, Jamal, “Moving
intellectual property offshore in a transfer pricing setting”,
48 Tax Notes Int’l (2007), p. 683.
Hill et al. (2008) Hill, Dale C.; Di Libero, Santino, “Migrating IP offshore
within a transfer pricing setting”, 9 Tax Planning Interna
tional Transfer Pricing 4 (2008), p. 22.
Hines (1990) Hines Jr., James R., “The Transfer Pricing Problem: Where
the Profits Are”, NBER Working Paper Series, No. 3528, Na
tional Bureau of Economic Research (1990).
Hines (2014) Hines Jr., James R., “How Serious Is the Problem of Base
Erosion and Profit Shifting?”, 62 Canadian Tax Journal 2
(2014), p. 443, available at https://repository.law.umich.edu/
cgi/viewcontent.cgi?article=2378&context=articles.
789
References
Hoecke (2011) Hoecke, Mark Van, “Legal doctrine: Which method for what
kind of discipline”, in “Methodologies of Legal Research:
Which Kind of Method for What Kind of Discipline?” (Hoe-
cke ed. 2011), p. 1.
Horst (1993) Horst, Thomas, “The Comparable Profits Method”, 6 Tax
Notes Int’l, (1993), p. 1443.
Horst (2009) Horst, Thomas, “Employee Stock Options and the Xilinx
Case”, 55 Tax Notes Int’l (2009).
Horst (2011) Horst, Thomas, “Credit Ratings and the Debt-related Costs
for a Subsidiary of a Multinational Firm”, in “Fundamentals
of International Transfer Pricing in Law and Economics”,
MPI Studies in Tax Law and Public Finance 1 (Schön, Wolf-
gang; Konrad, Kai A. eds. 2011), p. 159.
Hughes et al. Hughes, John; Khripounova, Elena, “Veritas and the aim of
(2010) a scientific approach to valuing intangible property in con-
trolled transactions”, 19 Tax Management Transfer Pricing
Report 13 (2010), p. 770.
Huibregtse et al. Huibregtse, Steef et al., “IP Companies and Substance: No-
(2011) Fly Zones?”, 12 International Transfer Pricing Journal 5
(2011), p. 4.
Hyde et al. (2005) Hyde, Mark; Dennis, Barry, “A Buyer’s Guide to Transfer
Pricing Software”, 16 Journal of International Taxation 16
(2005).
IFA (1971) IFA General Report, “Criteria for the Allocation of Items of
Income and Expense between Related Corporations in Dif-
ferent States, Whether or not parties to tax conventions”, Ca
hiers De Droit Fiscal International, Vol. 56b (1971).
IFA (1975) IFA General Report, “Allocation of Expenses in International
Arm’s Length Transactions of Related Companies”, Cahiers
De Droit Fiscal International, Vol. 60b (1975).
IFA (1992) International Fiscal Association, “Transfer pricing in the ab-
sence of comparable market prices”, Vol. LXXVIIa, Cahiers
de droit fiscal international (1992).
IFA (2004) International Fiscal Association, “Comments on Services
Transfer Pricing Regs”, available at Tax Analysts (www.taxan
alysts.com); Doc 2004-18756.
IFA (2007) International Fiscal Association, “Transfer pricing and intan-
gibles”, Cahiers de droit fiscal international, Vol. 92a (2007).
IFA (2011) International Fiscal Association, “Cross-border business re-
structuring”, Vol. 96a, Cahiers de droit fiscal international
(2011).
790
References
IFA (2012) International Fiscal Association, “The debt-equity conun-

drum”, Vol. 97b, Cahiers de droit fiscal international (2012).
Ihli (2008) Ihli, Uwe, “Transfer pricing, restructuring, apportionment
and other challenges for tax directors”, 36 Intertax 8/9
(2008), p. 346.
Innamorato Innamorato, Caterina, “The Concept of a Permanent Es-
(2008) tablishment within a Group of Multinational Enterprises”,
48 European Taxation 2 (2008), p. 81.
Jie-A-Joen et al. Jie-A-Joen, Clive; Brandt, Ramon, “U.S. Intangibles Regs
(2007) – A Comparison with OECD Guidelines”, 8 Tax Planning
International Transfer Pricing 6 (2007).
Kadet (2015) Kadet, Jeffery M., “Expansion of the Profit-Split Method:
The Wave of the Future”, 77 Tax Notes Int’l (2015), p. 1183.
Kane (2006) Kane, Mitchell A., “Risk and Redistribution in Open and
Closed Economies”, 92 Virginia Law Review (2006), p. 867.
Kane (2014) Kane, Mitchell A., “Transfer Pricing, Integration and Syner-
gy Intangibles: A Consensus Approach to the Arm’s Length
Standard”, 6 World Tax Journal 3 (2014), p. 282.
Kantowitz (2018) Kantowitz, Robert, “Altera, Transfer Pricing, and the Costs
of Stock-Based Compensation”, 89 Tax Notes Int’l (2018),
p. 709.
Kapoor (2008) Kapoor, Shiv, “Rolls Royce Decision: Income Attribution to
Permanent Establishments”, 15 International Transfer Pri
Kauder (1993) Kauder, Louis M., “Intercompany Pricing and Section 482:
A Proposal to Shift from Uncontrolled Comparables to For-
mulary Apportionment Now”, 58 Tax Notes Int’l (1993),
p. 485.
Kirschenbaum Kirschenbaum, Robert S. et al., “Intangibles/Services Regs –
et al. (2003) At Your Service(s): ‘Still Crazy After All These Years’”, 14
Journal of International Taxation 10 (2003).
Kirschenbaum Kirschenbaum, Robert S. et al., “Proposed Cost Sharing Regu
et al. (2005) lations Introduce ‘New’ Standards”, 16 Journal of Internation
al Taxation 12 (2005).
Kleinbard (2011) Kleinbard, Edward D., “Stateless Income”, 11 Florida Tax
Review 9 (2011), p. 699.
Kleinbard (2012a) Kleinbard, Edward D., “Stateless Income’s Challenge to Tax
Policy”, 68 Tax Notes Int’l, 29 October 2012, p. 499.
Kleinbard (2012b) Kleinbard, Edward D., “Stateless Income’s Challenge to Tax
Policy: Part 2”, 68 Tax Notes Int’l (2012), p. 671.
791
References
Kobetsky (2006) Kobetsky, Michael, “Article 7 of the OECD Model: Defining

the Personality of Permanent Establishments”, 60 Bulletin
for International Taxation 10 (2006), p. 411.
Kobetsky (2008) Kobetsky, Michael, “The Case for Unitary Taxation of Inter-
national Enterprises”, 62 Bulletin for International Taxation
5 (2008), p. 201.
Kobetsky (2011) Kobetsky, Michael, “International Taxation of Permanent
Establishments” (2011).
Kochman et al. Kochman, Neal M.; Smiley, Stafford, “The Temporary Cost
(2009) Sharing Regulations”, 36 Corporate Taxation 44 (2009).
Kofler (2013) Kofler, Georg W., “The BEPS Action Plan and Transfer Pri
cing: The Arm’s Length Standard under Pressure?”, British
Tax Review 5 (2013), p. 646.
Koomen (2015a) Koomen, Mirjam, “Transfer Pricing in a BEPS Era: Rethink-
ing the Arm’s Length Principle – Part I”, 22 International
Koomen (2015b) Koomen, Mirjam, “Transfer Pricing in a BEPS Era: Rethink-
ing the Arm’s Length Principle – Part II”, 22 International
Kotarba (2009) Kotarba, Christopher J., “Better than the ‘Best’: Transfer
Pricing Methodology in the Wake of Roche”, 48 Colombia
Journal of Transnational Law (2009), p. 140.
Koue et al. (2017) Koue, Lars Røikjær et al., “Defining and Calculating the
Arm’s-Length Range”, 86 Tax Notes Int’l (2017), p. 987.
Kroppen et al. Kroppen, Heinz-Klaus; Dawid, Roman; Schmidtke, Richard,
(2011) “Profit Split, the Future of Transfer Pricing? Arm’s Length
Principle and Formulary Apportionment Revisited from a
Theoretical and a Practical Perspective”, in “Fundamentals
of International Transfer Pricing in Law and Economics”,
MPI Studies in Tax Law and Public Finance 1 (Schön, Wolf-
gang; Konrad, Kai A. eds. 2011), p. 267.
Lagarden (2014) Lagarden, Martin, “Intangibles in a Transfer Pricing Con-
text: Where Does the Road Lead?”, 21 International Transfer
Pricing Journal 5 (2014), at p. 337.
Laine (2012) Laine, Anne, “Transfer pricing of intangible assets – Impli-
cations of Finnish intellectual property case law on valuation
of intangible assets” (2012).
Lang et al. (2011) Lang, Michael et al. (eds.), “Tax treaty case law around the
globe – 2011” (2012).
Langbein (1986) Langbein, Stanley, “The Unitary Method and the Myth of
Arm’s Length”, 30 Tax Notes Int’l (1986), p. 625.
792
References
Langbein (2005) Langbein, Stanley, “U.S. Transfer Pricing and the Outsour
cing Problem”, 37 Tax Notes Int’l (2005), p. 1065.
Lashbrooke Lashbrooke, Elvin C. Jr., “I.R.C. § 482 Commensurate With
(1989) Income Standard for Transfers of Intangibles”, 1 DePaul
Bus. L. J. 174 (1989), p. 189.
Lebowitz (1999) Lebowitz, Brian E., “Transfer Pricing and the End of Interna-
tional Taxation”, 19 Tax Notes Int’l (1999), p. 1201.
Lebowitz (2008) Lebowitz, B., “Profit Sharing as a New World Order in Inter-
national Taxation”, 52 Tax Notes Int’l (2008), p. 585.
Ledure et al. Ledure, David; van der Breggen, Michel; Hardy, Matthew,
(2010) “Financial Transactions in Today’s World: Observations
from a Transfer Pricing Perspective”, 38 Intertax 6/7 (2010),
p. 350.
Lemein (2005) Lemein, Gregg D., “Transfer pricing and related tax aspects
of global supply chain restructurings”, 37 Tax Notes Int’l
(2005), p. 715.
Lepard Lepard, Brian D., “Is the United States Obligated to Drive
(1999/2000) on the Right? A Multidisciplinary Inquiry into the Norma-
tive Authority of Contemporary International Law using the
Arm’s Length Standard as a Case Study”, 10 Duke Journal of
Comparative and International Law (1999/2000), p. 43.
Levey (2000) Levey, Marc M., “Tax Court Sends Messages to Taxpayers In
DHL”, PLI/TAX (2000).
Levey et al. Levey, Marc M.; Ruchelman, Stanley C., “Section 482 – The
(1987) Super Royalty Provisions Adopt The Commensurate Stand-
ard”, 41 Tax Lawyer 3 (1987/1988).
Levey et al. Levey, Mark M. et al., “DHL: Ninth Circuit sheds very little
(2002) light on bright-line test”, 13 Journal of International Taxa
tion 10 (2002).
Levey et al. Levey, Marc M. et al., “The Quest for Marketing Intangi-
(2006) bles”, 34 Intertax 1 (2006), p. 2.
Levey et al. Levey, Mark M.; Herksen, Monique. van, “General part of
(2008a) special issue on intangibles valuation”, 16 Tax Management
Transfer Pricing (Special) Report 18 (2008), p. S3.
Levey et al. Levey, Marc M.; van Herksen, Monique; Fletcher, Richard,
(2008b) “Intangibles valuation: Controversies, solutions”, 16 Tax
Management Transfer Pricing Report 18 (2008), p. S1.
Levey et al. Levey, M.; Herksen, M. van; Fletcher, R., “Identifying, valu-
(2008c) ing, and migrating intangibles: trouble ahead”, 9 Tax Plan
ning International Transfer Pricing 5 (2008), p. 16.
793
References
Levey et al. Levey, Marc M.; Wrappe, Steven C., “Transfer Pricing:
(2010) Rules, Compliance and Controversy”, 3rd ed. (2010).
Levey et al. Levey, Marc M.; Arthur, Brian P., “Indian Supreme court
(2011) sets aside high court ruling in Maruti Suzuki – what’s next
for marketing intangibles?”, 22 Journal of International Tax
ation 1 (2011).
Li (2011) Li, Jinyan, “Soft Law, Hard Realities and Pragmatic Sugges-
tions: Critiquing the OECD Transfer Pricing Guidelines”, in
“Fundamentals of International Transfer Pricing in Law and
Economics”, MPI Studies in Tax Law and Public Finance 1
(Schön, Wolfgang; Konrad, Kai A. eds. 2011), p. 71.
Lin et al. (2010) Lin, Alice; Wright, Deloris R., “The Tax Court Decision
in Veritas: A comment”, 17 International Transfer Pricing
Journal 2 (2010), p. 147.
Llinares (2005) Llinares, Emmanuel, “Profit Split Methods: Intangibles,
Market Structure and the Use of Profit Split Methods”, 24
International Tax Review (2005), Supplement, at p. 36.
Lipsey et al. Lipsey, Richard; Steiner, Peter, “Economics”, 6th ed. (1981).
(1981)
Lodin (2009) Lodin, Sven-Olof, “Professorn som blev näringslivstorped:
min tid i skattepolitiken” (2009).
Luckhaupt et al. Luckhaupt, Hagen; Overesch, Michael; Schreiber, Ulrich,
(2011) “The OECD Approach to Transfer Pricing: A Critical As-
sessment and Proposal”, in “Fundamentals of International
Transfer Pricing in Law and Economics” MPI Studies in Tax
Law and Public Finance 1 (Schön, Wolfgang; Konrad, Kai
A. eds., 2011), p. 91.
Lüdicke et al. Lüdicke, Jochen; Avery Jones, John F., “The Origins of Arti-
(2014) cle 5(5) and 5(6) of the OECD Model”, 6 World Tax Journal
3 (2014).
MacIsaac et al. MacIsaac, Shaun; Wall, Matthew, “A Lesson from Alberta
(2012) Printed Circuits: Intangibles Key to Proper TNMM Analy-
sis”, 20 BNA Tax Management Transfer Pricing Report 21
(2012), p. 4.
Maisto (1992) Maisto, Guglielmo, “General report – Transfer pricing in the
absence of comparable market prices”, Cahiers de droit fiscal
international, Vol. LXXVIIa (1992).
Malherbe (2010) Malherbe, Jacques; Daenen, Philip, “Permanent Estab-
lishments Claim Their Share of Profits: Does the Taxman
Agree?”, 64 Bulletin for International Taxation 7 (2010),
p. 359.
794
References
Markham (2004) Markham, Michelle, “Transfer Pricing of Intangible Assets

