Keywords

Author name

Accounting, Organizations and Society

Volume

Issue

Pages

Accounting, Organizations and Society

SUPPORTS OPEN ACCESS

Articles in press All issues About the journal Follow journal

Volume 56
Pages 1-84 (January 2017)

Previous vol/issue Next vol/issue

7 selected: Expand all article previews Download PDFs Export

Full text access

Editorial board and publication information
Page IFC

Download PDF

Research article Full text access

Mandatory management disclosure and mandatory independent audit of internal controls: Evidence of
configural information processing by investors
Khim Kelly, Hun-Tong Tan
Pages 1-20

Download PDF Article preview

Research article Full text access

Tournament group identity and performance: The moderating effect of winner proportion
Khim Kelly, Adam Presslee
Pages 21-34

Download PDF Article preview

Special themed section on Financial Accounting as Social and Organizational Practice: Exploring The Work of
Financial Reporting

Editorial Full text access

Themed Section on Financial Accounting as Social and Organizational Practice: Exploring the work of financial
reporting
Keith Robson, Joni Young, Mike Power
Pages 35-37

Download PDF

Research article Full text access

Interlingual translation of the International Financial Reporting Standards as institutional work
Jaana Kettunen
Pages 38-54

Download PDF Article preview

Research article Full text access

Representing the market perspective: Fair value measurement for non-financial assets
Richard Barker, Sebastian Schulte
Pages 55-67

Download PDF Article preview

Research article Full text access

Relative reliability and the recognisable firm: Calculating goodwill impairment value
Jari Huikku, Jan Mouritsen, Hanna Silvola
Pages 68-83

Download PDF Article preview

ISSN: 0361-3682
Copyright © 2018 Elsevier Ltd. All rights reserved

About ScienceDirect Remote access Shopping cart Contact and support Terms and conditions Privacy policy

Cookies are used by this site. For more information, visit the cookies page.
Copyright © 2018 Elsevier B.V. or its licensors or contributors. ScienceDirect ® is a registered trademark of Elsevier B.V.
 EDITOR IN CHIEF
Keith Robson, HEC Paris, Paris, France
EDITORS
R.J. Bloomfield R. Krishnan M. Shields
Cornell University, Ithaca, USA Michigan State University, East Lansing, Michigan State University, East Lansing, USA
Michigan, USA H. Tan
W.F. Chua
University of Sydney, New South Wales, Australia Nanyang Technological University, Singapore
P.B. Miller
London School of Economics, London, UK K. Trotman
C. Chapman University of New South Wales, Sydney, Australia
University of Bristol, Clifton, UK
M. Power M. Williamson
D.J. Cooper London School of Economics and Political Science, University of Illinois at Urbana-Champaign, Champaign,
University of Alberta, Edmonton, Canada London, UK Illinois, USA

EDITORIAL BOARD
C. Agoglia J. Fisher D. Larcker P. Quattrone
University of Massachusetts at Amherst, Indiana University, Bloomington, USA Stanford University, Stanford, USA University of Edinburgh, Edinburgh,
Amherst, Massachusetts, USA T.J. Fogarty R. Libby Scotland, UK
T. Ahrens Case Western Reserve University, Cleveland, Cornell University, Ithaca, USA
A. Richardson
United Arab Emirates University, Al-Ain, USA A.M. Lillis
United Arab Emirates University of Windsor, Windsor, Canada
C. Free University of Melbourne, Melbourne, Australia
S. Anderson University of New South Wales, Sydney, M. Lipe J. Roberts
University of California at Davis, Davis, USA New South Wales, Australia University of South Carolina, Columbia, South University of Sydney, Darlington, Australia
M. Annisette Y. Gendron Carolina, USA S. Salterio
York University, Toronto, Canada University Laval, Quebec, Canada
J. Luft Queen’s University, Kingston, Ontario,
M. Artz I. Grabner Michigan State University, East Lansing, USA Canada
Frankfurt School of Finance and Management, University of Maastricht, Maastricht,
K. Lukka K.L. Sedatole
Frankfurt, Germany Netherlands
Turku School of Economics, Turku, Finland
R. Balakrishnan C. Graham Michigan State University, East Lansing, USA
V. Maas
University of Iowa, Iowa City, Iowa, USA York University, Toronto, Canada G. Sprinkle
Universiteit van Amsterdam, Amsterdam,
A. Bhimani J. Hales Netherlands Indiana University, Bloomington, USA
London School of Economics, London, UK Georgia Institute of Technology, Atlanta,
D. MacKenzie W. Tayler
J. Birnberg Georgia, USA
University of Edinburgh, Edinburgh, UK Brigham Young University, Provo, Utah, USA
University of Pittsburgh, Pittsburgh, USA M. Hall
B. Malsch S. Toms
J.C. Bol London School of Economics, London, UK
Queens University, Canada, Kingston, Canada
Tulane University, New Orleans, Louisiana, USA J. Hammersley University of Leeds, Leeds, UK
C. McWatters
R. Bryer University of Georgia, Atlanta, USA K. Towry
University of Ottawa, Ontario, Canada
University of Warwick, Coventry, UK L. Hannan Emory University, Atlanta, USA
A. Mennicken
B. Carruthers Georgia State University, Atlanta, USA
London School of Economics, London, UK J. Unerman
Northwestern University, Evanston, Illinois, USA R. Hatfield
K. Merchant Royal Holloway, University of London,
C. Carter University of Alabama, Tuscaloosa, Alabama,
USA University of Southern California, Los Angeles, London, England, UK
University of Edinburgh, Edinburgh, Scotland, USA
UK F. Hodge W.A. Van der Stede
University of Washington, Seattle, Washington, M. Messner London School of Economics, London, UK
C. Chen Leopold-Franzens-Universität Innsbruck,
USA
University of Illinois at Urbana-Champaign, Innsbruck, Austria H. Vollmer
V.B. Hoffman
Champaign, Illinois, USA University of Leicester, Leicester, UK
University of Pittsburgh, Pittsburgh, USA J. Meyer
R. Chenhall Stanford University, Stanford, USA M. Walker
P. Hopkins
Monash University, Melbourne, Australia
Indiana University, Bloomington, Indiana, USA F. Moers University of Manchester, Manchester, UK
C. Cho University of Maastricht, Maastricht,
M. Holzhacker S.P. Walker
ESSEC Business School, Clergy Pontoise Netherlands
Michigan State University, Michigan, USA University of Edinburgh, Edinburgh,
Cedex, France
C. Ittner D. Moser
J. Choi Scotland, UK
University of Pennsylvania Wharton School, University of Pittsburgh, Pittsburgh, USA
University of Pittsburgh, Pennsylvania, USA M. Wouters
Philadelphia, USA J. Mouritsen
C. Cooper Copenhagen Business School, Copenhagen, Karlsruhe Institute of Technology (KIT),
J. Jollineau
University of Strathclyde, Glasgow, UK Denmark Karlsruhe, Germany
University of San Diego (USD), San Diego,
C. Dambrin California, USA M. Nelson J. Young
ESCP Europe, Paris, France Cornell University, Ithaca, USA
K. Kadous University of New Mexico, Albuquerque,
H.C. Dekker Emory University, Atlanta, USA B. O’Dwyer
VU University, Amsterdam, Netherlands USA
L. Koonce Universiteit van Amsterdam, Amsterdam,
S. Dikolli S. Young
University of Texas at Austin, Austin, USA Netherlands
Duke University, Durham, USA Lancaster University, Bailrigg, UK
J. Kuang M. Peecher
W. Espeland Georgia Institute of Technology, Atlanta, University of Illinois at Urbana-Champaign, M. Zimbelman
Northwestern University, Evanston, USA Georgia, USA Urbana, Illinois, USA Brigham Young University, Provo, USA

Publication information. Accounting, Organizations and Society (ISSN 0361-3682). For 2017, volumes 56–63 (8 issues) are scheduled for publication. Subscription prices are available upon request from the Publisher
or from the Elsevier Customer Service Department nearest you or from this journal’s website (http://www.elsevier.com/locate/aos). Further information is available on this journal and other Elsevier products through
Elsevier’s website: (http://www.elsevier.com). Subscriptions are accepted on a prepaid basis only and are entered on a calendar year basis. Issues are sent by standard mail (surface within Europe, air delivery outside
Europe). Priority rates are available upon request. Claims for missing issues should be made within six months of the date of dispatch.
Advertising information: If you are interested in advertising or other commercial opportunities please e-mail Commercialsales@elsevier.com and your inquiry will be passed to the correct person who will respond to
you within 48 hours.
Orders, claims, and journal inquiries: please contact the Elsevier Customer Service Department nearest you: St. Louis: Elsevier Customer Service Department, 3251 Riverport Lane, Maryland Heights, MO 63043,
USA; phone: (877) 8397126 [toll free within the USA]; (+1) (314) 4478878 [outside the USA]; fax: (+1) (314) 4478077; e-mail: JournalCustomerService-usa@elsevier.com. Oxford: Elsevier Customer Service
Department, The Boulevard, Langford Lane, Kidlington, Oxford OX5 1GB, UK; phone: (+44) (1865) 843434; fax: (+44) (1865) 843970; e-mail: JournalsCustomerServiceEMEA@elsevier.com.
Tokyo: Elsevier Customer Service Department, 4F Higashi-Azabu, 1-Chome Bldg, 1-9-15 Higashi-Azabu, Minato-ku, Tokyo 106-0044, Japan; phone: (+81) (3) 5561 5037; fax: (+81) (3) 5561 5047; e-mail:
JournalsCustomerServiceJapan@elsevier.com. The Philippines: Elsevier Customer Service Department, 2nd Floor, Building H, UP-Ayalaland Technohub, Commonwealth Avenue, Diliman, Quezon City,
Philippines 1101; phone: (+65) 6349 0222; fax: (+63) 2 352 1394; e-mail: JournalsCustomerServiceAPAC@elsevier.com.
∞ The paper used in this publication meets the requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper)
USA mailing notice: Accounting, Organizations and Society (ISSN 0361-3682) is published 8 issues per year (in January, February, April, May, July, August, October and November) by Elsevier Ltd
(The Boulevard, Langford Lane, Kidlington, Oxford OX5 1GB, UK). Periodical postage paid at Jamaica, NY 11431 and additional mailing offices (not valid for journal supplements). USA POSTMASTER:
Send change of address to Accounting, Organizations and Society, Elsevier Customer Service Department, 3251 Riverport Lane, Maryland Heights, MO 63043, USA.
AIRFREIGHT AND MAILING in USA by Air Business Ltd., c/o Worldnet Shipping Inc., 156-15, 146th Avenue, 2nd Floor, Jamaica, NY 11434, USA.

© 2017 Elsevier Ltd.

 Accounting, Organizations and Society 56 (2017) 1e20

Contents lists available at ScienceDirect

Accounting, Organizations and Society

journal homepage: www.elsevier.com/locate/aos

Mandatory management disclosure and mandatory independent audit

of internal controls: Evidence of configural information processing by
investors*
Khim Kelly a, *, Hun-Tong Tan b
a
University of Central Florida, 12744 Pegasus Dr, Orlando, FL 32816, United States
b
Nanyang Technological University, 50 Nanyang Avenue, Singapore 639798, Singapore

a r t i c l e i n f o a b s t r a c t

Article history: We conduct an experiment where alumni participants from a Canadian accounting and finance under-
Received 13 May 2014 graduate program assume they are in one of four regulatory regimes (manipulated between-subjects)
Received in revised form and make investment potential evaluations for two firms (manipulated within-subjects): a firm
6 December 2016
disclosing no material weaknesses (No-MW disclosure firm) and a firm disclosing material weaknesses
Accepted 11 December 2016
Available online 21 December 2016
(MW disclosure firm) in internal controls over financial reporting (ICFR). We find evidence of configural
information processing. For the No-MW disclosure firm, mandatory (versus voluntary) disclosure of ICFR
material weaknesses and mandatory (versus voluntary) independent ICFR audit are substitutes in
Keywords:
Internal controls over financial reporting
enhancing investment potential evaluations. However, for the MW disclosure firm, neither mandatory
Mandatory management disclosure disclosure nor mandatory audit has any effect on investment potential evaluations. Supplementary ex-
Mandatory independent audit periments with undergraduate participants suggest that the pattern of configural information processing
Configural information processing is a function of participants' knowledge of company disclosure incentives and the assurance value of an
audit, wherein undergraduates with lower levels of knowledge are less able to perceive the effects of
mandatory disclosure and mandatory audit on investment potential evaluations. Our findings have
implications for regulators who are concerned about balancing the costs and benefits of different reg-
ulatory mechanisms.
© 2016 Elsevier Ltd. All rights reserved.

1. Introduction
Management disclosure of material weaknesses in internal
controls over financial reporting (ICFR) is presently a salient aspect
of the disclosure environment in many countries, with variations as
to whether disclosure and an independent ICFR audit are voluntary
or mandatory across countries, companies, and time. United States
* periments to examine whether and why they, acting as investors,
We are indebted to Frances Houston for assisting with the recruitment of study
participants, and University of Waterloo School of Accounting and Finance alumni
and members of the Institute of Certified Public Accountants of Singapore who
participated in our study. We also thank Christine Earley, Vicky Hoffman, and
Jennifer Joe for sharing their experimental instrument with us. Comments from
Chris Agoglia, Wei Chen, Jun Han, Bill Kinney, Lisa Koonce, Terence Ng, Mark
Peecher, James Wainberg, Elaine Wang, Juan Zheng, Bo Zhou, and anonymous re-
viewers on earlier versions of the paper are appreciated. Financial support from the
Singapore Ministry of Education Academic Research Fund Tier 1 (RG3/10/NBS),
United Overseas Bank Endowed Chair, and Social Sciences and Humanities Research
Council of Canada (SSHRC SRG 410-2010-2589) is gratefully acknowledged.
* Corresponding author.
E-mail addresses: Khim.Kelly@ucf.edu (K. Kelly), ahttan@ntu.edu.sg (H.-T. Tan).
mandates both disclosure and audit for large issuers but mandates
only disclosure for small issuers, on the assumption that there is an
incremental benefit of having a mandatory audit that outweighs
the compliance cost for large issuers (i.e., mandatory audit and
mandatory disclosure have complementary effects) (SEC, 2009;
SEC, 2010). Canada decided to implement only mandatory disclo-
sure and indicated that it would consider in the future whether a
mandatory audit would have an incremental benefit that out-
weighs its costs (CSA, 2006). We use Canadian participants in ex-
may consider mandatory management disclosure of ICFR material
weaknesses and mandatory independent ICFR audit to be sub-
stitutory rather than complementary regulatory mechanisms in
terms of impact on the investment potential evaluation of com-
panies. Thus, we examine whether Canadian investors process ICFR
disclosures under different regulatory mechanisms in a configural
manner (i.e., as substitutes rather than as complements).
Our study demonstrates that at least in the context of Canadian
investors evaluating the investment potential of companies,
mandatory disclosure and mandatory audit are perceived to be

http://dx.doi.org/10.1016/j.aos.2016.12.002
0361-3682/© 2016 Elsevier Ltd. All rights reserved.
2 K. Kelly, H.-T. Tan / Accounting, Organizations and Society 56 (2017) 1e20
substitutes and mandatory audit has no incremental effect.
Complying with both mandatory disclosure and mandatory audit is
significantly more expensive than complying with only mandatory
disclosure.1 Understanding when and why mandatory disclosure
and mandatory audit can be substitutes for investors is helpful to
regulators who are concerned about balancing the costs and ben-
efits of different regulatory mechanisms.
We expect that the effects of mandatory disclosure and
mandatory audit on investment potential evaluations depend on
whether a firm disclosed that it has material weaknesses (MW
disclosure) or no material weaknesses (No-MW disclosure). A No-
MW disclosure compared to a MW disclosure is more consistent
with companies' incentives to disclose positive news, and thus may
be perceived as having more potential for bias. As such, mandatory
disclosure and mandatory audit are more likely to have an impact
on perceived reliability and relevance of a No-MW disclosure than a
MW disclosure. Therefore, we conduct our main experiment with a
2
2 (between-subjects)
2 (within-subjects) mixed design, using
alumni of an accounting and finance undergraduate program at a
major Canadian university. We manipulated on a between-subjects
basis mandatory (versus voluntary) management disclosure of ICFR
material weaknesses and mandatory (versus voluntary) indepen-
dent ICFR audit. Participants in each regulatory regime evaluate
two firms which are manipulated within-subjects: a MW disclosure
and a No-MW disclosure.2 We find that mandatory disclosure and
mandatory audit have substitutory rather than complementary
effects on investment potential evaluation of a No-MW disclosure
firm. Specifically, having both mandatory disclosure and mandatory
audit does not incrementally increase investment potential evalu-
ations beyond having each regulatory mechanism alone. With
respect to investment potential evaluation of a MW disclosure firm,
neither mandatory disclosure nor mandatory audit has any effects.
Additional analyses indicate that alumni participants believe
that both mandatory (versus voluntary) disclosure and mandatory
(versus voluntary) audit increases the reliability (i.e., free from er-
ror and bias) and the relevance (i.e., makes a difference to investors'
decisions) of a No-MW disclosure. Alumni participants also believe
that mandatory audit increases the reliability of a No-MW disclo-
sure more than mandatory disclosure, but they do not believe
mandatory audit increases the relevance of a No-MW disclosure
more than mandatory disclosure. This may explain why alumni
participants consider mandatory audit and mandatory disclosure to
be substitutes. If mandating disclosure alone or mandating audit
alone already increases the reliability of the No-MW disclosure
above a threshold level that makes a difference to investment po-
tential evaluations, adding the other regulatory mechanism may
not further increase the relevance of the No-MW disclosure to in-
vestment potential evaluations.
In order to better understand why investors consider mandatory
disclosure and mandatory audit to be substitutes, we further
conduct verbal protocol analyses with additional alumni
1
A survey conducted by the U.S. Securities Exchange Commission (SEC) reported
that the mean total compliance costs was $2.33 million for companies complying
with both Sections 404a (mandatory disclosure) and 404b (mandatory audit) of the
Sarbanes Oxley Act (of which $0.65 million was attributed to the independent ICFR
audit) and $0.34 million for companies complying with only Section 404a
(mandatory disclosure) (SEC, 2009).
2
A No-MW disclosure is more common in practice than a MW disclosure. A
survey by the SEC in the U.S. indicates that only about 22 percent of companies
complying with mandated disclosure under Section 404a disclosed an ineffective
ICFR with material weaknesses, while the remaining 78 percent disclosed an
effective ICFR with no material weaknesses (SEC, 2009). A No-MW (MW) disclosure
is also informationally equivalent to disclosing that the ICFR is effective (ineffective)
because U.S. and Canadian regulations require companies to disclose that their ICFR
is ineffective if there are one or more material weaknesses (CSA, 2009; SEC, 2003).
participations as well as two supplementary experiments with
first-year and third-year undergraduate participants from the same
accounting program as our alumni participants, all with the same
design as our main experiment. Undergraduate participants do not
exhibit the same configural information processing as alumni par-
ticipants. Only mandatory audit but not mandatory disclosure has
effects on investment potential evaluation of a No-MW disclosure
firm for third-year undergraduates, while neither mandatory
disclosure nor mandatory audit has any effects on investment po-
tential evaluation for first-year undergraduates. Alumni and third-
year undergraduates likely know more about the assurance value of
an independent audit compared to first-year undergraduates who
have not taken any auditing courses; and we speculate that this
explain why mandatory audit has effects for alumni and third-year
undergraduates but not for first-year undergraduates. Consistent
with the verbal protocols of additional alumni participants, we also
speculate that alumni, compared to undergraduates, better un-
derstand how mandatory disclosure increases the reliability of
positive disclosures of effective ICFR because they have more
exposure to companies' incentives for opportunistic voluntary
disclosures through their auditing/accounting work experience and
experience analyzing financial performance of firms. Alumni,
through their work experience, may also have more exposure to the
enforcement mechanisms associated with mandatory disclosure
that makes mandatory disclosure more reliable. These knowledge
differences may explain why investment potential evaluations are
affected by mandatory disclosure for alumni but not for third-year
and first-year undergraduates. Finally, verbal protocol analyses of
additional alumni participants suggest that the substitutory effects
are sub-conscious in that they stated that these two mechanisms
have complementary rather than substitutory roles.
Using experiments to examine the effects of mandatory disclo-
sure and mandatory audit on investor judgments complements
prior archival studies that have examined investor reactions under
particular regulatory regimes or between different regimes.3 Fig. 1
Panel A summarizes the various regulatory regimes in the U.S. and
Canada.4 The voluntary disclosure and voluntary audit regime (Cell
1) first occurred in the U.S. prior to the Sarbanes-Oxley Act (SOX),
and in Canada prior to National Instrument (NI) 52-109. The
mandatory disclosure and mandatory audit regime (Cell 4) was
next introduced in the U.S. in 2004 for large issuers under SOX
Sections 404a and 404b. Subsequently, the mandatory disclosure
and voluntary audit regime (Cell 3) came into effect in the U.S. in
2007 for small issuers subject to only Section 404a but not Section
3
For example, Ashbaugh-Skaife et al. (2009) find that firms that disclose Section
302 material weaknesses in “disclosure controls and procedures” (DCP) show an
increase in cost of equity, but cost of equity decreases when such firms subse-
quently disclose no Section 404b material weaknesses in ICFR. Ogneva,
Subramanyam, and Raghunandan (2007) find no direct association between dis-
closures of Section 404b material weaknesses in ICFR and cost of equity. Beneish
et al. (2008) find stronger negative market reactions to Section 302 material
weaknesses in DCP than Section 404b material weaknesses in ICFR. Section 302
(implemented in 2002 which preceded Section 404) required company manage-
ment to evaluate and disclose the effectiveness of DCP, but DCP is distinct from
ICFR. Under Section 302, there is ambiguity over whether disclosure of ICFR ma-
terial weaknesses is mandatory and no independent ICFR audit is required
(Ashbaugh-Skaife, Collins, & Kinney, 2007; Doyle et al., 2007; SEC, 2004). Further,
although some ICFR components will be included in DCP, some companies may
have DCP that exclude ICFR components that pertain to the accurate recording of
transactions and disposition of assets or to the safeguarding of assets (SEC, 2003).
4
Unlike the U.S. and Canada, many countries do not mandate ICFR disclosures
and audits, but instead rely on companies following the principle of “comply-or-
explain” with respect to voluntary codes (e.g., the Turnbull Guidance (2005) in the
United Kingdom and Dutch Corporate Governance Code (2009) in the Netherlands).
Financial Instruments and Exchange Law enacted by Japan Financial Services
Agency (2007) mandates both ICFR disclosures and audits.
 K. Kelly, H.-T. Tan / Accounting, Organizations and Society 56 (2017) 1e20 3

Panel A: Regulatory regimes present in the U.S. and Canada for mandating ICFR disclosure and
mandating ICFR audit and prior research of investor reactions in each regulatory regime

Voluntary audit regime Mandatory audit regime

Voluntary disclosure Pre-SOXa in the U.S. No such regime exists
regime Pre-NI 52-109b in Canada

Cell 1: Current study Cell 2: Current study

Mandatory disclosure SOX Section 404a in the U.S.
regime for small issuersc in 2007
NI 52-109 in Canada in 2008
SOX Sections 404a/404b in the
U.S. for large issuersd in 2004
Asbaugh-Skaife et al. 2009
Beneish et al. 2008
Ogneva et al.2007

Cell 3: Current study Cell 4: Current study

a
The Sarbanes-Oxley Act (SOX) was passed in 2002 in the U.S.
b
National Instrument 52-109 (NI 52-109) was approved in 2008 in Canada.
c
Small issuers are issuers with total market capitalization < $75m.
d
Large issuers are issuers with total market capitalization ≥$75m.

Panel B: Regulatory regimes and conditions examined in current study

Current study
Voluntary audit regime
Four possible settings:
1) Firm chooses to have an ICFR audit and chooses to explicitly Not examined
disclose that it has an ICFR audit (rare in real voluntary audit
regimes)
2) Firm chooses not to have an ICFR audit and chooses to explicitly Not examined
disclose that it does not have an ICFR audit (rare in real voluntary
audit regimes)
3) Firm is silent on whether there is an ICFR audit and chooses to Examined (consistent
have an ICFR audit (rare in real voluntary audit regimes) with real voluntary
4) Firm is silent on whether there is an ICFR audit and chooses not to audit regimes,
have an ICFR audit (most common in real voluntary audit regimes) majority of
participants believed
that the experimental
firm which is silent on
whether there is an
ICFR audit has not
hired an ICFR auditor
– see manipulation
check section for
details)
Mandatory audit regime
One possible setting:
1) Firm must have an ICFR audit Examined
Voluntary disclosure regime
Two possible settings:
1) Firm chooses to make an ICFR disclosure (rare in real voluntary Examined (to hold
disclosure regimes) constant the presence
of ICFR disclosure in
the experimental firm
between voluntary
versus mandatory
disclosure regimes)
2) Firm chooses not to make an ICFR disclosure (most common in Not examined
real voluntary disclosure regimes)
Mandatory disclosure regime
One possible setting:
1) Firm must make an ICFR disclosure Examined

Fig. 1. Panel B outlines all the possible settings that we could examine and the settings which we have chosen to examine in our study. In Fig.1 Panel A, we note that one of the conditions
we examine, VD_MA, is not observed in the real world. This condition is designed to rule out the alternative explanation that the lack of a significant incremental effect of a mandatory
audit in the presence of mandatory disclosure is due to the mandatory audit itself not having any significant positive effect on the No-MW firm's investment potential evaluation.

