Corruption and Fraud

in Financial Markets

1 Alexander42 177 1_ffirs01.indd 1

 Corruption and Fraud
in Financial Markets

Malpractice, Misconduct
and Manipulation

Carol Alexander

Douglas Cumming

WILEY

1 Alexander42 1771_ffirs01.indd 3 312{12020 8'19'26Pl\1 li rauta1a s

 This edition first published 2020
© 2020 Carol Alexander and Douglas Cumm ing

Registered office
John Wi ley & Sons Ltd, The Atrium, Southern Gate, Chichester, West Sussex, PO19 8SQ, United Kingdom

For details of our global editorial offices, for customer services and for information about how to apply
for permission to reuse the copyright material in this book please see our website at www.wiley.com.

Ali rights reserved. No part of this publ ication may be reproduced, stored in a retrieva l system, or
transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise,
except as permitted by the UK Copyright, Designs and Patents Act 1988, without the prior permission of
the publ isher.

W iley publishes in a variety of print and electronic formats and by print-on-demand. Some material
included with standard print versions of this book may not be included in e-books or in print-on-demand.
lf this book refers to media such as a CD or DVD that is not included in the version you purchased, you
may download this material at http://booksupport.wiley.com. For more information about W iley products,
visit www.wiley.com.

Designations used by companies to distinguish their products are often claimed as trademarks. AII brand
names and product names used in this book are trade names, service marks, trademarks or registered
trademarks of their respective owners. The publisher is not associated with any product or vendor
mentioned in this book.

Limit of Liability/Disclaimer ofWarranty: Wh ile the publisher and author have used their best efforts
in preparing this book, they make no representations or warranties with respect to the accuracy
or completeness of the contents of this book and specifically disclaim any implied warranties of
merchantabi lity or fitness for a particular purpose. lt is sold on the understanding that the publisher is not
engaged in rendering professional services and neither the publisher nor the author shall be liable for
damages arising herefrom. lf professional advice or other expert assistance is required, the services of a
competent professional should be sought.

Library of Congress Cataloging-in-Publication Data is Available:

ISBN 978-1-119-42177-1 (hardback)

Cover Design: Wiley

Cover lmage: © Comstock/Stockbyte/Getty lmages

Set in 1 O/l 2pt, OptimaLTStd by SPi G lobal, Chenrnai, lndia

Printed in Great Britain by TJ lnternational Ltd, Padstow, Cornwall, UK

10 9 8 7 6 S 4 3 2 1

Alexander42 177 1_ffirs01.indd 4 312{12020 8'19'26Pl\1 li rauta1a s

 Contents

Abo11t tbe Editors XV

List of Contributors xvii
Foreword xix
Acknowledgements xxi
Cha ter 1: lntroduction
Carol Alexander and Douglas Cumming

PARI 1 W HAT ARE MANIPlJI AJIQN ANO ERAI JD

ANO WHY DO THEY MATTER? 11
Chapter 2: An Overview of Market Manipulation 13
Tãlis}. Putnins
2 J lotrad1JCtiao 14
2.2 Definitions of Market Manipulation 16
2 .2.1 Legal lnterpretation and Provisio ns against Market Man ipulation 16
2 .2.2 Economics and Legal Studies Perspective 18
2.3 A Taxonomy of the Types of Market Manipulation 19
2 .3.1 Categories of Market Manipulation 19
2 .3.2 Market Manipulation Techniques 22
2.4 Research on Market Manipulation 26
2.4.1 Theoretical Literature 27
2 .4.2 Empirica l Literature 30
2.4.3 Conclusions from the Research on Market Manipulation 35
2.5 Summary and Conclusio ns 39
References 40

Alexander42 177 1_ftoc.indd 5 23-~~-2020 22:13:26 li ~ auta1a s

 VI CONTENTS

Chapter 3: A Taxonomy of Financia l Market Misconduct 45

Ai Deng and Priyank Gandhi
31 l □ trad11ctia□ 46
3.2 Challenges in Research on Financial Market M isconduct 50
3.3 Defining Financial Market Misconduct 51
3.3.1 Price Manipulation 53
3.3.2 Circu lar Trading 54
3.3.3 Collusion and lnformation Sharing 55
3 3.4 lnside lnformation 56
3 3 5 Reference Price lnflt1ence 56
3.3.6 lmproper Order Handling 57
3.3.7 Misleading Customers 58
3.4 Defining Financial Fraud 59
3 4 1 Credit Card Fra11d 59
3.4.2 Money Launderi_!!g 60
3 4 3 Financial Statement Fra11d 60
3.4.4 Computer lntrusion Fraud 61
3..5._Condu.~·~ - - - - - - - - - - - - - - - - - - - - - - - - -61