in the US, the OECD and Australia: Are Profit-Split Method-
ologies The Way Forward?”, 8 University of Western Sydney
Law Review 1 (2004), p. 56.
Markham (2005) Markham, Michelle, “The Transfer Pricing of Intangibles”
(2005).
Martínez (2010) Martínez, Bibiana A. Cruz, “The Arm’s Length Standard vs
the Commensurate with Income Standard: Transfer Pricing
Issues in the Valuation of Intangible Assets”, 2 U.P.R. Busi
ness Law Journal (2010), p. 302.
Mayer (2009) Mayer, Stefan, “Formulary Apportionment for the Internal
Market” (2009).
MBRM (2004) Mayer, Brown, “Rowe & Maw on behalf of Ford Motor Com-
pany, Comments on Proposed Regulations”, available at Tax
MBRM (2005) Mayer, Brown, “Rowe & Maw on behalf of the Services
Regulations Coalition, Comments on Proposed Regulations
Regarding the Treatment Under Section 482 of Controlled
Services Transactions and Intangible Transfers”, available at
Tax Analysts (www.taxanalysts.com); Doc 2005-13569.
McCarthy (1996) McCarthy, J. Thomas, “McCarthy on Trademarks and Unfair
Competition”, 4th ed. (1996).
McCawley (1990) McCawley, Harrison B., “Section 482: The Statute and the
Regulations”, Tax Management Portfolios (1990).
McClure (1994) McClure, William P., “Washington attorneys submit replace-
ment draft for ‘developer-assister rule’ in temporary transfer
pricing regs.”, 63 Tax Notes Int’l (1994), p. 996.
McCrodan et al. McCrodan, Andrew F.; Borraccia, Salvador M., “The Dilem-
(2009) ma of GSK – Unreasonable in the Circumstances?”, in “Re-
port of Proceedings of the Sixtieth Tax Conference”, 2008
Conference Report, Canadian Tax Foundation (2009).
McLure (2002) McLure, Charles E., Jr., “Replacing Separate Entity Account-
ing and the Arm’s Length Principle with Formulary Appor-
tionment”, 56 Bull. for Intl .Fiscal Docn. 12 (2002), p. 586.
McShan (1989) McShan, W. Scott et al., “Section 482 White Paper… A Re-
view of Third-Party License Agreements: Are Periodic Ad-
justments Arm’s Length?”, 41 Tax Executive (1988-1989),
p. 353.
Mentz (1999) Mentz, J. Roger, “Mentz recommends changes to US Treas-
ury’s developer-assister regs.”, Tax Notes Int’l (1999), p. 876.
795
References
Mentz et al. Mentz, J. Roger, Carlisle, Linda E., “The tax ownership of
(1997) intangibles under the arm’s length principle”, 15 Tax Notes
Int’l (1997), p. 1495.
Metrick (2010) Metrick, Andrew; Yasuda Ayako, “Venture Capital and the
Finance of Innovation”, 2nd ed. (2010).
Milewska (2017) Milewska, Marta, “A Mexican Perspective on Value Creation
under the OECD’s BEPS Approach to Transfer Pricing”, 24
Milewska et al. Milewska, Marta; Hurtado de Mendoza, Mauricio, “The
(2010) Increasing Importance of Intangible Assets and the Rise
of Profit Split Methods”, 17 International Transfer Pricing
Journal 2 (2010).
Miyatake et al. Miyatake, Toshio; Guttentag, Joseph H., “Transfer Pricing:
(1994) U.S. and Japanese Views”, 8 Tax Notes Int’l (1994), p. 375.
Mogle (2001) Mogle, J.R., “The Future of International Transfer Pricing:
Practical and Policy Opportunities Unique to Intellectual
Property, Economic Substance, and Entrepreneurial Risk in
the Allocation of Intangible Income”, 10 George Mason Law
Review (2001/2002), p. 925.
Monsenego (2014) Monsenego, Jérôme, “The Substance Requirement in the
OECD Transfer Pricing Guidelines: What is the substance of
the substance requirement?”, 21 International Transfer Pri
Morgan et al. Morgan, R. William, Meyer, Thomas E., “Risks, intangibles,
(2009) and business restructurings”, 18 Tax Management Transfer
Pricing Report 8 (2009), p. 475.
Morse (2010) Morse, Susan, “Revisiting Global Formulary Apportion-
ment”, 29 Virginia Tax Review 4 (2010), p. 593.
Murphy (2010) Murphy, W. Joe, “Services, intangibles, and exit charges:
The evolving views on intercompany services”, 21 Journal
of International Taxation 11 (2010), pp. 48-53.
Musselli (2006) Musselli, Andrea, “Arm’s Length Intragroup Intangible
Transfers: Economics, Regulations and Actual Behaviors”,
ECONPUBBLICA Working Paper No. 108 (2006), avail-
able at https://ssrn.com/abstract=878443 or http://dx.doi.org/
10.2139/ssrn.878443.
Musselli et al. Musselli, Andrea; Hunter, Donatella M., “Glaxo transfer
(2007a) pricing case: Economic Rationale, Legal Framework and In-
ternational Issues”, 14 International Transfer Pricing Jour
nal 3 (2007).
796
References
Musselli et al. Musselli, Andrea; Musselli, Alberto, “The arm’s length

(2007b) standard in multinationals’ taxation: Economics, regulations,
firms and administrations’ behavior”, 8 Tax Planning, Inter
national Transfer Pricing BNA 10 (2007), p. 3.
Musselli et al. Musselli, Andrea; Musselli, Alberto, “Stripping the Func-
(2008a) tions of Affiliated Distributors”, 15 International Transfer
Musselli et al. Musselli, Andrea; Musselli, Alberto, “Stripping the Func-
(2008b) tions of Producing Affiliates of a Multinational Group: Ad-
dressing Tax Implications via Economics of Contracts”, 15
Musselli et al. Musselli, Andrea; Musselli, Alberto, “Observations on
(2009) OECD Discussion Draft on Business Restructuring: Is the
Notion of Control over Risk at Arm’s Length?”, 16 Interna
tional Transfer Pricing Journal 4 (2009), p. 239.
Musselli et al. Musselli, Andrea; Musselli, Alberto, “Saving Arm’s Length
(2012) Pricing: From Economists’ Myths of Tax Avoidance by Tax-
payers, to the Reality of Uncertain Application of Rules”, 19
Musselli et al. Musselli, Andrea; Fusaro, Andrea, “Profit shifting and cer-
(2013) tainty on allocation of intangibles”, 14 Transfer Pricing In
ternational Journal BNA 4 (2013).
Musselli et al. Musselli, Andrea; Musselli, Alberto, “Rise of a New Stand-
(2017) ard: Profit Location in Countries of Important Intangible
Functions Managers”, 24 International Transfer Pricing
Journal 5 (2017), p. 331.
Nadal (2009) Nadal, Lisa M., “Cost-Sharing Periodic Adjustments Not
Automatic, Officials Say”, 53 Tax Notes Int’l (2009), p. 681.
Navaretti et al. Navaretti, Giorgio B. et al., “Multinational Firms in the World
(2004) Economy” (2004).
Navarro (2017) Navarro, Aitor, “Transactional Adjustments in Transfer Pri
cing” (2017).
NVCA, YB 2013 National Venture Capital Association Yearbook 2013, avail-
able at http://nvca.org/.
NYSBA (1991) New York State Bar Association Tax Section, “Report on
Proposed Legislation on the Amortization of Intangibles”,
Tax Notes Today (1991).
O’Brien et al. O’Brien, James M.; Oates, Mark A., “Transfer Pricing: Puer-
(2007) to Rico Transfer Pricing Emerges … Again … As an Emer
ging Issue”, 33 International Tax Journal 17 (2007).
797
References
Odintz et al. Odintz, Joshua D.; Rahim, Salim, “The future of transfer
(2017) pricing – United States branch report”, Cahiers De Droit Fis
cal International, Vol. 102b, Subject 2 (2017).
Oestreicher (2011) Oestreicher, Andreas, “Valuation issues in transfer pricing of
intangibles: Comments on the scoping of an OECD project”,
39 Intertax 3 (2011), p. 126.
Okten (2013) Okten, Rezan; Helderman, Loek; Sporken, Eduard, “The Re-
vised OECD Discussion Draft on Transfer Pricing Aspects
of Intangibles”, 21 International Transfer Pricing Journal 1
(2014).
Oosterhoff (2008) Oosterhoff, Danny, “The True Importance of Significant
People Functions”, 2 International Transfer Pricing Journal
2 (2008), p. 68.
Osborn et al. Osborn, Jason; Klein, Kenneth, “BEPS Actions 8-10: How
(2017) MNEs Can Take Control of Their Exposure by Taking Con-
trol of Their Risks”, 88 Tax Notes International (2017), p. 551.
Ossi (1999) Ossi, Gregory J., “The Significance of Intangible Property
Rights in Transfer Pricing”, 19 Tax Notes Int’l (1999), p. 993.
Pankiv (2016) Pankiv, Marta, “Post-BEPS Application of the Arm’s Length
Principle to Intangibles Structures”, 23 International Trans
fer Pricing Journal 6 (2016), p. 463.
Pankiv (2017) Pankiv, Marta, “Contemporary Application of the Arm’s
Length Principle in Transfer Pricing” (2017).
Pantelidaki et al. Pantelidaki, Stephanie; Thomas, Ryann, “Impact of OECD
(2013) intangibles developments in financial services”, Transfer
Pricing International Journal 11 (2013), p. 4.
Parekh (2015) Parekh, Siddharth, “The Concept of ‘Options Realistically
Available’ under the OECD Transfer Pricing Guidelines”, 22
Pedersen (2010) Pedersen, Jan, “Afkald på forpagtningsafgift – Armslængde-
princippet”, Revision & Regnskabsvæsen (2011).
Penelle (2014) Penelle, Philippe G., “The Economics of Business Restruc-
turing and Exit Charges”, 23 Tax Management Transfer Pri
cing Report 3 (2014), p. 238.
Peng (2016) Peng, Claire (Xue), “The Application of the Arm’s Length
Principle to the Allocation of Joint Efficiencies within MNEs”,
23 International Transfer Pricing Journal 5 (2016), p. 378.
Pereira (2009) Pereira, Neil, “First Thoughts on Draft OECD Model Tax
Convention Article 7: Business Profits”, 55 Tax Notes Int’l
(2009), p. 371.
798
References
Petruzzi (2016) Petruzzi, Raffaele, “The Arm’s Length Principle: Between

Legal Fiction and Economic Reality”, in “Transfer Pricing in
a Post-BEPS World” (Lang, Michael; Storck, Alfred; Petru-
zzi, Raffaele eds. 2016).
PG (2003) Pharmaceutical Group, “Comments on Proposed Transfer
Pricing Regs.”, available at Tax Analysts (www.taxanalysts.
com); Doc 2003-26027.
Pheiffer et al. Pheiffer, Sebastian; Ursprung-Steindl, Marlies (eds.), “Glob-
(2015) al trends in VAT/GST and Direct Taxes”, Series on interna
tional tax law, Vol. 93 (2015).
Picciotto (2014) Picciotto, Sol, “From Independent Entity Back to the Unitary
Principle”, 73 Tax Notes Int’l (2014), p. 13.
Pichhadze (2013) Pichhadze, Amir, “Canada’s Transfer Pricing Test in the Af-
termath of GlaxoSmithKline Inc.: A Critique of the Reason
able Business Person Test”, 20 International Transfer Pri
Pichhadze (2015) Pichhadze, Amir, “The Arm’s Length Comparable in Trans-
fer Pricing: A Search for an ‘Actual’ or a ‘Hypothetical’
Transaction?”, 7 World Tax Journal 3 (2015).
Pichhadze (2015a) Pichhadze, Amir, “Allocation of Income and Deductions
Among Taxpayers Under Section 482 of the United States
Internal Revenue Code: Alerting the Courts to the Role of
Contractual Interpretation Law in the Transfer Pricing Arm’s
Length Comparability Analysis”, 22 International Transfer
Pichhadze Pichhadze, Amir, “Exposing Unaddressed Issues in the
(2015b) OECD’s BEPS Project: What About the Roles and Implica-
tions of Contract Interpretation Law and Private Internation-
al Law in the Transfer Pricing Arm’s Length Comparability
Analysis?”, 7 World Tax Journal 1 (2015), p. 99.
Pinto (2006) Pinto, Dale, “The Need to Reconceptualise the Permanent Es-
tablishment Threshold”, 60 Bull. for Intl. Taxn. 7 (2006), p. 266.
Plumb et al. Plumb; Kapp, “Reallocation of Income and Deductions Un-
(1963) der Section 482”, 41 TAXEs (1963).
Polinsky (1993) Polinsky, Alexander, “New IRS Data Raise Further Doubts
About Revenue Gains From Increased Section 482 Enforce-
ment”, Tax Notes Int’l (1993), p. 1419.
Popper (1962) Popper, Karl, “Conjectures and Refutations: The Growth of
Scientific Knowledge” (1962).
Porter (1985) Porter, Michael E., “The Competitive Advantage: Creating
and Sustaining Superior Performance” (1985).
799
References
Przysuski et al. Przysuski, Martin; Lalapet, Srini; Swaneveld, Hendrik, “Se-