404b, and in Canada in 2008, under NI 52-109. Using archival data
to compare investor reactions across regimes can be challenging
when there is no available data in a particular regime (e.g., Cell 1's
voluntary disclosure/voluntary audit regime, where few firms
voluntarily disclose material weaknesses, or Cell 2's voluntary
disclosure/mandatory audit regime which does not exist); or when
different regimes are implemented for different firms, during
different time periods, and in different countries (e.g., Cell 3 versus
Cell 4). Experiments allow us to make comparisons between reg-
ulatory regimes and for different sets of participants with varying
knowledge levels, while keeping the information content of the
ICFR disclosures, firm characteristics, time period, and geographic
4 K. Kelly, H.-T. Tan / Accounting, Organizations and Society 56 (2017) 1e20
region constant. Also, to our knowledge, prior studies tend to focus
on examining disclosures of material weaknesses and no study has
specifically examined whether disclosures indicating an absence of
material weakness (or effective ICFR) is perceived differently by
investors under different regulatory regimes. Fig. 1 Panel B outlines
the specific settings in each regulatory regime that our study
examines.
Our findings are also informative for prior research that has
examined mandatory disclosure (Leuz & Wysocki, 2008) and
mandatory audit (Dopuch & King, 1991) each in isolation but not
their joint effects, and largely for financial statements disclosures. For
example, prior research has examined the complementary nature of
an audit on a voluntary disclosure (Ball, Jayaraman, & Shivakumar,
2012; Coram, Monroe, & Woodliff, 2009), but prior research has
not examined the incremental effect of an audit on a mandatory
disclosure. Further, by examining other investor perceptions (e.g.,
management trustworthiness and ICFR quality, reliability and rele-
vance of ICFR disclosures) in addition to investment potential eval-
uations as well as investors with varying knowledge levels, we
provide a better theoretical understanding of how mandatory ICFR
disclosures and audits influence investor judgments.
The remainder of the paper is organized as follows. We first
provide background on the associated literature and develop the
hypotheses. We next describe the design and results of the exper-
iments. Finally, we conclude with a discussion of the findings and
limitations of the study.
2. Theory and hypothesis development
Various archival studies have found that earnings quality and
market reactions are more strongly and negatively associated with
disclosures of unaudited material weaknesses in “disclosure con-
trols and procedures” (DCP) under Section 302 versus audited
material weaknesses in ICFR under Section 404b (Beneish, Billings,
& Hodder, 2008; Doyle, Ge, & McVay, 2007). Archival studies have
also examined how Section 302 versus Section 404 affects com-
panies' disclosure of material weaknesses (e.g., Hammersley,
Myers, & Zhou, 2012; Hermanson & Ye, 2009). Of particular inter-
est is the study by Kinney and Shepardson (2011), which finds that
compared to under Section 302, mandating both management
disclosure of ICFR effectiveness (Section 404a) and an independent
ICFR audit (Section 404b) for large issuers increases the disclosure
rate of material weaknesses to the same extent as mandating
management disclosure alone for small issuers (Section 404a).
Hence, they conclude that mandating disclosure alone is as effec-
tive as mandating both disclosure and audit in terms of encour-
aging their sample companies to disclose material weaknesses. The
findings of Kinney and Shepardson (2011) highlight the importance
of examining whether investors also behave as if mandatory
disclosure and mandatory audit are substitutes just as companies
do in terms of their disclosure rates.
2.1. Configural information processing of ICFR disclosures
Our study focuses on whether investor judgments are affected
by two regulatory mechanisms, mandatory disclosure and
mandatory audit, in a configural manner. Configural information
processing refers to how people's judgments and decisions are a
function of the pattern or configuration of information cues (Edgell,
1978; Slovic & Lichtenstein, 1971).
Prior accounting research on configural information processing
focuses on how auditors exhibit different patterns of configural
processing across different tasks (Brown & Solomon, 1990, 1991;
Maletta & Kida, 1993). For example, Brown and Solomon (1991)
find that auditors who are assessing the risk of misstatement for
an account considered evidence from two substantive audit pro-
cedures to be substitutes, producing an ordinal interaction where
the completion of either one procedure alone reduces misstatement
risk, and completing both procedures does not reduce misstate-
ment risk compared to completing only one procedure. However,
the pattern of configural processing is different in another task. In
this other task, auditors who are assessing the risk of misstatement
for the revenue and receipts transaction cycle considered two cues
from analytical procedures configurally to produce a disordinal
interaction where a material increase (versus material decrease) in
net accounts receivable decreases misstatement risk in the context
of a material increase in gross sales but increases misstatement risk
in the context of a material decrease in gross sales.
Prior research indicates that domain-specific knowledge is a
necessary determinant of whether decision makers employ con-
figural information processing and the exhibited pattern of such
configurality (Brown & Solomon, 1991). For investors assessing the
investment potential of firms under mandatory disclosure and/or
mandatory audit regulations, we believe that investors who have
domain-specific knowledge about companies' incentives for
opportunistic voluntary disclosures, impact of mandatory disclo-
sures on companies' disclosures, and the assurance value of ICFR
audits would engage in our predicted pattern of configural pro-
cessing. We also expect the nature of the ICFR disclosure, which is
either consistent or inconsistent with company disclosure in-
centives (i.e., No-MW versus MW disclosure), to affect the pattern
of configural processing. In the discussion that follows, we first
develop Hypothesis 1 relating to investors engaging in configural
processing in that mandatory disclosure and mandatory audit are
substitutes in terms of their positive effects on the assessed in-
vestment potential of a No-MW disclosure firm. We then develop
Hypothesis 2 relating to neither mandatory disclosure nor
mandatory audit having any effects on the assessed investment
potential of a MW disclosure firm.
2.2. Firms that disclose no material weaknesses (No-MW disclosure
firms)
Attribution theory and persuasion research (Eagly, Wood, &
Chaiken, 1978; Kelley & Michela, 1980) predict that the extent to
which investors perceive bias in the information disclosed by firms
depends on whether the disclosed information is consistent with
the perceived disclosure incentives of firms. As such, investors'
knowledge about companies' incentives to voluntarily disclose
positive news but withhold negative news is an important element
to our predicted pattern of configural information processing (see
Beyer, Cohen, Lys, & Walther, 2010 for review of literature on
companies' voluntary disclosures). Research in accounting has
documented that investors perceive incentive-consistent informa-
tion from firms to have more potential for bias and is therefore less
reliable than incentive-inconsistent information (Frederickson,
Hodge, & Pratt, 2006; Hirst, Koonce, & Simko, 1995; Hodge,
Hopkins, & Pratt, 2006). The reliability of a piece of information
influences the extent to which the information makes a difference
in investor decisions (i.e., relevance).5 Although prior research does
5
Both FASB's concept statement (SFAC No. 2) (FASB, 2008) and IASB Conceptual
Framework for Financial Reporting (2015) define the relevance construct as the
ability of information to make a difference in a decision. Koonce, Nelson, and
Shakespeare (2011) also used the phrase “makes a difference in a decision” in
their question to measure relevance. See also Footnote 10 in Koonce et al. where
they reported that the phrase “makes a difference in a decision” was considered the
most descriptive of the relevance concept amongst other phrases (important, free
from bias, can be confirmed, timely, represents what it is supposed to) by their
participants.
 K. Kelly, H.-T. Tan / Accounting, Organizations and Society 56 (2017) 1e20
not directly examine a setting where regulators and auditors affect
the reliability of a piece of information, the psychology theory on
attribute substitution used in prior research suggests that investors'
perception of whether a piece of information is likely to make a
difference in their decisions is influenced by their assessment of the
reliability of that information (Kadous, Koonce, & Thayer, 2012;
Kahneman & Frederick, 2002).6 In other words, a piece of theo-
retically relevant information would be practically relevant for
making a difference in investor decisions only if it is perceived as
reliable. In the following discussion, we argue that mandatory
disclosure and mandatory ICFR audit can increase the perceived
reliability of ICFR disclosures, and higher perceived reliability in
turn increases the perceived relevance of ICFR disclosures to in-
vestment potential evaluations.
2.3. Effect of mandatory versus voluntary disclosure regime on No-
MW disclosure
In our study, a voluntary disclosure regime allows a firm to
voluntarily issue a No-MW disclosure as a signal to investors about
the positive quality of its ICFR. However, a No-MW disclosure is
more consistent with companies' incentives to present themselves
in a positive light and may be perceived by investors as less reliable,
particularly when the disclosure is voluntary, which limits the
ability of firms to use a voluntary No-MW disclosure to signal good
ICFR quality. This is consistent with Crawford and Sobel’s (1982)
economic model that voluntary disclosure is untruthful and unin-
formative in equilibrium without a mechanism for the discloser to
credibly commit to being truthful.
Mandatory (versus voluntary) disclosure of material weaknesses
could increase the reliability of a No-MW disclosure. First, manda-
tory disclosure may signal to investors that the disclosure contains
reliable information that would be relevant for making a difference
to their decisions; otherwise, regulatory authorities would not
mandate the disclosure.7 Second, mandating disclosure allows
management to commit to reliable disclosures regardless of
whether the disclosures contain good news or bad news, because of
the costs of violating a mandated disclosure requirement (Beyer
et al., 2010; Leuz, 2010; Rock, 2002). The availability of regulatory
enforcement mechanisms for the mandatory disclosure require-
ment and the criminal and civil liabilities associated with violation
increase the reliability of the disclosures. Mandating disclosure
eliminates the possibility of investors interpreting that management
is selectively making only positive voluntary No-MW disclosures
and increases the reliability of the No-MW disclosure. Therefore, we
expect that a mandatory (versus voluntary) disclosure regime in-
creases the reliability of a No-MW disclosure and the extent to which
a No-MW disclosure makes a difference in investor decisions.
2.4. Effect of mandatory versus voluntary ICFR audit regime on No-
MW disclosure
Prior literature indicates that financial statement audits provide
6
Kadous et al. (2012) use the psychology theory on attribute substitution to
argue that reliability affects relevance because people substitute the more easily
accessible reliability assessment when they are making a relevance assessment.
Kadous et al. conducted two experiments where the source of reliability of infor-
mation was manipulated differently and their results were robust to different
sources of reliability.
7
Frederickson et al. (2006) find that users consider mandated recognition of
stock option expense to be more reliable than voluntary recognition. However,
Frederickson et al. examine voluntary versus mandatory recognition of a quantifi-
able accounting issue, rather than voluntary versus mandatory disclosure of a
qualitative accounting issue.
5
valuable assurance to investors and creditors about the risk of
material misstatement, reduce the costs of monitoring by investors
and creditors, and increase their willingness to invest or lend to the
firm (Allee & Yohn, 2009; Blackwell, Noland, & Winters, 1998;
Lennox & Pittman, 2011; Watts & Zimmerman, 1983). Economic
theory argues that a financial statement audit allows management
to commit to making credible voluntary disclosures of private in-
formation such as earnings forecasts (Ball et al., 2012; Gigler &
Hemmer, 1999). This is referred to as the “confirmatory role” of
audited financial statements because management is disciplined to
make truthful ex-ante voluntary disclosures because the disclosures
would be verified ex-post by the independent audit. Coram et al.
(2009) report similar findings that an audit improves the ability
of a positive voluntary non-financial disclosure to increase the
stock price estimates of financial statement users.
Without an ICFR audit, auditors are only required to review
management's No-MW disclosure for material inconsistencies with
the audited financial statements and they are not responsible for
performing additional audit procedures to detect misstatements
beyond the scope of inconsistency with the financial statement
audit (AU Section 550 issued by Public Company Accounting
Oversight Board (PCAOB, 2003); International Standard on Audit-
ing (ISA) 720 (IASB, 2010)). A mandatory ICFR audit provides in-
dependent assurance to investors beyond that provided by the
financial statement audit and should be useful for enhancing the
reliability of a potentially biased No-MW disclosure. Given the
significant cost of ICFR audits, voluntary ICFR audits in a voluntary
audit regime is rare in practice. Therefore, we expect that a
mandatory (versus voluntary) ICFR audit regime increases the
reliability of a No-MW disclosure and the extent to which a No-MW
disclosure makes a difference in investor decisions because of the
higher likelihood of an ICFR audit being conducted in a mandatory
(versus voluntary) audit regime.
2.5. Configural information processing for No-MW disclosure
We consider how investors would configurally process a No-
MW disclosure in the context of a mandatory (versus voluntary)
disclosure regime and a mandatory (versus voluntary) audit
regime. Prior studies find that there are negative stock price re-
actions to disclosures of unaudited Section 302 material weak-
nesses because material weaknesses signal higher risks of material
misstatements in a company's financial reports and investors dis-
count stock prices to compensate for the greater risk (Beneish et al.,
2008; Hammersley, Myers, & Shakespeare, 2008). Ashbaugh-
Skaife, Collins, Kinney, and LaFond (2009) find that when firms
which are expected to have poor quality ICFR disclose audited re-
ports of effective ICFR, there is a significant decrease in cost of
equity consistent with lower risk assessment. These prior findings
imply that a disclosure of an effective ICFR with no material
weaknesses signals a higher quality ICFR that increase investors'
confidence about the company's financial reports and lower the
risk of material misstatement in the company's financial reports,
which should then improve its investment potential. In other
words, all things equal, a MW disclosure signals poor quality ICFR
that should decrease the perceived investment potential of a
company whereas a No-MW disclosure signals better quality ICFR
that should increase the perceived investment potential of a
company.
The more reliable the No-MW disclosure is perceived to be, the
more likely the No-MW disclosure would make a difference (i.e., be
more relevant) in signaling better quality ICFR that increases the
perceived investment potential of a company. As we discuss above,
both a mandatory (versus voluntary) disclosure regime as well as a
mandatory (versus voluntary) audit regime increases the reliability
6 K. Kelly, H.-T. Tan / Accounting, Organizations and Society 56 (2017) 1e20
of a No-MW disclosure. Therefore, both mandatory disclosure
regime alone and mandatory audit regime alone are expected to
increase perceived ICFR quality and hence investment potential
evaluation for a No-MW disclosure firm. Investors who are cogni-
zant of the reliability-enhancing attributes of mandatory disclosure
regime alone and mandatory audit regime alone may perceive no
additional benefit from having both mandatory disclosure and
mandatory audit versus having only one regulatory mechanism
alone (i.e., mandatory disclosure and mandatory audit are
substitutes).8
Based on our discussion above, our hypothesis for how the in-
vestment potential evaluations of a No-MW disclosure firm are
affected by mandatory disclosure and mandatory audit is stated in
the alternative form as follows (see Fig. 2, Panel A for pictorial
depiction).
Hypothesis 1. For a No-MW disclosure firm, investors will perceive a
mandatory (versus voluntary) disclosure regime and a mandatory
(versus voluntary) audit regime to be substitutable regulatory mech-
anisms and the two regulatory mechanisms will have an ordinal
interactive effect on investors’ investment potential evaluations that
will take the form of:
a) Investment potential evaluation will be lower in a voluntary
disclosure/voluntary audit regime than in a mandatory disclosure
regime alone or in a mandatory audit regime alone.
b) Investment potential evaluation will not be higher in a mandatory
disclosure/mandatory audit regime than in a mandatory disclosure
regime alone or in a mandatory audit regime alone.
2.6. Firms that disclose material weaknesses (MW disclosure firms)
and configural information processing for MW disclosure
In contrast to a No-MW disclosure, a MW disclosure is generally
less consistent with companies' incentives to present themselves in
a positive light than a No-MW disclosure.9 Thus, based on attri-
bution theory, if management voluntarily discloses material
weaknesses despite incentives not to do so, investors should have
less reason to doubt the reliability of a MW disclosure than that of a
No-MW disclosure. Coram et al. (2009) report results consistent
with attribution theory where they find that an audit increases the
stock price estimates of financial statement users who receive a
positive voluntary non-financial disclosure; but an audit has no
effect on a negative voluntary non-financial disclosure.
Attribution theory suggests that the configural effects of
mandatory disclosure and mandatory audit in Hypothesis 1 for No-
MW disclosure firms would not manifest for MW disclosure firms.
The reliability and hence the ability of a voluntary MW disclosure to
8
The key difference between our study and those of Ball et al. (2012) and Coram
et al. (2009) is that we are studying the substitutory nature of mandatory audit and
mandatory disclosure, whereas Ball et al. and Coram et al. are studying the com-
plementary nature of audit and voluntary disclosure. Ball et al. and Coram et al. only
examine voluntary disclosure and find that the audit is a complementary regulatory
mechanism that verifies the credibility of a voluntary disclosure, which has no
inherent mechanisms to verify its credibility. Our study contrasts mandatory versus
voluntary disclosure. Mandatory disclosure is a regulatory mechanism that in-
creases disclosure credibility and thus may be perceived by investors as a substi-
tutable regulatory mechanism for mandatory audit. In fact, our study finds
consistent results with Ball et al. and Coram et al., in that when the No-MW
disclosure is voluntary, a mandatory (versus voluntary) audit regime has a posi-
tive effect on investment potential evaluations.
9
Although firms may have incentives to voluntarily disclose material weak-
nesses to signal that they are transparent (Ashbaugh-Skaife et al., 2007), there is a
cost to such disclosure of material weaknesses in terms of investors' perceptions of
the quality of the firm's internal control system and/or misstatement risk (Beneish
et al., 2008; Hammersley et al., 2008).
make a difference (i.e., be relevant) to investor decisions may not be
further enhanced by a mandatory disclosure regime alone, or by a
mandatory audit regime alone, or by a combined mandatory
disclosure and mandatory audit regime (unlike the case for No-MW
disclosures).
Based on our discussion above, our hypothesis for how the in-
vestment potential evaluations of a MW disclosure firm are affected
by mandatory disclosure and mandatory audit is stated in the null
form as follows (see Fig. 2, Panel A for pictorial depiction).
Hypothesis 2. For a MW disclosure firm, a mandatory (versus
voluntary) disclosure regime and a mandatory (versus voluntary)
audit regime will not have any effects on investors’ investment po-
tential evaluations. Specifically,
a) Investment potential evaluation will not be different in a voluntary
disclosure/voluntary audit regime than in a mandatory disclosure
regime alone or in a mandatory audit regime alone.
b) Investment potential evaluation will not be different in a manda-
tory disclosure/mandatory audit regime than in a mandatory
disclosure regime alone or in a mandatory audit regime alone.
3. Method
3.1. Experimental design and independent variables
We use a 2
2 (between-subjects)
2 (within-subjects) mixed
design to test our hypotheses. The within-subjects variable consists
of two firms: one disclosed that there are no material weaknesses
(No-MW disclosure firm) and one disclosed that there are material
weaknesses (MW disclosure firm). The two between-subjects in-
dependent variables are whether management disclosure of ICFR
material weaknesses is voluntary or mandatory (VD for voluntary
disclosure versus MD for mandatory disclosure) and whether an
ICFR audit is voluntary or mandatory (VA for voluntary audit versus
MA for mandatory audit). Detailed description of the manipulated
variables is provided in the “Procedures” sub-section.
In practice, while firms must make an ICFR disclosure in a
mandatory disclosure regime (MD), they can choose to make an
ICFR disclosure or not in a voluntary disclosure regime (VD). Our VD
condition examines a voluntary disclosure regime setting where
the firm has chosen to make an ICFR disclosure. This design choice
holds constant the presence of ICFR disclosures while manipulating
whether the disclosure is voluntary or mandatory.
Similarly, in practice, while firms must have an ICFR audit in a
mandatory audit regime (MA), they can choose to have an ICFR
audit or not in a voluntary audit regime (VA). Participants in our VA
condition are told that an independent ICFR audit is voluntary and
the experimental firm is silent on whether they had an ICFR audit,
which is consistent with real world voluntary audit regimes.10
Participants in the VA condition are further told that “some com-
panies engage in an independent auditor voluntarily and disclose
that; others choose not to engage an independent auditor and are
silent on this matter.” This design choice is intended to have par-
ticipants in the VA condition not only conclude that ICFR audit is
voluntary but to also increase the likelihood that they would
further infer that the silent experimental firm has not hired an ICFR
10
Our VA versus MA manipulation focuses on whether the ICFR audit is voluntary
or mandatory. We do not manipulate (nor indicate to participants) whether the
firm's choice about an ICFR audit in the VA condition is mandated to be disclosed or
not, and we also do not manipulate (nor indicate to participants) whether the ICFR
audit findings are mandated to be disclosed or not if the firm had an ICFR audit
(whether the audit is voluntary or mandatory).
 K. Kelly, H.-T. Tan / Accounting, Organizations and Society 56 (2017) 1e20 7

Panel A: Hypotheses
No-MW disclosure
firm
Investment Mandatory audit regime ____
potential
evaluations
Voluntary audit regime - - - -

MW disclosure
firm

Voluntary Mandatory
disclosure regime disclosure regime

Panel B: Results for investment potential evaluations rating

4
Mandatory audit regime for
No-MW disclosure firm
Voluntary audit regime for
3 No-MW disclosure firm
Mandatory audit regime for
MW disclosure firm
Voluntary audit regime for
2 MW disclosure firm

1
Voluntary disclosure regime Mandatory disclosure regime

Fig. 2. Effects of mandatory management disclosure and mandatory ICFR audit on investment potential evaluationsaa See Table 1 for definitions of investment potential evaluations
rating, the regulatory regimes (voluntary disclosure/voluntary audit, voluntary disclosure/mandatory audit, mandatory disclosure/voluntary audit, and mandatory disclosure/
mandatory audit), and the two firm types (No-MW disclosure firm and MW-disclosure firm).