Chapter 4: Financial Misconduct and Market-Based Penalties 65

Cbelsea Liu aod Alfred V,3wsoo
41 lntroduction 66
4.2 Notable Cases of Financial Reporting Fraud 69
4.3 Financial Reporting M isconduct and Lega l Redress 70
4.4 Evolution of US Financial Regulations 71
4.4.1 Private Securities Litigation Reform Act (1995) 72
4.4.2 Sarbanes-Oxley Act (2002) 72
4.4.3 Dodd- Frank Act (201 0l 73
4.5 Legal versus Market-Based Penalties for Financial Misconduct 74
4.5.1 Common Forms of Legal Penalties 74
4 5 2 Rale af M;irke!dl_as-e-d~P~e~□-a~lt~ie_s~-----------------~Z-5
4.6 Firm-Level Penalties for Corporate Financial Misconduct 75
4.6.1 Direct Economic Costs Captured in Loss of Market Value 83
4.6.2 Loss of Firm Reputation 83
4.6.3 Spi llover of Reputational Effect 84
4 6 4 Governance Risk and losurance Premiums 85
4.6.5 Reduced Liguidity 85
4.6.6 Access to Financing 85
4 6 Z Reduced lnnavatian 86
4.6.8 Mergers and Acguisitions 86
4.7 Individual-Levei Penalties for Corporate Financial Misconduct 87
4 Z 1 Executive aod Director Iurnaver BZ
4.7.2 lmpaired Career Progression 95
4.7.3 Loss of Reputation 96
4.7.4 Executive Compensation 97
4.7.5 Strengthened Monitoring 97
4 .8 Causes, Risks, and Moderators of Financial M isconduct 98
4 8 1 Eraud Incentives 98
4 8 2 Risk Eactars 11 3

Alexander42 l 77 1_ftoc.indd 6 23-~~-2020 22: 13:26 li ~ auta1a s

 Contents vii

4.8.3 Public Enforcement: Regulatory and Judicial Stringency 115

4 B 4 Public Enforcement· Detectioo and Surveillance 116
A..8 5 Priva.te_Eníorcemen-t_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _l~l~Z
4 9 Other Non-Einancial Misconduct 118
4.1 O Conclud ing Remarks 119
Refereoces 120

Cha ter 5: lnsider Trad in and Market Mani ulation 135

Jonathan A. Batten, lgor Loncarski, and Peter C. Szilagyi
5.1 lntroduction 135
5.2 Regulatory Framework on lnsider Trading and Market Manipulation 140
5.3 Recent Examples of Market Manipu lation and lnsider Trading 145
5 4 Conclusioos 148
References 149

Cha ter 6: Financia l Fraud and Re utational Ca ital 153

Jonathan M. Karpoff
61 Financial Fra11ds in tbe 2QQQs 154
6.2 The Effects of Fraud Revelation on Firm Value and Reputational Capital 156
62 J Market Yalue Losses When Financial Misconduct Is Revealed 156
6.2.2 Spil lover Effects 157
6.2.3 Reputational Losses for Financial Misconduct 158
6.2.4 Direct Measures of Lost Reputational Capital 159
6.2.5 Do M isconduct Firms Always Lose Reputational Capital? 160
6.2.6 Rebui lding Reputational Capital 161
6.3 The Effects of Fraud Revelation on Shareholders and Managers 162
6.3.1 Shou ld Shareholders Pay? Do Managers Pay? 162
6.3.2 Do Shareholders Pay Twice? 162
6 3 3 Are Eirm-I evel Peoalties Efficient2 163
6.3.4 Conseguences for Managers and Directors 163
6.4 Why Do Managers Do lt? Motives and Constraints 165
64 J Motives for Financial Misconduct 165
642 Caostraiots ao Eioaocial Miscood1JCt 167
6.5 Proxies and D atabases Used to ldentify Samples of Financial Statement
Miscooduct 168
6.6 Concl usion: Reputation, Enforcement, and Culture 170
References 171

PART li H O W AN D W HERE D O ES MISCONDUCT OCCUR? 179

Chapter 7: Manipulative and Collusive Practices in FX Markets 181
Al.exisS.tenio.cs.
ZJ lotrod11ctiao 181
7 .2 Different Types of FX Orders 183
7.3 The Unigue FX Market Structure 184
7.4 Examples of Manipulative and Collusive Practices in FX Markets 188
7.4. 1 Front Running 188
7.4.2 Triggering Stop-Loss Orders 190
7.4.3 'Banging the Close' 192

Alexander42 177 1_ftoc.indd 7 23-~~-2020 22: 13:26 li ~ auta1a s

 VII I CONTENTS

7.4.4 Collusion and Sharing of Confidential lnformation 193

7.4.5 Spoofing 195
7.4.6 Market Abuse via Electronic Trading Platforms 196
7 5 The Reforro Pracess 197
References 199

Chapter 8: Fraud and Manipulation within Cryptocurrency Markets 205

David Twomey and Andrew Mann
81 lotroductioo 206
8.2 Why Do fraud and Manipulation Occur in Cryptocurrency Markets? 212
8.2 .1 Lack of Consistent Regulation 212
8.2.2 Relative Anonymity 213
8.2.3 LowBarrierstoEntry 214
8.2.4 Exchange Standards and Sophisticat ion 214
8.3 Pump and Dumps 215
8 3 1 Case Studies 21 7
8.4 lnflated Trading Volume 21 7
8.4.1 Case Study: January 2017 and PBoC lnvolvement 219
8.5 Exchange DDoS Attacks 22 0
8.5.1 Case Study 223
8.6 Hacks and Exploitations 224
8.6.1 Exchange Hacks 224
8.6.2 Smart Contract Exploits 229
8.6.3 Protocol Exploitat ion 230
8.7 Flash Crasbes 23 0
B 7 1 GDAX-ETH/lu:l....
$..,D~E-..la°"s,.._h'-'C'-'ra=""sb"-- - - - - - - - - - - - - - - --"2__,3=4
8.8 Order Book-Based Man ipulations 235
8.8.1 Quote Stuffing 236
8.8.2 O rder Spoofing 237
89 Stablecoios and Tetber 239
8 9 1 Tether Historical JJmelioe 240
8.9.2 Tether Controversy and Criticism 242
8.9.3 Tether's Significance in Cryptocurrency G lobal Markets 245
8.1 O Summary and Conclusions 245
Refeuc..u,...,_,__ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ __ ,'-=-'-