(2003) cret Transfer Pricing Comparables Still Popular in Some Ju-
risdictions”, 32 Tax Notes Int’l (2003), p. 169.
Przysuski et al. Przysuski, Martin; Lalapet, Srini; Swaneveld, Hendrik, “De-
(2004a) termination of Intangible Property Ownership in Transfer
Pricing Analyses”, 33 Tax Notes Int’l (2004), p. 285.
Przysuski et al. Przysuski, Martin; Swaneveld, Hendrik; Lalapet, Srini, “Mul-
(2004b) tinational Enterprise Business Strategies and Transfer Pricing
in a Global Marketplace”, 33 Tax Notes Int’l (2004), p. 631.
Przysuski et al. Przysuski, Martin; Lalapet, Srini, “A Comprehensive Look
(2005) at the Berry Ratio in Transfer Pricing”, 40 Tax Notes Int’l
(2005), p. 759.
Purnell (1992) Purnell, E.A., “The Net Present Value Approach to Intangi-
ble Asset Pricing under Section 482: An Economic Model
Takes the BALRM Floor”, 45 Tax Lawyer (1992), p. 647.
PwC (2012) PricewaterhouseCoopers, “PWC Examines India Section of
U.N. Draft Manual On Developing Country Transfer Pri
cing”, 68 Tax Notes Int’l (2012), p. 379.
Quiquerez (2013) Quiquerez, Alexandre, “IP holding companies: Perfect match
or liaisons dangereuses between tax law and IP law? A cross-
border perspective”, available at http://atrip.org/wp-content/
uploads/2016/12/2013-3Alexandre-Quiquerez.pdf.
Rasch et al. (2011) Rasch, Stephan; Schmidtke, Richard, “OECD Guidelines on
Business Restructuring and German Transfer of Function Reg-
ulations: Do Both Jeopardize the Existing Arm’s Length Princi-
ple?”, 18 International Transfer Pricing Journal 1 (2011), p. 57.
Reams et al. Reams, Keith et al., “Proposed U.S. Cost-Sharing Regula-
(2005) tions: Are They a Realistic Alternative?”, 40 Tax Notes Int’l
(2005), p. 269.
Reyneveld et al. Reyneveld, Jaap; Bonekamp, Mark, “Recent Case Law on
(2012) Captive Insurance Companies: What Is the Right Transfer
Pricing Approach?”, 19 International Transfer Pricing Jour
nal 1 (2012), p. 93.
Rhee (2008) Rhee, Jennifer, “The Tax Court of Canada in GlaxoSmith
Kline: Drawing the Boundaries of the Arm’s-Length Fic-
tion”, 17 Transfer Pricing Report 16 (2008), p. 671.
Riedel et al. Riedel, Nadine et al., “Corporate Taxes and Strategic Pat-
(2015) ent Location within Multinational Firms” (2015), available at
https://www.cesifo-group.de/dms/ifodoc/docs/Akad_Conf/
CFP_CONF/CFP_CONF_2015/pse15-van-der-Ploeg/Papers/
pse15-Riedel.pdf.
800
References
Riedy et al. (2010) Riedy, James; Anwar, Nabeel; Wright, Deloris, “Pricing of a
Guarantee Fee in a Related-Party Context”, 17 International
Roberge (2013) Roberge, Chantal, “Transfer pricing in the Pharmaceutical
Industry: The remuneration of marketing intangibles”, 20 In
ternational Transfer Pricing Journal 4 (2013), p. 212.
Robillard (2017a) Robillard, Robert, “Profit-Split Methods and the OECD:
Leaning Toward Formulary Apportionment?”, 87 Tax Notes
Int’l (2017), p. 1005.
Robillard (2017b) Robillard, Robert, “‘Profits’ in Profit-Split Methods: Hazard-
ous Crossovers on the Way to Global Formulary Apportion-
ment”, 88 Tax Notes Int’l (2017), p. 167.
Rocha (2017) Rocha, Sergio André, “The future of transfer pricing – Gen-
eral Report”, Cahiers De Droit Fiscal International, Vol.
102b, Subject 2 (2017).
Roin (2008) Roin, Julie, “Can the Income Tax Be Saved? – The Promise
and Pitfalls of Adopting Worldwide Formulary Apportion-
ment”, 61 Tax Law Review 3 (2008), p. 169.
Roin (2011) Roin, Julie, “Transfer Pricing in the Courts: A Cross-Coun-
try Comparison”, in “Fundamentals of International Transfer
Pricing in Law and Economics” (Schön, Wolfgang, Konrad,
Kai A. eds. 2011), at p. 185.
Romano et al. Romano, Carlo; Russo, Raffaele, “Italian Tax Consequences
(2010) of Permanent Establishments”, 64 Bulletin for International
Taxation 8/9 (2010), p. 456.
Rosenbloom Rosenbloom, H. David, “The U.S. Approach to Transfer
(2006) Pricing: Benchmarks and Hallmarks”, 41 Tax Notes Int’l 4
(2006), p. 341.
Rosenbloom Rosenbloom, H. David, “Self-initiated transfer pricing ad-
(2007) justments”, 46 Tax Notes Int’l (2007), p. 1019.
Ross (1958) Ross, Alf, “On law and justice” (1958).
Ross et al. (1963) Ross, Stanford G.; Guttentag, Joseph H., “United States Tax-
ation of International Business Transactions”, in “A Lawyer’s
Guide to International Business Transactions” (Surrey, Wal-
ter S. ed., 1963).
Rotondaro (2000) Rotondaro, Carmine, “The Notion of ‘Associated Enterpri
ses’: Treaty Issues and Domestic Interpretations – An Over-
view”, 7 International Transfer Pricing Journal 1 (2000), p. 2.
Rozek et al. Rozek, Richard P.; Korenko, George G., “Transfer Prices for
(1999) the Intangible Property Embodied in Products with Extraor-
dinary Profit Potentials”, 19 Tax Notes Int’l (1999), p. 1553.
801
References
Rozek et al. Rozek, Richard P.; Korenko, George G., “Using In-Country
(2003) Comparables to Measure the Returns Due to Pharmaceuti-
cal Marketing and Distribution Affiliates”, 10 International
Russo (2005) Russo, Raffaele, “The Attribution of Profits to Permanent Es
tablishments: The taxation of intra-company dealings” (2005).
Russo et al. Russo, Raffaele; Pedrazzini, Edoardo, “Permanent Estab-
(2007) lishments under Italian Tax Law: An Overview”, 47 Euro
pean Taxation 8 (2007), p. 389.
Sadiq (2001) Sadiq, Kerrie, “Unitary Taxation – The Case for Global For-
mulary Apportionment”, 55 Bulletin for International Fiscal
Documentation 7 (2001), p. 275.
Samuel (2003) Samuel, Geoffrey, “Epistemology and method in law” (2003).
Samuel (2008) Samuel, Geoffrey, “Is law really a social science? A view
from comparative law”, 67 Cambridge Law Journal (2008),
pp. 288-321.
Sandgren (1995) Sandgren, Claes, “Om empiri och rättsvetenskap”, Juri
disk Tidsskrift, Del I och II (1995/1996), at pp. 726-748 and
pp. 1035-1057 respectively.
Schatan (2008) Schatan, Roberto, “Cross-border migration of intangibles:
Some tax policy considerations”, 34 International Tax Jour
nal (2008), p. 59.
Schmitz (2001) Schmitz, Patrick W., “The Hold-Up Problem and Incomplete
Contracts: A Survey of Recent Topics in Contract Theory”,
53 Bulletin of Economic Research 1 (2001), p. 1.
Schön (2007) Schön, Wolfgang, “Attribution of Profits to PEs and the
OECD 2006 Report”, 46 Tax Notes Int’l (2007), p. 1059.
Schön (2008) Schön, Wolfgang, “Losing out at the Snooker Table: Cross
Border Loss Compensation for PEs and the Fundamental
Freedoms”, in Hinnekens, Luc; Hinnekens, Philippe (eds.),
“A vision of Taxes within and outside European Borders”,
Festschrift in Honor of Prof. Dr. Frans Vanistendael (2008),
p. 813.
Schön (2010a) Schön, Wolfgang, “International Tax Coordination for a Sec-
ond-Best World (Part III)”, 2 World Tax Journal 3 (2010),
p. 227.
Schön (2010b) Schön, Wolfgang, “Persons and Territories: On the Interna-
tional Allocation of Taxing Rights”, British Tax Review 6
(2010), p. 554.
802
References
Schön (2011a) Schön, Wolfgang, “Transfer Pricing, the Arm’s Length

Standard and European Union Law”, Working Paper 2011-
08, Max Planck Institute for Tax Law and Public Finance
(2011).
Schön (2011b) Schön, Wolfgang, “Transfer Pricing – Business Incentives,
International Taxation and Corporate Law”, in “Fundamen-
tals of International Transfer Pricing in Law and Econom-
ics”, MPI Studies in Tax Law and Public Finance 1 (Schön,
Wolfgang; Konrad, Kai A. eds. 2011), p. 47.
Schön (2014) Schön, Wolfgang, “International Taxation of Risk”, Max
Planck Institute for Tax Law and Public Finance, Working
Paper 2014-03 (2014).
Schön et al. (2011) Schön, Wolfgang; Konrad, Kai A. (eds.), “Fundamentals of
international transfer pricing in law and economics”, MPI
Studies in Tax Law and Public Finance1 (2012).
Schoueri (2015) Schoueri, Luis Eduardo, “Arm’s Length: Beyond the Guide-
lines of the OECD”, 69 Bulletin for International Taxation
12 (2015), p. 690.
Schoueri et al. Schoueri, Luis Eduardo; Günther, Oliver-Christoph, “The
(2011) subsidiary as a permanent establishment”, 65 Bulletin for In
ternational Taxation 2 (2011), p. 69.
Schwarz (2015) Schwarz, Jonathan, “Are Legal Rights Irrelevant for Trans-
fer Pricing after BEPS Actions 8, 9, and 10?”, Kluwer In
ternational Tax Blog (2015), available at http://kluwer-
taxblog.com/2015/11/15/are-legal-rights-irrelevant-for-trans-
fer-pricing-after-beps-actions-8-9-and-10/.
Sharma (2008) Sharma, Sanjeev, “Attribution to Profits to Permanent Estab-
lishments – The Indian Experience”, 14 Asia-Pacific Tax Bul
letin 2 (2008), p. 93.
Sheppard (2011) Sheppard, Lee A., “Intangibles Migration From the United
States and Excess Profits”, 61 Tax Notes Int’l (2011), p. 983.
Sheppard (2015) Sheppard, Lee A., “Foreign Goodwill in Outbound Trans-
fers”, 80 Tax Notes Int’l (2015), p. 886.
Sheppard et al. Sheppard, Lee A.; Ayayo, Herman P., “ABA Tax Section
(2006) Meeting: IRS Reassures Practitioners on Cost-Sharing
Regs”, Tax Notes Int’l (2006), p. 537.
SIA (2004) Securities Industry Association, “Comments on Treatment of
Services Under Section 482 and Related Matters”, available
at Tax Analysts (www.taxanalysts.com); Doc 2004-431.
Sinclair (1985) Sinclair, M.B.W., “Law and language: The Role of Pragmat-
ics in Statutory Interpretation”, 46 U. Pitt. L. Rev. (1985).
803
References
Skaar (1991) Skaar, Arvid Aage, “Permanent establishment – erosion of a

tax treaty principle” (1991).
Sprague (2012) Sprague, Gary D., “The Italian Supreme Court weighs in on
commissionaire structures – is the fat lady singing yet?”, 41
Tax Management International Journal 7 (2012).
Steeves (2011) Steeves, Christopher J., “Business Restructurings: Recent
Changes to the OECD Transfer Pricing Guidelines”, 23 Ca
nadian Tax Journal 1 (2011), p. 151.
Stevens et al. Stevens, Elizabeth J.; Rosenbloom, H. David, “GILTI Pleas-
(2018) ures”, 89 Tax Notes Int’l (2018), p. 615.
Stewart (2009) Stewart, David D., “U.S. Issues Final Service Regs”, Tax
Notes Int’l (2009), p. 418.
Storck et al. Storck, Alfred et al., “Global Transfer Pricing Conference
(2016) ‘Transfer Pricing in a Post-BEPS World’”, 23 International
Surrey (1968) Surrey, Stanley S., “Treasury’s Need to Curb Tax Avoidance
in Foreign Business Through Use of 482”, 28 J. Tax (1968),
p. 75.
Swaneveld et al. Swaneveld, Hendrik et al., “Transfer Pricing and Characteri-
(2004) zation of Multinational Enterprise Operations with a Focus
on Canada and the U.S.”, 34 Tax Notes Int’l (2004), p. 743.
Taly (1996) Taly, Michel, “Comparison of CPM and TNMM Transfer
Pricing Methods: A Point of View”, 12 Tax Notes Int’l (1996),
p. 351.
Tan (2001) Tan, How Teck, “Singapore’s tax treaty policy”, 29 Intertax
5 (2001), p. 184.
Tavares (2016) Tavares, Romero J.S., “Multinational Firm Theory and Inter-
national Tax Law: Seeking Coherence”, 8 World Tax Journal
2 (2016).
TEI (2003) Tax Executives Institute, “Comments on the Proposed Regu-
lations Relating to the Treatment of Services Under Section
482 and the Allocation of Income and Deductions From In-
tangibles”, available at Tax Analysts (www.taxanalysts.com);
Doc 2003-26985.
TEI (2004a) Tax Executives Institute, “Comments on Proposed Regula-
tions Relating to the Treatment of Services Under Code Sec-
tion 482 and the Allocation of Income and Deductions From In-
tangibles”, available at Tax Analysts (www.taxanalysts.com);
Doc 2004-13800.
804
References
TEI (2004b) Tax Executives Institute, “Comments on List of Services for