auditor. We believe this is the most consistent with practice where
most firms in a voluntary audit regime choose not to have an in-
dependent ICFR audit and are silent on whether an independent
ICFR audit has been conducted. That said, as disclosed in our
manipulation check section, there are a minority of participants in
the VA condition who believed that the silent experimental firm has
voluntarily hired an ICFR auditor.
3.2. Participants
Participants are alumni of an accounting and finance under-
graduate program at a large Canadian university. Canadian partic-
ipants function in a regime where regulators have indicated that
they would consider adding mandatory audit to the existing
11
As an alternative, we could have used U.S. participants who are currently in a
regime that mandates both disclosure and audit for large issuers (which our
experimental firm is likely to be perceived as given the size of its sales and
stockholder's equity). However, U.S. regulators have not expressed any interest in
removing the audit requirement for large issuers, and U.S. participants may
therefore find it unrealistic to consider a disclosure only (i.e., MD_VA) regime for
large issuers.
disclosure requirement based on experience in Canada and other
countries. Hence, our reasoning is that they would not find it un-
realistic to consider a new regime where both audit and disclosure
are mandated (i.e., MD_MA).11 On behalf of the researchers, the
University's alumni office sent about 3000 emails with a brief
introduction of the study and a link to a website that provided
access to the experimental materials if people agreed to participate.
Participants who completed the study could enter their name in a
random drawing for one of 20 gift cards (of a major retailer) valued
at Canadian Dollars $100 each. A total of 89 alumni participated in
our experiment; 69 in the four conditions in the experiment and 20
in the fifth additional condition reported in Footnote 17, for a
response rate of about 3%.
Our participants have an average of 6.3 years of work experience
and 4.2 years of experience analyzing the financial performance of
firms. Almost 90 percent of the participants work in accounting and
finance-related professions (i.e., auditing, tax, accounting, and
finance, banking, or investing). On a scale of 0 (“never”) to 14 (“with
high frequency”), our participants indicated a mean of 3.1 with
8 K. Kelly, H.-T. Tan / Accounting, Organizations and Society 56 (2017) 1e20
respect to how often they invested in the stock market, and the
results are the same when we controlled for their investment fre-
quency.12 Given the educational background and work experience
of our participants, they would have the following requisite
domain-specific knowledge necessary for the configural informa-
tion processing that we predict: knowledge about companies' in-
centives for opportunistic voluntary disclosures, impact of
mandatory disclosures on companies' disclosures, and the assur-
ance value of ICFR audits.
3.3. Procedures
Participants were randomly assigned by the survey software to
one of the four between-subject experimental conditions. They
were asked to assume that they worked in the investment
department of a company and that they were assessing the in-
vestment potential of two firms on behalf of their company. This
design choice of asking participants to assess firms on behalf of
their company rather than for their own personal investment has
been used in prior studies such as Elliot, Hodge, and Sedor (2012).
This design choice likely increases participants' objectivity in that
they are less likely to bring their personal risk preferences into their
judgments.13
Participants were told that the two firms were listed on a hy-
pothetical stock exchange where regulations required management
to evaluate ICFR in a listed firm. Participants were also informed
that the external auditors of each firm had expressed unqualified
opinions on the financial statements.14 Then, participants read one
of the independent variable manipulations about ICFR disclosure
and audit regulations in the hypothetical stock exchange (VD_VA,
MD_VA, VD_VA, and MD_MA).15 Excerpts of our between-subjects
12
None of these demographic variables (years of work experience, years of
experience analyzing financial performance of firms, profession, investment fre-
quency) are significantly different across conditions nor are they significant cova-
riates in the ANCOVA for investment potential rating (all p values  0.210).
13
However, prior research suggests that findings on individuals making risky
financial decisions for themselves are consistent with findings on individuals
making those decisions for others (Stone, Yates, & Caruthers, 2002).
14
This provides a cleaner test of the effects of mandatory disclosure and
mandatory audit. We want to eliminate the situation where investors were not sure
whether management had evaluated the ICFR (particularly for the No-MW
disclosure firm) and whether the external auditors had expressed an unqualified
opinion on the financial statements (particularly for the MW disclosure firm).
15
Our manipulations did not include any language on the effectiveness of the
regulatory enforcement mechanisms associated with mandatory disclosure and
mandatory audit. We do not collect any evidence on participant perceptions about
the strength of investor protection laws relating to company misrepresentation and
auditor misrepresentation, although participants (and hence their perceptions) are
randomly assigned to conditions.
16
We created minor 2e4 percent variations in the financial data of the two firms
so that participants will see two firms with fairly similar financials but different
internal control disclosures. The financial data was crossed between the two firms
to avoid an effect from the differences in financial data.
17
The original study had a fifth additional VD_VA condition with a No-MW
disclosure firm and a firm that is silent on whether there is MW or not. Invest-
ment potential rating does not differ between the No-MW disclosure firm
(mean ¼ 3.85) and the silent firm (mean ¼ 3.60) (p ¼ 0.371), suggesting that a
Silent-MW disclosure firm is not penalized vis-a -vis a No-MW disclosure firm in
terms of investment potential evaluation in a VD_VA setting. More importantly,
participants gave a No-MW disclosure firm in this fifth condition (mean ¼ 3.85) a
marginally higher investment potential evaluation when it is contrasted against a
silent firm than when a No-MW disclosure firm (mean ¼ 2.97) is contrasted against
a MW disclosure firm in the VD_VA condition reported in text (one-tailed
p ¼ 0.054). This finding suggests that contrasting a No-MW disclosure firm against
a MW disclosure firm may serve to highlight the difference in consistency of the
disclosure with the firm's disclosure incentives, making the low reliability of the
No-MW disclosure more salient, and increasing the power of detecting effects of
regulatory mechanisms (such as mandatory disclosure and mandatory audit) aimed
at improving disclosure reliability.
manipulations are reproduced in Appendix 1.
Next, participants were given information about two firms,
comprising financial data (e.g., sales, net income, working capital,
long-term debt, new sales orders) and a note disclosure on controls
and procedures.16 The note disclosure on controls and procedures
for the No-MW disclosure firm disclosed that management did not
identify any material weaknesses in its ICFR whereas the MW
disclosure firm disclosed that management had identified various
material weaknesses in its ICFR.17 The No-MW disclosure firm was
presented either before or after the MW disclosure firm to avoid an
order effect and the results are the same controlling for the order in
which the two firms appear on the instrument. We adapted the
description of disclosed ICFR material weaknesses for the MW
disclosure firm from one of the cases in Earley, Hoffman, and Joe
(2008) rated by their participating auditors to be more severe
than a significant deficiency.
For VD_MA and MD_MA, management's note disclosure on
controls and procedures additionally disclosed that the firm's in-
dependent auditors had also conducted their own evaluation of the
firm's ICFR, and that their opinion concurred with the firm (i.e.,
either MW or No-MW). For VD_VA and MD_VA, management's note
disclosure did not indicate that there was an independent ICFR
audit in line with common practice. Excerpts of our within-subjects
manipulations for the NO-MW/MW disclosure firms are repro-
duced in Appendix 1.
Participants then made several judgments about the two firms
(during which they could access the cases), including investment
potential and other assessments discussed later in the paper.
Finally, participants answered manipulation check and de-
mographic questions during which the case could not be consulted.
3.4. Dependent variables
Participants judged the two firms' investment potential by rat-
ing on a 15 point scale (7 to þ7): (1) the attractiveness of the stock
as an investment, and (2) the company's earnings potential
(Spearman-Brown correlation for No-MW disclosure firm ¼ 0.62,
p < 0.001; Spearman-Brown correlation for MW disclosure
firm ¼ 0.65, p < 0.001).18 We averaged these two ratings to obtain
an overall investment potential rating for each firm because the
two individual ratings are highly correlated and the results for each
individual rating are similar to the averaged overall investment
potential rating.
We also asked participants four questions that measure their
perceptions of management credibility (Giffin, 1967; Mercer, 2005):
(1) whether management is trustworthy in its disclosures about
ICFR (on a scale of 0 “Not at all trustworthy” to 14 “Very trust-
worthy”) (Trust), (2) whether the company had undisclosed ma-
terial weaknesses in its ICFR (on a scale of 0 “Very certain that there
are no undisclosed material weaknesses” to 14 “Very certain that
there are undisclosed material weaknesses”, reverse coded)
(UndisclosedMW), (3) the quality of the company's ICFR (on a scale
of 7 “Extremely Poor” to 7 “Extremely Good”) (ICFR Quality), and
(4) whether management is competent in establishing and main-
taining adequate ICFR (on a scale of 0 “Not at all competent” to 14
“Very competent”) (Management Competence). Factor analyses for
these credibility questions indicate that the first two questions load
together on one factor with factor loadings 0.69 or higher for both
firms, while the last two questions load together on another factor
with factor loadings 0.65 or higher for both firms. The first two
questions relate generally to the trustworthiness of management
18
All p values are two-tailed unless otherwise stated.
 K. Kelly, H.-T. Tan / Accounting, Organizations and Society 56 (2017) 1e20
(e.g., did management disclose all material weaknesses?) and re-
sponses to these two questions (Trust and UndisclosedMW) are
averaged to obtain a TRUSTWORTHINESS measure (Spearman-
Brown correlation for No-MW disclosure firm ¼ 0.43, p < 0.001;
Spearman-Brown correlation for MW disclosure firm ¼ 0.37,
p ¼ 0.002). The last two questions relate generally to management's
ability to establish and maintain a good quality ICFR and responses
to these two questions (ICFR Quality and Management Compe-
tence) are averaged to obtain an ICFR-QUALITY measure (Spearman-
Brown correlation for No-MW disclosure firm ¼ 0.54, p < 0.001;
Spearman-Brown correlation for MW disclosure firm ¼ 0.65,
p < 0.001).
3.5. Manipulation checks
We asked participants whether under the hypothetical stock
exchange's regulations (1) management's disclosure of identified
material weaknesses in ICFR was voluntary or mandatory, and (2)
an independent audit of ICFR was voluntary or mandatory. Fifty-
one (51) out of the 69 participants (74 percent) correctly
answered both questions. The manipulation check failure rate does
not differ significantly across conditions (p ¼ 0.614). Participants
who failed manipulation checks might still have been influenced by
the manipulations but were not able to recall correctly the ma-
nipulations by the time they responded to the manipulation check
questions (Perdue & Summers, 1986). We retain all participants in
our analyses because differences in ICFR-QUALITY assessment
across the VD versus MD manipulation (F ¼ 6.61, p ¼ 0.012) and the
VA versus MA manipulation (F ¼ 10.15, p ¼ 0.002) (see Table 2 Panel
B) indicate that our manipulations were successful regardless of the
participants' recall ability (Sigall & Mills, 1998). The results for the
full sample of participants (N ¼ 69) are also similar to the results for
the smaller sample of participants (N ¼ 51) who passed both
manipulation checks.19
We also asked participants whether they thought the No-MW
disclosure firm and the MW-disclosure firm hired an indepen-
dent auditor to evaluate its ICFR. For participants who were in the
VA conditions and the firm was silent on whether it had hired an
auditor, 82% (27 out of 33) believed the No-MW disclosure firm had
not hired an auditor while 56% (18 out of 33) believed the MW
disclosure firm had not hired an ICFR auditor.20 Participants' beliefs
were consistent with practice where in a voluntary audit regime, if
the firm was silent on whether an ICFR auditor was hired, investors
tend to believe that the firm had not done so. For participants who
were in the MA conditions, 89% (32 out of 36) and 86% (31 out of 36)
indicated that the No-MW disclosure firm and the MW disclosure
firm, respectively, had hired an ICFR auditor. Given that our VA
versus MA manipulations were whether or not the regulatory
regime mandated an ICFR audit rather than whether or not par-
ticipants believed the firm hired an ICFR auditor, we do not exclude
VA participants who indicated that the firm hired an ICFR auditor or
MA participants who indicated that the firm did not hire an ICFR
auditor. Regardless, the results are similar whether or not we
19
According to Sigall and Mills (1998), differences between experimental condi-
tions on a measure of the conceptual independent variable or a dependent measure
provide evidence that participants notice the manipulated differences between
conditions (p. 221).
20
It is not surprising that a high percentage of participants (44% or 15 out of 33) in
the VA conditions believed the MW disclosure firm had hired an ICFR auditor. In
practice, a MW disclosure is a less common event than a No-MW disclosure (see
Footnote 2). Even though the firm was silent on whether it had hired an ICFR
auditor, participants may have assumed that it had voluntarily hired an ICFR auditor
which helped it identify those material weaknesses and pressured it to disclose
identified material weaknesses.
9
exclude these participants.
Finally, we asked participants whether they recalled if the firm
disclosed that there were material weaknesses or no material
weaknesses. About 84% (58 out of 69) participants recalled that the
MW disclosure firm disclosed material weaknesses, and 91% (63
out of 69) participants recalled that the No-MW disclosure firm
disclosed no material weaknesses. We retain all participants
because differences in investment potential rating and ICFR-
QUALITY assessment across the MW and No-MW disclosure firms
(reported in the paired t-tests below) indicate that our manipula-
tion was successful regardless of the participants' recall ability
(Sigall & Mills, 1998). The results are also similar whether or not
participants who indicated the MW disclosure firm disclosed no
material weaknesses and participants who indicated the No-MW
disclosure firm disclosed material weaknesses are excluded.
4. Results
Paired t-tests indicate that investment potential evaluation
(3.99 versus 2.14, t ¼ 6.31, one-tailed p < 0.001) and ICFR-QUALITY
(6.22 versus 2.57, t ¼ 8.43, one-tailed p < 0.001) are higher for the
No-MW disclosure firm than the MW disclosure firm. These results
are consistent with our hypotheses development that disclosing
material weaknesses decreases the perceived quality of ICFR in the
firm, which is associated with lower investment potential
evaluations.
We use MANOVA to first test if there are multivariate differences
(in the investment potential evaluations of both the No-MW
disclosure firm and the MW disclosure firm) across the between-
subjects conditions, followed by standardized canonical co-
efficients and univariate ANOVAs to test that the between-subjects
differences are driven by the No-MW disclosure firm as predicted in
H1 and not by the MW disclosure firm as predicted in H2 (Garson,
2015; Scheiner, 2001, pp. 99e115; Wilkinson, 1975). We use
MANOVA rather than ANOVA of the difference score in investment
potential evaluations between the two firms because the reliability
of a difference score and hence statistical power is low when the
components that make up that difference score are positively
correlated, as is the case for our study where the No-MW firm's
investment potential evaluation is strongly and positively corre-
lated with the MW firm's investment potential evaluation (Pearson
r ¼ 0.38, p ¼ 0.001) (Bergh & Fairbanks, 2002; Edwards, 1995, 2001;
Peter, Churchill, & Brown, 1993).21 Edwards (1995) recommends
that the individual components that make up the difference score
be used as separate dependent variables in a multivariate
21
Indeed, for the investment potential evaluations of the two firms, the re-
liabilities of the individual firms are higher than the reliability of the difference
score between the two firms. With respect to the two items in the investment
potential evaluation variable (i.e., stock attractiveness and earnings potential),
Cronbach alpha and Spearman-Brown correlation for the No-MW disclosure firm
(0.77 and 0.62) and the MW disclosure firm (0.72 and 0.65) are higher than the
respective Cronbach alpha and Spearman-Brown correlation for the difference
score (0.65 and 0.55). Low reliability and low power of a difference score is less of
an issue in cases where (1) the individual components are not correlated or
negatively correlated or (2) the individual components are one-item measures
(Bergh & Fairbanks, 2002). Edwards (1995, 2001), Peter, Churchill, and Brown
(1993), Hom, Griffeth, Palich, and Bracker (1999), Bergh and Fairbanks (2002),
Kristof-Brown, Zimmerman, and Johnson (2005), Klein, Jiang, and Cheney (2009)
discuss various cases in research (e.g., person-environment fit, gap between
employee/consumer expectations and experiences, change over time, difference
between wholesaler and broker perceptions) where using difference scores may
lead to misleading conclusions or conceal more complex relationships. Thus, it is
now common practice in literature in management, marketing, and psychology to
recognize the statistical limitations of using difference scores and to analyze indi-
vidual components jointly (e.g., Mullins et al., 2014; Ostroff et al., 2004; Seo,
Gamache, Devers, & Carpenter, 2015; Sturm et al., 2014).
10 K. Kelly, H.-T. Tan / Accounting, Organizations and Society 56 (2017) 1e20
regression analysis or MANOVA for improved reliability and power
(for examples of such studies, see Mullins, Ahearne, Lam, Hall, &
Boichuk, 2014; Ostroff, Atwater, & Feinberg, 2004; and Sturm,
Taylor, Atwater, & Braddy, 2014). Edwards (1995) also recom-
mends MANOVA because it allows unambiguous specification of
the individual effects of the independent variables on the No-MW
firm's investment potential evaluation (ordinal interaction effect)
and the MW firm's investment potential evaluation (no effect) as
prescribed by theory, whereas ANOVA of the difference score does
not allow for that. The results of the difference score are reported in
Footnote 22 for completeness.22
Given that H1 predicts an ordinal interaction for the pattern of
configural information processing for the No-MW disclosure firm
as shown in Fig. 2 Panel A, we maximize the power of our test by
using contrast coding (Buckless & Ravenscroft, 1990; Rosenthal &
Rosnow, 1985). The contrast weights to test H1 are 3 for
VD_VA, þ1 for VD_MA, þ1 for MD_VA, þ1 for MD_MA; and the
multivariate contrast for this ordinal interaction is significant
(Table 1 Panel C, F ¼ 4.18, one-tailed p ¼ 0.010). The multivariate
contrast to test if investment potential evaluations are higher in
MD_VA and VD_MA as compared to VD_VA (as predicted in H1a) is
also significant (Table 1 Panel C, F ¼ 3.93, one-tailed p ¼ 0.012). The
multivariate contrast to test that investment potential evaluations
are not higher in MD_MA compared to MD_VA and VD_MA (as
predicted in the null H1b) is not significant (Table 1 Panel C,
F ¼ 0.04, p ¼ 0.965). For each multivariate contrast, MANOVA
generates the standardized canonical coefficients for the No-MW
firm's investment potential evaluation and the MW firm's invest-
ment potential evaluation to enable comparisons of the contribu-
tion of each dependent variable to the extracted dependent
canonical variate. For all three multivariate contrasts, the stan-
dardized canonical coefficients (reported in Table 1 Panel C) are
always larger for the No-MW firm's investment potential evalua-
tion than those for the MW firm's investment potential evaluation
with respect to the dependent canonical variate: 1.08 versus 0.12
for H1, 1.08 versus 0.14 for H1a, and 1.07 versus 0.67 for H1b,
respectively. Standardized canonical coefficients are interpreted
similarly to beta weights in multiple regression analysis. For
example, in H1, one standard deviation increase in No-MW firm's
investment potential evaluation increases the dependent canonical
variate by 1.08 standard deviation, whereas one standard deviation
increase in the MW firm's investment potential evaluation
22
An alternative analysis is to use the difference in investment potential evalu-
ations of the two firms (DiffInvm) as the dependent variable to jointly test H1 and
H2. The results are directionally consistent with our hypotheses, with the smallest
DiffInvm in the VD_VA condition and with DiffInvm being higher in all other con-
ditions, although the differences are not statistically significant (VD_VA ¼ 1.24; VD_
MA ¼ 2.05; MD_VA ¼ 2.16; MD_MA ¼ 1.94) (Mandatory Disclosure: F ¼ 0.46,
p ¼ 0.498, df ¼ 1; Mandatory Audit: F ¼ 0.26, p ¼ 0.614, df ¼ 1; Mandatory
Disclosure
Mandatory Audit: F ¼ 0.76, p ¼ 0.388, df ¼ 1). The contrast for the
disordinal interaction predicted in H1 (with weights of 3, þ1, þ1, þ1) for DiffInvm
is also not significant (one-tailed p ¼ 0.120). The between-subject results using
difference scores are statistically identical to the within-subject results using a
repeated measures ANOVA (i.e., Firm: F ¼ 38.67, p < 0.001, df ¼ 1; Firm
Mandatory
Disclosure: F ¼ 0.46, p ¼ 0.498, df ¼ 1; Firm
Mandatory Audit: F ¼ 0.26, p ¼ 0.614,
df ¼ 1; Firm
Mandatory Disclosure
Mandatory Audit, F ¼ 0.76, p ¼ 0.388, df ¼ 1),
and a repeated measures ANOVA approach faces the same issues as a difference
score approach (Griffin, Murray, & Gonzalez, 1999; Wright, 1990). In the repeated
measures ANOVA, the between-subjects effects of Mandatory Disclosure (F ¼ 1.84,
p ¼ 0.179, df ¼ 1), Mandatory Audit (F ¼ 0.74, p ¼ 0.393, df ¼ 1), and Mandatory
Disclosure
Mandatory Audit are not significant (F ¼ 1.35, p ¼ 0.250, df ¼ 1).
Another alternative analysis used to overcome the statistical limitations of the
difference score is to analyze one dependent variable while controlling for the other
dependent variable as a covariate (e.g. Seo et al., 2015). Our results using this
method are consistent with our hypotheses and are qualitatively similar to the
results reported in text when the other dependent variable is not included as a
covariate.
decreases the dependent canonical variate by 0.12 standard devi-
ation. Therefore, the standardized canonical coefficients indicate
that the No-MW disclosure firm explains more of the multivariate
differences predicted in H1 and H1a than does the MW disclosure
firm (Garson, 2015; Scheiner, 2001, pp. 99e115).
We then use univariate ANOVAs and contrasts to further test if
these multivariate results are driven by the No-MW disclosure firm
(H1) and not the MW disclosure firm (H2). The univariate 2
2
ANOVA for the investment potential rating is significant for the No-
MW disclosure firm (F ¼ 2.85, p ¼ 0.044) but it is not significant for
the MW disclosure firm (F ¼ 0.21, p ¼ 0.888), consistent with our
hypotheses that manipulated regulatory regimes affect the in-
vestment potential evaluation of a No-MW disclosure firm but not
that of a MW disclosure firm. Results are summarized in Fig. 2 Panel
B and Table 1.
4.1. Firms that disclose no material weaknesses (No-MW disclosure
firms)
The univariate contrast to test the predicted ordinal interaction
effect in H1 for the No-MW disclosure firm (3 for VD_VA, þ1 for
VD_MA, þ1 for MD_VA, þ1 for MD_MA) is significant (Table 1 Panel
D, F ¼ 8.38, one-tailed p ¼ 0.003), supporting H1. H1a predicts that
-vis a voluntary disclosure/voluntary audit regime, a manda-
vis-a
tory disclosure regime alone and a mandatory audit regime alone
have positive effects on investment potential evaluations of a No-
MW disclosure firm. The univariate contrast for H1a is supported
in that investment potential evaluation for the No-MW disclosure
firm is higher in MD_VA and VD_MA as compared to VD_VA (Table 1
Panel D, F ¼ 7.86, one-tailed p ¼ 0.003).23 Investment potential
evaluation for the No-MW disclosure firm is also higher in MD_VA
versus VD_VA (t ¼ 2.61, one-tailed p ¼ 0.006), and in VD_MA versus
VD_VA (t ¼ 2.25, one-tailed p ¼ 0.014) (untabulated).24 H1b is in the
null form and predicts that for a No-MW disclosure firm, a
mandatory disclosure/mandatory audit regime does not signifi-
cantly increase investment potential evaluations compared to a
mandatory disclosure regime alone or mandatory audit regime
alone. There is no evidence to reject the null H1b. The univariate
contrast indicates that investment potential evaluation for the No-
MW disclosure firm is not higher in MD_MA compared to MD_VA
and VD_MA (Table 1 Panel D, F ¼ 0.04, p ¼ 0.836). Investment po-
tential evaluation for the No-MW disclosure firm is also not higher
in MD_MA versus MD_VA (t ¼ 0.40, p ¼ 0.690), and in MD_MA
versus VD_MA (t ¼ 0.05, p ¼ 0.961) (untabulated). Hence, alumni
participants exhibit configural information processing where
mandatory disclosure and mandatory audit are substitutes in terms
of their positive effects on the investment potential evaluations of a
No-MW disclosure firm.
23
Although there may be a concern that participants are not able to differentiate
between VD_MA and MD_MA, the failure rate for the manipulation check question
on whether ICFR disclosure is voluntary or mandatory is not significantly different
across the two conditions (p ¼ 0.905) which suggests that participants in VD_MA do
not face increased difficulty interpreting the voluntary nature of the disclosure.
Also, the results using only participants who pass both manipulation checks on
voluntary/mandatory disclosure and voluntary/mandatory audit are the same as
the results for the full sample.
24
Given that the investment potential rating for the No-MW disclosure firm is
higher in VD_MA versus VD_VA (4.24 versus 2.97, t ¼ 2.25, one-tailed p ¼ 0.014), the
mandatory audit itself has a positive effect on investment potential evaluations
under a VD setting. As such the null incremental effect of the mandatory audit
when mandatory disclosure is already in place is not because mandatory audit itself
has no effect.
 K. Kelly, H.-T. Tan / Accounting, Organizations and Society 56 (2017) 1e20 11

Table 1
Descriptive statistics on investment potential evaluations and planned contrasts.

Panel A: Means (standard deviations) of investment potential evaluations ratinga

Regulatory regime conditionb N No-MW disclosure firmc MW disclosure firmc

Voluntary disclosure/voluntary audit 17 2.97 (2.53) 1.74 (2.72)

[VD_VA]
Voluntary disclosure/mandatory audit 19 4.24 (1.21) 2.18 (2.30)
[VD_MA]
Mandatory disclosure/voluntary audit 16 4.50 (1.39) 2.34 (2.30)
[MD_VA]
Mandatory disclosure/mandatory audit 17 4.26 (1.29) 2.32 (2.76)
[MD_MA]

Panel B: Univariate ANOVAs of investment potential evaluations

ANOVA for No-MW disclosure firm

Source SS df MS F p

Mandatory Disclosure 10.42 1 10.42 3.67 0.060

Mandatory Audit 4.57 1 4.57 1.61 0.209
Mandatory Disclosure*Mandatory Audit 9.69 1 9.69 3.41 0.069
Error 184.48 65 2.84
Overall F ¼ 2.85, p ¼ 0.044

ANOVA for MW disclosure firm

Source SS df MS F p

Mandatory Disclosure 2.40 1 2.40 0.38 0.542

Mandatory Audit 0.79 1 0.79 0.12 0.726
Mandatory Disclosure*Mandatory Audit 0.95 1 0.95 0.15 0.702
Error 414.99 65 6.38
Overall F ¼ 0.21, p ¼ 0.888

Panel C: MANOVA planned contrasts for investment potential evaluations

F p Standardized canonical coefficient

No-MW disclosure firm MW disclosure firm

Overall contrast for H1: 4.18 0.010* 1.08 0.12

3 VD_VA, þ1 VD_MA, þ1 MD_VA, þ 1 MD_MA
H1a: MD_VA and VD_MA > VD_VA (i.e., 2 VD_VA, þ1 VD_MA, þ1 MD_VA, 0 MD_MA) 3.93 0.012* 1.08 0.14
H1b: MD_VA and VD_MA ¼ MD_MA (i.e., 0 VD_VA, 1 VD_MA, 1 MD_VA, þ2 MD_MA) 0.04 0.965 1.07 0.67

Panel D: Univariate planned contrasts for investment potential evaluations

F pd

For No-MW disclosure firm

Overall contrast for H1: 8.38 0.003*

3 VD_VA, þ1 VD_MA, þ1 MD_VA, þ 1 MD_MA
H1a: MD_VA and VD_MA > VD_VA 7.86 0.003*
H1b: MD_VA and VD_MA ¼ MD_MA 0.04 0.836

For MW disclosure firm

Overall contrast: 0.60 0.440

3 VD_VA, þ1 VD_MA, þ1 MD_VA, þ 1 MD_MA
H2a: MD_VA and VD_MA ¼ VD_VA 0.50 0.482
H2b: MD_VA and VD_MA ¼ MD_MA 0.01 0.937
a
Investment potential evaluations rating is the average responses to two questions regarding the stock's investment attractiveness and the company's earnings potential.
b
Participants in the VD_VA and VD_MA (MD_VA and MD_MA) conditions are told that management disclosure of identified material weaknesses in the internal controls over
financial reporting (ICFR) is voluntary (mandatory). Participants in the VD_VA and MD_VA (VD_MA MD_MA) conditions are told that an independent audit of the ICFR is
voluntary (mandatory). Participants in the VD_MA and MD_MA conditions are additionally told in the note disclosures of the two firms that an independent auditor had
conducted their own evaluation of the firms' ICFR, while there is no mention of an independent ICFR audit in the VD_VA and MD_VA conditions.
c
All participants are given financial data and note disclosures of two firms, a No-MW disclosure firm that discloses that management did not identify any material weakness
and a MW disclosure firm that discloses that management has identified material weaknesses described in the note.
d
One-tailed p values are reported for directional predictions in H1 and H1a (and indicated with *). Two-tailed p values are reported for all other contrasts.

4.2. Firms that disclose material weaknesses (MW disclosure firms)
H2 is in the null form and predicts that neither a mandatory
disclosure regime alone nor a mandatory audit regime alone will
have any effects on investment potential evaluation of a MW
disclosure firm (see Fig. 2 Panel A). H2 suggests that the significant
multivariate differences in the multivariate contrasts reported
earlier are driven by the No-MW disclosure firm rather than by the
MW disclosure firm. Univariate contrasts for the No-MW disclosure
firm discussed in the preceding sub-section indicate significant
contrasts for H1 and H1a, and a non-significant contrast for the null
H1b. We next conduct univariate contrasts to test H2.
Unlike the significant multivariate contrast for H1 (3 for
VD_VA, þ1 for VD_MA, þ1 for MD_VA, þ1 for MD_MA), the uni-
variate contrast for the MW disclosure firm is not significant
(Table 1 Panel D, F ¼ 0.60, one-tailed p ¼ 0.440). H2a predicts that
vis-
a-vis a voluntary disclosure/voluntary audit regime, neither a
mandatory disclosure regime alone nor a mandatory audit regime
12 K. Kelly, H.-T. Tan / Accounting, Organizations and Society 56 (2017) 1e20
alone has any significant effect on investment potential evaluation
of a MW disclosure firm. We find no evidence to reject the null H2a,
with a univariate contrast indicating that investment potential
evaluation of a MW-disclosure firm is not different in VD_VA versus
MD_VA and VD_MA (Table 1 Panel D, F ¼ 0.50, p ¼ 0.482). Invest-
ment potential evaluation is also not different in MD_VA versus
VD_VA (t ¼ 0.69, p ¼ 0.492), and in VD_MA versus VD_VA (t ¼ 0.53,
p ¼ 0.596) (untabulated). We also find no evidence to reject the null
H2b, with a univariate contrast indicating that investment potential
evaluation of a MW-disclosure firm is not different in MD_MA
versus MD_VA and VD_MA (Table 1 Panel D, F ¼ 0.01, p ¼ 0.937).
Investment potential evaluation is also not different in MD_MA
versus MD_VA (t ¼ 0.02, p ¼ 0.982), and in MD_MA versus VD_MA
(t ¼ 0.17, p ¼ 0.869) (untabulated). Therefore, in contrast to the
positive and substitutory effects of mandatory disclosure and
mandatory audit on the investment potential evaluation of a No-
MW disclosure firm, there are no effects of mandatory disclosure
or mandatory audit on the investment potential evaluation of a MW
disclosure firm.
4.3. Additional analyses on reliability and relevance assessments
H1 posits that mandatory disclosure and mandatory audit act as
substitutable regulatory mechanisms that increase investors'
perceived reliability of No-MW disclosures and their relevance to
investment potential evaluations. In contrast, H2 posits that
mandatory disclosure and mandatory audit do not increase in-
vestors' perceived reliability and relevance of MW disclosures.
Therefore, our post-experimental questions asked all participants
to indicate their agreement for each type of disclosure (No-MW and
MW) whether, in general, (a) making management disclosure of
ICFR material weaknesses mandatory, and (b) making an ICFR audit
mandatory, make the disclosure (i) more free from error and bias
(i.e., reliable) and (ii) more likely to make a difference in investors'
decisions (i.e., relevant) (on scales of 7 “Strongly disagree” to þ7
“Strongly agree”).
Participants agree that both mandatory audit and mandatory
disclosure increase the reliability and the relevance of No-MW
disclosures (all responses are greater than zero, all p
values  0.002). Paired t-tests show that participants believe that
mandating audit increases the reliability of a No-MW disclosure
more than mandating disclosure (2.58 versus 1.39, t ¼ 2.92,
p ¼ 0.005). However, they do not believe that mandating audit
increases the relevance of a No-MW disclosure to investors' de-
cisions more so than mandating disclosure (1.71 versus 1.74,
t ¼ 0.09, p ¼ 0.927).
We also find that participants believe that both mandatory audit
and mandatory disclosure increase the reliability and the relevance
of MW disclosures (all responses are greater than zero, all p
values  0.001). Paired t-tests show that participants believe that
mandating audit increases the reliability of a MW disclosure more
than mandating disclosure (3.23 versus 1.70, t ¼ 3.70, p < 0.001),
but they do not believe that mandating audit increases the rele-
vance of a MW disclosure more than mandating disclosure (1.48
versus 1.67, t ¼ 0.63, p ¼ 0.529).
Together, these results are consistent with participants treating
mandatory disclosure and mandatory audit as substitutes in terms
of making a difference to their investment potential evaluations for
No-MW disclosure firms.
4.4. Additional analyses on TRUSTWORTHINESS and ICFR-QUALITY
assessments
In our hypotheses development, we argue that the NO-MW
disclosure is consistent with company disclosure incentives, and
that both mandatory disclosure regime alone and mandatory audit
regime alone would increase the reliability of a No-MW disclosure,
and hence its relevance. In contrast, the MW disclosure is incon-
sistent with company disclosure incentives and hence already
reliable. Mandatory disclosure and/or mandatory audit are not
expected to have further incremental effects.
A reliable ICFR disclosure is perceived to be trustworthy, and we
use participants' assessments of TRUSTWORTHINESS of each firm as
a proxy for reliability. A relevant ICFR disclosure is one that ulti-
mately makes a difference in investor assessment of the firm's in-
vestment potential. To the extent that investors perceive ICFR
quality to be an important element in their assessments of the
firm's investment potential, ICFR-QUALITY can be interpreted as a
proxy for relevance.25 The univariate 2
2 ANOVA is significant for
the No-MW disclosure firm for both TRUSTWORTHINESS (F ¼ 4.03,
p ¼ 0.011) and ICFR-QUALITY (F ¼ 5.65, p ¼ 0.002), and it is not
significant for the MW disclosure firm for both TRUSTWORTHINESS
(F ¼ 1.06, p ¼ 0.372) and ICFR-QUALITY (F ¼ 0.06, p ¼ 0.981); see
Table 2 for results. Consistent with our expectations, our manipu-
lated regulatory regimes affect the perceived trustworthiness and
ICFR quality of the No-MW disclosure firm, but not the perceived
trustworthiness and ICFR quality of the MW disclosure firm. We
then conduct contrasts of differences in TRUSTWORTHINESS and
ICFR-QUALITY.
We use similar contrast weights used in H1 to test for the pre-
dicted ordinal interaction for the No-MW disclosure firm: 3 for
VD_VA, þ1 for VD_MA, þ1 for MD_VA, þ1 for MD_MA. The overall
contrasts for TRUSTWORTHINESS (F ¼ 9.08, one-tailed p ¼ 0.002)
and ICFR-QUALITY (F ¼ 8.72, one-tailed p ¼ 0.002) are significant,
consistent with the results for investment potential evaluations in
H1. However, contrasts with the same contrast weights are not
significant for both TRUSTWORTHINESS (F ¼ 0.63, p ¼ 0.430) and
ICFR-QUALITY (F ¼ 0.03, p ¼ 0.874) for the MW disclosure firm.26
We then test if the effects on investment potential evaluation of
the No-MW disclosure firm are mediated through perceived
TRUSTWORTHINESS and ICFR-QUALITY of the firm. Kenny, Kashy, and
Bolger (1998, 260) outline two essential steps to establish media-
tion: (1) the independent variable is related to the mediator, and (2)
the mediator is in turn related to the dependent variable, while
controlling for the independent variable. Fulfilling the first step, our
results above indicate that the regulatory regime independent
variables affect both TRUSTWORTHINESS (F ¼ 4.03, p ¼ 0.011) and
ICFR-QUALITY (F ¼ 5.65, p ¼ 0.002) of the No-MW disclosure firm.
For the second step, an ANCOVA analysis indicates that, controlling
for the regulatory regime independent variables, ICFR-QUALITY
(F ¼ 16.41, p < 0.001) is significantly related to investment potential
rating of the No-MW disclosure firm, but TRUSTWORTHINESS
(F ¼ 1.49, p ¼ 0.226) is not. The overall contrast for the predicted
ordinal interaction pattern in investment potential evaluation also
reduces in significance when ICFR-QUALITY and TRUSTWORTHINESS
are included in the ANCOVA (i.e., from one-tailed p ¼ 0.003 to one-
tailed p ¼ 0.03). These results suggest that ICFR-QUALITY partially
mediates the effect of regulatory regimes on investment potential
25
We conducted verbal protocol analyses with 9 additional alumni participants
(see details in a later sub-section), and additionally asked them to rate the extent to
which information about a company having effective ICFR with no material
weaknesses would positively affect their assessment of a company's investment
potential (on a scale of 0 “Not at all” to 14 “To a large extent”. In line with our
expectations, participants' average rating of 9.22 (s.d. ¼ 4.18) is significantly larger
than 0 (t ¼ 6.62, one-tailed p < 0.001) and larger than the mid-point 7 at marginal
significance (t ¼ 1.60, one-tailed p ¼ 0.075).
26
The multivariate overall contrasts for both TRUSTWORTHINESS (untabulated,
F ¼ 5.26, p ¼ 0.008) and ICFR-QUALITY (untabulated, F ¼ 4.30, p ¼ 0.018) are
significant.
 K. Kelly, H.-T. Tan / Accounting, Organizations and Society 56 (2017) 1e20 13

Table 2
Descriptive statistics on trustworthiness and ICFR quality assessments and contrasts.

Panel A: Means (standard deviations) of trustworthiness and ICFR quality assessmentsa

Regulatory regime conditiona N No-MW disclosure firm MW disclosure firm

a a
TRUST-WORTHINESS (std dev.) ICFR-QUALITY (std dev.) TRUST-WORTHINESSa (std dev.) ICFR-QUALITYa (std dev.)