Chapter 9: The lntegrity of Closing Prices 251

Ryan }. Davíes
9.1 Why Closing Prices Matter 251
9.2 Painting the Tape and Portfolio Pumping 252
9.3 ' Bang-the-Close' Manipulation: The Response of Financial lntermediaries 255
9.4 Stock Price Pinning on Option Expiration Dates 259
9.5 Conclusion: Lessons for the Regulation and Design of Financial Markets 263
.RefeLenCJ....,_____________________________ ..,.,,_
Chapter 1 O: A Trader's Perspective on Market Abuse Regu lations 275
Sam Baker
1Q 1 lotrad11ctia□ 275
10.2 Getting the Trading Edge 278

Alexander42 177 1_ftoc.indd 8 23-~~-2020 22: 13:26 li ~ auta1a s

 Contents ix

10.3 A Typical Trader's Market Window 281

1O 4 Wash Trades 282
10.5 High Ticking/Low Ticking - Momentum lgnition 284
1 0.6 Spoofing 286
10.7 Layering 290
10.8 Smoking 292
10.9 Case Study: Paul Rotter a.k.a. 'The Fl ipper' 295
10.1 O The lnnocent and the Guilty 299
1 0.11 What Are Exchanges Doing to Prevent Market Abuse? 301
10.11.1 CME Group 301
10.11.2 ICE 302
10.12 What Are Trading Companies Doing to Prevent Abuse? 302
1O 13 Will There Be ao Eod to Market Abuse? 303

PARI Ili W HQ ARE THESE 5Cül JN DREI 57 305

Chapter 11: Misconduct in Banking: Governance and the Board of Directors 307
Duc Duy Nguyen, /ens Hagendorff, and Arman Eshraghi
11.1 lntroduction 307
11 2 1iteratme Review 31 1
11.3 Research Design 312
11.3.1 Data 312
11 .3.2 Empirical Design 313
11 .3.3 Variables 314
11.4 Empirical Results 316
11 4 1 Maio Results 316
11 .4.2 Results for D ifferent Classes of Enforcement Actions 320
11.4.3 Does Better Board Quality Alleviate Shareholder Wealth Losses? 323
11.5 Conclusion 323
Refereoces 32 5
Cha ter 12: Misconduct and Fraud b lnvestment Mana ers 327
Stephen C. Dimmock, Joseph D. Farizo, and William C. Cerken
12 1 lotroductioo 327
12 2 Related R==..u....- - - - - - - - - - - - - - - - - - - - - - -..J.L.J..
12.3 The lnvestment Advisers Act of 1940 and Mandatory Disclosures 331
12.4 Data 332
12 .4.1 lnvestment Fraud 332
12 4 2 Earm ADV Data aod Variables 337
12.5 Predicting Fraud and Misconduct 340
12.5.1 Predicting Fraud by lnvestment Managers 340
12 .5 .2 lnterpreting the Predictive Content of the Models 345
12..5 3 K-Eold..Cross-Validatia □Jes -U-----------------=-'-'
12.6 Predicting the ln itiation vs. the Continuance of Fraud 347
12.7 Firm-Wide Fraud vs. Fraud by a Rogue Employee 349
12.8 Out-of-Sample Prediction and Model Stability 351
12.9 Pol icy lmplications and Conclusions 352
References 355

Alexander42 l 77 1_ftoc.indd 9 23-~~-2020 22:13:26 li ~ auta1a s

 X CONTENTS

Chapter 13: Options Backdating and Shareholders 359

[ohan Sulaeman and Cennaro Bernile
13 1 l □ trad11ctia□ 359
13 .2 Stock Return Patterns around Option Grants 360
13.3 The Backdating Practice 361
13 .4 Media Coverage, Restatement, and I nvestigation 362
13.5 Stock Market Reaction to Public Revelations of Backdating 363
13.6 lnvestor Reaction to (and Anticipation of) Public Revelations 364
13 .7 Other Types of Misbehaviour Related to Option Grants 365
13.7.1 Forward Dating 365
13.7.2 Selective Disclosure 366
13.7.3 Option Exercise Backdating 366
13.7.4 lndependent Director Backdating 366
13.8 Connections with Questionable Practices by Corporate Executives
and Other Agents 366
13 9 Caocl11siao 367
References 368