Cost Safe Harbor Under IRS Rules”, 13 BNA Transfer Pri
cing Rep. (2004), p. 215.
TEI (2004c) Tax Executives Institute, “Comments on Proposed Section
482 Services Rules”, 12 BNA Transfer Pricing Rep. (2004),
p. 743.
Terr (2004) Terr, Leonard B., “Proposed U.S. Transfer Pricing Services
Regulations”, 33 Tax Notes Int’l (2004), p. 547.
Terr (2006) Terr, Leonard B., “Temporary U.S. Transfer Pricing Services
Regulations”, 44 Tax Notes Int’l (2006), p. 861.
Thomas et al. Thomas, Mark P.; Novak, Kristina L.; Lowell, Cym H., “Evo-
(2018) lution of Advance Pricing Agreement Processes: Current and
Future Experience in the United States”, 25 International
Transfer Pricing Journal 2 (2018).
Ting (2014) Ting, Antony, “iTax – Apple’s International Tax Structure
and the Double Non-Taxation Issue”, 1 British Tax Review
40 (2014).
TMTPR (1992) Comments by officials of the Japanese Ministry of Finance
and National Tax Administration, in Tax Management Trans
fer Pricing Report (1992).
Toro (2009) Toro, Carlos Eduardo Costa M.A., “Different Methods of At-
tributing Profits to Agency PEs”, 53 Tax Notes Int’l (2009),
p. 421.
Treidler et al. Treidler, Oliver; Jung, Samuel, “Economic Valuation Tech-
(2017) niques for Transfer Pricing”, 85 Tax Notes Int’l (2017), p. 561.
Turro (1990) Turro, John, “U.S. Congressional Committee Blasts Foreign
Firms for Tax Dodging”, Tax Notes Int’l (1990), p. 797.
Turro (1992) Turro, John, “Toyota denies recent report of transfer-pricing
settlement”, Tax Notes Int’l (1992), p. 1379.
Turro (1994) Turro, John, “The Battle Over Arm’s Length and Formulary
Apportionment”, 65 Tax Notes Int’l (1994), p. 1595.
Ungerman (1989) Ungerman, Josh O., “The White Paper: The Stealth Bomb-
er of the Section 482 Arsenal”, 42 Southwest Law Journal
(1989), p. 1107.
Van Breda et al. Breda, Michael F. van; Hendriksen, Eldon S., “Accounting
(1992) theory”, 5th ed. (1992).
Vann (2003) Vann, Richard, “Reflections on Business Profits and the
Arm’s-Length Principle”, in “The Taxation of Business Prof-
its Under Tax Treaties” (Arnold, Brian J.; Sasseville, Jacques;
Zolt, Eric M. eds. 2003), p. 133.
805
References
Vann (2006) Vann, Richard, “Tax Treaties: The Secret Agent’s Secrets”,
British Tax Review 3 (50th Anniversary Edition) (2006),
p. 345.
Vann (2010) Vann, Richard, “Taxing International Business Income:
Hard-Boiled Wonderland or the End of the World”, 2 World
Tax Journal 3 (2010), p. 291.
Vega (2012) Vega, Alberto, “International governance through soft law:
The case of the OECD transfer pricing guidelines”, TranState
Working Papers, no. 163 (2012).
Velarde (2015) Velarde, Andrew, “Treasury Outlines Road Not Taken in
Foreign Goodwill Regs”, 80 Tax Notes Int’l (2015), p. 327.
Velarde (2016) Velarde, Andrew, “Routine Return Foreign Goodwill Safe
Harbor Not Off the Table”, 81 Tax Notes Int’l (2016), p. 236.
Verbeek (2005) Verbeek, Patricia, “GlaxoSmithKline and the IRS: Will $8
Billion Change Hands?”, Tax Notes Int’l (2005), p. 330.
Verdoner (2011) Verdoner, Louan, “The Concept of Dependent Agent Perma-
nent Establishment in Transfer Pricing Theory”, 18 Interna
tional Transfer Pricing Journal 2 (2011), p. 128.
Verlinden et al. Verlinden, Isabel; Mondelaers, Yoko, “Transfer Pricing As-
(2010) pects of Intangibles: At the Crossroads between Legal, Valu-
ation and Transfer Pricing Issues”, 17 International Transfer
Vidal (2002) Vidal, Jean-Pierre, “Taking Imperfect Competition, Returns
to Scale Into Account in Performing Transfer Pricing Analy-
ses”, 11 Tax Management Transfer Pricing Report 9 (2002),
p. 413.
Vidal (2009) Vidal, Jean-Pierre, “The Achilles’ Heel of the Arm’s Length
Principle and the Canadian GlaxoSmithKline Case”, 37 Inter
tax 10 (2009), p. 512.
Vincent (2005) Vincent, François, “Transfer Pricing and Attribution of In-
come to Permanent Establishments: The Case for Systematic
Global Profit Splits (Just Don’t Say Formulary Apportion-
ment)”, 53 Canadian Tax Journal 2 (2005), p. 409.
Vincent (2006) Vincent, François, “Marketing Intangibles in Canada: Myth
or Reality?”, in “Report of Proceedings of Fifty-Seventh Tax
Conference”, 2005 Conference Report Canadian Tax Foun-
dation (2006), at 28:1-13.
Vincent (2008) Vincent, François, “GlaxoSmithKline Inc. v. The Queen: A
Commentary”, 17 Transfer Pricing Report 4 (2008), p. 177.
806
References
Vogel et al. (1993) Vogel, Klaus; Prokisch, Rainer G., “General report – Inter-
pretation of double taxation conventions”, Cahiers De Droit
Fiscal International, Vol. LXXVIIIa (1993).
Ward (2010) Ward, David A., “Commentary, General Electric Capital Can
ada Inc. v. Her Majesty The Queen”, 12 ITLR (2010), at p. 509.
Warner (2004) Warner, John P., “Too Much Ado About Something – The
Proposed Regulations Governing Controlled Services Trans-
actions”, 33 Tax Mgmt. Int’l J. 67 (2004).
Wartzman (1993) Wartzman, Rich, “Foreign Firm’s Income in U.S. is Under-
stated: IRS Briefing Papers Show Japan, U.K., and Canada
Top List of 10 Nations”, Wall Street Journal (4 June 1993),
at A-2.
Weber et al. Weber, Dennis; Weeghel, Stef van (eds.), “The 2010 OECD
(2011) Updates – Model Tax Convention & Transfer Pricing Guide-
lines: A critical review” (2011).
Weisbach (2004) Weisbach, David A., “The (Non)Taxation of Risk”, 58 Tax Law
Review (2004), p. 1.
White (2016) White, Sienna C., “Cost Sharing Agreements & The Arm’s
Length Standard: A Matter of Statutory Interpretation?”, 19
Florida Tax Review 3 (2016), p. 191.
Wilkie (2011) Wilkie, J. Scott, “Reflecting on the ‘Arm’s Length Principle’:
What is the ‘Principle’? Where Next?”, in “Fundamentals of
International Transfer Pricing in Law and Economics”, MPI
Studies in Tax Law and Public Finance 1 (Schön, Wolfgang;
Konrad, Kai A. eds. 2011), p. 137.
Wilkie (2012) Wilkie, J. Scott, “The Definition and Ownership of Intangi-
bles: Inside the Box? Outside the Box? What is the Box?”, 4
World Tax Journal 3 (2012), p. 222.
Wilkie (2014a) Wilkie, J. Scott, “Intangibles and Location Benefits (Custom-
er Base)”, 68 Bulletin for International Taxation 6/7 (2014),
p. 352.
Wilkie (2014b) Wilkie, J. Scott, “Policy Forum: BEPS One Year In – Taking
Stock”, 62 Canadian Tax Journal 2 (2014), p. 455.
Wilkie (2016) Wilkie, J. Scott, “Transfer Pricing Aspects of Intangibles: The
License Model”, in “Transfer Pricing in a Post-BEPS World”
(Lang, Michael; Storck, Alfred; Petruzzi, Raffaele eds., 2016).
Wittendorff Wittendorff, Jens, “Transfer Pricing and the Arm’s Length
(2010a) Principle in International Tax Law” (2010).
Wittendorff Wittendorff, Jens, “Valuation of intangibles under income-
(2010b) based methods – Part I”, 17 International Transfer Pricing
Journal 5 (2010), p. 323.
807
References
Wittendorff Wittendorff, Jens, “Valuation of intangibles under income-

(2010c) based methods – Part II”, 17 International Transfer Pricing
Journal 6 (2010), p. 383.
Wittendorff Wittendorff, Jens, “International skatteret” 2010, S R – Skat.
(2010d) (2010).
Wittendorff Wittendorff, Jens, “U.N. Transfer Pricing Manual – The
(2012a) Choice Between International Consistency and Conflict”, 68
Tax Notes Int’l (2012), p. 569.
Wittendorff Wittendorff, Jens, “Darkness Descends on Danish Tax Law:
(2012b) Supreme Court Decision in Swiss Re”, 19 International Trans
fer Pricing Journal 3 (2012), p. 217.
Wittendorff Wittendorff, Jens, “Transfer Pricing Timing Issues Revis-
(2012c) ited”, 68 Tax Notes Int’l (2012), p. 87.
Wittendorff Wittendorff, Jens, “‘Shadowlands’: The OECD on Intangi-
(2012d) bles”, 67 Tax Notes Int’l (2012), p. 935.
Wittendorff Wittendorff, Jens, “Triangular Cases: The Interaction Be-
(2012e) tween Transfer Pricing and PEs”, 66 Tax Notes Int’l (2012),
p. 545.
Wittendorff Wittendorff, Jens, “Aggregation of Transactions in Transfer
(2013) Pricing: Glaxo and Other Cases”, 69 Tax Notes Int’l (2013),
p. 183.
Wittendorff Wittendorff, Jens, “BEPS Actions 8-10: Birth of a New
(2016) Arm’s-Length Principle”, 81 Tax Notes Int’l (2016), p. 331.
Wolosoff et al. Wolosoff, Todd; Ryan, James; Eisenhauer, Susan, “The pro-
(2004) posed Section 482 Regs: A path to confusion, uncertainty,
and endless controversy”, 34 Tax Notes Int’l (2004), p. 625.
Wu (2005) Wu, Tim, “Intellectual Property, Innovation, and Decentral-
ized Decisions”, 92 Virginia Law Review 1 (2005), p. 101.
Yamakawa (2007) Yamakawa, Hiroki, “A Practical Analysis of Transfer Pricing
Methodologies for Bilateral Advance Pricing Arrangements”
(2007), Part 250, available at https://core.ac.uk/download/
pdf/27288414.pdf.
Yoost (1996) Yoost, Dean A., “Dealing with Japan’s tax administration”,
13 Tax Notes Int’l (1996), p. 1837.
Zammit (2015) Zammit, Ivan, “Centralized Intellectual Property Business
Models – Tax Implications of EU Patent Box Regimes”, 69
Bulletin for International Taxation 9 (2015), p. 540.
808
References
Zollo (2010) Zollo, Thomas M., “Clarification or Modification? The Tax

Treatment of the Outbound Transfer of Goodwill, Going
Concern Value, and Workforce in Place to a Foreign Corpo-
ration”, 39 Tax Mgmt. Int’l J. 2 (2010).
Zollo (2011) Zollo, Thomas M., US Chapter in “Cross-border business re-
structuring”, Cahiers de droit fiscal international, Vol. 96a
(2011), p. 757.
Zuurbier (2016) Zuurbier, Niels, “Transfer Pricing in a Post-BEPS World –
Transfer Pricing Aspects of Intangibles” (2016), available at
http://arno.uvt.nl/show.cgi?fid=141892.
Case law
Australia
Roche Products Pty Ltd v. FCT (2008) AATA 639; 2008 ATC 10-036.
Canada
GlaxoSmithKline Inc. v. R., 2008 TCC 324 (T.C.C., 2008), reversed by 2010 CAF
201 (F.C.A., 2010), application/notice of appeal filed in 2010 CarswellNat 4089
(S.C.C., 2010), and leave to appeal allowed by 2011 CarswellNat 682 (S.C.C.,
2011), and affirmed by 2012 SCC 52 (S.C.C., 2012).
GSW Appliances Ltd. v. R., 1985 CarswellNat 346 (T.C.C., 1985).
Koffler Stores Ltd. v. R., 1975 CarswellNat 336 (Fed. T.D., 1975), affirmed by 1976
CarswellNat 200 (Fed. C.A., 1976).
McClurg v. R., 1984 CarswellNat 369 (T.C.C., 1984), reversed by 1986 Carswell-
Nat 244 (Fed. T.D., 1986), affirmed by 1987 CarswellNat 556 (Fed. C.A., 1987),
leave to appeal allowed by 1988WL876606 (S.C.C., 1988), and affirmed by 1990
CarswellNat 520 (S.C.C., 1990).
Shell Canada Ltd. v. R., 1997 CarswellNat 401 (T.C.C., 1997), reversed by 1998
CarswellNat 170 (Fed. C.A., 1998), leave to appeal allowed by 1998WL1727462
(S.C.C., 1998), and leave to appeal allowed by 1999WL33183768 (S.C.C., 1999),
and reversed by 1999 CarswellNat 951 (S.C.C., 1999), additional reasons in 1999
CarswellNat 1808 (S.C.C., 1999).
809
References
Singleton v. R., 1996 CarswellNat 1816 (T.C.C., 1996), reversed by 1999 Carswell-
Nat 1009 (Fed. C.A., 1999), leave to appeal allowed by 2000 CarswellNat 653
(S.C.C., 2000), and affirmed by 2001 SCC 61 (S.C.C., 2001).
Alberta Printed Circuits Ltd. v. The Queen (2011 TCC 232).
Denmark
H1 A/S v. Skatteministeriet, SKM2010.455.VLR (Vestre Landsret, 2010).
H1.1.1 A/S v. Skatteministeriet, SKM2012.92.HR (Supreme Court of Denmark,