Voluntary disclosure/voluntary audit 17 6.38 (2.32) 4.94 (2.28) 8.97 (2.50) 2.47 (2.71)
[VD_ VA]
Voluntary disclosure/mandatory audit 19 8.68 (2.63) 6.21 (1.88) 7.84 (3.38) 2.47 (2.60)
[VD_MA]
Mandatory disclosure/voluntary audit 16 7.59 (1.78) 5.91 (2.33) 9.25 (2.37) 2.84 (3.51)
[MD_ VA]
Mandatory disclosure/mandatory audit 17 8.85 (2.58) 7.79 (1.69) 7.82 (3.47) 2.50 (3.35)
[MD_MA]

Panel B: Univariate ANOVAs for No-MW disclosure firm

TRUSTWORTHINESS

Source SS df MS F p

Mandatory Disclosure 8.18 1 8.18 1.46 0.231

Mandatory Audit 54.48 1 54.48 9.72 0.003
Mandatory Disclosure*Mandatory Audit 4.67 1 4.67 0.83 0.365
Error 364.36 65 5.61
Overall F ¼ 4.03, p ¼ 0.011

ICFR-QUALITY

Source SS df MS F p

Mandatory Disclosure 27.91 1 27.91 6.61 0.012

Mandatory Audit 42.82 1 42.82 10.15 0.002
Mandatory Disclosure*Mandatory Audit 1.64 1 1.64 0.39 0.535
Error 274.24 65 4.22
Overall F ¼ 5.65, p ¼ 0.002

Panel C: Univariate ANOVAs for MW disclosure firm

TRUSTWORTHINESS

Source SS df MS F p

Mandatory Disclosure 0.29 1 0.29 0.03 0.857

Mandatory Audit 28.04 1 28.04 3.13 0.082
Mandatory Disclosure*Mandatory Audit 0.38 1 0.38 0.04 0.837
Error 583.23 65 8.97
Overall F ¼ 1.06, p ¼ 0.372

ICFR-QUALITY

Source SS df MS F p

Mandatory Disclosure 0.69 1 0.69 0.07 0.787

Mandatory Audit 0.50 1 0.50 0.05 0.818
Mandatory Disclosure*Mandatory Audit 0.52 1 0.52 0.06 0.814
Error 604.33 65 9.30
Overall F ¼ 0.06, p ¼ 0.981

Panel D: Univariate planned contrasts TRUSTWORTHINESS and ICFR-QUALITY

TRUSTWORTHINESS F pb

For No-MW disclosure firm

Overall contrast for H1: 9.08 0.002*

3 VD_VA, þ1 VD_MA, þ1 MD_VA, þ 1 MD_MA
H1a: MD_VA and VD_MA > VD_VA 6.28 0.007*
H1b: MD_VA and VD_MA ¼ MD_MA 1.04 0.312

For MW disclosure firm

Overall contrast: 0.63 0.430

3 VD_VA, þ1 VD_MA, þ1 MD_VA, þ 1 MD_MA
H2a: MD_VA and VD_MA ¼ VD_VA 0.23 0.634
H2b: MD_VA and VD_MA ¼ MD_MA 0.66 0.418

ICFR-QUALITY F pb

For No-MW disclosure firm

Overall contrast for H1: 8.72 0.002*

3 VD_VA, þ1 VD_MA, þ1 MD_VA, þ 1 MD_MA
H1a: MD_VA and VD_MA > VD_VA 3.38 0.035*
H1b: MD_VA and VD_MA ¼ MD_MA 8.15 0.006

For MW disclosure firm

(continued on next page)

14 K. Kelly, H.-T. Tan / Accounting, Organizations and Society 56 (2017) 1e20
Table 2 (continued )
Panel D: Univariate planned contrasts TRUSTWORTHINESS and ICFR-QUALITY
TRUSTWORTHINESS
Overall contrast:
3 VD_VA, þ1 VD_MA, þ1 MD_VA, þ 1 MD_MA
H2a: MD_VA and VD_MA ¼ VD_VA
H2b: MD_VA and VD_MA ¼ MD_MA
a
TRUSTWORTHINESS is measured using two questions: (1) whether management is trustworthy in its disclosures about ICFR and (2) whether the company had undisclosed
material weaknesses in its ICFR (reverse coded). ICFR-QUALITY is measured using two questions: (1) quality of the company's ICFR and (2) whether management is competent
in establishing and maintaining adequate ICFR. See Table 1 for definitions of the four regulatory regime conditions (VD_VA, MD_VA, MD_MA, VD_MA) and the two firm types
(No-MW disclosure firm and MW-disclosure firm).
b
One-tailed p values are reported for directional predictions in H1 and H1a (and indicated with *). Two-tailed p values are reported for all other contrasts.
evaluations of the No-MW disclosure firm.
4.5. Verbal protocol analyses for additional alumni participants
In order to obtain more direct evidence of the thought processes
of participants when they are making investment potential evalu-
ations, we conducted verbal protocol analyses with 9 additional
alumni participants (2 each in VD_VA, VD_MA, and MD_MA, and 3 in
MD_VA). Alumni for the verbal protocol study were recruited via
instructors in the program who were in contact with these alumni.
Participants verbalized their thoughts as they completed a paper
instrument similar to the online instrument in the main experi-
ment, except that the paper instrument only included the main
dependent variables to keep the verbal protocol session to a
manageable time. Verbal protocol participants were paid $20 each.
The means for the dependent variables for the No-MW disclosure
firm are directionally similar to that in our main experiment, and all
results are qualitatively unchanged when we include these 9
additional alumni participants in the analyses. All 9 participants
verbally noted the regulatory regime manipulations.
After the main dependent variables, we asked additional ques-
tions not in the main experiment to get direct evidence on whether
participants consciously considered mandatory disclosure and
mandatory audit as substitutes. Participants agreed (on scales of 7
“Strongly disagree” to þ7 “Strongly agree”) more strongly with a
statement that they considered mandating management disclosure
of ICFR effectiveness and mandating independent ICFR audit as
complementary (mean ¼ 3.22, s.d. ¼ 2.77) than a statement that
they considered the two to be substitutory (mean ¼ 2.67,
s.d. ¼ 4.53) (t ¼ 2.99, p ¼ 0.017). Participants also indicated (on
scales of 0 “Not at all” to þ14 “To a large extent”) that additionally
mandating an ICFR audit when disclosure is already mandatory
further increases their investment potential evaluation
(mean ¼ 11.67, s.d. ¼ 1.87) to a larger extent than additionally
mandating disclosure when an ICFR audit is already mandatory
(mean ¼ 9.56, s.d. ¼ 4.28) (t ¼ 1.94, p ¼ 0.088). Thus, they believed
that the incremental effect of adding mandatory audit to manda-
tory disclosure is stronger than the incremental effect of adding
mandatory disclosure to mandatory audit. These results suggest
that notwithstanding existing regulations in Canada that mandate
only disclosure but not audit, verbal protocol participants indicated
a preference for mandatory audit over mandatory disclosure. To the
extent that the verbal protocols here are generalizable to findings in
our main experiment, this suggests that the substitutory effects of
mandatory disclosure and mandatory audit on investment poten-
tial evaluations in the main experiment are sub-conscious in that
participants in the verbal protocol study indicated that the two
mechanisms are complements rather than substitutes when asked
directly (contrary to what we observed in the main experiment).
We also analyzed the verbal protocols for evidence of why
participants perceive mandatory disclosure and mandatory audit to
F pb
0.03 0.874
0.04 0.836
0.03 0.861
have positive effects on the investment potential of a No-MW
disclosure firm. The positive effects of mandatory disclosure
verbalized included: (1) enabling companies to credibly commit to
consistently disclose both good news and bad news, and (2)
providing increased oversight over management disclosures. The
positive effect of mandatory audit verbalized is predominantly
about having an independent party counter management bias.
Excerpts of the verbal protocol quotes are reproduced in Appendix
2.
4.6. Supplementary experiments using undergraduates with less
knowledge
A major premise in our hypotheses development is that our
predicted pattern of configural information processing is a function
of investors' knowledge about companies' incentives for opportu-
nistic voluntary disclosures, impact of mandatory disclosures on
disclosure reliability, and the assurance value of ICFR audits. Our
main experiment used alumni participants from an accounting and
finance undergraduate program for whom we believe possess this
set of knowledge. This suggests that participants with less of this
knowledge would either not employ configural information pro-
cessing, or, if they do, in a different form. To test this prediction, we
repeated the same experiment with third-year and first-year un-
dergraduate students from the same accounting and finance un-
dergraduate program as our alumni participants.
4.7. Participants
To proxy for investors with moderate level of knowledge, 83
undergraduate students were recruited from a third year audit
strategy course. These students had already completed their
introductory management accounting, introductory financial ac-
counting, intermediate financial accounting, and introduction to
auditing and tax courses at the time they participated in the
experiment. The third-year undergraduate participants report an
average of 1.0 years of work experience and 1.3 years of experience
analyzing the financial performance of firms.
To proxy for investors with the lowest level of knowledge, 76
undergraduate participants were recruited from a first-year intro-
ductory management accounting course. These students had
already completed their introductory financial accounting course at
the time they participated in the experiment and have not had any
audit courses. The first-year undergraduate participants report an
average of 1.2 years of work experience and 0.5 years of experience
analyzing the financial performance of firms.
In contrast to our third-year and first-year undergraduate par-
ticipants, our alumni participants have completed all introductory,
intermediate, and advanced financial accounting, managerial ac-
counting, auditing, and finance courses, and they have an average
of 6.3 years of work experience and 4.2 years of experience
 K. Kelly, H.-T. Tan / Accounting, Organizations and Society 56 (2017) 1e20
analyzing the financial performance of firms. Neither third-year nor
first-year undergraduates have been exposed to ICFR disclosures
and ICFR audits in any of their courses at the point of participation
(although third-year undergraduates have exposure to auditing),
whereas alumni would have exposure to ICFR disclosures and ICFR
audits through their courses and work experience. Alumni partic-
ipants have significantly more work experience than first-year and
third-year undergraduates (all p values < 0.001), but first-year and
third-year undergraduates do not differ in their work experience
(p ¼ 0.819). Alumni participants have more experience analyzing
financial performance of firms than first-year and third-year un-
dergraduates (all p values < 0.001), and third-year undergraduates
have more experience analyzing the performance of firms than
first-year undergraduates (p ¼ 0.054).
Given that both third-year undergraduates and alumni have
taken auditing courses, we expect them to have more knowledge
about the assurance value of ICFR audits than first-year un-
dergraduates who have not taken any auditing courses. This sug-
gests that alumni and third-year undergraduates may be more
likely than first-year undergraduates to perceive a positive effect of
mandatory ICFR audit. Alumni are also expected to have more
knowledge about company disclosure incentives and the impact of
mandatory disclosures on disclosure reliability than third-year and
first-year undergraduates because their increased experience
analyzing the financial performance of firms and auditing/ac-
counting work experience likely expose them to instances where
companies are opportunistic in disclosing good news and with-
holding negative news. Alumni, through their auditing/accounting
work experience, may also have more exposure to the enforcement
mechanisms associated with mandatory disclosure that makes
mandatory disclosure more reliable (e.g., regulatory oversight and
independent auditors who review mandatory ICFR disclosures as
part of financial statement audits). If so, alumni may be more likely
than third-year and first-year undergraduates to perceive manda-
tory disclosure as a mechanism that increases the reliability of
positive ICFR disclosures that may otherwise be perceived as
opportunistic voluntary disclosures.
Participants who completed the study were paid $10 each. Un-
like the main experiment which was conducted via an online in-
strument, the two experiments with undergraduates were
conducted using a paper instrument.27A total of 65 of the 83 (78%)
third-year undergraduates and 46 of the 76 (61%) first-year un-
dergraduates passed both the manipulation checks on the
mandatory (versus voluntary) audit and mandatory (versus
voluntary) disclosure regulations. Similar to the main experiment,
we report the results for all participants, although the results are
similar for participants who passed both the manipulation checks.
4.8. Results
The results for undergraduate participants are reported in
Table 3. For third-year undergraduates, the ANOVA and post-hoc
contrasts for the No-MW disclosure firm indicate that mandatory
(versus voluntary) audit has a positive effect on the investment
potential evaluation (4.76 vs. 4.10, F ¼ 3.57, p ¼ 0.062), whereas
mandatory (versus voluntary) disclosure has no significant effect
(4.38 vs. 4.48, F ¼ 0.09, p ¼ 0.766) and neither does the interaction
effect (F ¼ 1.21, p ¼ 0.275). For first-year undergraduates, manda-
tory audit, mandatory disclosure, and the interaction effect are all
not significant for the No-MW disclosure firm. The overall contrast
27
Undergraduates were recruited in-class and completed the experiment in an
experimental laboratory. The authors were not course instructors for these
students.
15
for the predicted ordinal interaction pattern for the No-MW
disclosure firm (contrast weights: 3 for VD_VA, þ1 for
VD_MA, þ1 for MD_VA, þ1 for MD_MA) is not significant for both
third-year undergraduates (F ¼ 0.08, one-tailed p ¼ 0.386) and
first-year undergraduates (F ¼ 0.07, one-tailed p ¼ 0.395). Consis-
tent with the main experiment, mandatory audit, mandatory
disclosure, and the interaction effect are all not significant for the
MW disclosure firm for both third-year and first-year
undergraduates.
These results indicate that whilst moderately knowledgeable
third-year undergraduates perceive a positive effect of mandatory
audit on investment potential evaluation of a No-MW disclosure
firm like alumni participants, they do not perceive the positive ef-
fect of mandatory disclosure on investment potential evaluation of
a No-MW disclosure firm in contrast to alumni participants. The
least knowledgeable first-year undergraduate participants are not
able to perceive the positive effects of mandatory audit or
mandatory disclosure on investment potential evaluation of a No-
MW disclosure firm. These results are consistent with the differ-
ences in knowledge amongst the three types of participants dis-
cussed in the preceding sub-section. Prior research indicates that
people use counterfactual reasoning to assess causality, in that “to
understand whether X caused Y, a person may imagine that X had
not occurred” and “the easier it is to imagine that Y would not now
follow, the more likely the person is to view X as a cause of Y”
(Macrae, Milne, & Griffiths, 1993; Spellman & Mandel, 1999, p. 121;
Wells & Gavanski, 1989). In our context, investors are trying to
assess if mandatory disclosure (X) increases disclosure reliability
(Y). The easier it is for investors to imagine that increased reliability
would not follow if there had been no mandatory disclosure (i.e.,
under voluntary disclosure), the more likely investors would
conclude that mandatory disclosure increases reliability. Compared
to first-year and third-year undergraduates, alumni who have more
exposure to companies' incentives for opportunistic voluntary
disclosures may find it easier to imagine low reliability under
voluntary disclosure and hence they may be more likely to
conclude that mandating disclosure increases reliability.
5. Conclusion
This study examines how judgments of investors about a firm's
ICFR and investment potential are affected by regulations
mandating management disclosure of material weaknesses in the
ICFR and the independent audit of the ICFR. Specifically, we
examine whether investors consider mandatory disclosure and
mandatory audit to be substitutes rather than complements in
terms of their effects on investment potential evaluations. In our
main experiment, participants are alumni of a Canadian accounting
and finance undergraduate program and thus are more likely to
have domain-specific knowledge about companies' incentives
regarding voluntary disclosures, impact of mandatory disclosures
on disclosure reliability, and the assurance value of ICFR audits. In a
setting where management is required to evaluate the firm's ICFR
and the auditor has already given an unqualified opinion on the
financial statements, we establish that material weakness disclo-
sures affect investment potential evaluations with our alumni
participants judging a firm that discloses it has no material weak-
nesses (No-MW disclosure firm) to have a higher investment po-
tential and ICFR quality than a firm that discloses that it has
material weaknesses (MW disclosure firm). We find that
mandating ICFR material weakness disclosure alone and mandating
ICFR audit alone both positively affect investment potential evalu-
ations for the No-MW disclosure firm but not significantly so for the
MW disclosure firm.
Our results also suggest that mandating disclosure and
16 K. Kelly, H.-T. Tan / Accounting, Organizations and Society 56 (2017) 1e20

Table 3
Results for first-year undergraduates and third-year undergraduates.

Panel A: Means (standard deviations) of investment potential evaluations ratinga

First-year undergraduatesb Third-year undergraduatesb

a a a
Regulatory regime condition N No-MW disclosure firm MW disclosure firm N No-MW disclosure firma MW disclosure firma

Voluntary disclosure/voluntary audit 18 3.97 (2.21) 2.64 (3.32) 22 4.34 (1.58) 3.05 (2.31)
[VD_VA]
Voluntary disclosure/mandatory audit 20 4.03 (0.83) 2.53 (2.04) 21 4.62 (1.67) 2.71 (2.19)
[VD_MA]
Mandatory disclosure/voluntary audit 20 3.98 (2.19) 2.38 (2.06) 20 3.85 (1.52) 2.83 (2.30)
[MD_VA]
Mandatory disclosure/mandatory audit 18 3.50 (2.16) 1.86 (2.54) 20 4.90 (1.63) 3.08 (2.30)
[MD_MA]

Panel B: Univariate ANOVA of investment potential evaluations rating for No-MW disclosure firm

FIRST-YEAR UNDERGRADUATES

Source SS df MS F p

Mandatory Disclosure 1.29 1 1.29 0.35 0.557

Mandatory Audit 0.84 1 0.84 0.23 0.635
Mandatory Disclosure*Mandatory Audit 1.32 1 1.32 0.36 0.553
Error 266.71 72 3.70
Overall F ¼ 0.30, p ¼ 0.824

THIRD-YEAR UNDERGRADUATES

Source SS df MS F p

Mandatory Disclosure 0.23 1 0.23 0.09 0.766

Mandatory Audit 9.14 1 9.14 3.57 0.062
Mandatory Disclosure*Mandatory Audit 3.09 1 3.09 1.21 0.275
Error 202.00 79 2.56
Overall F ¼ 1.57, p ¼ 0.202

Panel C: Univariate ANOVA of investment potential evaluations rating for MW disclosure firm

FIRST-YEAR UNDERGRADUATES

Source SS df MS F p

Mandatory Disclosure 4.08 1 4.08 0.64 0.426

Mandatory Audit 1.87 1 1.87 0.29 0.590
Mandatory Disclosure*Mandatory Audit 0.76 1 0.76 0.12 0.731
Error 457.48 72 6.35
Overall F ¼ 0.34, p ¼ 0.799

THIRD-YEAR UNDERGRADUATES

Source SS df MS F p

Mandatory Disclosure 0.10 1 0.10 0.02 0.889

Mandatory Audit 0.03 1 0.03 0.03 0.936
Mandatory Disclosure*Mandatory Audit 1.75 1 1.75 0.34 0.563
Error 409.02 79 5.18
Overall F ¼ 0.12, p ¼ 0.947

Panel D: Univariate planned contrasts for investment potential evaluations rating

FIRST-YEAR UNDERGRADUATES pc

For No-MW disclosure firm

Overall contrast for H1: F ¼ 0.07 0.395*
3 VD_VA, þ1 VD_MA, þ1 MD_VA, þ 1 MD_MA

THIRD-YEAR UNDERGRADUATES pc

For No-MW disclosure firm

Overall contrast for H1: F ¼ 0.08 0.386*
3 VD_VA, þ1 VD_MA, þ1 MD_VA, þ 1 MD_MA
MA vs VA (4.76 vs 4.10) t ¼ 1.89 0.062
MD vs VD (4.38 vs 4.48) t ¼ 0.30 0.766
a
See Table 1 for definitions of the four regulatory regime conditions (VD_VA, MD_VA, MD_MA, VD_MA), the two firm types (No-MW disclosure firm and MW-disclosure
firm), and investment potential evaluations rating.
b
First-year undergraduates are recruited from the first-year introductory management accounting course. Third-year undergraduates are recruited from the third-year
audit strategy course.
c
One-tailed p values are reported for directional prediction in H1 (and indicated with *). Two-tailed p values are reported for all other contrasts.