Chapter 14: The Strategic Behaviour of Underwriters in

Valu ing IPOs 371
Stefano Paleari, Andrea Signori, and Sílvio Vismara
14.1 ValuinglPOs 371
14 2 The lJnderwriter's Incentives in theValiiatiao of iPOs 373
14.3 Literature Review 374
14.4 Sample, Data, and Methodology 376
14.4.1 Sample and Data 376
14.4.2 Alternative Selection Criteria of Comparable Fi rms 380
14.4.3 Valuation Bias and IPO Premium 380
14 5 Res1dts 381
14.5.1 Algorithm ic Selections 381
14.5.2 Affiliated and Unaffiliated Analysts 386
14.5.3 Underwriters' Selection of Comparable Firms Pre- vs.
Post-lPO 390
14.5.4 Pre- vs. Post-lPO Selections and lndustry Effects 394
14 6 Caocl11sians 396
RefeD=n...s:;.i_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _L:LJ._

Chapter 15: Governance of Financial Services Outsourci ng: Managing

Misconduct and Third-Party Risks 399
Joseph A. McCahery and F. Alexander de Roode
1 5 1 lotrod1 ,ctian 399
15 .2 The Four Components in Outsourcing 402
15.2.1 Efficientüutsourcing 402
15.2.2 The Four-Factor Governance Model 404
15.2.3 Misconduct in Outsourcing and the Ability of Financial
lnstitutions to Monitor 407

Alexander42 177 1_ftoc.indd 10 23-~~-2020 22:13:26 li ~ auta1a s

 Contents XI

15.3 The lnteraction between Contracting and Monitoring 408

15.3.1 Characterization of Financial lnstitutions 409
15.3.2 Risks in Outsourcing Services 412
15.4 Governance Mechanisms to Detect Misconduct in Financial
Outsourcing 413
15.4.1 Screening and Detection 414
15 5 Condusion 416
References 417

.eARLJ',' DETECTION AND-5.UfilLEI.L.LAN.Cf OE FINANCIAL

MISCONDUCT 423
Chapter 16: ldentifying Security Market Manipulation 425
Mike Aitken, Ann Leduc, and Shan li
16 1 lotrad11Ctiao 425
16.2 Background Legislation 427
trai ia 427

16.2.3 Hong Kong 428

16 2 4 Canada 429
16.2.5
Singapore 430
16.2.6
Malaysia 430
16.2.7 NewZealand 431
16.3 Attributes of Manipulation 431
16.3.1 How Traders Minimize the Resources Needed for Manipulative Trading 432
16.3.2 Difficulties in Determining Whether Trading Behaviour
Is Manipulative 433
16.3.3 Surveillance Systems 434
16.4 Detection Algorithms 436
16 5 CaocltJsiao 439

Chapter 17: The Analytics of Financial Market Misconduct 441

Ai Deng and Priyank Candhi
J Z 1 lntr_o_duc.ti.oLJ___ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _......:::r..::u...
17.2 Financial Economic Analysis 446
17 .2.1 Benchmarking to Historical or Past Data 447
17.2.2 Benchmarking to Alternate Proxies 451
17.2.3 Benchmarking to a Model 454
17.3 Quantitative Technigues 456
17 .3 .1 The Princ ipies of Fraud Detection 457
17.3.2 Popular Supervised LearningTechniques for Fraud Detection 458
17.3.3 Popular Unsupervised Learning Techniques for Fraud Detection 460
17.3.4 Dynamic Misconduct Detection 462
1 Z 4 CaocltJsiao 464
References 466

Alexander42 177 1_ftoc.indd 11 23-~~-2020 22: 13:26 li ~ auta1a s

 XII CONTENTS

Chapter 18: Benford's Law and lts Appl ication to Detecting Financial
Fraud and Manipulation 473
Christina Bannier, Corinna Ewelt-Knauer, Johannes Lips, and Peter Winker
18.1 lntroduction 47 4
18.2 Benford's Law and Generalizations 476
18.2.1 The Basic Principie of Benford's Law 476
18.2.2 lllustration of Benford's Law 477
18.2.3 Testing for Conformity with Benford's Law 478
18.2.4 Considering Further Digits with Benford's Law 480
18...2..5_Wben.Jla Data Caofarm ta B.~eo~f~n~cd~'sI,
~a~w
~?~ - - - - - - - - - -~48~2
18.2 .6 Limitations of Using Benford's Law for ldentification of Manipulations 483
18.2.7 General izations of Benford's Law for ldentification of Manipulations 484
18.3 Usage of Benford's Law for Detecting Fraud and Deviant Behaviour 485
18.3.1 Forensic Accounting in the Context of Auditing, Internai Contrai
Systems, and Taxation 486
18.3 .2 Finance 487
18.3.3 Surveys and Research 490
18.4 A Case Study: Benford's Law and the LI BOR 491
18.5 Policy lmplications 498
18.6 Summary, Limitations, and Outlook 498
References 499
18.A Appendix 504

PARTY REGULAJION ANO ENFORCEMENJ 505

Chapter 19: The Enforcement of Financ ial Market Crimes in Canada
and the United Kingdom 507
Anita Indica Anand
19 1 lotradiictiao 507
19.2 Existing Scholarship 508
19.3 Comparative Analysis 512
19 3 J Can
19.3.2 The United Kingdom 513
19 4 Refarm 51 5
19.AJ Resamc.e..Alln'-.U-'....,_..,..__ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _____,__15
19.4.2 Principles-Based Regulation 516
19.4.3 Targeted Regulatory Reforms 518
19 5 Caod11siao 520
Refereoces 520
Chapter 20: A Pyramid ora Labyrinth? Enforcement of Registrant
Misconduct Requirements in Canada 527
Mary Condon
20.1 lntroduction 527
20.2 Definitional and lnstitutional Quagmires 529
20.3 The Compliance/Enforcement Continuum 531