2012). See also lower court rulings in SKM2010.134.ØLR (Østre Landsret, 2010),
and SKM2008.975.LSR (Landsskatteretten, 2008).
France
Société Zimmer Limited v. Ministre de l’Économie, des Finances et de l’Industrie,
Nos. 304715 and 308525 (Supreme Administrative Court, 2010); reversing No. 05-
2361 (Administrative Court of Appeals of Paris, 2007).
India
Ahmed Bhain Umar Bhai & Co. v. CIT, 18 ITR 472 (Supreme Court of India,
1950).
Amadeus Global Travel Distribution v. CIT, 113 TTJ Delhi 767 (Income Tax Ap-
pellate Tribunal – Delhi, 2007).
Amadeus India Pvt Ltd v. CIT, ITA 938/2011 (Delhi High Court, 2011).
Annamalais Timber Trust And Co. v. CIT, 1961 41 ITR 781 Mad (Madras High
Court, 1960).
Aztec Software & Technology Services Limited v. CIT, 2007 107 ITD 141 Bang
(Income Tax Appellate Tribunal – Bangalore, 2007).
Global Vantedge Pvt. Ltd. v. CIT, ITA Nos.116 & 323/Del/2011 (Income Tax Ap-
pellate Tribunal – Delhi, 2010).
Intel Asia Electronics Inc. v. Asst Director Of Income Tax, ITA No. 131/Bang/2010
(Income Tax Appellate Tribunal – Bangalore, 2009).
810
References
M/S. Carborandum Co v. CIT, 1977 AIR 1259 (Supreme Court of India, 1977).
Maruti Suzuki India Ltd v. CIT, W.P.(C) 6876/2008 (Delhi High Court, 2010).
Rolls Royce Plc v. Dy. Director Of Income-Tax, (2008) 113 TTJ Delhi 446 (Income
Tax Appellate Tribunal – Delhi, 2007).
Sony India (P.) Ltd. v. CBDT, (2006) 206 CTR Del 157 (Delhi High Court, 2006).
Symantec Software Solutions Pvt Ltd v. CIT, ITA No. 7894 (Income Tax Appellate
Tribunal – Mumbai, 2010).
UCB India (P) Ltd. v. CIT, 30 SOT 95 (Income Tax Appellate Tribunal – Mumbai,
2009).
Italy
Ministry of Finance (Tax Office) v. Boston Scientific, Case No. 3769 (Supreme
Court of Italy, 2012).
Ministry of Finance (Tax Office) v. ITCO, Case No. 11949 (Supreme Court of Italy,
2012).
Ministry of Finance (Tax Office) v. Philip Morris (GmbH), Case No. 7682/05 (Su-
preme Court of Italy, 2002).
Norway
Dell Products v. the State, Utv. 2012 s. 1 (Norwegian Supreme Court, 2011), re-
versing Utv. 2011 p. 807 (Borgarting Appellant Court, 2011).
Norsk Agip AS v. the State, Rt. 2001 p. 1265 (Norwegian Supreme Court, 2001).
Norske Conocophillips AS and ConocoPhillips Skandinavia AS vs. the State, Utv.

2008 p. 889 (Oslo City Court, 2008).
Tore Rynning-Nielsen, Sverre Morten Blix, Gert Wilhelm Munthe and Herkules
Capital I AS v. the State, Rt. 2015 p. 1260 (Norwegian Supreme Court, 2015).
Vingcard Elsafe AS v. the State, Utv. 2012 s. 1191, (Borgarting Appellant Court,
2012), reversing in part Utv. 2010 p. 1690 (Oslo City Court, 2010).
811
References
Spain
Roche Vitamins Europe Ltd. v. Administración General del Estado, Case No.
STS/202/2012 (Supreme Court of Spain, 2012).
United States
Introduction
The categorization of US case law is based on which first-tier court assessed the
case. The author has included references to subsequent appeal rulings.
Using this system, the author avoids “double counting” in the sense that the first
and second-tier court assessments of the same case are not listed twice (once un-
der the first-tier court and once under the second-tier court). The author has only
listed a few “outlier” cases separately under the US Appellant Court (e.g. a case
brought directly from the US Board of Tax Appeals to the Appellant Court).
The author has also included references to the court documents referred to in the
book pertaining to the settled case GlaxoSmithKline Holdings (Americas), Inc. v.
CIR, 117 TC 1.
US Tax Court
Altama Delta Corp. v. CIR, 104 T.C. 424 (Tax Ct., 1995).
Amazon.com Inc. v. CIR, Tax Court Docket No. 31197-12.
Bausch & Lomb Inc. v. CIR, 92 T.C. No. 33 (Tax Ct., 1989), affirmed by 933 F.2d
1084 (2nd Cir., 1991).
Bergersen v. CIR, T.C. Memo. 1995-424 (Tax Ct., 1995), affirmed by 109 F.3d 56
(1st Cir., 1997).
Black Industries v. CIR, 38 T.C.M. 242 (Tax Ct., 1979).
BMC Software Inc. v. CIR, 141 T.C. No. 5 (Tax Ct., 2013), reversed by 780 F.3d
669 (5th Cir., 2015).
Boe v. CIR, 35 T.C. 720 (Tax Ct., 1961), affirmed by 307 F.2d 339 (9th Cir., 1962).
812
References
Boston Scientific Corporation & Subsidiaries v. CIR, Tax Court Docket No.
26876-11.
Cardiac Pacemakers, Inc. v. CIR, Tax Court Docket No. 5990-11.
Charles Schwab Corp. v. CIR, 122 T.C. No. 10 (Tax Ct., 2004), supplemented,
clarified and superseded in part by 123 T.C. No. 18 (Tax Ct., 2004), affirmed by
495 F.3d 1115 (9th Cir., 2007).
Ciba-Geigy Corp. v. CIR, 85 T.C. No. 11 (Tax Ct., 1985), acquiescence recom-
mended by AOD-1987-21 (IRS AOD, 1987), and acquiescence 1987 WL 857882
(IRS ACQ, 1987).
Citizens and Southern Corp. v. CIR, 91 T.C. No. 35 (Tax Ct., 1988), affirmed by
900 F.2d 266 (11th Cir., 1990), published in full at 919 F.2d 1492 (11th Cir., 1990).
Coca-Cola Co. and Includible Subsidiaries v. CIR, 106 T.C. 1 (Tax Ct., 1996).
Computing & Software, Inc. v. CIR, 64 T.C. 223 (Tax Ct., 1975), acquiescence
1976 WL 175506 (IRS ACQ, 1976).
DHL Corp. & Subsidiaries v. CIR, T.C. Memo. 1998-461 (Tax Ct., 1998), affirmed
in part, reversed in part by 285 F.3d 1210 (9th Cir., 2002).
Differential Steel Car Co. v. CIR, 16 T.C. 41388 (Tax Ct., 1951).
Dittler Brothers, Inc. v. CIR, 72 T.C. 896 (Tax Ct., 1979), affirmed by 642 F.2d
1211 (5th Cir.(Ga.), 1981).
Eaton Corporation and Subsidiaries v. CIR, Tax Court Docket No. 5576-12.
Edwards v. CIR, 67 T.C. 224 (Tax Ct., 1976), acquiescence recommended by

AOD-1977-155 (IRS AOD, 1977), and acquiescence, 1977 WL 185611 (IRS ACQ,
1977), and acquiescence, 1977 WL 185630 (IRS ACQ, 1977).
Electronic Arts, Inc. v. CIR, 118 T.C. 226 (Tax Ct., 2002).
Eli Lilly and Company and Subsidiaries v. CIR, 84 T.C. No. 65 (Tax Ct., 1985),
affirmed in part, reversed in part by 856 F.2d 855 (7th Cir., 1988).
Essex Broadcasters, Inc. v. CIR, 2 T.C. 523 (Tax Ct., 1943).
Exxon Corp. v. CIR, T.C. Memo. 1993-616 (Tax Ct., 1993), affirmed by 98 F.3d 825
(5th Cir., 1996), certiorari denied by 520 U.S. 1185 (S.Ct., 1997).
First Data Corp. v. CIR, Tax Court Docket No. 007042-09, case settled.
813
References
G.D. Searle & Co. v. CIR, 88 T.C. 252 (Tax Ct., 1987).
GlaxoSmithKline Holdings (Americas), Inc. v. CIR, 117 TC 1 (Tax Ct., Docket No.
5750-04, 2004), settled 2006.
Granberg Equipment, Inc. v. CIR, 11 T.C. 70479 (Tax Ct., 1948).
Guidant LLC v. CIR, Tax Court Docket No. 5989-11.
Hospital Corporation of America v. CIR, 81 T.C. No. 31 (Tax Ct., 1983), nonacqui-
escence recommended by AOD-1987-22 (IRS AOD, 1987), and nonacquiescence
1987 WL 857897 (IRS ACQ, 1987).
InverWorld, Inc. v. CIR, T.C. Memo. 1996-301 (Tax Ct., 1996).
Ithaca Industries, Inc. v. CIR, 97 T.C. No. 16 (Tax Ct., 1991), affirmed by 17 F.3d
684 (4th Cir., 1994), certiorari denied by 513 U.S. 821 (S.Ct.,1994).
Kraft Foods Company v. CIR, 21 TC 513 (Tax Ct., 1954).
Lufkin Foundry and Mach. Co. v. CIR, T.C. Memo. 1971-101 (Tax Ct., 1971), re-
versed by 468 F.2d 805 (5th Cir., 1972), and acquiescence in part and nonacquies-
cence in part recommended by 1974 WL 36323 (IRS AOD, 1974).
Massey-Ferguson, Inc. v. CIR, 59 T.C. 220 (Tax Ct., 1972), acquiescence, 1973
WL 157476 (IRS ACQ, 1973).
Medchem (P.R.), Inc. v. CIR, 116 T.C. 308 (Tax Ct., 2001), affirmed by 295 F.3d
118 (1st Cir., 2002).
Medtronic, Inc. & Consolidated Subsidiaries v. CIR, (Tax Court Docket No. 6944-
11).
Microsoft Corp. v. CIR, T.C. Memo. 1998-54 (Tax Ct., 1998).
National Securities Corp. v. CIR, 46 B.T.A. 562 (B.T.A., 1942), affirmed by 137
F.2d 600 (C.C.A.3, 1943), certiorari denied by 320 U.S. 794 (S.Ct., 1943).
Nestlé Co., Inc. v. CIR, T.C. Memo. 1963-14 (Tax Ct., 1963).
Nissho Iwai American Corp. v. CIR, T.C. Memo. 1985-578 (Tax Ct., 1985), af-
firmed by 812 F.2d 712 (2nd Cir., 1987).
Perkin-Elmer Corp. and Subsidiaries v. CIR, T.C. Memo. 1993-414 (Tax Ct.,
1993).
814
References
Philip Morris Inc. v. CIR, 96 T.C. No. 23 (Tax Ct., 1991), affirmed by 970 F.2d 897
(2nd Cir., 1992).
Polak’s Frutal Works, Inc. v. CIR, 21 T.C. 953 (Tax Ct., 1954), acquiescence, 1955
WL 45485 (IRS ACQ, 1955), and acquiescence withdrawn, nonacquiescence.,
1972 WL 124383 (IRS ACQ, 1972).
PPG Industries Inc. v. CIR, 55 T.C. 928 (Tax Ct., 1970).
R. T. French Co. v. CIR, 60 T.C. 836 (Tax Ct., 1973), distinguished by 1977 WL
191022 (IRS TAM, 1977), and distinguished by 1979 WL 56002 (IRS TAM,
1979), and distinguished by 1992 WL 1354859 (IRS FSA, 1992).
Rollman v. CIR, 25 T.C. 481 (Tax Ct., 1955), acquiescence, 1956 WL 56488 (IRS
ACQ, 1956), reversed by 244 F.2d 634 (4th Cir., 1957), on remand to T.C. Memo.
1957-182 (Tax Ct., 1957).
Ross Glove Co. v. CIR, 60 T.C. 569 (Tax Ct., 1973), acquiescence in part recom-
mended by 1974 WL 36011 (IRS AOD, 1974), and acquiescence, 1974 WL 172643
(IRS ACQ, 1974).
Seagate Technology, Inc. v. CIR, 102 T.C. No. 9 (Tax Ct., 1994), acquiescence in
result only recommended by AOD-1995-09 (IRS AOD, 1995), and acquiescence,
1995-33 I.R.B. 4 (IRS ACQ, 1995).
Seminole Flavor Co. v. CIR, 4 T.C. 1215 (Tax Ct., 1945), nonacquiescence rec-
ommended by 1972 WL 32906 (IRS AOD, 1972), and acquiescence withdrawn,
nonacquiescence, 1972 WL 124385 (IRS ACQ, 1972).
South Texas Rice Warehouse Co. v. CIR, 43 T.C. 540 (Tax Ct., 1965), affirmed
by 366 F.2d 890 (5th Cir., 1966), certiorari denied by 386 U.S. 1016 (S.Ct., 1967).
Southern Bancorporation v. CIR, 67 T.C. 1022 (Tax Ct., 1977).
Sundstrand Corp. and Subsidiaries v. CIR, 96 T.C. No. 12 (Tax Ct., 1991).
Taisei Fire and Marine Ins. Co., Ltd. v. CIR, 104 T.C. No. 27 (Tax Ct., 1995), ac-
quiescence, 1995-44 I.R.B. 4 (IRS ACQ, 1995), and acquiescence recommended
by AOD-1995-12 (IRS AOD, 1995).
U.S. Steel Corp. v. CIR, T.C. Memo. 1977-140 (Tax Ct., 1977), reversed by 617
F.2d 942 (2nd Cir., 1980), and nonacquiescence recommended by AOD-1980-179.
Veritas Software Corporation & Subsidiaries v. CIR, 133 T.C. No. 14 (U.S. Tax
Ct. 2009), IRS nonacquiescence in AOD-2010-05.
815
References
VGS Corp. v. CIR, 68 T.C. 563 (Tax Ct., 1977), acquiescence recommended by
AOD-1978-186 (IRS AOD, 1978), and acquiescence in 1979 WL 194041 (IRS
ACQ, 1979).
Xilinx Inc. and Subsidiaries v. CIR, 125 T.C. No. 4 (Tax Ct., 2005), affirmed by
598 F.3d 1191 (9th Cir., 2010), recommendation regarding acquiescence AOD-
2010-03 (IRS AOD, 2010), and acquiescence in result, 2010-33 I.R.B. 240 (IRS
ACQ, 2010).
US Court of Claims
Bell Intercontinental Corp. v. US, 1967 WL 156523 (Ct.Cl., 1967), report and rec-
ommendation Adopted by 180 Ct.Cl. 1071 (Ct.Cl., 1967).
E.I. Du Pont de Nemours and Co. v. US, 1978 WL 3449 (Cl.Ct., 1978), adopted by
221 Ct.Cl. 333 (Ct.Cl., 1979), certiorari denied by 445 U.S. 962 (S.Ct., 1980), and
judgment entered by 226 Ct.Cl. 720 (Ct.Cl., 1980).
E.I. Du Pont de Nemours and Co. v. US, 296 F.Supp. 823, (D.Del., 1969), affirmed
in part, reversed in part by 432 F.2d 1052 (3rd Cir., 1970).
Eli Lilly & Co. v. US, 372 F.2d 990 (Cl.Ct, 1967).
Hooker Chem. & Plastics Corp. v. US, 1978 WL 21534 (Ct.Cl. Trial Div., 1978),
affirmed by 219 Ct.Cl. 161 (Ct.Cl., 1979), supplemented by 221 Ct.Cl. 988 (Ct.Cl.,
1979).
Keene Corp. v. US, 17 Cl.Ct. 146 (Cl.Ct., 1989), reversed by 911 F.2d 654 (Fed.Cir.,
1990), modified on rehearing by 920 F.2d 916 (Fed.Cir., 1990) and opinion with-
drawn on grant of rehearing by 926 F.2d 1109 (Fed.Cir., 1990), and on rehearing
962 F.2d 1013 (Fed.Cir., 1992), certiorari granted by 506 U.S. 939 (U.S., 1992) and
judgment affirmed by 508 U.S. 200 (U.S., 1993).
Merck & Co., Inc. v. US, 24 Cl.Ct. 73 (Cl.Ct., 1991).
National Westminster Bank, PLC v. US, 58 Fed.Cl. 491, 92 A.F.T.R.2d 2003-7013,