mandating audit are substitutes for these alumni participants, in one regulatory mechanism alone. Therefore, if policy makers
that having both regulatory mechanisms does not significantly believe that investment potential evaluations should normatively
affect investment potential evaluations compared to having just be affected by adding mandatory audit to mandatory disclosure,
 K. Kelly, H.-T. Tan / Accounting, Organizations and Society 56 (2017) 1e20
more investor education may be needed to help investors under-
stand the incremental benefit of an ICFR audit.
We also conduct supplementary experiments with first-year
undergraduates and third-year undergraduates of the same ac-
counting and finance program. We find that our undergraduate
participants do not display configural information processing as our
alumni participants. Specifically, neither mandatory audit nor
mandatory disclosure affects investment potential evaluation of a
No-MW disclosure firm for the least knowledgeable first-year un-
dergraduates, whereas only mandatory audit but not mandatory
disclosure positively affects investment potential evaluation of a
No-MW disclosure firm for the moderately knowledgeable third-
year undergraduates.
There are several limitations to our study. First, management
ICFR disclosures and ICFR quality are exogenous to the regulatory
regimes manipulated in our experiments. However, mandating
disclosure and mandating audit could benefit investors further if
they actually improve management disclosure behavior or the ICFR
quality in companies (but see Kinney & Shepardson, 2011). Second,
we acknowledge that the complementary effect of ICFR audit and
ICFR disclosure may be more apparent in other settings which
future research can examine, such as (1) when investors perceive
that there are weaker investor protection laws with respect to
company misrepresentation but stronger investor protection laws
with respect to auditor misrepresentation, (2) when investors are
concurrently comparing no material weakness disclosures with
and without a voluntary ICFR audit on a within-subject basis, and
(3) for firms with poorer financials. Our participants are from
Canada where recent legislative changes have strengthened
investor protection laws with respect to both company misrepre-
sentation and auditor misrepresentation.28 The relative impact of
mandating ICFR audit and ICFR disclosure may vary across regimes
with differences in the strength of investor protection laws, auditor
liability, and regulatory enforcement capabilities. Third, our within-
subjects manipulation of the No-MW versus MW disclosure en-
hances the salience of these conditions. However, we do not believe
that it would create demand effects for our main results, which
contrast participants' responses across regulatory regimes that are
manipulated between-subjects. However, we note that because
MW disclosure firms are not common, having participants
concurrently review a No-MW disclosure firm and a MW disclosure
firm may increase participants' perception of the effectiveness of
the hypothetical mandatory disclosure regime. The greater doubt
investors hold about companies' compliance with a mandatory
disclosure regulation, the less likely mandatory disclosure would be
perceived as an effective substitute for mandatory audit. Future
research can examine whether investor perceptions of the effec-
tiveness of mandatory disclosure and mandatory audit are influ-
enced by the type of firms they review concurrently, in a configural
manner. Lastly, our participants are Canadians. The existing regu-
latory regime in Canada (i.e., one that mandates only disclosure)
may, however, bias them towards favoring only mandating disclo-
sure rather than mandating both audit and disclosure. Our verbal
protocol results do not support this view, with participants indi-
cating that they consider mandatory audit and mandatory
28
In Canada, changes were made in 2005 to securities legislation to make it easier
for investors to sue both the company and its auditors for negligent misrepresen-
tation (Emerson & Clarke, 2003; Lam, Marques, & O0 Brien 2013). Specifically, the
Ontario Securities Act (Section 138.3 (1)), with respect to an issuer releasing a
document that contains a misrepresentation, allows an investor to a right of action
for damages against the issuer, directors and officers of the issuer, and third-party
experts such as auditors, regardless of whether the investor relied on the misrep-
resentation (http://www.osc.gov.on.ca/en/SecuritiesLaw_ar_20050805_notice-
amend-act.jsp).
17
disclosure to be complements rather than substitutes. As an alter-
native, we could have used U.S. participants who are accustomed to
a regime that mandates both disclosure and audit for large issuers.
Parallel to the situation with Canadian participants, there may be a
reverse bias where participants favor mandating both audit and
disclosure rather than mandating only disclosure. In addition, given
that few countries (e.g., U.S. and Japan) mandate the ICFR audit,
using Canadian participants may have broader generalizability to
many regimes across the world that have no mandatory audits. We
also acknowledge that investors who are more experienced in
investing compared to our participants may not respond similarly if
their personal investment experience gives them more information
to assess how mandatory ICFR disclosures and audits actually
impact a firm's stock prices and investment potential.
Appendix 1
Excerpts of experiment materials
Between-subjects Manipulations of VD_VA (Voluntary
Disclosure and Voluntary Audit), MD_VA (Mandatory Disclosure
and Voluntary Audit), VD_MA (Voluntary Disclosure and
Mandatory Audit), and MD_MA (Mandatory Disclosure and
Mandatory Audit) conditions:
Please take note of the following regulations in this hypo-
thetical stock exchange with respect to management disclosures
and independent audits relating to firms' internal control over
financial reporting:
[VD_VA and VD_MA conditions state:]Management disclosure of
identified material weaknesses in the internal control over financial
reporting is voluntary. Some management voluntarily discloses
whether or not there are material weaknesses; others choose to be
silent on this matter.
[Versus MD_VA and MD_MA conditions state:]Management
disclosure of identified material weaknesses in the internal control
over financial reporting is mandatory.
[VD_VA and MD_VA conditions state:] An independent audit of
the internal control over financial reporting is voluntary. Some
companies engage an independent auditor voluntarily and
disclose that; others choose not to engage an independent auditor
and are silent on this matter.
[Versus VD_MA and MD_MA conditions state:]An independent
audit of the internal control over financial reporting is mandatory.
Within-subjects Manipulations of NO-MW disclosure firm
versus MW disclosure firm and between-subjects Manipulations
of VD_MA and MD_MA versus VD_VA and MD_VA conditions:
[MW disclosure firm states:] Selected notes to financial
statements
1) Controls and procedures
Management is responsible for establishing and maintaining
adequate internal control over financial reporting. Our internal
control over financial reporting is designed with the objective of
providing reasonable assurance regarding the reliability of financial
reporting and the preparation of financial statements for external
purposes in accordance with generally accepted accounting prin-
ciples. Based on an evaluation of the effectiveness of our internal
control over financial reporting, management identified the
following material weaknesses:
Several weaknesses in the accounting for fixed assets were
identified. First, there was inadequate documentation of fixed as-
sets to assist in the identification and location of certain fixed as-
sets. Specifically, several fixed assets could not be located at the
sites that were recorded in the accounting records. Second, the
18 K. Kelly, H.-T. Tan / Accounting, Organizations and Society 56 (2017) 1e20
company failed to affix ID tags to fixed assets in locations outside of
corporate headquarters. In addition, the company did not docu-
ment the policies, practices, and procedures pertaining to the
classification of certain expenditures as fixed assets. Finally, the
requisite documentation supporting the authorization and cate-
gorization of fixed assets could not be obtained. These control
weaknesses resulted in an overstatement of fixed assets. Specif-
ically, Repairs and Maintenance Expense amounting to 2.4% of net
income was erroneously capitalized as Property, Plant, and Equip-
ment. While the error was detected and corrected during the in-
ternal control evaluation, the company acknowledges that it could
have resulted in a misstatement in the annual financial statements
that otherwise would not have been prevented or detected.
[VD_MA/MW disclosure firm and MD_MA/MW disclosure firm
conditions additionally state the following, versus VD_VA/MW
disclosure firm and MD_VA/MW disclosure firm conditions which do
not have this statement:] Our independent auditor, ABC LLP, has
also conducted their own evaluation of our internal control over
financial reporting. In their opinion, the company had material
weaknesses in internal control over financial reporting as of June
30, 2010, as described above
[Versus No-MW disclosure firm states:] Selected notes to finan-
cial statements
1) Controls and procedures
Management is responsible for establishing and maintaining
adequate internal control over financial reporting. Our internal
control over financial reporting is designed with the objective of
providing reasonable assurance regarding the reliability of financial
reporting and the preparation of financial statements for external
purposes in accordance with generally accepted accounting prin-
ciples. Based on an evaluation of the effectiveness of our internal
control over financial reporting, management did not identify any
material weaknesses.
[VD_MA/No-MW disclosure firm and MD_MA/No-MW disclosure
firm conditions additionally state the following, versus VD_VA/No-MW
disclosure firm and MD_VA/No-MW disclosure firm conditions which
do not have this statement:] Our independent auditor, DEF LLP, has
also conducted their own evaluation of our internal control over
financial reporting. In their opinion, the company had no material
weaknesses in internal control over financial reporting as of June
30, 2010.
Appendix 2
A) Examples of quotes on why mandatory ICFR disclosure increases
investment potential evaluations:
(1) Enables companies to credibly commit to consistently
disclose both good news and bad news
Participant 6 (VD_MA): “It [No-MW disclosure] wouldn't have a
higher effect on me from an investment potential perspective if
management was telling about their ICFR on a voluntary basis
subject to that voluntary disclosure being present in all, I guess,
reviewable years. … when it's voluntary, …yeah, so they could not
disclose in ineffective years and disclose in effective years … I
would assume that if it's mandatory to disclose ICFR effectiveness,
then management must disclose in poor years and in favourable
years for ICFR effectiveness, so that would make me happy.”
Participant 7 (VD_MA) “Management disclosure of identified
material weaknesses and the internal control over the financial
reporting is voluntary. So that sounds important to me. Since we
[management] can decide not to disclose, it kind of weakens sort of
the point, I guess, of [management] being required to do sort of
internal evaluation of internal controls over financial reporting …
this [No-MW disclosure] is assertion by management is not,
necessarily, valid or carries any weight because they're not required
to disclose anything”.
Participant 9 (VD_VA): “So, if it's mandatory versus voluntary
disclosure, that does really impact my assessment. For me, I think
they [companies under mandatory disclosure] would have pro-
cesses. If it [disclosure] is voluntary, if there's silence on it, there's
no trust there for me. I guess the regulation kind of builds a bit of
trust, a bit of confidence. There's a bit more consistency in each
company. I don't have to wonder is there any disclosure missing
here.”
(2) Provides increased oversight over management disclosures
Participant 1 (MD_VA): “if regulation makes management
disclosure mandatory … I think I would feel more comfortable if it
[disclosure] is mandatory because … versus voluntary. There's no
faith. There's no oversight, so I could lie and nobody would catch
me, so it [mandatory disclosure] does positively affect it [my
assessment of the company's investment potential] … [for
mandatory disclosure] there would still be some sort of oversight,
whether it's from an independent auditor or regulatory, or what-
ever oversight there is.”
B) Examples of quotes on why mandatory ICFR audit increases
investment potential evaluations:
(1) Provides an independent opinion that counters management
bias
Participant 3 (MD_VA): “There is the margin for error that
management could be hiding potential control weaknesses just as a
result of an independent party not performing the control pro-
cedures because with internal controls what you can have is there
can be a lot of judgment involved in determining whether or not in
a specific instance a control is going to pass or a control is going to
fail; and, in my mind, if you have management attempting to
objectively evaluate their controls, they're going to err on the side
of a control being effective as opposed to being ineffective.”
Participant 6 (VD_MA): “Management can tell me whatever they
want, right? If I have an auditor, that's also corroborating that or
telling me that's okay, and, you know, they've independently come
to the same conclusion management has, and I've got better in-
formation than when just management is telling me something
because if management is just telling me something, then there's
the element of bias that needs to be considered, whereas when the
auditors tell me the same thing, there's less bias to affect my
judgment, or to consider when I'm judging.”
C) Examples of quotes on preference for mandatory ICFR audit over
mandatory ICFR disclosure:
Participant 2 (MD_VA): “I think that management can do it [ICFR
evaluation]. … but I think an auditor should be … if I want credi-
bility to it, somebody else independent has to do it.”
Participant 4 (MD_MA): “The independent firm to assess the
internal controls over financial reporting, I think that's such a
stronger sort of assurance to investors, so, in my opinion, I think
that … mandatory disclosure over internal weaknesses of their
internal controls, I don't think that's enough. I think that I would
feel way more comfortable having that independent audit over
internal controls.”
Participant 7 (VD_MA): “I definitely agree that a mandatory
 K. Kelly, H.-T. Tan / Accounting, Organizations and Society 56 (2017) 1e20
disclosure is important, but I think sort of the ability to trust
management and to keep the sort of professional scepticism alive,
the fact that this doesn't seem to make any indication that it’s, you
know, independently verified or sort of tested, it doesn't really add
any weight to management's disclosure, I guess. So I think it's
definitely stronger than sort of completely voluntary disclosure, but
it wouldn't affect my assessment that extensively without an in-
dependent verification because sort of I guess, from my experience
sometimes management thinks that something is fine and dandy;
but when we go in, it's, like, “Nah, maybe not”.”
D) Examples of quotes on the limits of mandatory ICFR audit:
Participant 1 ((MD_VA): “If regulation makes an independent
ICFR audit mandatory … I'm running into, like, a pros and con
situation. There are pros to always adding things on; but from a cost
benefit analysis or an effectiveness situation, sometimes doing
these things is just more work, and it's not … it doesn't, necessarily,
end up accomplishing the purpose.”
Participant 2 (MD_VA): “so I'm going to feel better about having
management and auditors doing it [ICFR evaluation]. The only
problem I would have with that would be is if the auditors don't go
any further than what management has, so I think there might be a
little bit of a risk … with the auditors getting a little bit of familiarity
by just working off management's document.”
Participant 7 (VD_MA): “I think mandatory independent audit is
… stronger than lack of independent ICFR audit … I don't think it is
a make-or-break issue. … Yeah, so I guess in comparing environ-
ments, so, say, you know, they're working in different jurisdictions,
and one jurisdiction requires an independent auditor and the other
does not. I would definitely take that into consideration; but if sort
of the financials for one company which asserts that they have
effective ICFR and they have an independent auditor saying that
their financial statements are unqualified, and sort of that company
is sort of, I guess … yeah, investment potential is significantly
stronger than another company who has sort of an independent
ICFR audit and says that they're fine. I wouldn't, necessarily, dis-
count the strength of the investment opportunity just because the
… because they haven't been independently audited for internal
controls over financial reporting because the independent auditor
has audited the financial statements, which I'm using to form my
opinion on the investment opportunity, at least those are good or
free of material misstatement.”
Note: Words in square brackets [ ] are added for clarification and
they are not in the verbatim quotes of participants.
References
Allee, K. D., & Yohn, T. L. (2009). The demand for financial statements in an un-
regulated environment: An examination of the production and use of financial
statements by privately held small businesses. The Accounting Review, 84, 1e25.
Ashbaugh-Skaife, H., Collins, D. W., & Kinney, W. R. (2007). The discovery and
reporting of internal control deficiencies prior to SOX-mandated audits. Journal
of Accounting and Economics, 44, 166e192.
Ashbaugh-Skaife, H., Collins, D. W., Kinney, W. R., & LaFond, R. (2009). The effect of
SOX internal control deficiencies on firm risk and cost of equity. Journal of Ac-
counting Research, 47, 1e43.
Ball, R., Jayaraman, S., & Shivakumar, L. (2012). Audited financial reporting and
voluntary disclosure as complements: A test of the confirmation hypothesis.
Journal of Accounting and Economics, 53, 136e166.
Beneish, M. D., Billings, M. B., & Hodder, L. D. (2008). Internal control weaknesses
and information uncertainty. The Accounting Review, 83, 665e703.
Bergh, D. D., & Fairbanks, J. F. (2002). Measuring and testing change in strategic
management research. Strategic Management Journal, 23(4), 359e366.
Beyer, A., Cohen, D. A., Lys, T. Z., & Walther, B. R. (2010). The financial reporting
environment: Review of the recent literature. Journal of Accounting and Eco-
nomics, 50, 296e343.
Blackwell, D. W., Noland, T. R., & Winters, D. B. (1998). The value of auditor assur-
ance: Evidence from loan pricing. Journal of Accounting Research, 36, 57e70.
Brown, C. E., & Solomon, I. (1990). Auditor configural information processing in
19
control risk assessment. Auditing: A Journal of Practice & Theory, 9, 17e38.
Brown, C. E., & Solomon, I. (1991). Configural information processing in auditing:
The role of domain-specific knowledge. The Accounting Review, 66, 100e119.
Buckless, F. A., & Ravenscroft, S. P. (1990). Contrast coding: A refinement of ANOVA
in behavioral analysis. The Accounting Review, 65(4), 933e945.
Canadian Securities Administrators (CSA) Staff Notice 52-313. (2006). Status of
proposed MI 52-111 reporting on internal control over financial reporting and
proposed amended and restated MI 52-109 certification of dislcosure in issuers'
annual and interim findings. Montreal, Quebec: CSA.
Canadian Securities Administrators (CSA) Staff Notice 52-325. (2009). Certification
compliance review. Montreal, Quebec: CSA.
Coram, P. J., Monroe, G. S., & Woodliff, R. R. (2009). The value of assurance on
voluntary nonfinancial disclosure: An experimental evaluation. Auditing: A
Journal of Practice & Theory, 28, 137e151.
Crawford, V. P., & Sobel, J. (1982). Strategic information transmission. Econometrica,
50, 1431e1451.
Dopuch, N., & King, R. R. (1991). Experimental tests of auditing as a credibility
generating mechanism. Vancouver, British Columbia: Canadian Certified General
Accountants’ Research Foundation.
Doyle, J. T., Ge, W., & McVay, S. (2007). Accruals quality and internal control over
financial reporting. The Accounting Review, 82, 1141e1170.
Dutch Corporate Governance Code. (2009). Dutch corporate governance code moni-
toring committee. Available at: http://www.commissiecorporategovernance.nl/
Dutch_Corporate_Governance_Code.
Eagly, A. H., Wood, W., & Chaiken, S. (1978). Causal inferences about communicators
and their effect on opinion change. Journal of Personality and Social Psychology,
36, 424e435.
Earley, C. E., Hoffman, V. B., & Joe, J. R. (2008). Reducing management's influence on
auditors' judgments: An experimental investigation of SOX 404 assessments.
The Accounting Review, 83, 1461e1485.
Edgell, S. E. (1978). Configural information processing in two-cue probability
learning. Organizational Behavior and Human Performance, 22, 404e416.
Edwards, J. R. (1995). Alternatives to difference scores as dependent variables in the
study of congruence in organizational research. Organizational Behavior and
Human Decision Processes, 64(3), 307e324.
Edwards, J. R. (2001). Ten difference scores myths. Organizational Research Methods,
4(3), 265e287.
Elliot, W. B., Hodge, F. D., & Sedor, L. M. (2012). Using online video to announce a
restatement: Influences on investment decisions and the mediating role of
trust. The Accounting Review, 87, 513e535.
Emerson, G., & Clarke, G. A. (2003). Bill 198 and Ontario's securities Act: Giving in-
vestors and the OSC added muscle. Fasken and Martineau.
Financial Accounting Standards Board (FASB). (2008). Statement of financial ac-
counting concepts No. 2: Qualitative characteristics of accounting information.
Norwalk, CT: FASB. Available at: http://www.fasb.org/jsp/FASB/Document_C/
DocumentPage?cid¼1218220132599&acceptedDisclaimer¼true.
Frederickson, J. R., Hodge, F. D., & Pratt, J. H. (2006). The evolution of stock option
accounting: Disclosure, voluntary recognition, mandated recognition, and
management disavowals. The Accounting Review, 81, 1073e1093.
Garson, D. G. (2015). GLM multivarate, MANOVA, and canonical correlation. Asheboro,
NC: Statistical Associates Publishers.
Giffin, K. (1967). The contribution of studies of source credibility to a theory of
interpersonal trust in the communication process. Psychological Bulletin, 68,
104e120.
Gigler, F., & Hemmer, T. (1999). On the frequency, quality, and informational role of
mandatory financial reports. Journal of Accounting Research, 36(Supplement),
117e147.
Griffin, D., Murray, S., & Gonzalez, R. (1999). Difference score correlations in rela-
tionship research: A conceptual primer. Personal Relationships, 6, 505e518.
Hammersley, J. S., Myers, L. A., & Shakespeare, C. (2008). Market reactions to the
disclosure of internal control weaknesses and to the characteristics of those
weaknesses under section 302 of the Sarbanes Oxley Act of 2002. Review of
Accounting Studies, 13, 141e165.
Hammersley, J. S., Myers, L. A., & Zhou, J. (2012). The failure to remediate previously
disclosed material weaknesses. Auditing: A Journal of Practice & Theory, 31,
73e111.
Hermanson, D. R., & Ye, Z. (2009). Why do some accelerated filers with SOX Section
404 material weaknesses provide early warning under Section 302. Auditing: A
Journal of Practice & Theory, 28, 247e271.
Hirst, E. D., Koonce, L., & Simko, P. J. (1995). Investor reactions to financial analysts'
research reports. Journal of Accounting Research, 33, 335e351.
Hodge, F. D., Hopkins, P., & Pratt, J. (2006). Management reporting incentives and
classification credibility: The effects of reporting discretion and reputation.
Accounting, Organizations and Society, 31, 623e634.
Hom, P. W., Griffeth, R. W., Palich, L. E., & Bracker, J. S. (1999). Revisiting met ex-
pectations a a reason why realistic job previews work. Personnel Psychology, 52,
97e112.
International Accounting Standards Board. (2015). The conceptual Framework for
financial reporting. London, United Kingdom: IASB.
International Auditing and Assurance Standards Board. (2010). The auditor's re-
sponsibilities relating to other information in documents containing audited
financial statements. International Standard on Auditing No. 720. New York, New
York: International Auditing and Assurance Standards Board.
Japan Financial Services Agency. (2007). Provisional translation of “On the setting of
the standards and practice standards for management assessment and audit
20 K. Kelly, H.-T. Tan / Accounting, Organizations and Society 56 (2017) 1e20
concerning internal control over financial reporting (council opinions)”. Tokyo,
Japan: Japan Financial Services Agency. Available at: http://www.fsa.go.jp/en/
news/2007/20070420.html.
Kadous, K., Koonce, L., & Thayer, J. M. (2012). Do financial statement users judge
relevance based on properties of reliability. The Accounting Review, 87,
1335e1356.
Kahneman, D., & Frederick, S. (2002). Representativeness revisited: Attribute sub-
stitution in intuitive judgment. In T. Gilovich, D. Griffin, & D. Kahneman (Eds.),
Heuristics and biases: The psychology of intuitive judgment (pp. 49e81). New
York, New York: Cambridge University Press.
Kelley, H. H., & Michela, J. L. (1980). Attribution theory and research. Annual Review
of Psychology, 31, 457e501.
Kenny, D. A., Kashy, D. A., & Bolger, N. (1998). Data analysis in social psychology. In
D. Gilbert, S. Fiske, & G. Lindzey (Eds.), The handbook of social psychology Vol. 1
(pp. 233e265). Boston, Massachusetts: McGraw-Hill.
Kinney, W. R., & Shepardson, M. L. (2011). Do control effectiveness disclosures
require SOX 404 (b) internal control audits? A natural experiment with small
U.S. public companies. Journal of Accounting Research, 49, 413e448.
Klein, G., Jiang, J. J., & Cheney, P. (2009). Resolving difference score issues in in-
formation systems research. MIS Quarterly, 22(4), 811e826.
Koonce, L., Nelson, K. K., & Shakespeare, C. M. (2011). Judging the relevance of fair
value for financial instruments. The Accounting Review, 86, 2075e2098.
Kristof-Brown, A. L., Zimmerman, R. D., & Johnson, E. C. (2005). Consequences of
individuals' fit at work: A meta-analysis of person-job, person-organization,
person-group, and person-supervisor fit. Personnel Psychology, 58, 281e342.
Lam, M., Marques, E. C., & O’ Brien, M. J. P. (2013). Ontario court approves $117m
auditor liability settlement in Sino-Forest class action. McCarthy Tetrault. Available
at: http://www.mccarthy.ca/article_detail.aspx?id¼6236.
Lennox, C., & Pittman, J. A. (2011). Voluntary audits versus mandatory audits. The
Accounting Review, 86, 1655e1678.
Leuz, C. (2010). Different approaches to corporate reporting regulation: How ju-
risdictions differ and why. Accounting and Business Research, 40, 229e256.
Leuz, C., & Wysocki, P. (2008). Economic consequences of financial reporting and
disclosure regulation: A review and suggestions for future research. Working pa-
per. University of Chicago. Available at: SSRN http://papers.ssrn.com/sol3/
papers.cfm?abstract_id¼1105398.
Macrae, C. N., Milne, A. B., & Griffiths, R. J. (1993). Counterfactual thinking and the
perception of criminal behavior. British Journal of Psychology, 84, 221e226.
Maletta, M. J., & Kida, T. (1993). The effect of risk factors on auditors' configural
information processing. The Accounting Review, 68(3), 681e691.
Mercer, M. (2005). .The fleeting effects of disclosure forthcomingness on manage-
ment's reporting credibility. The Accounting Review, 80, 723e744.
Mullins, R. R., Ahearne, M., Lam, S. K., Hall, Z. R., & Boichuk, J. P. (2014). Know your
customer: How salesperson perceptions of customer relationship quality form
and influence account profitability. Journal of Marketing, 78, 38e58.
National Instrument 52-109. (2008). Certification of disclosure in issuers' annual and
interim filings. Montreal, Quebec: Canadian Securities Administrators.
Ogneva, M., Subramanyam, K. R., & Raghunandan, K. (2007). Internal control
weaknesses and cost of equity: Evidence from SOX Section 404 disclosures. The
Accounting Review, 80, 1255e1297.
Ostroff, C., Atwater, L. E., & Feinberg, B. J. (2004). Understanding self-other agree-
ment: A look at rater and ratee characteristics, context, and outcomes. Personnel
Psychology, 57, 333e375.
Perdue, B. C., & Summers, J. O. (1986). Checking the success of manipulations in
marketing experiments. Journal of Marketing Research, 23, 317e326.
Peter, J. P., Churchill, G. A., & Brown, T. J. (1993). Caution in the use of difference
scores in consumer research. Journal of Consumer Research, 19(4), 655e662.
Public Company Accounting Oversight Board (PCAOB). (2003). Interim auditing
standards: Other information in documents containing audited financial state-
ments. AU Section 550. (Washington, DC).
Rock, E. (2002). Securities regulation as lobster trap: A credible commitment theory
of mandatory disclosure. Cardozo Law Review, 23, 675e704.
Rosenthal, R., & Rosnow, R. L. (1985). Contrast Analysis: Focused comparisons in the
analysis of variance. Cambridge, UK: Cambridge University Press.
Scheiner, S. M. (2001). MANOVA: Multiple response variables and multi-species
interactions. In S. M. Scheiner, & J. Gurevitch (Eds.), Design and analysis of
ecological experiments (pp. 99e115). New York, New York: Oxford University
Press.
Securities and Exchange Commission (SEC). (2003). Final rule: Management's report
on internal control over financial reporting and certification of disclosure in ex-
change act periodic reports. Washington, DC: SEC.
Securities and Exchange Commission (SEC). (2004). Management's report on internal
control over financial reporting and certification of disclosure in Exchange Act
periodic reports: Frequently asked questions. Washington, DC: SEC.
Securities and Exchange Commission (SEC). (2009). Study of the Sarbanes-Oxley Act
of 2002 Section 404 internal control over financial reporting requirements.
Washington, DC: SEC.
Securities and Exchange Commission (SEC). (2010). Final rule: Internal control over
financial reporting in exchange act periodic reports of non-accelerated filers.
Washington, DC: SEC.
Seo, J., Gamache, D. L., Devers, C. A., & Carpenter, M. A. (2015). The role of CEO
relative standing in acquisition behavior and CEO pay. Strategic Management
Journal, 36, 1877e1894.
Sigall, H., & Mills, J. (1998). Measures of independent variables and mediators are
useful in social psychology experiments: But are they necessary? Personality
and Social Psychology Review, 2(3), 218e226.
Slovic, P., & Lichtenstein, S. (1971). Comparison of Bayesian and regression ap-
proaches to the study of information processing in judgment. Organizational
Behavior and Human Performance, 6, 649e744.
Spellman, B. A., & Mandel, D. R. (1999). When possibility informs reality: Coun-
terfactual thinking as a cue to causality. Current Directions in Psychological Sci-
ence, 8(4), 120e123.
Stone, E. R., Yates, A. J., & Caruthers, A. S. (2002). Risk taking in decision making for
others versus the self. Journal of Applied Social Psychology, 32, 1797e1824.
Sturm, R. E., Taylor, S. N., Atwater, L. E., & Braddy, P. W. (2014). Leader self-
awareness: An examination and implications of women's under-prediction.
Journal of Organizational Behavior, 35, 657e677.
Turnbull Guidance. (2005). Internal control: Revised guidance for directors on the
combined code. London, UK: Financial Reporting Council.
Watts, R. L., & Zimmerman, J. L. (1983). Agency problems, auditing, and the theory
of the firm: Some evidence. Journal of Law and Economics, 26, 613e633.
Wells, G. L., & Gavanski, I. (1989). Mental simulation of causality. Journal of Per-
sonality and Social Psychology, 56(2), 161e169.
Wilkinson, L. (1975). Response variable hypotheses in the multivariate analysis of
variance. Psychological Bulletin, 82(3), 408e412.
Wright, T. A. (1990). A comparison of difference score and pre-test post-test
repeated measures ANOVA: Implications for research in management infor-
mation systems. Journal of Applied Business Research, 6(4), 88e92.
 Accounting, Organizations and Society 56 (2017) 21e34

Contents lists available at ScienceDirect

Accounting, Organizations and Society

journal homepage: www.elsevier.com/locate/aos

Tournament group identity and performance: The moderating effect

of winner proportion
Khim Kelly a, Adam Presslee b, *
a
Kenneth G. Dixon School of Accounting, University of Central Florida, Orlando, FL, 32816-1400, USA
b
Katz Graduate School of Business, University of Pittsburgh, Pittsburgh, PA, 15260, USA

a r t i c l e i n f o a b s t r a c t

Article history: Tournament incentives are common in organizations, and how characteristics of the tournament group
Received 19 November 2015 (e.g., tournament group identity) and the tournament incentives (e.g., winner proportion) affect tour-
Received in revised form nament performance are of both practical and theoretical importance. We conduct two experiments in
6 December 2016
which participants compete for tournament rewards against others in their group. In both experiments,
Accepted 8 December 2016
Available online 31 January 2017
we manipulate the strength of participants' identity with their fellow group members and whether the
tournament has a small winner proportion with a single reward or a large winner proportion with
multiple rewards. In Experiment 1, we find increasing tournament group identity leads to higher other-
Keywords:
Competitiveness
regarding preference. We also find other-regarding preference decreases competitiveness more in a large
Group identity winner proportion tournament compared to a small winner proportion tournament. In Experiment 2, we
Other-regarding preference find increasing tournament group identity decreases performance in a real-effort task under a large
Tournament incentives winner proportion tournament, but it has no effect on performance under a small winner proportion
Winner proportion tournament. Together, the two experiments suggest that increasing tournament group identity increases
other-regarding preference, and other-regarding preference has a larger negative impact on competi-
tiveness and hence, tournament performance when the winner proportion is large than when it is small.
Our results highlight for managers the importance of considering group identity when determining
tournament winner proportions.
© 2016 Elsevier Ltd. All rights reserved.

1. Introduction variation in tournament group identity can be the result of differ-

Organizations commonly use tournament incentive systems in
which employees are rewarded based on their relative individual
performance (Hazels & Sasse, 2008; Kwoh, 2012). Examining how
tournament group identity affects performance is important
because the strength of tournament group identity can vary widely
in practice and can be influenced by organizational decisions. This
* Corresponding author.
E-mail addresses: Khim.Kelly@ucf.edu (K. Kelly), presslee@pitt.edu (A. Presslee).
1
Although it may seem counter intuitive, many companies that are known for
fostering high group identity amongst employees utilize tournament incentive
systems. For example, salesforce.com is ranked 8th on the Fortune's list of 2015's
top 100 places to work because of its great workplace atmosphere and its focus on
teamwork (Fortune, 2015). The company offers its top sales representatives tour-
nament rewards such as travel and shopping incentives. Baird & Co. is another
example that is ranked 5th on the list and it stresses the shared values employees
have. Baird & Co. allows employees to thank one another with Blue Chip Awards for
exceeding peer expectations on a project, and recognizes the top 2% of Blue Chip
winners at its annual meeting.
ences in organizational structures and opportunities that are either
conscious attempts by the organization to build a strong group
identity or unconscious organic developments.1 For example, a
group of competing employees may be in a smaller work group,
share common characteristics (e.g., similar age, educational back-
ground), work closely together in the same office, and/or interact
often; all of which tend to encourage stronger group identity. In
contrast, a group of competing employees may be in a larger work
group, have dissimilar characteristics, work in different offices or
telecommute, and/or interact infrequently.
A key design choice regarding tournament incentive systems is
the proportion of tournament participants to reward. Many com-
panies such as Avis and Samsung choose to reward only their top
employees (e.g., top 10%e20%) with a large tournament reward
(Keenan, 1995; Samsung, 2013). Other companies such as AIG
follow the advice of some compensation consultants to “share the
wealth” by rewarding the majority (e.g., top 80%e90%) of their
employees with smaller-sized tournament rewards (Demos, 2010;
Kwoh, 2012).

http://dx.doi.org/10.1016/j.aos.2016.12.001
0361-3682/© 2016 Elsevier Ltd. All rights reserved.
22 K. Kelly, A. Presslee / Accounting, Organizations and Society 56 (2017) 21e34
It is important for management accountants who are involved in
determining the winner proportion in tournaments to understand
how this choice can influence the effect of tournament group
identity on employee performance. Our study examines how the
strength of identity with fellow tournament participants affects
performance in a large winner proportion tournament versus a
small winner proportion tournament, filling a gap in extant liter-
ature (see Dechenaux, Kovenock, and Roman (2012) and Sisak
(2009) for reviews).
We posit that increasing the strength of tournament group
identity increases tournament participants' other-regarding pref-
erence, which decreases their competitiveness and hence, tourna-
ment performance. Compared to a small winner proportion, a large
winner proportion results in smaller valued rewards, which re-
duces the cost of acting on one's other-regarding preference by
lowering effort so others could earn a reward. Therefore, we further
posit that other-regarding preference decreases competitiveness
and hence, performance more in a large winner proportion tour-
nament than in a small winner proportion tournament.
We test our predictions with two experiments. We chose not to
test our predictions with a single experiment to avoid influencing
participants' performance by asking them about their other-
regarding preference and competitiveness before they exert effort
(Spencer, Zanna, & Fong, 2005). In both experiments, we manipu-
late winner proportion and tournament group identity on a
between-groups basis. Winner proportion in a tournament is either
small with a single larger reward or large with multiple smaller
rewards, holding constant the total value of rewards between these
two winner proportions. Tournament group identity is either at a
moderate or strong level. In Experiment 1, tournament participants
imagine competing against their fellow group members over eight
sequential tournament rounds, and the dependent variables are
participants' other-regarding preference and their competitiveness.
In Experiment 2, tournament participants actually compete against
their fellow group members over eight sequential tournament
rounds, and the dependent variable is their individual performance
in each tournament round.
The results of our two experiments are consistent with our
predictions. In Experiment 1, we find increasing tournament group
identity increases participants' other-regarding preference. We also
find that other-regarding preference is negatively associated with
competitiveness, and this effect is more negative under the large
winner proportion tournament than under the small winner pro-
portion tournament. In Experiment 2, we find increasing tourna-
ment group identity leads to lower performance under the large
winner proportion tournament, but it does not significantly affect
performance under the small winner proportion tournament.
Further, this interactive effect of tournament group identity and
winner proportion persists over the eight tournament rounds.
Our study contributes to management accounting research
examining tournament incentives systems (e.g. Berger, Klassen,
Libby, & Webb, 2013; Casas-Arce & Martinez-Jerez, 2009; Choi,
Newman, & Tafkov, 2016; Matsumura & Shin, 2006; Newman &
Tafkov, 2014), as well as management accounting research exam-
ining how the quality of the relationships among a group of
workers affects performance (Chen, Williamson, & Zhou, 2012;
Towry, 2003). Prior research that has examined the performance
effects of tournaments with different winner proportions notes
that many tournaments in practice are embedded in social envi-
ronments where competitors know each other and know the out-
comes of the competition, and that it is important for future
research to consider how social environments affect tournament
performance (Chen, Ham, & Lim, 2011; Lim, 2010). Our study ex-
amines one aspect of the social environments of tournaments,
namely tournament group identity, in a setting where people know
the tournament outcomes of fellow competitors. We are the first
study that we know of to examine how identity with fellow
competing tournament participants interacts with winner pro-
portions to affect the performance of tournament participants.
Given that group identity varies across tournament groups in
practice, our study furthers the understanding of performance
consequences of tournaments with different winner proportions.
Further, in contrast with some prior research that illustrates the
positive effects of increasing group identity on cooperation and
performance (e.g., De Cremer & Van Vugt, 1999; Ellemers, Gilder, &
Haslam, 2004; Knippenberg, 2000; Towry, 2003; Wit & Wilke,
1992), our study illustrates a setting where increasing group
identity decreases performance (i.e., when people are rewarded
with a larger winner proportion tournament).
Our results may help organizations with the difficult decision of
choosing an appropriate winner proportion for their tournament
incentives (Welch, 2013). While there are arguments for using a
larger winner proportion to motivate weaker performers, organi-
zations should consider how different organizational-level or
group-level factors such as the degree of job interdependencies,
group size and demographics, and group culture affect tournament
participants' identity with fellow tournament participants. Our
results caution firms using larger winner proportion tournaments
when there is stronger tournament group identity because the
performance in a larger winner proportion tournament may be
lower when tournament group identity is stronger, and thus, other-
regarding preference is higher. In contrast, our results indicate that
the performance in a smaller winner proportion tournament is less
sensitive to variations in tournament group identity. So, firms that
are concerned with employees deliberately lowering their tour-
nament performance out of the concern for the earnings of others
might consider increasing the financial cost of such lowering of
performance by using a smaller winner proportion tournament.
Thus, our study provides a plausible explanation for why organi-
zations frequently use tournaments with small winner proportions
despite prior research suggesting that a larger winner proportion
motivates higher effort, particularly from less able tournament
participants (Chen et al., 2011; Knauer, Sommer, & Wohrmann,
2016; Szymanski & Valletti, 2005).
The remainder of the paper is organized as follows. Section 2
develops our hypotheses. Section 3 describes our Experiment 1
design and results, and Section 4 describes our Experiment 2 design
and results. Section 5 concludes with a discussion of our findings
and the limitations of our study.
2. Research setting and hypotheses development
2.1. Research setting
We conduct two experiments using a research setting in which
groups of four participants with heterogeneous abilities on an
effort sensitive task compete for tournament rewards. We focus on
such a setting since heterogeneity in employee abilities is common
in organizations. In both experiments, we manipulate whether
participants share strong versus moderate tournament group
identity and whether the tournament winner proportion is small
(only the top performer in a group wins a larger reward) versus
large (top three performers in a group each win a smaller reward).
The total value of the reward(s) under the two winner proportion
conditions is held constant. Participants in Experiment 1 imagine a
scenario in which they will compete against the other three
members of their group in the task over eight sequential tourna-
ments, whereas participants in Experiment 2 actually compete and
hence receive rewards based on their performance and assigned
winner proportion condition. Our dependent variables in
 K. Kelly, A. Presslee / Accounting, Organizations and Society 56 (2017) 21e34
Experiment 1 are participants' other-regarding preference and
competitiveness. Our dependent variable in Experiment 2 is the
individual performance of each participant in each round.
2.2. Hypotheses development
2.2.1. Effect of group identity on other-regarding preference
Tournament participants can have varying levels of group
identity with their fellow participants. Group identity refers to an
individual categorizing and defining himself/herself as a member of
a group (Tajfel & Turner, 1986; Turner, 1982). Strong group identity
is characterized by a sense of belonging to the group, perceived
relationship closeness with others in the group, and positive af-
fective feelings towards others in the group (Aron, Aron, & Smollan,
1992; Buunk, Dijkstra, Bosch, Dijkstra, & Barelds, 2012; Heider,
1958; Tesser & Campbell, 1982). Given strong group identity cre-
ates greater perceived overlap between self and others in the group,
it can result in people experiencing the outcomes of others in the
group as their own (Gino & Galinksy, 2012).
As such, stronger group identity can increase tournament par-
ticipants' other-regarding preference with respect to other partic-
ipants in their group. Other-regarding preference refers to a general
concern for the payoffs of others and a willingness to make per-
sonal sacrifices for the benefit of others (see Cooper and Kegal
(2016) and Fehr and Schmidt (2006) for reviews). Studies find
increasing the level of identity between interacting individuals
tends to increase their other-regarding preference such that they
are more willing to cooperate with one another and that higher-
paid individuals are more willing to sacrifice personal wealth to
increase the wealth of lower-paid individuals (e.g., Balliet, Wu, & De
Dreu, 2014; Chen & Li, 2009; Van Lange, Rusbult, Drigotas, Arriaga,
& Witcher, 1997). Thus, we test the following hypothesis:
H1. Stronger group identity with fellow tournament participants
leads to higher other-regarding preference.
2.2.2. Effect of other-regarding preference on competitiveness and
the moderating effect of winner proportion
Competitiveness refers to the degree to which an individual
desires to compare favorably against others (Garcia, Tor, & Schiff,
2013), and arises because being superior to others boosts one's
self-esteem (Beach & Tesser, 1995; Tesser & Campbell, 1980, 1982).
In a tournament, competitiveness is evidenced by a willingness to
exert effort to win (Hannan, Towry, & Zhang, 2013).
2
A possible alternative explanation for why group identity might reduce
competitiveness and performance is the effect group identity can have on strategic
collusion, wherein tournament contestants implicitly or explicitly agree to collec-
tively reduce effort in order to maximize their joint payoffs at the expense of the
firm (Harbring, 2006). Indeed, studies show higher group identity can lead to
greater cooperation (De Cremer & Van Vugt, 1999; Towry, 2003; Wit & Wilke,
1992), and within a tournament context, this cooperation may take the form of
collusion, which reduces competitiveness and performance (Hannan et al., 2013).
Two features of our experiment setting reduce the likelihood of strategic collusion
as an alternative explanation for our results. First, participants in our study are
unable to communicate with one another, making it difficult to organize and form
collusive agreements (Harbring, 2006). Second, although strategic collusion may
explain why group identity has a negative main effect on competitiveness, it is
unlikely to explain why group identity has a more negative effect on competi-
tiveness in a larger winner proportion tournament than a small winner proportion
tournament. In fact, because all participants with the same winning rank receive
the same reward in the event of a tie, the economic reward for strategic collusion is
larger in the smaller winner proportion tournament. If stronger group identity
makes strategic collusion easier, and the desire for strategic collusion is higher
under a smaller winner proportion because of the larger economic reward, then
group identity should have a more negative effect on competitiveness under a
smaller winner proportion, which is inconsistent with H2 and H3.
23
Competitiveness varies to the extent an individual finds it relevant
or important that he or she performs well on the dimension of
comparison (Garcia & Tor, 2007). Individuals tend to find it less
important to compare favorably against others whom they care
about, and instead, are more willing to trade off personal wealth
and status for the benefit of others (Garcia, Tor, Bazerman, & Miller,
2005; Garcia et al., 2013; Lount & Phillips, 2007; Stapel & Koomen,
2001). Thus, other-regarding preference likely has a negative effect
on individuals' competitiveness.2
Further, we predict winner proportion as a moderator such that
other-regarding preference has a stronger negative effect on
competitiveness with others when the proportion of tournament
winners is larger. We do not predict the direct effect of winner
proportion on competitiveness because there are conflicting argu-
ments regarding the direct effect of winner proportion on compet-
itiveness and performance when tournament participants have
heterogeneous abilities. The direct effect of winner proportion on
competitiveness and performance is not only a function of the dif-
ferences in the size of the rewards, but depends on the perceptions of
less versus more able participants about their probabilities of win-
ning. Less able participants may be more encouraged to compete by
the perceived higher probabilities of winning with a larger winner
proportion (Casas-Arce & Martinez-Jerez, 2009; Chen et al., 2011;
Szymanski & Valletti, 2005). The effect of winner proportion on
more able participants is less certain because it depends on their
perception of the effort needed in order to win a reward. If more able
participants perceive more competition from less able participants
who are working harder under a larger winner proportion, then
more able participants would be less complacent and increase effort
under a larger winner proportion. However, more able participants
could also be more complacent and exert lower effort if they
perceive higher probability of winning one of the multiple rewards
even with the lower effort when the winner proportion is larger.
Given these conflicting arguments, we do not predict the direct ef-
fect of winner proportion on competitiveness.
Instead, our hypothesis about the moderating effect of winner
proportion is based on the argument that both the perceived eco-
nomic cost and psychological cost of lowering competitiveness to
help others in response to other-regarding preference is likely
lower under a larger winner proportion. First, holding constant the
total tournament payout, increasing the number of winners also
decreases the amount of reward each winner receives. Thus, as the
proportion of tournament winners increases, it becomes econom-
ically less costly for participants to be less competitive and lower
their effort to help others. In other words, people are more likely to
respond to their other-regarding preference by lowering their effort
to help others win a reward because the potential economic cost of
losing the reward themselves is lower. Second, the psychological
cost to one's self-esteem from losing the reward is lower when
winner proportion is higher. People seek information about their
superiority to boost their self-esteem, and performance in a
competition is more critical to one's self-esteem if the outcomes of
the competition convey better information about one's superiority
(Festinger, 1954; Garcia et al., 2013). Winning a tournament with a
larger winner proportion provides less meaningful information
about one's superiority, and thus, losing that tournament is less
costly for one's self-esteem. Therefore, we predict that other-
regarding preference is more likely to reduce people's competi-
tiveness (or willingness to exert effort to win) when the winner
proportion is larger because the economic and psychological costs
of reducing competitiveness to help others are lower. This predic-
tion is consistent with the cost-reward analysis of helping behavior
and charitable donation where people are more inclined to help or
donate when the costs of helping or donating are lower, holding
other factors constant (Bekkers & Wiepking, 2010; Dovidio &
24 K. Kelly, A. Presslee / Accounting, Organizations and Society 56 (2017) 21e34