Alexander42 177 1_ftoc.indd 12 23-~~-2020 22: 13:26 li ~ auta1a s

 Contents xiii

20.4 Enforcement Options Available to Sanction Registrant Misconduct 5.33.

20.5 Empirical lnformation Available about Registrant Misconduct in Canada 5..3.5.
20..5.J__Criminalinf.orcement 535.
20.5.2 CSA Non-Criminal Enforcement 536
20.5.3 Director's Decision Data in Ontario 537
20.5.4 SRO Enforcement 538
20.6 Analysis 538

Chapter 21; Judi cial Local Protectionism anel Home Court Bias
in Corporate Litigation 541
Michael Firth, Oliver M. Rui, and Wenfeng Wu
21.1 lntroduction 542
21 . ~ itutional Baclsg_li.l~nç!_ 5M:
21.2.1 Decentra iization and Local Protectiooism 5M:
21 .2.2 Judicial lndependence 5A5.
21.2.3 The Heterogeneity of the Legal Environment across Regions 548
21 .3 Empirical Evidence 548
21 .3.1 Sample 549
21.3.2 Basic Statistics 550
21.3.3 The Wealth Effect for Defendants and Plaintiffs around the
Filing Announcements at Different Courts 556
21 .3.4 The lmpact of Court Location on the Wealth Effect 560
21.3.5 Regression Analysis of the Wealth Effects from a Filing Announcement 560
21.3.6 Heckman Two-Step Analysis of Sample Selection Bi as 567
21.3.7 The lmpact of Court Location on the Likelihood to Appeal 573
2'1.3.8 Sensitivity Tests .52fi
21 4 Canclusiao 5.29.
References 580

Jruiex 5.83.

Ale.xander42 l 771_ftoc.indd 13
 An Overview of Market Manipulation 19

should not be prohibited as manipulative; but fictitious trades (e.g. trades in which the
buyer and seller is the sarne person) and spreading false information shou ld be clas-
sified as fraud. Their reasoning is that (i) purely trade based manipulation is unlikely
to be successfu l; and (ii) rules that prohibit manipulation deter some legitimate trad-
ing. Dozens of subsequent prosecution cases that show purely trade-based market
manipulation is not justa theoretica l possibility, but can be highly profitable, casts
doubt over their reasoning.
Thel (1994) delivers a strong rebuttal. Based on evidence in the economics lit-
erature Thel argues that manipulation is easier to accomplish than Fischel and Ross
claim and provides some striking examples. Thel points out that manipulators can
sometimes contrai prices with trades and in doing so profit either from pre-existing
contracts that are contingent on prices, or by inducing other market participants to
trade at manipulated prices.
Thel uses the term "manipulation" to mean trading undertaken with the intent of
increasing or decreasing the reported price of a security. Cherian and Jarrow (1995)
define manipulation as trading by an individual (or group of individuais) in a man-
ner such that the share price is influenced to his advantage. Many subsequent papers
implicitly use the term "market manipulation " to refer to trading strategies or actions
taken to influence the price to one's advantage.
ln forming their view about what forms of trading should or should not be pro-
hibited, Kyle and Viswanathan (2008) consider the welfare effects of trading. They
propose that trading strategies should on ly be illegal if they undermine economic effi-
ciency both by decreasing price accuracy and reducing liquidity. Unless both of these
conditions are satisfied, the trading strategy is not unambiguously social ly harmful
and therefore, according to Kyle and Viswanathan, should not be prohibited.

2.3 A Taxonomy of the Types of Market Manipulation

The term "market manipulation" encampasses a wide variety of different strate-
gies. To map out the relations between the various forms of market manipulation,
Figure 2.1 provides a taxonomy of the most common types of market manipulation.
The taxonomy is an expanded version of the framework set out in Putnir,is (2012),
accounting for recent forms of market manipulation. This section first describes the
two broad leveis on which manipulation can be grouped into categories and then
defines the individual techniques.

2.3.1 Categories of Market Manipulatio n

At the broadest levei, manipulation can be divided into four categories: runs, contract-
based or benchmark manipulations, spoofing, and market power techniques. Within
these groups, manipu lation can be further broken down according to the main mech-
anism used to facilitate the manipulation: trade-based, information-based, action-
based, submission-based, and order-based forms. And within these mechanisms are a
number of individual manipulation techniques. These categories and mechanisms are
not mutually exclusive and there are also hybr id manipulation strategies that combine
severa! of the individual techniques or elements of the techniques.