2004-1 USTC P 50,105 (Fed.Cl. 14 November 2003) (NO. 95-758T), affirmed,
National Westminster Bank, PLC v. U.S., 512 F.3d 1347, 101 A.F.T.R.2d 2008-490,
2008-1 USTC P 50,140 (Fed.Cir. 15 January 2008) (NO. 2007-5028), rehearing en
banc denied (21 April 2008).
Richard S. Miller & Sons, Inc. v. US, 537 F.2d 446 (Ct.cl., 1976).
816
References
US District Court
Bank of America v. US, 1978 WL 1257 (N.D. Cal., 1979).
Cotton Ginny, Ltd. v. Cotton Gin, Inc., 691 F.Supp. 1347 (S.D.Fla., 1988).
Frank v. International Canadian Corp., 1961 WL 12345 (W.D.Wash., 1961), af-

firmed by 308 F.2d 520 (9th Cir., 1962).
Houston Chronicle Pub. Co. v. US, 339 F.Supp. 1314 (S.D.Tex., 1972), affirmed
by 481 F.2d 1240 (5th Cir., 1973), certiorari denied by 414 U.S. 1129 (S.Ct., 1974).
Massey Motors, Inc. v. US, 156 F.Supp. 516 (S.D.Fla., 1957), reversed by 264 F.2d
552 (5th Cir., 1959), and affirmed by 364 U.S. 92 (S.Ct., 1960).
Metropolitan Nat. Bank v. St. Louis Dispatch Co., 36 F. 722 (C.C.E.D.Mo., 1888),
affirmed by 149 U.S. 436 (S.Ct., 1893).
Newark Morning Ledger Co. v. US, 734 F.Supp. 176 (D.N.J., 1990), reversed by
945 F.2d 555 (3rd Cir., 1991), reversed by 507 U.S. 546 (S.Ct., 1993).
Northwestern Nat. Bank of Minneapolis v. US, 1976 WL 1016 (D.Minn. 1976),

affirmed by 556 F.2d 889 (8th Cir.[Minn.], 1977).
Stanfield v. Osborne Industries, Inc., 839 F.Supp. 1499 (D.Kan., 1993), order af-
firmed by 52 F.3d 867 (10th Cir., 1995), certiorari denied by 516 U.S. 920 (S.Ct.,
1995).
Toyota Motor Corp. v. US, 561 F.Supp. 354 (C.D.Cal., 1983).
United States of America v. Microsoft Corporation, W.D. Wash., Docket No.

2:14-mc-00117-RSM, motion filed 18 Dec. 2014.
US v. Merrill, 258 F.R.D. 302 (E.D.N.C., 2009).
US v. Toyota Motor Corp., 569 F.Supp. 1158 (C.D.Cal., 1983).
Waterman v. MacKenzie, 29 F. 316 (1886), affirmed by 138 U.S. 252 (Circuit

Court, S.D. New York, 1891).
US Court of Customs and Patents Appeals
Haymaker Sports, Inc. v. Turian, 581 F.2d 257 (C.C.P.A., 1978).
817
References
US Court of Appeal
Donrey, Inc. v. US, 809 F.2d 534 (8th Cir., 1987).
Northern Natural Gas Co. v. US, 470 F.2d 1107 (8th Cir., 1973), certiorari denied
by 412 U.S. 939 (S.Ct., 1973).
Russello v. US, 648 F.2d 367 (5th Cir.[Fla.], 1981), on reconsideration in part, 650
F.2d 651 (5th Cir.[Fla.], 1981), and on rehearing 681 F.2d 952 (5th Cir.[Fla.], 1982),
and judgment affirmed by 464 U.S. 16 (S.Ct., 1983).
U.S. Indus. Alcohol Co. (West Virginia) v. CIR, 42 B.T.A. 1323 (B.T.A., 1940), af-
firmed in part, reversed in Part by 137 F.2d 511 (C.C.A.2, 1943).
US Supreme Court
Mayo Found. v. US, 562 US (2011).
Chevron USA Inc. v. Natural Resources Defence Council Inc., 467 US 837.
Case documents
The following documents filed with the US Tax Court are referred to in the dis-
cussions of GlaxoSmithKline Holdings (Americas), Inc. v. CIR, 117 TC 1 (Tax Ct.,
Docket No. 5750-04, 2004).
Year Description Reference

2004 GlaxoSmithKline’s 7 January 2004 Available at Tax Analysts (www.
court documents. taxanalysts.com); Doc 2004-363.
2004 Press release from GlaxoSmith- Available at Tax Analysts (www.
Kline. taxanalysts.com); Doc 2004-363.
2004 2 April 2004 GlaxoSmithKline Tax Available at Tax Analysts (www.
Court petition (contained the notice taxanalysts.com); Doc 2004-7600.
of deficiency as an appendix).
2005 GlaxoSmithKline’s 22 February Available at Tax Analysts (www.
2005 memorandum in opposition to taxanalysts.com); Doc 2005-3635.
the IRS motion for partial summary
judgment.
818
References
2005 GlaxoSmithKline’s 26 January Available at Tax Analysts (www.

2005 court documents. taxanalysts.com); Doc 2005-1656.
2006 IRS press release. IR-2006-142
omestic statutes, preparatory works, regulations,

D
administrative guidelines and decisions, etc.
US Code

1913 Tariff Act of 1913 38 Stat. 166
1917 War Revenue Act of 1917 40 Stat. 300
1918 Revenue Act of 1918 40 Stat. 1058
1933 Securities Act of 1933 48 Stat. 74
1954 Internal Revenue Code of 1954 68 Stat. 1112
1954 Industrial Incentive Act of 1954 P.R. Laws Ann. tit. 13,
Sections 241-251
1963 Industrial Incentive Act of 1963 P.R. Laws Ann. tit. 13,
Sections 252-252j
1976 Tax Reform Act of 1976 90 Stat. 1834
1982 Tax Equity and Fiscal Responsibility Act 96 Stat. 324
of 1982
1984 Deficit Reduction Act of 1984 98 Stat. 494
1986 Tax Reform Act of 1986 100 Stat. 2085
1993 The Omnibus Budget Reconciliation Act 107 Stat. 312
of 1993
1997 Taxpayer Relief Act of 1997 111 Stat. 788
2009 US Freedom of Information Act 5 U.S.C.A. § 552
2012 Regulation S–K (Standard Instructions for 17 CFR Part 229
Filing Forms Under Securities Act of 1933)
Regulation S–K
2017 Tax Cuts and Jobs Act, Pub. L. 115–97, title 131 Stat. 2219
I, §  14221(b)(2), 22 Dec., 2017
819
References
US Treasury Regulations
The US regulations under IRC section 482 are comprehensive and have un-
dergone several revisions. In order to make the references to this material
more accessible, the author has sorted the different documents (proposed,
temporary and final regulations) in chronological order with respect to
each main revision of the regulations.

1918 T.D. 2694, 20 Treas. Dec. Int. Rev. 294, 321 (1918).
Regulations to War Revenue Act of 1917, Ch. 63, 40
Stat. 300 (1917).
1935 Reg, 86, Art. 45 (1935). Regulations to Section 45 of
the Revenue Act of 1928.
The 1968 regulations

1965 Allocation of Income and Deductions Among Tax- 30 Fed. Reg.
payers, proposed regulations (1 April 1965). 4256
payers; Determination of Sources of Income, pro- 10394
posed regulations (2 August 1966).
payers, final regulations (16 April 1968). 5848
The 1994 regulations

1992 Intercompany Transfer Pricing and Cost Sharing 57 FR 3571-01
Regulations Under Section 482, proposed regula-
tions (30 January 1992).
1993 Intercompany Transfer Pricing Regulations Under 58 FR 5310-01
Section 482, Notice of proposed rulemaking (21 Jan-
uary 1993).
Section 482, temporary regulations (21 January 1993).
Section 482, final regulations (8 July 1994).
The cost-sharing regulation

1992 Intercompany Transfer Pricing and Cost-Sharing 57 FR 3571-01
Regulations Under Section 482, proposed regula-
tions (30 January 1992).
820
References

1995 Section 482 Cost-Sharing Regulations, final regula- 60 FR 65553-01
tions (20 December 1995).
1996 Revision of Section 482 Cost-Sharing Regulations, 61 FR 21955-01
final regulations (13 May 1996).
2003 Compensatory Stock Options Under Section 482, 68 FR 51171-02
final regulations (26 August 2003).
2005 Section 482: Methods To Determine Taxable Income 70 FR 51116-01
in Connection With a Cost-Sharing Arrangement,
proposed regulations (29 August 2005).
temporary regulations (5 January 2009).
final regulations (22 December 2011).
2013 Use of Differential Income Stream as an Applica- 78 FR 52854-01
tion of the Income Method and as a Consideration
in Assessing the Best Method, final regulations
(27 August 2013).
The services regulations

2003 Treatment of Services Under Section 482; Alloca- 68 FR 53448-01
tion of Income and Deductions From Intangibles,
proposed regulations (10 September 2003).
tion of Income and Deductions From Intangibles;
Stewardship Expense, temporary regulations (4 Au-
gust 2006).
tion of Income and Deductions From Intangible
Property; Stewardship Expense, final regulations
(4 August 2009).
2015 best-method rule and valuation regulations

2015 Clarification of the Coordination of the Transfer 80 FR 55538-01
Pricing Rules With Other Code Provisions, tempo-
rary regulations (16 September 2015).
821
References

Outbound section 367 transfers and relationship
to section 482 regulations
2013 Certain Outbound Property Transfers by Domestic 78 FR 17024-01
Corporations; Certain Stock Distributions by Do-
mestic Corporations.
2015 Treatment of Certain Transfers of Property to For- 80 FR 55568-01
eign Corporations, proposed regulations (16 Sep-
tember 2015).
2016 Treatment of Certain Transfers of Property to For- T.D. 9803
eign Corporations (final regulations).
2017 Transfers of Certain Property by U.S. Persons to T.D. 9814
Partnerships with Related Foreign Partners.
US “check the box” regulations

1996 Simplification of entity classification rules. 61 FR 66584-01
IRS documents

1963 Guidelines for cases involving the allocation of in- Rev. Proc. 63-10
come and expenses between United States companies
and their manufacturing affiliates in Puerto Rico.
1968 IRS Revenue Procedure with guidelines regarding Rev. Proc. 68-23
certain transactions involving foreign corporations
requiring an advance ruling under Section 367 of the
Internal Revenue Code of 1954.
1974 IRS revenue ruling on the depreciation of insurance Rev. Rul. 74-456
expirations, customer and subscription lists, and
similar assets.
1975 Revenue procedure with respect to conditions for Rev. Proc. 75-29
issuing favorable rulings pursuant to Section 375 of
the Internal Revenue Code of 1954.
1976 Revenue procedure explaining the effect of Rev. Rev. Proc. 76-20
Rul. 75–561, 1975–2 C.B. 129, on transactions de-
scribed in Section 3.03(1)(c) of Rev. Proc. 68–23,
1968–1 C.B. 821, 827.
822
References

1977 On issuing rulings under Section 367 of the Code. Rev. Proc. 77-17
1980 On guidelines that will be used by the IRS in con- Rev. Proc. 80-14
sidering requests for rulings under Section 367(a)(1)
of the IRC.
1991 IRS Coordinated Issue Paper on Amortization of 91 TNT 90-35
Assembled Workforce.
1996 IRS issue paper on the amortization of assembled 96 TNT 49-27
workforce.
2003 Ruling publishing a list of previously published rul- Rev. Rul. 2003-
ings that have been identified as no longer determi- 99
native with respect to future transactions.
2005 IRS guidance on Section 936 termination. Notice 2005-21
2006 Procedures for requesting competent authority as- Rev. Proc. 2006-
sistance under tax treaties. 54
2007 Industry Director Directive No. 1 on Section 936 L M S B - 0 4 - 01-
Exit Strategies. 07-002
2007 Coordinated Issue Paper on IRC Section 482 CSA LMSB-04-0907-
buy-in adjustments. 62
2007 IRS advice memorandum on taxpayer use of Section IRS AM 2007-
482 and the commensurate with income standard. 007
2008 IRS industry directive on IRC Section 936 exit strat- LMSB-04-0108-
egies. 001
2009 IRS Generic Legal Advice Memorandum (GLAM) IRS AM 2009-
on lending in the US by foreign person giving rise to 010
effectively connected income.
2012 Internal Revenue Bulletin containing 2011 final IRB 2012-12
cost-sharing regulations (T.D. 9568).
2012 Internal Revenue Bulletin on outbound transfers of Notice 2012-39
intangibles in certain asset reorganizations.
2015 Updates to the procedures for requesting assistance Rev. Proc. 2015-
from the U.S. competent authority. 40
2017 U.S. Treasury Department, Second Report to the TNT Doc 2017-
President on Identifying and Reducing Tax Regula- 72131
tory Burdens, Executive Order 13789, 2 October
2017.
823
References
US legislative preparatory works, reports, etc.