Penner, 2001; Dovidio, Piliavin, Gaertner, Schroeder, & Clark, 1991;

Peloza & Steel, 2005). Thus, we test the following interaction
hypothesis:
H2. Winner proportion moderates the negative effect of tourna-
ment participants’ other-regarding preference on competitiveness
such that other-regarding preference reduces competitiveness to a
larger degree in a tournament with a larger winner proportion.

2.2.3. Effect of group identity on performance and the moderating

effect of winner proportion
We predict a stronger negative effect of group identity on perfor-
mance under a larger winner proportion through the positive effect of
group identity on other-regarding preference (H1) and larger winner
proportion exacerbating the negative effect of other-regarding pref-
erence on competitiveness (H2). Higher levels of competitiveness
typically lead to higher effort and performance in tournaments as
individuals are driven to compete over rank and rewards (Hannan
et al., 2013). Given H1 and H2, and given that higher levels of
competitiveness lead to higher tournament performance, increased
other-regarding preference caused by stronger tournament group
identity is more likely to reduce competitiveness and hence perfor-
mance in a larger winner proportion tournament. As previously dis-
cussed, people are willing to make costly personal sacrifices in order
to benefit others whom they care about. We expect people to be more
willing to make that sacrifice by reducing their competitiveness and
hence performance when there are a larger number of smaller valued
rewards at stake under a larger winner proportion. Thus, we test the
following interaction hypothesis:
H3. Winner proportion moderates the negative effect of group
identity with fellow tournament participants on performance such Fig. 1. Conceptual model and operationalizations.

that stronger identity reduces performance to a larger degree in a

tournament with a larger winner proportion.
We summarize our hypothesized conceptual model in Fig. 1. We
use two experiments to test for theoretical mediation (as opposed to
testing statistical mediation using a single experiment) to avoid
influencing participants’ performance by asking them about their
other-regarding preference and competitiveness prior to exerting
effort (Spencer et al., 2005). An alternative design would be to use a
single experiment with ex post measures of other-regarding prefer-
ence and competitiveness. However ex post measures of other-
regarding preference and competitiveness could be influenced by
tournament outcomes. To avoid these potential demand effects, we
use Experiment 1 to test H1 and H2 regarding process measures of
other-regarding preference and competitiveness, whereas we use
Experiment 2 to test H3 regarding performance.3 In Experiment 2,
tournament participants are provided detailed individual perfor-
mance feedback on themselves and fellow participants after each
tournament round. We are interested in whether the interactive effect
of tournament group identity and winner proportion persists over
multiple rounds, but we make no prediction ex ante because theory is
not clear on how our predicted effects might evolve over time.
3. Experiment 1: Method and results
Large)
two (Group Identity: Moderate versus Strong) experiment
design. Winner Proportion and Group Identity were manipulated
between-subjects at the group level. Each experiment session was
randomly assigned to one of the two Group Identity conditions.
Participants within each session were assigned into groups of four
and each group within a session was randomly assigned to one of
the two Winner Proportion conditions. We measured participants'
Other-regarding Preference and their Competitiveness as it relates to
competing in eight hypothetical tournaments.
3.1.2. Participants
We recruited 104 undergraduate accounting students from a
large U.S. public university to participate in one of six experiment
sessions. Each session consisted of 12e20 participants resulting in
3e5 groups per session. The mean (standard deviation) participant
age was 20.13 years (1.71) and the number (percentage) of females
that participated was 52 (50%).4, 5 A summary of the number of
participants in each condition is shown in Panel A of Table 1.
3.1.3. Procedures
On arrival for an experiment session, participants were assigned
to a group of four persons with each group seated at a table
together, and told their assigned group number and player ID. We

3.1. Method
4
All p-values reported in the paper are two-tailed unless otherwise labeled.
5
3.1.1. Experiment design Age and gender are uncorrelated with Competitiveness (p ¼ 0.28 and p ¼ 0.19,
respectively). Conversely, age (r ¼ 0.34, p < 0.01) and gender (Female ¼ 0, Male ¼ 1;
In Experiment 1, we used a two (Winner Proportion: Small versus
r ¼ 0.31, p < 0.01) are correlated with Other-regarding Preference. Also, age differs
by Winner Proportion condition (p ¼ 0.08) and gender differs by Group Identity
condition (p < 0.01). Path results and fit statistics are inferentially identical whether
3
Experiment 1 and Experiment 2 received ethics clearance through the research or not we control for age and/or gender in our analyses. Thus, age and gender are
ethics committee at the university that each experiment was conducted. excluded from further analyses.
 K. Kelly, A. Presslee / Accounting, Organizations and Society 56 (2017) 21e34 25

Table 1
Experiment 1 descriptives.a

Panel A: Total number of participants

Moderate Group Identity Strong Group Identity Total

Small Winner Proportion 28 24 52

Large Winner Proportion 28 24 52

Total 56 48 104

Panel B: Mean (Standard Deviation) Tournament Group Identity Measureb

Moderate Group Identity Strong Group Identity Total

Small Winner Proportion 4.14 (0.88) 6.00 (0.87) 5.00 (1.27)

Large Winner Proportion 4.16 (1.01) 6.01 (0.65) 5.01 (1.28)

Total 4.15 (0.94) 6.00 (0.76) 5.00 (1.26)

Panel C: Mean (Standard Deviation) Other-regarding Preferencec

Moderate Group Identity Strong Group Identity Total

Small Winner Proportion 2.53 (1.46) 3.69 (1.10) 3.06 (1.42)

Large Winner Proportion 3.07 (1.53) 3.85 (1.60) 3.43 (1.60)

Total 2.80 (1.51) 3.77 (1.36) 3.25 (1.52)

Panel D: Mean (Standard Deviation) Competitivenessd

Moderate Group Identity Strong Group Identity Total

Small Winner Proportion 5.73 (1.23) 5.93 (0.81) 5.82 (1.05)

Large Winner Proportion 5.46 (1.07) 5.20 (1.10) 5.34 (1.05)

Total 5.60 (1.15) 5.58 (1.03) 5.58 (1.09)

a
One hundred and four participants made up 26 groups, each consisting of four participants, and participated in Experiment 1. Each session consisted of 12e20 participants
resulting in 3e5 groups per session. Group Identity is manipulated between-sessions as either moderate or strong tournament group identity amongst participants in a group.
Winner Proportion is manipulated between-groups as either a Small Winner Proportion versus a Large Winner Proportion. In the Small Winner Proportion condition, participants
were asked to imagine a scenario where they are competing in eight rounds of decoding where only the top ranked performer in their group of four wins a reward of $3 in
each round. In the Large Winner Proportion condition, participants were asked to imagine the same scenario except that the top three ranked performers in a group of four each
win a reward of $1 in each round. The total hypothetical value of the reward(s) is held constant between the two Winner Proportions.
b
The Tournament Group Identity measure is the average responses by individuals to three questions on whether they are happy to be a part of the group (Happy), whether
they feel that they are a member of the group (Belong), and whether they like their group members (Like), all measured on a 7-point scale (1-strongly disagree to 7-strongly
agree). Participants are asked these questions after experiencing the Group Identity and Winner Proportion manipulations.
c
Other-regarding Preferences is the average responses of participants' rating of the personal importance (on scales of 1 “Not important” to 7 “Extremely important) of two
statements: “Winning the $3/$1 bonus in fewer rounds so that other members of the group have a chance to win the bonus” and “Ensuring all members of the group have a
chance to earn the $3 bonus”.
d
Competitiveness is the average responses of participants' rating of the personal importance (on scales of 1 “Not important” to 7 “Extremely important) of two statements:
“Winning the $3/$1 bonus in each round at all cost” and “Competing as hard as possible in each round”.

then applied the Group Identity manipulation using a slogan
guessing game, with participants in half of the sessions assigned to
the Strong Group Identity condition and participants in the other
half of the sessions assigned to the Moderate Group Identity con-
dition (described in the “independent variables” subsection).
Once participants experienced the first half of the slogan
guessing game, they turned their attention to their individual paper
workbook that described a version of Chow’s (1983) decoding task.
The decoding task required participants to use a provided decoding
key to decode a series of single three-digit numbers into a series of
single letters of the alphabet. Participants then completed a two-
minute decoding round to familiarize themselves with the task.
After completing the decoding round, participants were asked
to imagine a scenario in which they will complete eight decoding
rounds and will compete with their fellow group members under
an assigned Winner Proportion tournament in each round. Half of
the participant-groups were provided a description of the Small
Winner Proportion condition and the other half were provided a
description of the Large Winner Proportion condition (described in
the “independent variables” subsection).
Participants next provided responses to measures of other-
regarding preference and competitiveness (described in the
“dependent variables” subsection). They also responded to three
statements, which we use to measure tournament group identity.
Specifically, participants indicated on a seven-point scale (1-
strongly disagree to 7-strongly agree) if they were happy to be a
part of their group (Happy), if they felt that they were members of
their group (Belong), and if they liked their group members (Like).
Recall that strong group identity is characterized by a sense of
belonging to the same group as others and positive affective feel-
ings towards others in the group (Aron et al., 1992; Buunk et al.,
2012; Heider, 1958; Tesser & Campbell, 1982). Participants then
provided demographic information, finished the remaining guesses
in the slogan game, and were paid $10 for their participation as
they left the room.
3.1.4. Independent variables
Our first independent variable, Group Identity, was manipulated
between sessions at the group level; that is, all groups in the same
session experienced the same group identity manipulation. Par-
ticipants in the Strong Group Identity condition had more oppor-
tunities to interact and collaborate with their fellow group
26 K. Kelly, A. Presslee / Accounting, Organizations and Society 56 (2017) 21e34
members (Dawes, Van de Kragt, & Orbell, 1988; King, 2002). Each
group sat at a table with a different colored tablecloth (e.g., red,
green, blue) and a colored placard for their group name, which
helps to create strong identity with the group (Towry, 2003). Once
seated, group members were asked to create and write down a
unique group name on the colored placard. Groups then played a
slogan guessing game in which each group wrote down the names
of companies/products they believed were associated with 16
company/product slogans (e.g., “Have it your way” e Burger King).
The groups competed against other groups to get the most number
of correct company/product names and the winning group received
a basket of chocolate bars and candy (valued at $20) to be shared
between all group members. Playing a group game helps develop
strong group identity through collaborating on a task, sharing a
common outcome (i.e., the shared reward), and competing together
against other groups (Eckel & Grossman, 2005; Friedkin & Simpson,
1985). Groups guessed 10 slogans and then put the game aside
while they worked on their individual decode workbook. The
slogan guessing game was interrupted by the decode workbook to
ensure that tournament group identity was not affected by per-
formance in the game (Snyder, Lassegard, & Ford, 1986). Partici-
pants remained seated together as a group at a table with their
colored placard and their group name on the table as they worked
on their decode workbooks. To sustain the saliency of the Group
Identity manipulation throughout the decode workbook, we
replaced the word “group” with “team”. Finally, once the decode
workbook was completed, the groups returned to finish guessing
the remaining six slogans.
Participants in the Moderate Group Identity condition experi-
enced little interaction and collaboration with their fellow group
members.6 Members of the same group sat together at a table with
a white placard for their group number. Participants played the
same slogan guessing game as in the Strong Group Identity condi-
tion except that they competed as individuals instead of as a group.
That is, participants competed against all other participants in the
room (including those in their group) for rewards. The four par-
ticipants with the highest number of correct company/product
names each won a reward of a basket of chocolate bars and candy
(valued at $5). Participants guessed 10 slogans and then put the
game aside while they worked on their individual decode work-
book. Participants returned to guess the remaining six slogans after
their decode workbook was completed.
Our second independent variable is Winner Proportion. Groups
were randomly assigned to either a Small Winner Proportion tour-
nament or a Large Winner Proportion tournament. This manipula-
tion was applied after the Group Identity manipulation, but before
Group Identity was measured. In both Winner Proportion conditions,
participants were asked to imagine a scenario in which they earned
$1.00 per round of decode and competed in a tournament each
round to earn additional rewards. In the Small Winner Proportion
condition, participants imagined a scenario in which the top ranked
performer in their group of four participants for that round won a
single tournament reward of $3.00. In the Large Winner Proportion
condition, participants imagined a scenario in which the top three
ranked performers in their group of four participants for that round
won a $1.00 reward each. In both imagined scenarios, participants
were told to assume that performance rankings would be based on
the number of correct decodes and that at the end of each
6
We refer to our manipulation of tournament group identity as either Moderate
versus Strong rather than Weak versus Strong for expositional purposes. It is un-
likely to achieve a weak level of tournament group identity in our measures
because participants are not likely to indicate strong disagreement with the three
statements we use to measure tournament group identity.
tournament round, all persons in the group would receive infor-
mation on the performance rank, whether a bonus was awarded or
not, number of correct decodes, and earnings for every person in
the group. Further, participants were told to assume that all tied
participants for the top rank would receive $3 each in the Small
Winner Proportion condition, and all tied participants for any of the
top 3 ranks would receive $1 each in the Large Winner Proportion
condition.
3.1.5. Dependent variables
The Other-regarding Preference measure is the average partici-
pants' rating of the personal importance (on scales of 1 “Not
important at all” to 7 “Extremely important”) of two statements
(“Winning the $3/$1 bonus in fewer rounds so that other members
of the group have a chance to win the bonus” and “Ensuring all
members of the group have a chance to earn the $3/$1 bonus”)
given the imagined scenario. The Competitiveness measure is the
average participants' rating of the personal importance (on scales of
1 “Not important at all” to 7 “Extremely important”) of two state-
ments (“Winning the $3/$1 bonus in each round at all cost” and
“Competing as hard as possible in each round”) given the imagined
scenario.
Confirmatory factor analysis indicates that these four questions
represent two distinct factors, with the two other-regarding pref-
erence questions loading as a factor and the two competitiveness
questions loading as another factor. All items that load on a factor
have an absolute loading greater than 0.65 and all items said to not
load on a factor have an absolute loading less than 0.25. Further, the
combined eigenvalue (variance explained) of the two factors is 3.15
(79%) (Stevens, 1996) and their Cronbach alphas are greater than
0.71 (Kline, 2011). Therefore, we capture two distinct constructs
with our measures of other-regarding preference and
competitiveness.
3.2. Results
3.2.1. Group identity manipulation check
To evaluate the effectiveness of our manipulation of Group
Identity, we asked participants their level of agreement to three
statements (Happy, Belong, Like, described earlier in the “Proced-
ures” sub-section) on a seven-point scale. We averaged each par-
ticipant's responses to these three questions to obtain a measure of
Tournament Group Identity. Confirmatory factor analysis indicates
that these three questions represent a uni-dimensional construct,
with all loadings greater than 0.84, an eigenvalue (variances
explained) of 2.30 (77%), and a Cronbach alpha of 0.91 (Kline, 2011;
Stevens, 1996). As shown in Panel B of Table 1, the Tournament
Group Identity is higher in the Strong Identity condition
(mean ¼ 6.00) than in the Moderate Identity condition
(mean ¼ 4.15) (t ¼ 10.78, one-tailed p < 0.01).
Given that individual participants are nested within tournament
groups and our manipulation of Group Identity is applied at the
tournament group level, we further examine the intra-class cor-
relation coefficient (ICC), which is the proportion of variance in
Tournament Group Identity that is accounted for by tournament
groups (Luke, 2004). We find that tournament groups account for a
significant portion of the variability in the Tournament Group
Identity (ICC ¼ 54%, F ¼ 5.68, p < 0.01). Based on the significant ICC
and the average Tournament Group Identity across the two Group
Identity conditions, our manipulation was effective.
3.2.2. Hypotheses testing
We test H1 and H2 using path analysis. Path analysis facilitates
the simultaneous analysis of the relationships between manipu-
lated variables and multiple dependent variables, offering the
 K. Kelly, A. Presslee / Accounting, Organizations and Society 56 (2017) 21e34
opportunity to analyze the predicted dependencies of constructs
while reducing measurement error (Kline, 2011). We test our path
model using STATA 14 with maximum likelihood estimation
method and using robust clustered standard errors by tournament
group to address the potential for correlated error terms amongst
participants from the same group. Theory underlying H1 and H2
suggest Winner Proportion does not moderate the effect of Group
Identity on Other-regarding Preference, but does moderate the effect
of Other-regarding Preference on Competitiveness. Thus, we constrain
the path estimate from Group Identity to Other-regarding Preference
to be equal across Winner Proportion conditions, whereas we use
unconstrained path estimates from Other-regarding Preference to
Competitiveness by Winner Proportion such that a different coeffi-
cient is calculated for the Small Winner Proportion and the Large
Winner Proportion conditions.7 Our path model fits the data well,
and results of our path model are presented in Fig. 2.8
H1 predicts that increasing tournament group identity increases
participants' other-regarding preference. As shown in Panel C of
Table 1, the average (standard deviation) Other-regarding Preference
is 3.77 (1.36) in the Strong Group Identity condition and 2.80 (1.15) in
the Moderate Group Identity condition. Path results presented in
Table 2 show the difference in Other-regarding Preference by Group
Identity condition is significant (b ¼ 0.96, one-tailed p < 0.01). Thus,
we find support for H1.
H2 predicts an interaction such that the effect of participants'
other-regarding preference on their competitiveness is more
negative in a large winner proportion tournament than a small
winner proportion tournament. Panel D of Table 1 shows that
the average Competitiveness across all conditions is 5.58, which is
significantly greater than the scale mid-point of 4.00 (t ¼ 14.85,
p < 0.01). When constrained to Large Winner Proportion tour-
naments, Other-regarding Preference has a significant negative
effect on Competitiveness (b ¼ 0.27, one-tailed p < 0.01). When
constrained to Small Winner Proportion, Other-regarding Prefer-
ence has a marginally significant negative effect on Competitive-
ness (b ¼ 0.11, one-tailed p ¼ 0.09). We conduct a post-
estimation non-linear combination test of whether these two
paths differ significantly from one another. As shown in Panel B
of Table 2, the Large Winner Proportion path coefficient is
significantly more negative than the Small Winner Proportion
path coefficient (b ¼ 0.16, one-tailed p < 0.01), which provides
support for H2.
To summarize, Experiment 1 provides support for our hy-
potheses that increasing group identity increases other-
regarding preference (H1) and that the effect of other-
regarding preference on competitiveness is more negative with
a larger proportion of tournament winners (H2). Next, we
conduct Experiment 2 to test whether winner proportion mod-
erates the negative effect of increasing group identity on tour-
nament performance (H3).
7
Supporting our constraint that the Group Identity to Other-regarding Preference
path be equal across Winner Proportion conditions, Other-regarding Preference does
not differ by Winner Proportion condition (p ¼ 0.21) nor does Winner Proportion
interact with Group Identity to affect Other-regarding Preference (p ¼ 0.52).
Furthermore, if we do allow this path to vary across Winner Proportion condition,
the two path coefficients do not differ from one another (p ¼ 0.50).
8
We are unable to produce RMSEA, CFI, or chi-square model fit statistics because
we robust cluster standard errors by tournament group. However, a similar path
model that does not robust cluster standard errors produces almost identical co-
efficients and p-values, and also suggests that our path model fits the data well
(RMSEA ¼ 0.04; CFI ¼ 0.97; c2(7) ¼ 7.66, p ¼ 0.36) (Kline, 2011).
27
Fig. 2. Experiment 1 Results. See Table 1 for definition of variables and see Fig. 1 for
operationalization of variables. See Table 2 for full results. The Tournament Group
Identity / Other-regarding Preference path is constrained to be equal across Winner
Proportion conditions. The Other-regarding Preference / Competitiveness path is
allowed to vary across the two Winner Proportion conditions. Standard errors are
robust clustered on tournament group to address the potential for correlated error
terms amongst participants from the same group. We are unable to produce RMSEA,
CFI, or chi-square model fit statistics because we robust cluster standard errors by
tournament group. However, a similar path model that does not robust cluster stan-
dard errors produces almost identical coefficients and p-values, while suggesting our
path model fits the data well (RMSEA ¼ 0.04; CFI ¼ 0.97; c2(7) ¼ 7.66, p ¼ 0.36) (Kline,
2011). Other-regarding Preference R2 ¼ 10.2%, Competitiveness in the Large Winner
Proportion conditions R2 ¼ 14.8%, and Competitiveness in the Small Winner Proportion
conditions R2 ¼ 2.8%. p-values are one-tailed; ***p < 0.01 and **p < 0.05.
4. Experiment 2: Method and results
4.1. Method
4.1.1. Experiment design
In Experiment 2, we used a two (Winner Proportion: Small versus
Large)
two (Group Identity: Moderate versus Strong)
eight
(Rounds) mixed-factor experiment design. Winner Proportion and
Group Identity were manipulated between-subjects at the group
level, while Rounds was manipulated within-subjects. Like Exper-
iment 1, each experiment session in Experiment 2 was randomly
assigned to one of the two Group Identity conditions and groups of
four individuals within a session were randomly assigned to one of
the two Winner Proportion conditions. Participants remained in the
same group and same experimental condition for all eight tour-
nament rounds. However, unlike Experiment 1, participants in
Experiment 2 actually competed in performing the decoding task
against the other participants in their assigned group for eight
sequential tournament rounds.
4.1.2. Participants
We recruited 368 undergraduate accounting students from a
large Canadian public university to participate in one of 15 exper-
iment sessions. Each session consisted of 16e24 participants
resulting in 4e6 groups per session. A total of 92 groups completed
Experiment 2. However, participants from one group (from the
Strong Group Identity, Small Winner Proportion condition) violated
our instructions by communicating with each other during the
tournament rounds; so, all data relating to these participants were
removed from our analysis.9 Our final data set comprises 2912
9
The experiment administrator noticed that participants in that group were
talking with each other. The administrator instructed the group not to have any
further communication with each other, but allowed the group to complete the
study. Participants from the removed group had a mean Individual Performance of
11.75 decodes per round, which was 3.0 standard deviations below the mean In-
dividual Performance for all other participants. Thus, their communication during
the tournament rounds dramatically reduced their performance. As described in
the “Independent Variables” sub-section, if there was a tie in performance for a
particular rank, then those tied participants were assigned the same rank. As such,
in a Small Winner Proportion tournament, if all participants in the group are able to
coordinate their effort and reduce their performance to zero decodes, then they
would all be assigned the top rank and be awarded the large reward of $3 per
person. This explains why in each round from rounds 5 to 8, each participant in this
omitted group only completed one decode, even though these participants
completed a mean of 22.50 decodes per round from rounds 1 to 4 and a mean of
24.10 decodes per round in the two practice rounds.
28 K. Kelly, A. Presslee / Accounting, Organizations and Society 56 (2017) 21e34

Table 2
Experiment 1 path analysis results.a

Panel A: Path Analysisb

Paths Unstandardized Estimates Standard Error p-value

Tournament Group Identity / Other-regarding Preference 0.96 0.24 < 0.01

Small Winner Proportion:
Other-regarding Preference / Competitiveness 0.11 0.08 0.09
Large Winner Proportion:
Other-regarding Preference / Competitiveness 0.27 0.07 < 0.01

Panel B: Path Comparisonc

Unstandardized Estimates Standard Error p-value

Other-regarding Preference / Competitiveness:

Small Winner Proportion vs. Large Winner Proportion 0.16 0.05 < 0.01
a
N ¼ 104. See Table 1 for definition of variables and see Fig. 1 for operationalization of variables.
b
The Tournament Group Identity / Other-regarding Preference path is constrained to be equal across Winner Proportion conditions. Standard errors are robust clustered on
tournament group to address the potential for correlated error terms amongst participants from the same group; p-values are one-tailed because directional predictions are
made. We are unable to produce RMSEA, CFI, or chi-square model fit statistics because we robust cluster standard errors. However, a similar path model that does not robust
cluster standard errors produces almost identical coefficients and p-values, while suggesting our path model fits the data well (RMSEA ¼ 0.04; CFI ¼ 0.97; c2 (7) ¼ 7.66,
p ¼ 0.36) (Kline, 2011). Other-regarding Preference R2 ¼ 10.2%, Competitiveness in the Large Winner Proportion conditions R2 ¼ 14.8%, and Competitiveness in the Small Winner
Proportion conditions R2 ¼ 2.8%.
c
Path comparison test uses a post-estimation non-linear combination test to examine whether the Other-regarding Preference / Competitiveness path differs by Winner
Proportion condition. This test uses the path results shown in Panel A. The p-value is reported one-tailed.