Alexander42 177 1_c02.indd 19

 20 CORRUPTION ANO FRAUD IN FINANCIAL MARKETS

'Painting the tape'

'Wash trades'

Trade-based 'Matched orders'

'Pools'

Runs 'Momentum ignition'

('Pump-and-
dump' /'Bear
raids') ' Hype-and-dump'
Info-based ---[
'Slur-and-dump'

Action-based Action-based

' Marking the close/open/set'

Contract-based - { Trade-based
manipulations/
-[ 'Capping/pegging'
Benchmark
manipulations
Market Subm1ssion-
manipulation 'Benchmark rigging'
based

'Layering'

'Advancing the bid/offer'

Spoofing/
- - - Order-based
Order-based 'Quote stuffing'

'Abusive liquidity detection' /

' pinging' /'phishing '

'Comer'
Market power
techniques
- - - Trade-based
--[ 'Squeeze'

Figure 2.1 Taxonomy of manipulation techniques.

ln the first category of market manipulation, "runs", the manipulator takes either
a long ora short position in a stock, inflates or deflates the stock's price while attract-
ing other traders, and fina lly reverses his position at the inflated or deflated price.
While runs are commonly observed in the stock market, in principie they could be
conducted in any financial security. Runs that involve the manipulator taking a long
position and then inflating a stock's price are often referred to as "pump-and-dump"
manipulation, whereas the reverse strategy of first taking a short position and then
manipulating the price downwards is known as a "bear raid" . The stock "pumping"
or "raiding" can take anywhere from a matter of seconds to severa! years and involve
techniques such as spreading rumours, executing wash trades, and coordinated pool-
ing by severa! manipulators.

Alexander42 177 l_c02.indd 20

 An Overview of Market Manipulation 21

A feature of runs is that the manipulator profits directly from the manipulated mar-
ket by exploiting other investors that buy at inflated prices or sell at depressed prices.
The main challenge in conducting this form of manipulation is having to induce other
market participants to buy or sell the manipulated security.
The second category of market manipulation, "contract-based manipulation" or
"benchmark manipulation" or "reference rate manipulation" contrasts with a run in
that it involves the manipulator profiting from a contract or market that is externai to
the manipulated market. For example, a manipulator might take a position in a deriva-
tives contract and then manipulate the underl ying stock price to profit from the deriv-
atives position. Ora manipulator might have positions in securities whose cash flows
or prices are determined by a financial benchmark (e.g. floating rate loans, swaps,
futures) and then manipulate the benchmark to profit from the externai positions.
An important difference between runs and contract-based or reference rate
manipulation is that the latter does not require the manipulator to induce others to
trade at manipulated prices. This category of manipulation therefore tends to be more
mechanical.
The third category of manipulation is "spoofing" and other order-based tech-
niques. General ly speaking, spoofing involves submitting orders to a market with
the intention to cancel the orders before they execute. The defining feature of this
category is that orders play a central role in the manipulation, although they are often
accompanied by trades as part of the strategy. Although spoofing can be performed
manual ly, this type of market manipulation naturally lends itself to being imple-
mented by a computer algorithm that manages the entry, amendment, and cancel-
lation of the manipulative orders. Automating the strategy allows it to be scaled and
repeated many times. While each instance of spoofing might make only a small profit,
repetition can be used to accumulate a sizable profit. This category of manipulation
is often implemented at relatively high frequencies, with positions typically opened
and closed intraday.
Dueto the growth in algorithmic and high-frequency trading, spoofing and order-
based market manipulation techniques have gained increasing attention in recent
years. Regulators have brought a number of prosecution cases against such manip-
ulation, lawmakers have made legislative amendments to account for this form of
manipulation, and market/surveillance system operators have developed algorithms
to detect such strategies.
The fourth broad category of manipulation techniques involves the manipulator
exploiting market power by, for example, taking a controlling position in the supply
of a security. Like contract-based manipulation, market power techniques are more
mechanical in nature than runs. However, they are similar to runs in that the manipu-
lator profits by exploiting participants of the manipulated market.
Within the four broad categories, manipul ation techniques can be further grouped
according to the five main mechanisms used in their implementation. Allen and
Gale (1992) define three of the five techniques: trade-based, information-based, and
action-based techniques, which I augment by adding order-based and submission-
based techniques. Trade-based manipulation involves influencing the price of a finan-
cial instrument through trading. ln information-based manipulation, a manipulator

Alexander42 177 1_c02.indd 21

 24 CORRUPTION ANO FRAUO IN FINANCIAL MARKETS

(x) Marking the open

Is similar to marking the close, but involves influencing the opening price rather than
closing price.

(xi) Marking the setor banging the set

Is also similar to closing price manipulation but involves trading to influence a par-
ticular reference rate set, which might not be at the close or open. For example, in a
24-hour market, such as the global foreign exchange market, key reference rates are
set at predetermined times, which are not necessarily the open or close of trading.