1921 H.R. Rep. No. 350, 67th Cong., 1st Sess. 1 (1921),
1939-1 C.B. (Part 2) 168, 174.
1932 H.R. Rep. No. 708, 72d Cong., 2d Sess. (1932),
1939-1 (Part 2) C.B. 471.
1962 H.R. Rep. No. 87-1447 (1962). 1962-3 CB 405
1984 House Ways and Means Committee Report on the H.R. Rep. No.
Tax Reform Act of 1984. 98-432, Pt. 2
(1984)
1984 General Explanation of the Revenue Provisions of JCS-41-84
the Deficit Reduction Act of 1984, H.R. 4170, 98th
Congress, Public Law 98-369 (Joint Committee on
Taxation).
1984 Senate Finance Committee Report on the Deficit S. Rep. 98-169,
Reduction Act of 1984. Vol. 1 (1984)
1985 H.R. Rep. No. 426, 99th Cong., 1st Sess. 424 (1985). 1985 House
Report
1986 H.R. Conf. Rep. No. 841, 99th Cong., 2d Sess. II-638 1986 Committee
(1986). Report
1988 A study of intercompany pricing under Section 482 Notice 88-123
of the Code, 5 December 1988. (White Paper)
2009 Description Of Revenue Provisions Contained In JCS-4-09
The President’s Fiscal Year 2010 Budget Proposal
Part Three: Provisions Related To The Taxation Of
Cross-Border Income And Investment (The Joint
Committee on Taxation).
2009 General explanations of the Administration’s Fiscal FY 2010 Green-
Year 2010 Revenue Proposals (US Department of book
the Treasury).
2010 Joint Committee on Taxation Report: Present law JCX-37-10
and background related to possible income shifting
and transfer pricing, 20 July 2010.
2011 Description of revenue provisions contained in the JCS-3-11 No. 11
President’s fiscal year 2012 budget proposal (Joint
Committee on Taxation).
2013 General Explanations of the Administration’s Fiscal FY 2014 Green-
Year 2014 Revenue Proposals (US Department of book
the Treasury).
2017 Description of the Chairman’s mark of the “Tax Cuts JCX-51-17
and Jobs Act”.
824
References
Norwegian Tax Act
Year Description Abbreviation

1999 Lov om skatt av formue og inntekt (skatteloven), NTA
LOV-1999-03-26-14.
Official international documents
Model treaties for the avoidance of double taxation

2011 United Nations Model Double Taxation Convention UN MTC
between Developed and Developing Countries.
2014 OECD Model Tax Convention on Income and on OECD MTC
Capital.
OECD documents

1963 Draft Double Taxation Convention on Income and
Capital, Report of the OECD Fiscal Committee
(1963).
1979 Transfer Pricing and Multinational Enterprises 1979 OECD
(1979). Report
1984 Transfer Pricing and Multinational Enterprises: 1984 OECD
Three Taxation Issues (1984). Report
1987 Issues in International Taxation, No. 2, Thin capi- 1987 OECD
talisation, Taxation of Entertainers, Artistes and Report
Sportsmen (1987).
1993 Attribution of Income to Permanent Establishments 1993 OECD
(1993). Report
1993 Intercompany transfer pricing regulations under US
Section 482 temporary and proposed regulations
(1993).
825
References

1993 Tax aspects of transfer pricing within multinational Task Force
enterprises – The United States proposed regula- Report
tions: A report by the committee on fiscal affairs
on the proposed regulations under section 482 IRC
(1993).
1993 Intercompany transfer pricing regulations under US 2nd Task Force
Section 482 temporary and proposed regulations Report
(1993).
1994 Issues in International Taxation 5: Model Tax Con-
vention: Attribution of Profits to Permanent Estab-
lishments (1994).
1994 Transfer pricing guidelines for multinational enter- 1994 OECD
prises and tax administrations, Part 1: Principles and discussion draft
methods, discussion draft (1994).
1995 Transfer pricing guidelines for multinational enter- 1995 OECD
prises and tax administrations, Draft text of Part II discussion draft
(1995).
1995 OECD Transfer Pricing Guidelines for Multinational 1995 OECD
Enterprises and Tax Administrations (1995). TPG
1997 Report on cost contribution arrangements, adopted
by the Committee on Fiscal Affairs on 25 June 1997.
1998 Harmful Tax Competition: An emerging global is-
sue (1998).
2001 Discussion draft on the attribution of profits to per- 2001 OECD PE
manent establishments (2001). Draft
2004 Discussion draft on the attribution of profits to per- 2004 OECD PE
manent establishments – Part I: General considera- draft
tions (2004).
2005 OECD tax policy studies, e-commerce: Transfer
pricing and business profits taxation, No. 10 (2005).
2006 Report on the Attribution of Profits to Permanent
Establishments, Part I: General Considerations
(2006).
2006 Report on the Attribution of Profits to Permanent
Establishments, Part II: Special Considerations for
Applying the Authorized Approach to Permanent
Establishments (PEs) of Banks (2006).
826
References

2006 Comparability: Public invitation to comment on a 2006 OECD
series of draft issue notes (2006). comparability
report
2008 Attribution of Profits to Permanent Establishments 2008 OECD
(2008). Report
2008 Transfer pricing aspects of business restructurings:
Discussion draft for public comment (Paris, 2008).
2008 Transactional profit methods, discussion draft for 2008 OECD
public comment (2008). discussion draft
2008 The 2008 update to the OECD Model Tax Conven-
tion (2008).
2008 OECD Benchmark Definition of Foreign Direct
Investment, 4th ed. (2008).
2009 Chapter 1 Proposed revision of Chapters I-III of the 2009 OECD
Transfer Pricing Guidelines (2009). proposed
revisions
2009 Chapter 2 Proposed revision of Chapters I-III of the
Transfer Pricing Guidelines, discussion draft (2009).
2010 2010 Report on the attribution of profits to perma- 2010 OECD
nent establishments (2010). Report
2010 Model Tax Convention on Income and on Capital, OECD
condensed version (2010). Commentaries
2010 OECD Transfer Pricing Guidelines for Multinational 2010 TPG
Enterprises and Tax Administrations (2010).
2011 Tackling aggressive tax planning through improved
transparency and disclosure, report on disclosure
initiatives (2011).
2012 Discussion draft Revision of the special considera- 2012D
tions for intangibles in Chapter VI of the OECD
Transfer Pricing Guidelines and Related Provisions
(2012).
2012 OECD Transfer Pricing Guidelines for Multinational
Enterprises and Tax Administrations, Dealing Effec-
tively with the Challenges of Transfer Pricing (2012).
2012 Draft on timing issues relating to transfer pricing 2012 OECD
(2012). timing draft
2013 Action Plan on Base Erosion and Profit Shifting OECD BEPS
(2013). Action Plan
827
References

2013 Revised discussion draft on transfer pricing aspects 2013 RDD
of intangibles (2013).
2013 Addressing Base Erosion and Profit Shifting (2013).
2013 A step change in tax transparency, OECD Report for
the G8 Summit Lough Erne (2013).
2014 Guidance on Transfer Pricing Aspects of Intangi- 2014D
bles, OECD/G20 Base Erosion and Profit Shifting
Project, Action 8: 2014 Deliverable (2014).
2014 Public discussion draft BEPS Action 10: Discus- PSMDD
sion Draft on the use of profit splits in the context of
global value chains (2014).
2014 Public discussion draft, BEPS Actions 8, 9 and 10: 2014 OECD risk
Discussion draft on revisions to chapter I of the draft
Transfer Pricing Guidelines, including risk, rechar-
acterisation, and special measures (Paris, 2014).
2015 Explanatory Paper Agreement on Modified Nexus
Approach for IP Regimes 2015 (2015).
2015 Addressing the Tax Challenges of the Digital Econo-
my, Action 1: 2015 Final Report (2015).
2015 Limiting Base Erosion Involving Interest Deduc- 2015 Action 4
tions and Other Financial Payments, Action 4: 2015 Report
Final Report (Paris, 2015).
2015 OECD/G20 Base Erosion and Profit Shifting Project,
Action 5: Agreement on Modified Nexus Approach
for IP Regimes (2015).
2015 Countering Harmful Tax Practices More Effectively, 2015 Action 5
Taking into Account Transparency and Substance, Report
Action 5: 2015 Final Report (2015).
2015 Preventing the Artificial Avoidance of Permanent 2015 Action 7
Establishment Status, Action 7: 2015 Final Report Report
(2015).
2015 Aligning Transfer Pricing Outcomes with Value 2015TPR
Creation, Actions 8-10: 2015 Final Report (2015).
2015 Measuring and Monitoring BEPS, Action 11: 2015
Final Report (2015).
2015 Developing a Multilateral Instrument to Modify BEPS Action
Bilateral Tax Treaties, Action 15: 2015 Final Report 15: 2015 Final
(2015). Report
828
References

2015 Public Discussion Draft, BEPS Action 8: Revisions 2015 OECD
to Chapter VIII of the Transfer Pricing Guidelines CSA draft
on Cost Contribution Arrangements (2015).
2016 Document for public review, Conforming Amend-
ments to Chapter IX of the Transfer Pricing Guide-
lines (2016).
2016 Public discussion draft, BEPS Action 7, Additional
Guidance on the Attribution of Profits to Permanent
Establishments (2016).
2016 Public discussion draft, BEPS Actions 8-10 Revised
Guidance on Profit Splits (2016).
2016 Public Discussion Draft, BEPS Action 8, Implemen-
tation Guidance on Hard-to-Value Intangibles (2017).
2016 Public discussion draft, BEPS Action 7, Additional
Guidance on Attribution of Profits to Permanent
Establishments (2017).
2016 Public discussion draft, BEPS Action 10, Revised
Guidance on Profit Splits (2017).
2017 OECD Transfer Pricing Guidelines for Multinational 2017 OECD
Enterprises and Tax Administrations (Paris, 2017). TPG
2018 Report, Additional Guidance on the Attribution of
Profits to a Permanent Establishment under BEPS
Action 7 (2018).
UN documents

2011 United Nations Model Double Taxation Convention UN MTC
between Developed and Developing Countries.
2013 United Nations Practical Manual on Transfer Pricing UNTPM
for Developing Countries.
2015 World Investment Report 2015: Reforming Inter- UNCTAD
national Investment Governance (United Nations (2015)
Conference on Trade and Development).
2017 United Nations Practical Manual on Transfer Pricing UN PMTP 2017
for Developing Countries.
829
References
EU documents

2011 Member States’ responses to questionnaire on com- JTPF/019/
pensating adjustments/year-end adjustments (EU REV1/2011/EN
Joint Transfer Pricing Forum).
2013 Report on Compensating Adjustments (EU Joint JTPF/009/FI-
Transfer Pricing Forum). NAL/2013/EN
2014 Communication from the Commission to the Eu- COM(2014) 315
ropean Parliament, the Council and the European Final
Economic and Social Committee on the Work of
the EU Joint Transfer Pricing Forum in the Period
July 2012 to January 2014 (European Commis-
sion).
2015 Communication from the Commission to the Eu- 7249/15
ropean Parliament, the Council and the European
Economic and Social Committee on the Work of the
EU Joint Transfer Pricing Forum in the Period July
2012 to January 2014 – Council Conclusions (Coun-
cil of the European Union).
2016 European Commission Study on the Application of
Economic Valuation Techniques for Determining
Transfer Prices of Cross Border Transactions be-
tween Members of Multinational Enterprise Groups
in the EU
Treaties and conventions
Treaties for the avoidance of double taxation

1966 Convention between the Swiss Confederation and Spain-
Spain for the Avoidance of Double Taxation with Switzerland
Respect to Taxes on Income and Fortune, entered treaty (1966)
into 26 April 1966. (In force. Amended by subse-
quent protocols.)
830
References