rounds of Individual Performance by the 364 participants within the
91 groups. The mean (standard deviation) participant age was 18.3
years (1.31) and the number (percentage) of females that partici-
pated was 193 (53%).10 A summary of the number of groups (in-
dividuals) in each condition is shown in Panel A of Table 3.
4.1.3. Task
We used a version of Chow's (1983) decoding task to test per-
formance. However, unlike Experiment 1, Experiment 2 used a
computerized version of the decoding task. Using a decoding key
provided on a computer screen, participants decoded either a sin-
gle three-digit number into a single letter of the alphabet (a single
decode) or two three-digit numbers into two letters of the alphabet
(a double decode) by using a keyboard. The program immediately
informs the participant if a decode is incorrect and that the number
of correct decodes would be reset to zero if the participant makes
five consecutive errors. Once participants accurately completed a
decode, the program provided them with another decode. Each
round lasted two minutes. During a round, the participant's cu-
mulative performance (i.e., number of correct decodes) for that
round and the time remaining for that round were shown onscreen
in real-time. Similar to Kelly, Webb, and Vance (2015), we oper-
ationalized common environmental uncertainty by varying the
ratio of single-to-double decodes in each round.11 This environ-
mental uncertainty was common to all participants because
everyone experienced the same ratio and order of decodes.
We required participants to answer correctly several informa-
tion check questions before starting their decoding rounds to
ensure that all participants had common knowledge of the task
setting. These questions tested participants' knowledge of the
10
Age and gender do not differ by condition (p > 0.27 and p > 0.34 respectively).
Gender is uncorrelated with the average of eight rounds of Individual Performance
(p ¼ 0.50), but age is correlated with the average of eight rounds of Individual
Performance (r ¼ 0.13, p ¼ 0.01). If age is included as a fixed effect in our multi-
level model analysis, then the main interaction result becomes more significant.
We report conservative results by excluding age from our analysis.
11
The ratio of decodes in each of the eight rounds was as follows: Rounds 1 and 8:
1 single decode to 3 double decodes (toughest), Rounds 2 and 7: 3 single decodes to
1 double decode (easiest), Rounds 3 and 6: 1 single decode to 2 double decodes
(tough), Rounds 4 and 5: 2 single decodes to 1 double decode (easy).
number of rounds, the varying ratio of decodes, that everyone in
the group faced the same ratio of decodes in each round, and the
information that each person in the group would receive at the end
of each round.
4.1.4. Procedures
On arrival for an experiment session, participants were seated at
a computer. Then, participants individually completed two practice
rounds of the decoding task. Each practice round lasted two mi-
nutes and consisted of a 1:1 ratio of single to double decodes.
Participants earned $0.05 per single decode and $0.10 per double
decode to motivate effort in the practice periods. During a practice
round, cumulative performance (i.e., number of correct decodes)
was shown onscreen in real-time. After each practice round, the
program informed participants of their final performance and their
total dollar amount earned in that round.
Our objective is to test H3 in a setting that mirrors many real
world settings. So, we created a setting where participants in the
tournaments were actually heterogeneous in abilities, yet they
were not explicitly informed of that heterogeneity other than what
they can infer from the relative performance feedback that they
receive after each tournament round. The following procedures
were used to achieve actual participant heterogeneity. The program
first sorted all the participants in a session into four quartiles
(below the 26th percentile, 26th to 50th, 51st to 75th, and above
the 75th percentile) based on their performance in the second
practice round, which we refer to as Ability.12 Then, the program
created groups of four participants by selecting one individual from
12
We expected participants to familiarize themselves with the task in the first
practice round and thus their performance in the second practice round is more
indicative of ability. So, we designed our program to use the second practice round
performance to sort participants into four ability quartiles. Participants' second
practice round performance is lower than their first practice round performance
(mean difference ¼ 1.36, t ¼ 2.92, p < 0.01). This suggests that the performance in
the first practice round was not sustainable in the second practice round, even with
a piece rate incentive, possibly due to exhaustion or boredom. Since we used the
second practice round performance to sort participants into heterogeneous groups
and the second practice round performance also seems indicative of a more sus-
tainable performance level, we use the second practice round performance to
control for participants' ability in our analyses.
 K. Kelly, A. Presslee / Accounting, Organizations and Society 56 (2017) 21e34 29

Table 3
Experiment 2 descriptives.a

Panel A: Total Number of Participants

Moderate Group Identity Strong Group Identity Total

Small Winner Proportion 92 84 176

Large Winner Proportion 92 96 188

Total 184 180 364

Panel B: Mean (Standard Deviation) Average Expectancy over Eight Roundsb

Moderate Group Identity Strong Group Identity Total

Small Winner Proportion 39.42% (10.10%) 40.41% (9.36%) 39.89% (9.65%)

Large Winner Proportion 64.04% (10.65%) 62.63% (10.37%) 63.32% (10.42%)

Total 51.73% (16.13%) 52.26% (14.89%) 51.99% (15.44%)

Panel C: Mean (Standard Deviation) Tournament Group Identity Measurec

Pre-tournaments Post-tournaments Change

Moderate Group Identity

Small Winner Proportion 4.90 (0.69) 4.57 (0.78) 0.33 (0.88)
Large Winner Proportion 4.89 (0.67) 4.74 (0.60) 0.15 (0.46)

Total 4.90 (0.67) 4.64 (0.69) 0.25 (0.69)

Strong Group Identity
Small Winner Proportion 6.06 (0.61) 5.48 (0.84) 0.59 (0.49)
Large Winner Proportion 5.98 (0.60) 5.72 (0.54) 0.26 (0.61)

Total 6.01 (0.60) 5.61 (0.70) 0.41 (0.72)

Panel D: Mean (Standard Deviation) Individual Performanced

Ability Round 1 Round 2 Round 3 Round 4 Round 5 Round 6 Round 7 Round 8

Moderate Group Identity

Small Winner Proportion 23.65 (6.26) 29.36 (6.65) 30.66 (6.65) 25.27 (7.11) 27.14 (7.05) 28.72 (7.05) 26.89 (7.63) 24.79 (6.25) 26.57 (6.73)
Large Winner Proportion 23.40 (7.01) 29.68 (5.79) 31.00 (5.79) 25.62 (6.88) 27.63 (5.79) 28.57 (5.98) 28.07 (6.56) 24.87 (5.57) 26.92 (6.84)
Strong Group Identity
Small Winner Proportion 24.34 (5.78) 30.22 (7.13) 31.55 (7.13) 25.70 (7.01) 27.69 (5.92) 29.04 (5.69) 28.49 (8.19) 25.39 (5.82) 27.90 (6.75)
Large Winner Proportion 24.11 (5.78) 28.74 (6.51) 30.48 (6.51) 25.80 (6.48) 27.44 (6.38) 28.23 (6.57) 27.78 (6.22) 25.04 (5.64) 26.70 (5.88)
a
Three hundred sixty eight participants made up ninety-two groups, each consisting of four participants, and participated in our study. However, one group of four
participants is removed from analysis because group members communicated with each other during the tournament rounds in violation of our instructions. Thus, our data
analysis is based on N ¼ 364 participants making up 91 groups. Winner Proportion is manipulated between-groups as either a Small Winner Proportion versus a Large Winner
Proportion. In the Small Winner Proportion condition, only the top ranked performer in a group of four wins a reward. In the Large Winner Proportion condition, the top three
ranked performers in a group of four each win a smaller reward. The total value of the reward(s) is held constant between the two Winner Proportions. Group Identity is
manipulated between-sessions as either moderate or strong tournament group identity amongst participants in a group.
b
Prior to each tournament round, participants report their level of expectancy (i.e., probability) of winning a reward in the tournament on a scale of 0%e100%. Average
Expectancy over eight rounds is the average of Expectancy for an individual participant across the eight tournament rounds.
c
The Tournament Group Identity measure is the average responses of individuals to three questions on whether they are happy to be a part of the group (Happy), whether
they feel that they are a member of the group (Belong), and whether they like their group members (Like), all measured on a 7-point scale (1-strongly disagree to 7-strongly
agree). Pre-tournament rounds, participants are asked these questions after experiencing the Group Identity manipulation but before the Winner Proportion manipulation.
Post-tournament rounds, participants are asked the same three questions.
d
Individual Performance in each round is an individual's total equivalent unit correct decodes (i.e., number of single decodes þ 2*number of double decodes). Ability is the
performance of an individual in the second practice round and is included as a covariate because it is highly correlated with average Individual Performance over the eight
rounds (r ¼ 0.75, p < 0.01).

each quartile. The highest performers from each quartile formed
the first group of four participants, the second highest performers
from each quartile formed the next group, and so on, until all
participants in the session were grouped into four-person groups.
13
With respect to the second practice round's performance, participants sorted
into quartile 1 (mean ¼ 31.5) outperformed those sorted into quartile 2
(mean ¼ 25.7) (t ¼ 11.28, p < 0.01), those sorted into quartile 2 outperformed those
sorted into quartile 3 (mean ¼ 21.2) (t ¼ 12.15, p < 0.01), and those sorted into
quartile 3 outperformed those sorted into quartile 4 (mean ¼ 17.1) (t ¼ 9.27,
p < 0.01). Performance in tournament round 1 shows a similar pattern, with par-
ticipants drawn from quartile 1 (mean ¼ 34.16) outperforming those drawn from
quartile 2 (mean ¼ 30.26) (t ¼ 5.14, p < 0.01), participants drawn from quartile 2
outperforming those drawn from quartile 3 (mean ¼ 28.11) (t ¼ 2.80, p < 0.01), and
participants drawn from quartile 3 outperforming those drawn from quartile 4
(mean ¼ 25.37) (t ¼ 3.42, p < 0.01). This evidence indicates tournament groups
comprise participants who are heterogeneous in ability.
In this way, a group was always comprised of four participants with
heterogeneous performances in the second practice round. Partic-
ipants were not told how the program determined group
assignments.13
Once the program assigned participants into their groups, par-
ticipants transitioned to another room with their assigned group
and they experienced the first half of the slogan guessing game
used to manipulate Group Identity (described in the “independent
variables” subsection). Then, participants returned to the original
room and responded to same three statements used in Experiment
1 to measure tournament group identity (Happy, Belong, and Like).
After responding to these statements, participants learned that
they would compete in eight sequential tournament rounds against
their assigned group members. Consistent with the description of
the incentives in Experiment 1, participants in Experiment 2 actu-
ally received a fixed pay of $1.00 for each round and an additional
30 K. Kelly, A. Presslee / Accounting, Organizations and Society 56 (2017) 21e34
reward depending on their Winner Proportion condition. Before
starting the first tournament round, participants had to correctly
answer manipulation check questions to ensure they understood
their fixed pay and their tournament rewards. Participants were
also provided with a school newspaper and told that they could
read the newspaper at any point during the rounds. This gave
participants an alternative way of spending their time rather than
sitting idle at their computer should they choose not to exert effort
at decoding either because they had given up or they were
complacent when they were far ahead (Van Dijk, Sonnemans, &
Van Winden, 2001). Alternatively, if participants did not wish to
read the newspaper because of social desirability concerns, they
could deliberately slow down their decoding.
Before beginning each round, we asked participants for their
Expectancy (i.e., probability) that they would earn a bonus in the
upcoming round on a scale of 0%e100% (Kelly et al., 2015). After
each round, all persons in a group received information on the
performance rank, whether a bonus was awarded or not, produc-
tion units, and earnings for every person in the group for that round
and all prior rounds. Finally, after completing the eight tournament
rounds, participants responded to the same three statements about
their tournament group identity (Happy, Belong, and Like) and other
post-experiment questions. The program then told participants
how much they earned for their two practice rounds and all eight
tournament rounds (with average earnings being $18.95), and they
collected their earnings as they left the room.
4.1.5. Independent variables
In Experiment 2, we manipulated Group Identity in a similar
14
We recruited, predominantly, first year undergraduate students to increase the
likelihood that groups are made up of strangers with levels of identity with fellow
group members that could be more strongly influenced by our Group Identity
manipulation. We also asked participants not to reveal their actual names during
their interaction. Analysis of a post experiment question shows that 76 of the 91
groups (84%) consisted of strangers who did not know each other before the study.
Our multi-level analysis results are inferentially identical if we control for whether
or not participants competed in a group consisting of strangers. Thus, whether or
not participants competed in a group consisting of strangers is excluded from
further analyses.
15
Of the 352 tournament rounds in the Small Winner Proportion condition, there
were 39 (11%) ties for the first rank (which is the only rank that will win a reward).
Of the 376 tournament rounds in the Large Winner Proportion condition, there were
48 (13%) ties for the third rank (which is the lowest rank that will win a reward).
Hence, there does not appear to be a significant difference in the proportion of
tournaments with ties such that participants could win a reward across the Small
and Large Winner Proportion conditions (z ¼ 0.69, p ¼ 0.45). Of the 368 tournament
rounds in Moderate Group Identity condition, there were 44 (12%) ties for the final
available reward or the only available reward. Of the 360 tournament rounds in the
Strong Group Identity condition, there were 43 (12%) ties for the final available
reward or the only available reward. Hence, there does not appear to be a signifi-
cant difference in the proportion of tournaments with ties such that participants
could win a reward across the Moderate and Strong Group Identity conditions
(z ¼ 0.56, p ¼ 0.58). If a higher proportion of tournament rounds with ties on the
winning rank is an indicator of greater strategic collusion, then these results sug-
gest that the degree of strategic collusion does not differ across Winner Proportion
or Group Identity conditions, supporting our argument in footnote 2 that strategic
collusion is not a likely alternative explanation for our results.
16
The mean earnings per participant for the four participants in a group are $14
for both the Winner Proportion conditions. Small Winner Proportion: The top
performer won a reward of $3.00 þ fixed pay of $1.00. The remaining 3 performers
received only the fixed pay of $1.00. The maximum earnings per participant ¼
($3.00 þ $1.00) x 8 rounds ¼ $32. The minimum earnings per
participant ¼ $1.00
8 rounds ¼ $8. The mean earnings per participant ¼
($32 þ 3*$8)/4 ¼ $14. Large Winner Proportion: The top 3 performers won a reward
of $1.00 þ fixed pay of $1.00. The bottom performer received only the fixed pay of
$1.00. The maximum earnings per participant ¼ ($1.00 þ $1.00) x 8 rounds ¼ $16.
The minimum earnings per participant ¼ $1.00
8 rounds ¼ $8. The mean earnings
per participant ¼ ($16*3 þ $8)/4 ¼ $14. Individual expected values will vary within
a group depending on an individual's perception of relative abilities in the group.
manner as in Experiment 1.14 We also manipulated Winner Pro-
portion using the same tournament descriptions used in Experi-
ment 1. However, unlike Experiment 1, participants in Experiment
2 actually earned the rewards each round based on their perfor-
mance rank and their Winner Proportion condition.15 Average
earnings per participant were held constant between the Winner
Proportion conditions.16
4.1.6. Dependent variable
Our dependent variable is Individual Performance in each tour-
nament round. Individual Performance is the equivalent units (EUs),
calculated as the number of single decodes plus two times the
number of double decodes, correctly decoded in a round by a
participant. Based on pilot test results, the time required to com-
plete two single decodes was approximately the same as the time
required to complete one double decode.
4.2. Results
4.2.1. Manipulation checks
For our Winner Proportion manipulation, we required partici-
pants to correctly answer manipulation check questions prior to
beginning the eight tournament rounds. We also evaluate the
effectiveness of our Winner Proportion manipulation by examining
participants' Average Expectancy over eight rounds. Expectancy was
a participant's assessment of the probability that s/he would earn a
reward in an upcoming round on a scale of 0%e100%. As shown in
Panel B of Table 3, the Average Expectancy over eight rounds of
63.32% in the Large Winner Proportion condition is larger than that
of 39.89% in the Small Winner Proportion condition (t ¼ 8.89,
p < 0.01).17 This is consistent with participants perceiving a greater
probability of winning a reward in the Large Winner Proportion
tournament than in the Small Winner Proportion tournament. Thus,
our manipulation of Winner Proportion appears effective.
To evaluate the effectiveness of our manipulation of Group
Identity, we asked participants their level of agreement to three
statements (Happy, Belong, Like, described earlier in the “Proced-
ures” sub-section) on a seven-point scale after they experienced
the Group Identity manipulation but before they experienced the
Winner Proportion manipulation. We also asked the same three
questions after participants completed all eight tournament
rounds. We averaged a participant's responses to these three
questions to obtain our measure for Tournament Group Identity pre-
tournaments and post-tournaments. Confirmatory factor analysis
indicates that these three questions represent a uni-dimensional
17
There are no significant round-by-condition interactive effects on Expectancy.
The Average Expectancy over eight rounds does not differ by the Group Identity
condition (p ¼ 0.93), nor is the interaction between Group Identity and Winner
Proportion significant (p ¼ 0.65). Average Expectancy over eight rounds is positively
associated with average Individual Performance over eight rounds (p < 0.01). In the
Large Winner Proportion tournament, Average Expectancy over eight rounds is not
significantly different for the two most able participants drawn from quartiles 1 and
2 (p ¼ 0.57); and it is lower for the participant drawn from quartile 3 compared to
participants drawn from quartiles 1 and 2 (p < 0.01); and it is lower for the
participant drawn from quartile 4 compared to the participant drawn from quartile
3 (p ¼ 0.09) (i.e., Q1 ¼ Q2 > Q3 > Q4). This is consistent with the design of the Large
Winner Proportion tournament where everyone but the least able participant wins a
reward. In the Small Winner Proportion tournament, Average Expectancy over eight
rounds is higher for the quartile 1 participant than the quartile 2 participant
(p < 0.01), but not significantly different between the quartile 2 participant and the
quartile 3 participant (p ¼ 0.95), and not significantly different between the
quartile 3 participant and quartile 4 participant (p ¼ 0.37) (i.e., Q1 > Q2 ¼ Q3 ¼ Q4).
This is consistent with the design of the Small Winner Proportion tournament where
only the most able participant wins a reward. These results indicate that our par-
ticipants' expectancies of winning a reward are consistent with their relative
(heterogeneous) abilities and the manipulated Winner Proportion.
 K. Kelly, A. Presslee / Accounting, Organizations and Society 56 (2017) 21e34 31

Table 4
Experiment 2 marginal means and multilevel model analysis.a

Panel A: Marginal Means (Delta-method Standard Error) Individual Performance

Small Winner Proportion Large Winner Proportion

Moderate Group Identity 27.57 28.06

(0.35) (0.36)
Strong Group Identity 27.92 27.35
(0.38) (0.35)

Panel B: Multilevel Model Analysisb

Model 1 Model 2

Fixed Effects
Intercept 27.74*** 16.53***
(0.28) (1.28)
Group Identity 4.62**
(1.99)
Winner Proportion 2.69
(0.48)
Group Identity X Winner Proportion 4.82**
(2.95)
Variance (Covariance) Estimates
Individual Variance (Covariance) 0.59*** 0.40***
(0.02) (0.02)
Group Variance (Covariance) 1.07* 1.12***
(0.77) (0.40)
Model Fit
AIC 18,199.00 17,236.68
BIC 18,209.00 17,308.40

Panel C: Planned Contrast

Contrast Estimate (S.E.) t-value p-valuec

Small Winner Proportion:

Strong Group Identity vs. Moderate Group Identity 0.20 (2.18) 0.09 0.93
Large Winner Proportion:
Strong Group Identity vs. Moderate Group Identity 4.62 (1.99) 2.33 0.01
a
See Table 3 for variable definitions. Large Winner Proportion is coded as “1” and Small Winner Proportion is coded as “0”. Strong Group
Identity is coded as “1” and Moderate Group Identity is coded as “0”. The number of Individual Performance observations is 2,912, which is
based on each individual (N ¼ 364) completing 8 rounds.
b
Round and Ability fixed effects are included in the model; however, the effect of Round, Ability, their interactions with condition are
not reported in Panel B to make the presentation of the results more concise. ML estimates are used. p-values are two-tailed except for the
interaction between Group Identity and Winner Proportion, which is shown one-tailed because a directional prediction is made.
***p < 0.01. **p < 0.05. The covariance structure for the group random effect is the default structure variance component because there is
no ex-ante prediction of the structure. The covariance structure for the individual random effect is auto-regressive because there is strong
evidence of serial correlation (Woolridge test: F ¼ 7.92, p < 0.01). Model 1 reflects the intra-class correlation and confirms that a
multilevel analysis is appropriate (i.e., covariance estimates at the individual and group levels are significant). Model 2 is the full
multilevel model reporting the unstandardized regression coefficients and the standard error in brackets.
c
p-value for the contrast in the Small Winner Proportion tournament is two-tailed because the direction of the effect is positive and
hence, opposite of our expectation.

construct in both the pre-tournaments and post-tournaments Identity appears effective.