(xii) Pegging and capping

Refer to placing orders that effectively prevent a price from moving up or down
beyond a particular threshold, thereby setting a floor or ceil ing price. This is often
done to ensure a derivatives contract expires in or out of the money, or that a refer-
ence rate or benchmark does not move in an unfavourable direction. For example, a
manipulator that wants to prevent a stock price from moving below $5.00 might place
large buy orders ata price of $5.00 to prevent any trades from occurring below this
price. Pegging and/or capping can be implemented dynamically with a threshold that
is adjusted by the market manipulator in response to changing conditions. For exam-
ple, if the manipulator in the above example encounters very heavy selling against his
$5.00 order, he might retreat (amend his buy arder) to a price of, say, $4.95 and seek
to prevent further declines. Such a retrea,ting defence can be used to reduce the extent
of adverse price movements.
Benchmarks that are based on submissions of a select group of benchmark pan-
ellists (e.g. the mechanism used for setting LIBOR at the time of the recent LIBOR
manipulations) rather than an actual underlying market can be manipulated through
submission-based manipulation techniques:

(xiii) Benchmark rigging

Achieved through submission-based manipulation involves making false or mislead-
ing submissions as inputs to a financial benchmark calculation. When there are many
parties whose submissions are used in calculating the benchmark, and/or when the
benchmark calculation involves a trimmed mean whereby the highest and lowest
submissions is excluded from the average, submitters might col lude to influence the
benchmark as the impact of one individual submission can be smal l. 14
Spoofing and other order-based manipulation strategies, which are often but not
always implemented via a computer algorithm, include the fol lowing techniques.

(xiv) Layering
Is a form of spoofing that involves placing one or severa! orders on one side of a
visible limit arder book (the bidor the offer side) atone or severa! price steps to cre-
ate a false or misleading impression with respect to the real demand or supply in a
given security. Being a form of spoofing, the manipulator's intention is for the orders
not to execute and therefore most spoofing orders result in cancellations. The term

14
For example, at the time of the recent manipu lations, USO LIBOR was based on submissions from
around 20 banks, taking an average that excludes the highest one-quarter and lowest one-quarter of the
submissions.

Alexander42 177 1_c02.indd 24

 An Overview of Market Manipulation 25

"layering" stems from the fact that often the manipulator's orders are placed in lay-
ers across several price steps or on top of one another ata given price step. Layering
could also be done with a single large orderr. Collectively, the layering orders often
represent a substantial proportion of the orders on one side of the limit arder book.
They are often placed close to the best quotes at the time, and sometimes dynami-
cally amended (or cancelled) as the market moves closer to the layering orders to
avoid execution. Layering can be used to help obtain a more favourable execution
price for a trade that the market participant wants to execute for an unrelated reason.
For example, a trader wanting to sel l out of a long position could use layering buy
orders to temporarily inflate the market price by misleading other market participants
and then sell the actual position at an inflated price. However, layering can be, and
often is, used repeatedly in a cycle together with other orders that profit from distorted
prices. A typical layering cycle is as follows: (i) place a small sell order ator near the
best ask price, (ii) layer the bid side of the order book until the market moves up and
the smal l sei! arder executes, (i ii) cancel the layering bid orders and repeat the above
steps in the opposite direction.

(xv) Advancing the bid/offer

lnvolves placing a buy or sell order within the prevailing best quotes for the purpose
of setting a new best bid or best offer price. These orders are often not intended to
execute and are therefore much like an aggressive layering strategy. Advancing the bid
or offer can be used to give other market participants a false signal about the security's
demand or supp ly and therefore can be used in conjunction with other techniques
to cause a run. Alternatively, advancing the bid or offer can be used in manipulating
a benchmark or reference price. The bid and ask quotes, or their half-way point (the
"midquote") are often used as reference prices for pegged dark/hidden orders, dark
pools, crossing systems, and some alternative trading venues. Therefore, advancing
the bidor offer on a transparent reference market can be dane as part of a strategy that
exploits orders or trading venues that reference the manipulated market's quotes. For
example, a manipulator that detects the presence of a dark sel l order pegged to the
midquote can place an arder that lowers the offer and thus midquote on the transpar-
ent reference market, then buy in the dark at the artificial ly low midquote, and finally
cancel the arder that lowered the offer.

(xvi) Quote stuffing

lnvolves jamming a financial market's infrastructure, such as the matching engine that
processes incoming order messages or the systems that disseminate market data to
participants, by submitting an enormous number of arder submission, amendment,
and cancellation messages in a short period of time (typically, second or sub-second
hori zons). By overwhelming the financial market's processing capacity, quote stuffing
can increase latency for anyone trying to submit legitimate orders to the market, or
deny other participants timely information on the actual state of the orders in a mar-
ket. lt appears possib le to slow down an entire market (such as the NYSE) by quote
stuffing on ly one or a subset of the stocks traded on that market. Thus, quote stuff-
ing allows a manipulator to control the latency of markets to their advantage. While
intense bursts of arder submissions and cancellations that appear consistent with
quote stuffing are frequently observed in today's markets (e.g. Egginton et ai., 2016),

Alexander42 177 1_c02.indd 25

 26 CORRUPTION ANO FRAUD IN FINANCIAL MARKETS

it is less clear exactly how this manipulation technique is used in conjunction with
other strategies to make a profit. Quote stuffing has nevertheless been acknowledged
as a form of market manipulation in severa! jurisdictions, in particu lar if the orders are
submitted with the intention of being cancelled before execution. 15