1968 Convention between the Government of the United 1968 France-
Kingdom of Great Britain and Northern Ireland and United Kingdom
France for the Avoidance of Double Taxation and the treaty
Prevention of Fiscal Evasion with Respect to Taxes
on Income, entered into 22 May 1968. (Not in force.
Current France-United Kingdom treaty was entered
into in 2008.)
1971 Convention between the Government of the Repub- 1971 Japan-
lic of Singapore and the Government of Japan for the Singapore treaty
Avoidance of Double Taxation and the Prevention
of Fiscal Evasion with Respect to Taxes on Income,
entered into 29 January 1971. (Not in force. Current
Japan-Singapore treaty was entered into in 1994,
and amended by subsequent protocols.)
1975 Convention between the Government of the United United
States of America and the Government of the United Kingdom-United
Kingdom of Great Britain and Northern Ireland for States treaty
the Avoidance of Double Taxation and the Preven-
tion of Fiscal Evasion with Respect to Taxes on In-
come and Capital Gains, entered into 31 December
1975. (Not in force. Current United Kingdom-United
States treaty was entered into in 2001, and amended
by subsequent protocols.)
1978 Convention between the Government of the United Canada-United
Kingdom of Great Britain and Northern Ireland and Kingdom treaty
the Government of Canada for the Avoidance of
Double Taxation and the Prevention of Fiscal Eva-
sion with Respect to Taxes on Income and Capital
Gains, entered into 8 September 1978. (In force.
Amended by a 2013 protocol.)
1980 Convention between the United States of America 1980 Canada-
and Canada with Respect to Taxes on Income and on United States
Capital, entered into 26 September 1980. (In force.) treaty
2000 Convention between Ireland and the Kingdom of Ireland-Norway
Norway for the Avoidance of Double Taxation and treaty (2000)
the Prevention of Fiscal Evasion with Respect to
Taxes on Income and on Capital, entered into 22 No-
vember 2000. (In force.)
831
References

2013 Convention Between the Kingdom of Norway and Norway-United
the United Kingdom of Great Britain and Northern Kingdom treaty
Ireland for the Avoidance of Double Taxation and (2013)
the Prevention of Fiscal Evasion with Respect to
Taxes on Income and on Capital Gains, entered into
14 March 2013. (In force.)
Conventions
1969 Vienna Convention on the Law of Treaties, conclud- VCLT
ed in Vienna on 23 May 1969.
Financial accounting standards

2001 International Accounting Standard 1: “Presentation of IAS 1
Financial Statements” (available at http://www.ifrs.org/).
2001 International Accounting Standard 2: “Inventories” IAS 2
(available at http://www.ifrs.org/).
2001 International Accounting Standard 18: “Revenue” IAS 18
(available at http://www.ifrs.org/).
2001 Statement of Financial Accounting Standards No. FAS 142
142: “Goodwill and Other Intangible Assets” (avail-
able at http://www.fasb.org/).
2004 International Financial Reporting Standards 3: “Busi- IFRS 3
ness Combinations” (available at http://www.ifrs.org/).
2007 Statement of Financial Accounting Standards No. FAS 141
141: (revised 2007) “Business Combinations” (avail-
able at http://www.fasb.org/).
2010 International Financial Reporting Standards Concep-
tual Framework (available at http://www.ifrs.org/).
2014 International Financial Reporting Standards 15: IFRS 15
“Revenue from contracts with customers” (available
at http://www.ifrs.org/).
832
Other Titles in the IBFD Doctoral Series
Vol. 1 – The Concept of Income: A Multi-Disciplinary Analysis

by Kevin Holmes
Vol. 2 – Taxing Portfolio Income in Global Financial Markets

by Doron Herman
Vol. 3 – Free Movement of Persons and Income Tax Law:

The European Court in Search of Principles
by Servaas van Thiel
Vol. 4 – Advance Tax Rulings and Principles of Law:

Towards a European Tax Rulings System?
by Carlo Romano
Vol. 5 – Arbitration under Tax Treaties: Improving Legal Protection

in International Tax Law
by Mario Züger
Vol. 6 – E-Commerce and Source-Based Income Taxation

by Dale Pinto
Vol. 7 – Interpretation of Tax Treaties under International Law

by Frank Engelen
Vol. 8 – Taxation of Investment Funds in the European Union

by Tomi Viitala
Vol. 9 – Taxation of Cross-Border Partnerships: Double Tax Relief

in Hybrid and Reverse Hybrid Situations
by Jesper Barenfeld
Vol. 10 – Taxation of International Performing Artistes: The Problems

with Article 17 OECD and How To Correct Them
by Dick Molenaar
Vol. 11 – Dispute Resolution under Tax Treaties

by Zvi Daniel Altman
Vol. 12 – Income from International Private Employment: An Analysis
of Article 15 of the OECD Model
by Frank Pötgens
Vol. 13 – Taxation of Investment Derivatives

by Antti Laukkanen
Vol. 14 – International Tax Planning and Prevention of Abuse

by Luc De Broe
Vol. 15 – EC Law Aspects of Hybrid Entities

by Gijs Fibbe
Vol. 16 – The EU VAT System and the Internal Market

by Rita Aguiar de Sousa e Melo de la Feria
Vol. 17 – Formulary Apportionment for the Internal Market

by Stefan Mayer
Vol. 18 – Cross-Border Consumption Taxation of Digital Supplies

by Pernilla Rendahl
Vol. 19 – EC Law and the Sovereignty of the Member States in Direct

Taxation
by Mathieu Isenbaert
Vol. 20 – Arm’s Length Transaction Structures – Recognizing and

Restructuring Controlled Transactions in Transfer Pricing
by Andreas Bullen
Vol. 21 – Optimization of Tax Sovereignty and Free Movement

by Sjoerd Douma
Vol. 22 – Taxation of Foreign Business Income within the European

Internal Market – An Analysis of the Conflict between
the Objective of Achievement of the European Internal
Market and the Principles of Territoriality and Worldwide
Taxation
by Jérôme Monsenego
Vol. 23 – The Missing Keystone of Income Tax Treaties

by Joanna Wheeler
Vol. 24 – The Principle of Non-Discrimination in International and
European Tax Law
by Niels Bammens
Vol. 25 – A VAT/GST Model Convention

by Thomas Ecker
Vol. 26 – The Interface of International Trade Law and Taxation –

Defining the Role of the WTO
by Jennifer Emma Farrell
Vol. 27 – The Structure and Organization of EU Law in the Field of

Direct Taxes
by Marcel Schaper
Vol. 28 – Taxpayer Participation in Tax Treaty Dispute Resolution

by Katerina Perrou
Vol. 29 – Triangular Cases: The Application of Bilateral Income Tax

Treaties in Multilateral Situations
by Emily Fett
Vol. 30 – Taxation of International Sportsmen

by Karolina Tetłak
Vol. 31 – On the Legitimacy of International Tax Law

by Cees Peters
Vol. 32 – The Sources of EU Law and Their Relationships: Lessons for

the Field of Taxation
by Rita Szudoczky
Vol. 33 – Multilingual Tax Treaties: Interpretation, Semantic Analysis

and Legal Theory
by Paolo Arginelli
Vol. 34 – VAT Grouping from a European Perspective

by Sebastian Pfeiffer
Vol. 35 – European Capital Movements and Corporate Taxation –

From Transaction-Based Origins to More Principle-Based
Capital and Taxation Policies
by Nana Šumrada Slavnić
Vol. 36 – European Value Added Tax in the Digital Era: A Critical
Analysis and Proposals for Reform
by Marie Lamensch
Vol. 37 – Taxing Cross-Border Services: Current Worldwide Practices

and the Need for Change
by Angharad Miller
Vol. 38 – Cooperative Compliance: A New Approach to Managing

Taxpayer Relations
by Katarzyna Bronżewska
Vol. 39 – Taxation of Services in Treaties between Developed and

Developing Countries: A Proposal for New Guidelines
by Fernando Souza de Man
Vol. 40 – Transactional Adjustments in Transfer Pricing

by Aitor Navarro
Vol. 41 – Nexus Requirements for Taxation of Non-Residents’ Business

Income: A Normative Evaluation in the Context of the Global
Economy
by Stjepan Gadžo
Vol. 42 – Dispute Resolution in the EU: The EU Arbitration

Convention and the Dispute Resolution Directive
by Harm Mark Pit
Vol. 43 – The Interpretation of Tax Treaties in Relation to Domestic

GAARs
by Eivind Furuseth
Vol. 44 – The Tax Sparing Mechanism and Foreign Direct Investment

by Na Li

Transfer Pricing and Intangibles: Oddleif Torvik

Uploaded by

Copyright:

Available Formats

You might also like

Transfer Pricing and Intangibles: Oddleif Torvik

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Transfer Pricing and Intangibles: Oddleif Torvik

Uploaded by

Copyright:

Available Formats

45

Distribution of Operating Profits from IP Value Chains

Distribution of Operating Profits

Oddleif Torvik is a tax lawyer with

Your Portal to Cross-Border Tax Expertise

18-10-2018 Doctoral 45 - 21,9 mm rug indd.indd 1 02-11-18 14:53

Editor-in-Chief: Pasquale Pistone

True to its mission of disseminating knowledge of international taxation and

Books published in the IBFD Doctoral Series should:

The IBFD Doctoral Series is indexed in Thomson Reuters’ Book Citation

US and OECD arm’s length distribution of

Degree awarded on 9 December 2016

Telephone: 31-20-554 0100

All rights reserved. No part of this publication may be reproduced, stored

ISBN 978-90-8722-495-0 (print)

Chapter 1: Research Questions, Methodology and Sources

1.1. Introductory comments 3

1.2. Key terminology and contextualization 6

1.3. Research questions and structure 12

1.5. The relevant OECD sources of law 22

1.6. The relevant US sources of law 30

1.7. A few words on the 2017 US tax reform 37

1.8. The relationship between the book and other

1.9. Reference register and source abbreviations 45

Chapter 2: Business and Tax Motivations for Intangible

2.2. Horizontal and vertical FDI 48

2.3. To stay home (outsource) or to go out (FDI)? 50

2.4. The centralized principal model for profit allocation 52

2.5. IP regimes and the 2015 OECD nexus approach 54

Chapter 3: Controlled Intangibles Transfers 63

3.2. The US intangibles definition 64

3.3. The OECD intangibles concept 85

3.4. Useful distinctions on the intangibles concept 88

3.5. Controlled intangibles transfers subject to

3.6. Controlled intangibles transfers subject to

3.7. Concluding comments 117

Chapter 4: Introduction to Part 2 119

Chapter 5: The Historical Development of ­

5.1. Introduction 125

5.2. Development of the US PSM and the contract

5.2.3.2. Nestlé (1963) 130

5.3. US legislative and regulatory implementation of

5.4. OECD implementation of “profit-based” transfer

Chapter 6: Metaconcepts Underlying the US and OECD

6.1. Introduction 165

6.2. The relationship between operating profits and

6.2.3.1. Sales 169

6.4. Which transfer pricing method should govern

6.5. The arm’s length range 190

6.6. Comparability 194

6.6.5. Comparability factors 200

6.7. The aggregation of controlled transactions 217

Chapter 7: Direct Transaction-Based Allocation of

7.1. Introduction 237

7.2. The US CUT method 238

7.2.3. The purported CUT pertains to a transfer of

7.3. The OECD CUT method 247

Chapter 8: Indirect Profit-Based Allocation of Residual

1.1. Introductory comments 3

1.2. Key terminology and contextualization 6

1.3. Research questions and structure 12

1.5. The relevant OECD sources of law 22

1.6. The relevant US sources of law 30

1.7. A few words on the 2017 US tax reform 37

1.8. The relationship between the book and other

1.9. Reference register and source abbreviations 45

2.2. Horizontal and vertical FDI 48

2.3. To stay home (outsource) or to go out (FDI)? 50

2.4. The centralized principal model for profit allocation 52

2.5. IP regimes and the 2015 OECD nexus approach 54

Chapter 3: Controlled Intangibles Transfers 63

3.2. The US intangibles definition 64

3.3. The OECD intangibles concept 85

3.4. Useful distinctions on the intangibles concept 88

3.5. Controlled intangibles transfers subject to

3.6. Controlled intangibles transfers subject to

3.7. Concluding comments 117

Chapter 4: Introduction to Part 2 119

Chapter 5: The Historical Development of

5.1. Introduction 125

5.2. Development of the US PSM and the contract

5.2.3.2. Nestlé (1963) 130

5.3. US legislative and regulatory implementation of

5.4. OECD implementation of “profit-based” transfer

6.1. Introduction 165

6.2. The relationship between operating profits and

6.2.3.1. Sales 169

6.4. Which transfer pricing method should govern

6.5. The arm’s length range 190

6.6. Comparability 194

6.6.5. Comparability factors 200

6.7. The aggregation of controlled transactions 217

7.1. Introduction 237

7.2. The US CUT method 238

7.2.3. The purported CUT pertains to a transfer of

7.3. The OECD CUT method 247

8.1. Introduction 253

8.2. A lead-in to the methodology 254

8.3. The scope of application of the methodology 257

8.4. How operating profits may be allocated to

8.5. Comparability under the CPM 265

8.6. Comparability under the TNMM 267

8.6.2. The concept of blended profits illustrated

8.7. Has the TNMM converged towards the 1988

8.8. Is reduced transactional comparability under

8.9. Closing comments on comparability under

9.1. Introduction 301

9.2. The scope of application of the methodology 302

9.2.1. Introduction 302

9.3. How operating profits may be split under

9.4. The PSM in valuation scenarios 320

9.5. The OECD PSM is limited to information known

10.1. Introduction 327

10.2. A lead-in to the topic: The incremental operating

10.3. Which jurisdiction should be entitled to tax

10.4. What are location savings? 331

10.5. The allocation of cost savings 333

10.6. What are other LSAs? 341

10.7. The allocation of location rents 343

10.8. Synergies 344

11.1. Introduction 347

11.2. A transfer pricing paradigm under pressure 347

11.3. Shall something more now be allocated to source

11.4. The relative roles of the CUT method, TNMM

12.1. Introductory comments 359

12.2. Unspecified methods under the US regulations 361

12.3. Unspecified methods under the OECD TPG 363

13.1. Introduction 367

13.2. The valuation techniques accepted under

13.3. The valuation parameters in DCF-based valuation 371

13.4. Allocation of the valuation amount among

13.5. The options realistically available as a restriction

13.6. Is there a legal basis for applying a discount

14.1. Introduction 383

14.2. Buy-in pricing under the US regulations 385

14.3. Buy-in pricing under the OECD TPG 430

14.4. The relationship between US and OECD

15.1. Introduction 437