(xvii) Abusive liquidity detection, pinging, or phishing

lnvolves submitting small probing orders for the purpose of detecting hidden or latent
liquidity. While such orders can result in trade executions, that is not their dominant
purpose. Rather, their dominant purpose is to gather information about other market
participants' trading intentions, which is then exploited to make a profit (often at
the expense of those traders w hose intentions were inferred). For example, to gauge
whether there are buyers, sel lers, or neither buyers nor sel lers waiting to trade in a
dark pool or trade w ith dark orders pegged to the midquote, a trader might submit
a smal l probing order (for as little as one share if allowed by the trading protocol) to
buy at the midquote. lf the order executes, the trader knows that others are trying to
sel l at the midquote. lf the order does not execute, the trader might cancel the order
and place a smal I probing order to sei! at the midquote. lf that order executes, the
trader knows there are buyers wanting to buy at the midquote. And if that order does
not execute, the trader knows that there are no resting dark orders waiting to trade
at the midquote. The trader then uses the information about others' tradi ng inten-
tions to earn a profit. This often involves trading in the direction of others' intentions,
but before them, exploiting the impact that their future orders will have on market
prices. Such techniques have gained popularity with the growth in algorithmic and
high-frequency trading and regu latory responses are still emerging. So far, Canadian
and European regulators have expressed views that such trading is considered market
manipulation. 16
The final category of market manipulation exploiting market power involves two
trade-based manipulation techniques, w hich have been used in markets for decades
if not centuries.

(xviii) Corners and squeezes

Are techniques in which the manipulator secures a controlling position in the supply
of an asset and/or a derivative contract. The manipulator then uses this position to
manipulate the price by exploiting investors that need the underlying asset to close
out short positions or deliver on a derivative contract.

2.4 Research on Market Manipulation

This section provides an overview of the academic research on market manipulation.
lf focuses on identifying what we know about market manipulation from the litera-
ture, as well as highlighting deficiencies and future research directions. The section

15
For example, the lnvestment lndustry Regu latory Organization of Canada (IIROC) Rules Notice Guid-
ance Note 13-0053 confirms IIROC's position that quote stuffing is "considered a manipulative and decep-
tive trading practice". Similarly, the US Commodity Futures Trading Commission (CFTC) lnterpretative
Guidance and Policy Statement (RIN 3088-AD96) concerning the anti-spoofing provisions introduced into
US legislation with the Dodd-Frank Act provides quote stuffing as one of the examples of trading that is
considered spoofing and therefore prohibited.
16
See IIROC Ru les Notice Guidance Note 13-0053 and ESMA Final Repor! 2015/224.

Alexander42 177 1_c02.indd 26

 An Overview of Market Manipulation 27

starts with an overview of the theoretical literature, followed by the empirical litera-
ture. lt then summarizes what we know and don't know about market manipulation
based on the literature and points out some future research directions.

2.4.1 Theoretical Literature

The theoretical market manipulation I iterature provides insights about the conditions
under which manipulation is possible and profitable. The literature, wh ich spans the
past 30 years, is fairly extensive, particularly regarding trade-based manipulation,
and to a lesser extent information-based techniques. Although action-based manipu-
lation is not explicitly studied in the literature, it can often be viewed as a type of
information-based strategy because the manipulator's actions create a false signa l
similar to false information. Consequently, many of the findings about information-
based manipulation are relevant for action-based manipulation. There is a deficiency
when it comes to theoretical models of order-based manipulation techniques such as
spoofing and layering. This deficiency is in part because these techniques have only in
recent years gained significant regulatory attention, but also because of the difficulties
in forming tractable and realistic limit order book models.

(i) Trade-based manipulation

Early theoretical trade-based manipulation literature establishes very general condi-
tions under which pure trade-based manipulation in a single market (e.g. a series of
buys followed by a series of sei Is) is and is 111ot profitable. Fischel and Ross (1991 ),
among others, argue that trade-based manipu lation is not possible in an efficient mar-
ket. Jarrow (1992), Cherian and Kuriyan (1995) and Cherian and Jarrow (1995) build
on the model of Hart (1977) and derive conditions under which trade-based manipu-
lation is not possible. ln Cherian and Kuriyan's model, manipulation is not possible
with rational agents when price responses to trades are symmetric. Jarrow demon-
strates that a sufficient condition to exclude market manipulation strategies is that the
price response function depends only on a trader's aggregate stock holdings and not
on his past sequence of trades, in other words, when prices do not exhibit momen-
tum. Huberman and Stanzl (2004) demonstrate that uninformed trading strategies that
generate infinite expected profits (in effect, trade-based market manipulation strate-
gies) are ruled out when the price impacts of trades are time independent and linear.
Empirically, price impacts are not linear, nor are they time-invariant, nor are they
always symmetrical. Thus, the conditions shown by theory to be necessary to rule out
profitable trade-based market manipulation are not satisfied in practice, suggesting
that some trade-based manipulation strategies should be profitable in real markets.
Actual trade-based strategies can exploit non-linearity, time variation, or asymmetry
in price impacts.
Many theoretical studies seek to prove that trade-based manipulation is possible
in variations of the seminal models of Kyle (1985) and Glosten and Milgrom (1985).
For example, Allen and Gorton (1992) argue that the natu ral asymmetry between
liquidity purchases and liquidity sales gives rise to profitable trade-based man ipula-
tion. lf liquidity-motivated sales are more likely than liquidity-motivated purchases,
buy orders are more informed on average and therefore have a larger effect on prices.
ln a Glosten and Milgrom (1985) model, thi s asymmetry allows an uninformed
manipulator to generate a profit by executing a series of buys to bid up the price and

Alexander42 177 1_c02.indd 27