measures, with all loadings greater than 0.78 and eigenvalues
(variances explained) greater than 2.10 (74%) (Stevens, 1996).
4.2.2. Hypothesis test
Further, both the pre-tournaments and post-tournaments mea-
We test H3 using multi-level model analysis. A multi-level
sures have Cronbach alphas greater than 0.90 (Kline, 2011). Both
model addresses the potential for correlated error terms caused
the pre-tournaments and the post-tournaments Tournament Group
by having eight rounds of Individual Performance data nested within
Identity measures in the Strong Group Identity condition are larger
each individual participant, and then by having data from four in-
than those in the Moderate Group Identity condition (Panel C of
dividual participants nested within each tournament group. That is,
Table 3: Pre-tournaments: 6.02 versus 4.90, t ¼ 8.87, p < 0.01; Post-
there are eight observations drawn from the same individual
tournaments: 5.61 versus 4.64. t ¼ 6.30, p < 0.01).
participant over the eight tournament rounds, and then the ob-
As in Experiment 1, we further examine the intra-class corre-
servations drawn from four individual participants belong to the
lation coefficient (ICC) because individual participants are nested
same tournament group. Thus, multi-level model analysis allow us
within tournament groups and our manipulation of Group Identity
to examine the fixed effects of Winner Proportion, Group Identity,
is applied at the tournament group level. We find that tournament
Round, and Ability, while controlling for the two random effects
groups account for a significant portion of the variability in the pre-
caused by our nested design (Klein & Kozlowski, 2000; Luke, 2004).
tournaments Tournament Group Identity measure (ICC ¼ 51%,
H3 predicts that increasing the proportion of tournament win-
F ¼ 2.06, p < 0.01) and the post-tournaments Tournament Group
ners increases the negative effect of group identity on tournament
Identity measure (ICC ¼ 24%, F ¼ 1.32, p ¼ 0.05). Based on the
performance. Table 3, Panel D presents the mean (standard devia-
significant ICCs and the mean Tournament Group Identity measures
tion) of individual Ability and Individual Performance in each round.
across the two Group Identity conditions, our manipulation of Group
Indeed, we find that the Group Identity and Winner Proportion
32 K. Kelly, A. Presslee / Accounting, Organizations and Society 56 (2017) 21e34
Fig. 3. Experiment 2 Results. See Table 3 for definitions of Winner Proportion, Group
Identity, and Individual Performance. Marginal mean Individual Performance from Panel
A of Table 4 are shown and are based on the multilevel analysis results presented in
Panel B of Table 4.
interaction in our multi-level model is negative and significant
(Table 4, Panel B: b ¼ 4.82, t ¼ 1.63, one-tailed p ¼ 0.05). We
graphically present marginal mean Individual Performance results
by condition in Fig. 3.18
In the Large Winner Proportion condition, marginal mean Indi-
vidual Performance is significantly worse by 0.71 decodes when
participants share Strong Group Identity rather than Moderate Group
Identity (see Panel A of Table 4 for means) (Panel C of Table 4:
t ¼ 2.33, one-tailed p ¼ 0.01). In the Small Winner Proportion con-
dition, marginal mean Individual Performance is insignificantly
better by 0.35 decodes when participants share Strong Group
Identity rather than Moderate Group Identity (see Panel A of Table 4
for means) (Panel C of Table 4: t ¼ 0.20, p ¼ 0.93). Collectively, these
results provide support for H3 such that the negative effect of
Strong Group Identity versus Moderate Group Identity on perfor-
mance is greater in Large Winner Proportion tournaments versus
Small Winner Proportion tournaments.19
We also consider whether the effects of Group Identity and
Winner Proportion differ over repeated tournament rounds. Despite
significant Rounds effects (not tabulated: b ¼ 0.46, t ¼ 13.76,
p < 0.01), the result of a decrease in performance over time, there is
no evidence of any Rounds-by-condition interactions (all p > 0.16).
This decrease in performance is consistent with participants
experiencing boredom or fatigue over time. Thus, the interactive
effect of Group Identity and Winner Proportion persists over multiple
tournament rounds.
18
Prior studies have shown that, when participants are heterogeneous in their
abilities on a real effort task, increasing winner proportion can increase tournament
performance (e.g., Chen et al., 2011; Freeman & Gelber, 2010; Knauer et al., 2016).
These studies appear to examine settings similar to our Moderate Group Identity
condition as participants experience little interaction and do not collaborate with
their fellow group members. Consistent with the results from these studies, we find
in the Moderate Group Identity condition that those in the Large Winner Proportion
marginally outperform those in the Small Winner Proportion condition (not tabu-
lated: t ¼ 1.41, one-tailed p ¼ 0.08).
19
The inferences from the results for Group Performance (calculated as the sum of
all Individual Performance of participants in a group) are similar. Multi-level model
analysis indicates that the interaction between Group Identity and Winner Propor-
tion is negative and significant such that in the Large (Small) Winner Proportion
condition, marginal mean Group Performance is 2.60 (1.92) decodes worse (better)
when participants share Strong Group Identity rather than Moderate Group Identity
(not tabulated:b ¼ 63.75, t ¼ 2.52, one-tailed p ¼ 0.01).
5. Conclusion
Tournament incentive systems are commonly used by organi-
zations to motivate employee performance (Hazels & Sasse, 2008;
Kwoh, 2012). We examine the effects of two independent vari-
ables, tournament group identity and winner proportion, on tour-
nament performance. The degree to which tournament
participants identify with each other can vary significantly in
practice and the proportion of tournament winners is a funda-
mental design choice management accountants are involved in
making. Using two experiments, we find that the proportion of
tournament winners moderates the degree to which tournament
participants' identity with each other affects performance. In
Experiment 1, we find increasing tournament group identity in-
creases other-regarding preference, and other-regarding prefer-
ence decreases competitiveness more when participants imagine
competing in a large winner proportion tournament compared to a
small winner proportion tournament. In Experiment 2, we find that
increasing tournament group identity decreases performance un-
der a large winner proportion tournament, but it has no effect on
performance under a small winner proportion tournament.
Our study contributes to the existing tournament incentive
literature by extending the understanding of the performance ef-
fects of using tournaments of different winner proportions when a
key tournament group attribute (i.e., tournament group identity)
varies. We provide evidence consistent with our theoretical argu-
ments that stronger identity with fellow tournament participants
increases other-regarding preference, which is more likely to
reduce effort and performance in less competitive large winner
proportion tournaments but less likely to do so in more competitive
small winner proportion tournaments. Despite participants in
Experiment 2 competing in multiple rounds against one another
and receiving detailed feedback after each round, the interactive
effect of tournament group identity and winner proportion on
performance persists over multiple rounds. Our findings have sig-
nificant practical implications for designing a tournament reward
structure. Organizations should consider how the variation in
identity amongst tournament participants affects performance in
tournaments with different winner proportions. Our results sug-
gest that when a large proportion of winners is used in tourna-
ments, stronger identity amongst competing employees may
induce lower competitiveness and performance from employees as
other-regarding preference increases.
Our study has limitations, which present opportunities for
future research. First, we only examine one type of tournament
reward structure where winners of a tournament are given rewards
of the same size. Future research could examine the interactive
effects of tournament group identity and winner proportion under
alternative tournament reward structures. Some examples include
one where the size of the rewards is proportional to the perfor-
mance ranking (Freeman & Gelber, 2010) and one that rewards
winners and punishes losers (Newman & Tafkov, 2014). Second, we
did not allow tournament participants to communicate during or
between tournament rounds for internal validity purposes. Future
research could examine how the opportunity for participants to
communicate affects effort because of its effect on collusion or
sabotage, and whether the levels of collusion or sabotage differ
depending on tournament group identity and winner proportion.
Third, our experiment using students shows that stronger group
identity resulted in lower performance in the larger winner pro-
portion condition. Although there is research suggesting that high
performing employees may deliberately lower their performance to
avoid violating performance norms (e.g., Jensen, Patel, & Raver,
2014), it would be useful for future research to use field settings
to examine if employees in the workplace would deliberately lower
 K. Kelly, A. Presslee / Accounting, Organizations and Society 56 (2017) 21e34
their performance to benefit fellow employees when economic
rewards are at stake. Our study suggests that employees would be
more willing to do so if the associated economic cost and psycho-
logical cost are lower, as in larger winner proportion tournaments.
Finally, our study has tournament participants competing against
members of their own group. Many firms use tournaments to
encourage competition between groups of employees rather than
within groups (Chen et al., 2012). Future research could investigate
the effects of between-group winner proportion and within-group
identity on performance. Despite these limitations, we believe our
study makes an important contribution to the literature on tour-
nament incentives.
Acknowledgements
We thank Nicole Cade, Eddy Cardinaels, Willie Choi, Bart Dier-
ynck, Harry Evans, Lan Guo, Gary Hecht, Theresa Libby, Elaine
Mauldin, Andrew Newman, Todd Thornock, Kristy Towry, Alan
Webb, Michael Williamson, anonymous reviewers, seminar par-
ticipants at Tilburg University, and reviewers and conference par-
ticipants at the 2013 Accounting, Behavior, and Organizations
Research Conference and 2014 Management Accounting Section
Research and Case Conference for their helpful comments on earlier
versions of this paper. Thanks to Ian Burt, Jeffrey Clark, and Kun Huo
for their research assistance. We are grateful to the University of
Waterloo's School of Accounting and Finance Experimental
Research Grant Program and to the University of Pittsburgh's
Central Development Research Fund for their financial support.
Khim Kelly gratefully acknowledges financial support from the CPA
Ontario Research Fellow Award.
References
Aron, A., Aron, E., & Smollan, D. (1992). Inclusion of other in the self scale and the
structure of interpersonal closeness. Journal of Personality and Social Psychology,
63(4), 596e612.
Balliet, D., Wu, J., & De Dreu, C. K. W. (2014). Ingroup favoritism in cooperation: A
meta-analysis. Psychological Bulletin, 140(6), 1556e1581.
Beach, S., & Tesser, A. (1995). Self-esteem and the extended self-evaluation main-
tenance model: The self in social context. In M. Kernis (Ed.), Efficacy, agency, and
self-esteem (pp. 145e170). New York: Plenum Press.
Bekkers, R., & Wiepking, P. (2010). A literature review of empirical studies of phi-
lanthropy: Eight mechanisms that drive charitable giving. Nonprofit and
Voluntary Sector Quarterly, 40(5), 924e973.
Berger, L., Klassen, K., Libby, T., & Webb, A. (2013). Complacency and giving up
across repeated tournaments: Evidence from the field. Journal of Management
Accounting Research, 25(1), 143e167.
Buunk, A. P., Dijkstra, P., Bosch, Z. A., Dijkstra, A., & Barelds, D. P. H. (2012). Social
comparison orientation as related to two types of closeness. Journal of Research
in Personality, 46(3), 279e285.
Casas-Arce, P., & Martinez-Jerez, F. A. (2009). Relative performance compensation,
contests, and dynamic incentives. Management Science, 55(8), 1306e1320.
Chen, H., Ham, S. H., & Lim, N. (2011). Designing multiperson tournaments with
asymmetric participants: An experimental study. Management Science, 57,
864e883.
Chen, Y., & Li, S. X. (2009). Group identity and social preferences. The American
Economic Review, 99(1), 431e457.
Chen, X., Williamson, M., & Zhou, H. (2012). Reward system design and group
creativity: An experimental investigation. The Accounting Review, 87, 1885e1911.
Choi, J., Newman, A., & Tafkov, I. (2016). A marathon, a series of sprints, or both?
Tournament horizon and task complexity in multi-period settings. The Ac-
counting Review, 91(5), 1391e1410.
Chow, C. W. (1983). The effects of job standard tightness and compensation scheme
on performance: An exploration of linkages. The Accounting Review, 58,
667e685.
Cooper, D., & Kegal, J. H. (2016). Other regarding preferences: A selective survey of
experimental results. In J. Kagel, & A. Roth (Eds.), Handbook of experimental
economics (Vol. 2, pp. 217e289). Princeton, NJ: Princeton University Press.
Dawes, R. M., Van de Kragt, A., & Orbell, J. M. (1988). Not me or thee but we: The
importance of group identity in eliciting cooperation in dilemma situations:
Experimental manipulations. Acta Psychologica, 68, 83e97.
De Cremer, D., & Van Vugt, M. (1999). Social identification effects in social di-
lemmas: A transformation of motives. European Journal of Social Psychology,
29(7), 871e893.
Dechenaux, E., Kovenock, D., & Roman, M. (2012). A survey of experimental research
33
on contests, all-pay auctions and tournaments. Experimental Economics, 18(4),
609e669.
Demos, T. (2010). The way we work: Motivate without spending millions. Fortune,
161(5), 37e38.
Dovidio, J. F., & Penner, L. A. (2001). Helping and altruism. In G. J. O. Fletcher, &
M. S. Clark (Eds.), Blackwell handbook of social psychology: Interpersonal pro-
cesses (pp. 162e195). Malden, MA: Blackwell Publishing.
Dovidio, J. F., Piliavin, J. A., Gaertner, S. L., Schroeder, D. A., & Clark, R. D. (1991). The
arousal: Cost-reward model and the process of intervention: A review of the
evidence. In M. Clark (Ed.), Prosocial behavior: Review of personality and social
psychology (Vol. 12, pp. 86e118). Thousand Oaks, CA, US: Sage Publications, Inc.
Eckel, C., & Grossman, P. (2005). Managing diversity by creating team identity.
Journal of Economic Behavior & Organization, 58, 371e392.
Ellemers, N., Gilder, D. D., & Haslam, S. A. (2004). Motivating individuals and groups
at work: A social identity perspective on leadership and group performance.
The Academy of Management Review, 29(3), 459e578.
Fehr, E., & Schmidt, K. M. (2006). The economics of fairness, reciprocity and
altruismeexperimental evidence and new theories. In S. C. Kolm, & J. M. Ythier
(Eds.), Handbook of the economics of giving, altruism and reciprocity (pp.
615e694). Amsterdam: North Holland.
Festinger, L. (1954). A theory of social comparison processes. Human Relations, 7,
117e140.
Fortune. (2015). 100 best companies to work for. Fortune. Retrieved from http://
fortune.com/best-companies/.
Freeman, R. B., & Gelber, A. M. (2010). Winner proportion and information in
tournaments: Experimental evidence. American Economic Journal: Applied Eco-
nomics, 2, 149e164.
Friedkin, N. E., & Simpson, M. J. (1985). Effects of competition on members' iden-
tification with their subunit. Administrative Science Quarterly, 30, 377e394.
Garcia, S. M., & Tor, A. (2007). Rankings, standards, and competition: Task vs. scale
comparisons. Organizational Behavior and Human Decision Processes, 102,
95e108.
Garcia, S. M., Tor, A., Bazerman, M. H., & Miller, D. T. (2005). Profit maximization
versus disadvantageous inequality: The impact of self-categorization. Journal of
Behavioral Decision Making, 18, 187e198.
Garcia, S. M., Tor, A., & Schiff, T. M. (2013). The psychology of competition: A social
comparison perspective. Perspectives on Psychological Science, 8(6), 634e650.
Gino, F., & Galinsky, A. (2012). Vicarious dishonesty: When psychological closeness
creates distance from one's moral compass. Organizational Behavior and Human
Decision Processes, 119(1), 15e26.
Hannan, R. L., Towry, K., & Zhang, Y. (2013). Turning up the volume: An experi-
mental investigation of the role of mutual monitoring in tournaments.
Contemporary Accounting Research, 30(4), 1401e1426.
Harbring, C. (2006). The effect of communication in incentive systems e an
experimental study. Managerial and Decision Economics, 27, 333e353.
Hazels, B., & Sasse, C. (2008). Forced ranking: A review. SAM Advanced Management
Journal, (Spring).
Heider, F. (1958). The psychology of interpersonal relations. New York: Wiley.
Jensen, J. M., Patel, P. C., & Raver, J. L. (2014). Is it better to be average? High and low
performers as predictors of employee victimization. Journal of Applied Psy-
chology, 99(2), 296e309.
Keenan, W. (1995). Selling with sales contests. Sales & Marketing Management,
147(6), 35.
Kelly, K., Webb, A., & Vance, T. W. (2015). The interactive effects of ex post goal
adjustment and goal difficulty on performance. Journal of Management Ac-
counting Research, 27(1), 1e25.
King, R. (2002). An experimental investigation of self-serving biases in an auditing
trust game: The effect of group affiliation. The Accounting Review, 77(2),
265e284.
Klein, K. J., & Kozlowski, S. W. J. (2000). Multilevel theory, research, and methods in
organizations: Foundations, extensions, and new directions. San Francisco, CA:
Jossey-Bass.
Kline, R. (2011). Principles and practice of structural equation modeling (3rd ed.). New
York: Guilford Press.
Knauer, T., Sommer, F., & Wohrmann, A. (2016). Tournament winner proportion and
its effect on effort: An investigation of the underlying psychological mecha-
nisms. European Accounting Review (in press).
Knippenberg, D. V. (2000). Work motivation and performance: A social identity
perspective. Applied Psychology: An International Review, 49(3), 357e371.
Kwoh, L. (2012, January 31). Rank and yank’ retains vocal fans. Wall Street Journal.
Retrieved from http://online.wsj.com/article/
SB10001424052970203363504577186970064375222.html. B6.
Lim, N. (2010). Social loss aversion and optimal contest design. Journal of Marketing
Research, 47(4), 777e787.
Lount, R., & Phillips, K. (2007). Working harder with the out-group: The impact of
social category diversity on motivation gains. Organizational Behavior and Hu-
man Decision Processes, 103(2), 214e224.
Luke, D. (2004). Multilevel modeling. New Delhi, India: Sage Publication.
Matsumura, E. M., & Shin, J. (2006). Empirical analysis of an incentive plan with
relative performance measures: Evidence from a postal service. The Accounting
Review, 81(3), 533e566.
Newman, A., & Tafkov, I. (2014). Relative performance information in tournaments
with different prize structures. Accounting, Organizations and Society, 39(5),
348e361.
Peloza, J., & Steel, P. (2005). The price elasticities of charitable contributions: A
34 K. Kelly, A. Presslee / Accounting, Organizations and Society 56 (2017) 21e34
meta-analysis. Journal of Public Policy & Marketing, 24(2), 260e272.
Samsung. (2013). Contest terms can be viewed at http://ertonline.co.uk/samsung-
sales-person-of-the-year.htm or once registered at http://www.
samsungrewards.co.uk.
Sisak, D. (2009). Multiple-Prize contests - the optimal allocation of prizes. Journal of
Economic Surveys, 23, 82e114.
Snyder, C. R., Lassegard, M. A., & Ford, C. E. (1986). Distancing after group success
and failure: Basking in reflected glory and cutting off reflected failure. Journal of
Personality and Social Psychology, 51(2), 382e388.
Spencer, S., Zanna, M., & Fong, G. (2005). Establishing a causal chain: Why exper-
iments are often more effective than mediational analyses in examining psy-
chological processes. Journal of Personality and Social Psychology, 89(6),
845e851.
Stapel, D. A., & Koomen, W. (2001). I, we, and the effects of others on me: How self-
construal level moderates social comparison effects. Journal of Personality and
Social Psychology, 80(5), 766e781.
Stevens, J. (1996). In Applied multivariate statistics for the social sciences (3rd ed.).
Hillsdale, N.J.: Erlbaum Associates.
Szymanski, S., & Valletti, T. M. (2005). Incentive effects of second prizes. European
Journal of Political Economy, 21, 467e481.
Tajfel, H., & Turner, J. C. (1986). The social identity theory of intergroup behavior. In
S. Worchel, & W. Austin (Eds.), Psychology of intergroup relations (pp. 7e24).
Chicago, IL: Nelson-Hall.
Tesser, A., & Campbell, J. (1980). Self-definition: The impact of the relative perfor-
mance and similarity of others. Social Psychology Quarterly, 43(3), 341e347.
Tesser, A., & Campbell, J. (1982). Self-evaluation maintenance and the perception of
friends and strangers. Journal of Personality, 50(3), 261e279.
Towry, K. (2003). Control in a teamwork environment: The impact of social ties on
the effectiveness of mutual monitoring contracts. The Accounting Review, 78(4),
1069e1095.
Turner, J. C. (1982). Towards a cognitive redefinition of the social group. In H. Tajfel
(Ed.), Social identity and intergroup relations (pp. 15e40). Cambridge, England:
Cambridge University Press.
Van Dijk, F., Sonnemans, J., & Van Winden, F. (2001). Incentive systems in a real
effort experiment. European Economic Review, 45, 187e214.
Van Lange, P. A. M., Rusbult, C. E., Drigotas, S. M., Arriaga, X. B., & Witcher, B. S.
(1997). Willingness to sacrifice in close relationships. Journal of Personality and
Social Psychology, 72(6), 1373e1395.
Welch, J. (2013, November 14). Jack welch: ‘Rank-and-yank’? that's not how it's
done. Wall Street Journal. Retrieved from http://www.wsj.com/articles/
SB10001424052702303789604579198281053673534. Opinion section.
Wit, A., & Wilke, H. A. (1992). The effect of social categorization on cooperation in
three types of social dilemmas. Journal of Economic Psychology, 13(1), 135e151.
 Accounting, Organizations and Society 56 (2017) 35e37

Contents lists available at ScienceDirect

Accounting, Organizations and Society

journal homepage: www.elsevier.com/locate/aos

Editorial

Themed Section on Financial Accounting as Social and Organizational

Practice: Exploring the work of financial reporting

The institutional and social aspects of financial accounting are still increasingly turn to the vehicle of the ‘special issue’ e hoping
relatively unexplored. Compared with our insights into the eco- that calls for papers with a specific focus may effect a shift in the
nomic theory of income calculation and the economic de- attention and ideas of scholars, and create a platform for risk-
terminants and consequences of modes of corporate financial taking and new ideas. However, such ventures create a new kind
reporting, our knowledge of how forms of financial accounting of problem: the unit of innovation is effectively a critical mass of pa-
emerge from, sustain and modify wider institutional and social pers in which the whole is greater than the sum of the parts, and
structures is modest (Anthony Hopwood) where individual papers can be evaluated both in themselves and
also for what they bring to what amounts to an ‘intellectual’ ven-
ture capital portfolio. Indeed, the growth and prevalence of such
Over the years innovation in accounting research has taken
special issues across the management and organization sciences,
place in a number of different ways: seeing accounting through
including accounting, is interesting given the apparent disdain
new theoretical lenses; describing and interpreting new accounting
with which edited book collections are now regarded in
practices; finding new associations between accounting phenom-
university-ranking methodologies and individual academic ap-
ena and aspects of the wider social and economic environment;
praisals. It is as if the underlying ideas and aspirations of the edited
exploiting new data sources and methodologies; and building
collection remain relevant but now find expression within the jour-
new theoretical models of decision making involving accounting
nal arena.
information. In some of these cases, connections between account-
As the history of Accounting, Organizations and Society reveals,
ing research and emerging issues in adjacent fields have been
Anthony Hopwood understood this portfolio concept of innovation
made. Indeed, innovation is occurring as other regions of the social
and its associated risks very well. In 2009, shortly before he passed
sciences begin to see the significance of mechanisms of control and
away, he commissioned and published a special issue on carbon ac-
performance accountability. And, as these new forms of account-
counting which exemplified the experimental and eclectic nature
ability have mushroomed, scholars have expanded their view of
of first steps to highlight and explore an emerging ‘accounting’
what counts as accounting. For example, there is currently intense
issue. Yet it is perhaps the mid-1980s which marks a peculiarly
interest in understanding such things as: ranking systems and their
rich period in the history of the journal, with the appearance of
effects; and the role of non-financial information in performance
landmark papers which continue to influence scholars today. In a
management systems and business model reporting. Furthermore,
sense, while the journal was born in 1976, the AOS field did not
recent work on the linkages between conventional accounting
take off until the mid-1980s as Hopwood's own pioneering ideas
practices and ideals of sustainability and social impact revisit
were taken up and adapted. However, to a large extent the field
some of the very earliest preoccupations of Accounting, Organiza-
at that time developed with a distinctive emphasis on managerial
tions and Society.
accounting and control. Space does not permit a full exploration
Yet if the vectors of possible research innovation are diverse e
of the reasons for this but two can be suggested (See also Miller
and only a few have been listed above - innovation is also chal-
& Power, 2013; Robson & Young, 2009).
lenging and, in these days of increasingly normalized notions of
On the one hand, the field of management control in the UK was
quality, also very risky. The dilemma for many journal editors is
particularly receptive to a range of social and organizational the-
simply this: how to encourage a reviewing environment receptive
ories, and the connection between accounting and management
to forms of innovation which, almost by definition, operate at the
studies was easy to make e although it required a few key catalysts.
margins of the conventional methodological and/or theoretical
On the other hand, financial accounting as an object of academic in-
norms of their chosen genre. New research areas, almost of neces-
terest somehow became associated with a very specific mode of
sity, reach beyond their data, and are more loosely linked to theo-
empirical study that focused on its role and effects in capital mar-
retical frames or are linked to frames sometimes significantly
kets. It is always intriguing how some fields of study acquire such
different than the “norm.” But it is exactly these characteristics
a strong association, almost identification, with the methodological
that help to open up new spaces for thinking and empirical
orientation through which they are investigated. For example, uni-
exploration.
versity departments of Finance police, and are allowed to police,
To meet the challenge of seeding innovation, editors
what may be intellectually associated with that label, despite the

http://dx.doi.org/10.1016/j.aos.2017.01.001
0361-3682/© 2017 Published by Elsevier Ltd.
36 Editorial / Accounting, Organizations and Society 56 (2017) 35e37
fact that “financial economics” is but one way to study the field of
finance. To a lesser extent, financial accounting has acquired a
similar kind of association with the methods of capital markets
research. Although exceptions exist (e.g. Burchell, Clubb, &
Hopwood, 1985; Gendron & Be dard, 2006; Ravenscroft &
Williams, 2009; Young, 2006), this work had been carved out by in-
dividuals working largely against the scholarly tide of the time.
These two, somewhat abbreviated, explanations for the mana-
gerial control emphasis in AOS in the 1980s and 1990s provide
part of the rationale for this themed section, and the three papers
it contains. But world events played a role too. The financial crisis
of late 2008, and the highly public and political debates about the
role of fair value measurement, energised enquiry about financial
accounting. Hopwood sensed this early on in 2009 when he
commissioned a number of short pieces on the financial crisis,
including the much cited paper by Laux and Leuz (2009). But other
things were happening too. The significance of the rise of financial
economics, even prior to the crisis, was not something that could be
comprehended by financial economics itself in anything other than
progressive and functional terms. While Whitley’s (1986) analysis
was ahead of its time, the crisis gave a new impetus and voice
both to the emerging field of social studies of finance (Abolafia,
1997), and also to sociological scholarship about values and value
(The online journal Valuation Studies began life in 2013). All these
things made it possible for accounting scholars to think about
financial accounting in new ways in relation to finance and valua-
tion (cf Vollmer, Mennicken, & Preda, 2009).
Sensing this wave, we initiated a call for papers in 2011 to
explore ‘Financial Reporting as a Social and Organizational Practice’
(FRASOP). The call attracted nearly 60 submissions of which there
was space for 14 in a workshop at the London School of Economics
in December 2011. This themed section contains three papers from
that original workshop e of which more below. The interest in, and
success of, ‘FRASOP 1’ as we now call it, generated the confidence
that there was sufficient commitment and capability to develop
and sustain this emerging sub-field, and so the workshop has
become institutionalized and expanded to include auditing. FRA-
SOPs 2 and 3 took place in 2013 and 2016 respectively.
A good call for papers is a kind of mini version of the innovation
which it hopes to stimulate and encourage. Indeed, some of Hop-
wood's calls for papers could be more innovative than the papers
they attracted. And this is as it should be. A call for papers should
be risky, exciting and attractive enough to catalyse interest. It
should be the statement of an agenda which cannot possibly be
realised in one workshop or one themed section like this; it is
more a statement of hope and aspiration, intended to excite and
attract scholarship even in these intellectually conservative times.
We cannot be sure that our original call in 2011 lived up to this
ideal, but we are happy enough with the outcome. And while we
publish only three papers in what follows, several other papers pre-
sented at the FRASOP workshops are also in the pipeline at AOS and
elsewhere.
In the first paper, Jaana Kettunen deals with the process by
which IFRSs are translated between different languages. Her anal-
ysis focuses on the work of specific actors who engage in transla-
tion in the Finnish context, and it draws on translation theory
and on notions of institutional work to capture the processual na-
ture of translation activity. While much has been written about
the interpretation of regulation in general, and accounting regula-
tions specifically, very little has been done on the work and impact
of translation, perhaps because many scholars assume the univer-
sality of the English language and therefore the unproblematic na-
ture of translation. Yet at the centre of Kettunen's study is a careful
account of how key actors deal with the problem of linguistic equiv-
alence, a problem which is often compounded by the essential
ambiguity of accounting terms themselves. Drawing on ideas of
inter-textuality, Kettunen shows how this seemingly marginal
and largely invisible activity constructs translations which sustain
the position of IASB as a global, and supposedly uni-lingual
institution.
In the second paper, Barker and Schulte develop a case-based
critique of fair value accounting which draws on the philosopher
John Searle's notion of an “institutional fact”, namely something
which can be known and represented objectively within, and rela-
tive to, an institution. Using the example of several case companies
complying with IFRS 1, Barker and Schulte argue that the prepara-
tion of financial statements involves either the reporting of institu-
tional facts already in existence, or else the creation of data that
cannot themselves constitute new institutional facts. In essence,
they suggest that in cases where fair values are determined by
reference to a hypothetical market participant, preparers are left
to wrestle with a problem of compliance in the absence of institu-
tional facts. Preparers solve this problem by various expedient and
pragmatic means that further a shift from reliability to claims of
representational faithfulness, a shift that contributes to funda-
mental incoherence at the level of practice.
In the third paper, by Hiukku, Mouritsen and Silvola, there is also
a focus on the pragmatism needed to operationalize fair value e
this time in the context of the impairment testing of goodwill. Their
analysis takes us to the heart of how financial statements are pro-
duced and collectively constructed. In the case of goodwill and its
possible impairment, organizational actors must look for evidence
e traces - of its future value. Based on case and interview material
in Finland, the authors argue that the facts of goodwill impairment,
or otherwise, are produced by a distributed network of heteroge-
neous valuation practices and, furthermore, it is the shape of this
network of human and non human actors which creates the
perceived reliability of the goodwill numbers. As with Barker and
Schulte's analysis, the key actors in the network have little concrete
guidance, other than the accounting standard itself, and must work
hard to make the numbers real.
These three papers stand alone in their own right, but they also
represent an emerging agenda for research into financial account-
ing and auditing, not least in their common emphasis on the
work involved in the production of financial reporting at the orga-
nizational level, a theme we had highlighted in our original call for
papers. Yet despite their contribution, we still know relatively little
about the wider organizational environment which conditions the
production of financial statements within and beyond the finance
function. The papers by Kettunen and Hiukku et al. in effect point
to a new object of potential analysis, namely the infrastructures
that support financial reporting, not least the data required to be
assembled in order to test fair values (Power, 2015). Not only do
such infrastructures span the managerial/financial accounting
divide but, as a prospective unit of analysis, they also embody ac-
counting ‘work’ being done by non-accountants within and outside
the organization. At stake in these analyses is therefore the nature
of the connectivity of the finance function to the productive parts of
an organization and also to external agencies.
These three papers also speak to another theme that we high-
lighted in the original call of 2011 and which is enjoying promi-
nence outside the accounting field, namely valuation, networks
and financial reporting. Each paper touches on different networks
of expertise which constitute the financial reporting process, and
two of them show how accounting valuation is the result of effort-
ful accomplishments in these largely invisible networks. But there
are still many questions for the FRASOP agenda to pursue, largely
because it aims to investigate exactly the space that capital market
based studies of accounting ignore or assume away, namely the
process by which financial accounts are collectively constructed
 Editorial / Accounting, Organizations and Society 56 (2017) 35e37
(their production) and the mechanisms by which such accounts
travel to, and are utilized by, different audiences, such as analysts
(their consumption). This is not the text book world of accounting
for decision making about capital allocation by clearly identified,
rational users. It involves sub- and trans- organizational worlds of
messy, collective endeavors and different sites characterized by
their own institutionalized rhythms of production: from the places
where standards are written to the places where standards are
interpreted, translated and applied to produce accounts, to the pla-
ces where these accounts are analysed. There is undoubtedly much
more scholarly work to do in this space using many different
methods and approaches, not least those which take a less norma-
tive approach to sustainability accounting, integrated reporting and
related developments. We think that there remains a need for work
which focuses more on the processes and routines by which these
new accounting forms do and don't take shape, rather than, as Hop-
wood put it many times, assuming and presuming their potential.
We hope that the three papers which follow will give an idea of
what is possible.
References
Abolafia, M. Y. (1997). Making markets: Opportunism and restraint on wall street.
Cambridge (Massachusetts): Harvard University Press.
Burchell, S., Clubb, C., & Hopwood, A. G. (1985). Accounting in its social context: To-
wards a history of value added in the United Kingdom. Accounting, Organiza-
tions and Society, 10(4), 381e413.
Gendron, Y., & Be dard, J. (2006). On the constitution of audit committee effective-
ness. Accounting, Organizations and Society, 31(3), 211e239.
Laux, C., & Leuz, C. (2009). The crisis of fair-value accounting: Making sense of the
37
recent debate. Accounting, Oganizations and society, 34(6), 826e834.
Miller, P., & Power, M. (2013). Accounting, organizing, and economizing: Connecting
accounting research and organization theory. The Academy of Management
Annals, 7(1), 557e605.
Power, M. (2015). How accounting begins: Object formation and the accretion of
infrastructure. Accounting, Organizations and Society, 47, 43e55.
Ravenscroft, S., & Williams, P. F. (2009). Making imaginary worlds real: The case of
expensing employee stock options. Accounting, Organizations and Society, 34(6),
770e786.
Robson, K., & Young, J. (2009). Socio-political studies of financial reporting and stan-
dard-setting. In C. S. Chapman, D. J. Cooper, & P. Miller (Eds.), Accounting, orga-
nizations, and institutions: Essays in honour of Anthony Hopwood (pp. 341e366).
Oxford: Oxford University Press.
Vollmer, H., Mennicken, A., & Preda, A. (2009). Tracking the numbers: Across ac-
counting and finance, organizations and markets. Accounting, Organizations
and Society, 34(5), 619e637.
Whitley, R. (1986). The transformation of business finance into financial economics:
The roles of academic expansion and changes in US capital markets. Accounting,
Organizations and Society, 11(2), 171e192.
Young, J. J. (2006). Making up users. Accounting, Organizations and Society, 31(6),
579e600.
Keith Robson*
HEC, Paris, France
Joni Young
University of New Mexico, United States
Mike Power
London School of Economics and Political Science, United Kingdom
*
Corresponding author.
E-mail address: robson.keith@gmail.com (K. Robson).
 Accounting, Organizations and Society 56 (2017) 38e54

Contents lists available at ScienceDirect

Accounting, Organizations and Society

journal homepage: www.elsevier.com/locate/aos

Interlingual translation of the International Financial Reporting

Standards as institutional work
Jaana Kettunen
€skyla
Jyva € University School of Business and Economics, Finland

a r t i c l e i n f o a b s t r a c t

Article history: The International Financial Reporting Standards (IFRS) have been widely adopted well beyond English-
Received 18 October 2012 speaking jurisdictions. Using the Finnish translation of the IFRS as a primary object of investigation, this
Received in revised form article analyses the way in which the standards are translated into another language. Drawing on in-
7 September 2016
terviews with translators and translation review committee members and on an analysis of archival
Accepted 3 October 2016
materials, it provides an empirically grounded understanding of practical problems of linguistic equiv-
Available online 22 November 2016
alence, and the institutional work required to maintain the IFRS as a global, translingual institution.
Accordingly, the article highlights the constructed and negotiated nature of the linguistic equivalence
between the IFRS and their translations. As translation lies at the interface between transnational
standard-setting and local implementation, examining the IFRS translation offers further insight into the
complex institutional interactions and practices that support transnational regulation. The article
identifies the commonalities and discrepancies between the translation policies of the EU and the IFRS
Foundation, and analyses the translation as a contested area of expertise involving multiple, recurrently
changing constituents.
© 2016 Elsevier Ltd. All rights reserved.

1. Interlingual translation of International Financial
Reporting Standards as institutional work
As standard-setting and regulation take place in an increasingly
transnational and multilingual context, accounting concepts travel
across borders and languages. On a global scale, the transnational
standards originally written in the English language are ultimately
translated into financial statements prepared in dozens of different
languages. When regulatory texts such as accounting standards are
initially drafted, accounting facts are constructed in the regulation
process by labelling particular matters with either newly coined or
existing accounting terms (Gro €jer, 2001; Hines, 1988; Young, 2003).
Thereafter, accounting standards are translated into other lan-
guages. Despite the rise and spread of English-language regulations
(Botzem & Dobusch, 2012; Chua & Taylor, 2008; Djelic & Sahlin-
Anderssen, 2006; Mennicken, 2008), we know very little about
the processes and related practices that facilitate the use of these
standards in non-Anglophone countries.
Using the Finnish translation of the International Financial
Reporting Standards (IFRS1) as a primary object of investigation,
this paper analyses how the IFRS2 originally drafted in English are
translated from one language (the source-language) into another
(the target-language), and how the problems of linguistic equiva-
lence that arise during the translation are handled in their social
and institutional context. The article also shows that the translation
of the IFRS is a contested area of expertise, and the translation work
is governed by different regulators. In this paper, the concept of
translation refers to the rendering of a source-language text into the
target-language.
The limits of interlingual translation are well recognised in the
academic field of translation studies, and scholars have argued that
very rarely will a translation both render the original text word-for-
word into another language and convey its meaning unchanged
(e.g., Catford, 1978; Nida, 1964; Toury, 1995). Moreover, several
studies have suggested that accounting is conceptualised in
different ways in different languages and their related cultures,
which may impede the translation of transnational standards (e.g.,
Evans, Baskerville, & Nara, 2015; Zeff, 2007). Indeed, the perceived
lack of equivalence between languages is manifest in the concerns

2
In this paper, the acronym IFRS is used to refer both to a single standard and to
more than one standard or the entire set.
E-mail address: jaana.m.kettunen@jyu.fi.
1
For a list of abbreviations, see Appendix 1.

http://dx.doi.org/10.1016/j.aos.2016.10.001
0361-3682/© 2016 Elsevier Ltd. All rights reserved.
 J. Kettunen / Accounting, Organizations and Society 56 (2017) 38e54
expressed by academics, professional accountants,3 and some
representatives of European Union (EU) Member States4 over the
adequacy, readability and comprehensibility of the translated IFRS
(Dahlgren & Nilsson, 2012; Hellmann, Perera, & Patel, 2010; Nobes,
2006; Sunder, 2011; Wong, 2004).
This article makes the following contributions to the field. First,
it extends the literature examining transnational accounting
regulation (e.g., Botzem & Dobusch, 2012; Cooper & Robson, 2006;
Erb & Pelger, 2015; Gillis, Petty, & Suddaby, 2014; Mennicken,
2008; Pelger, 2016). The translation of the IFRS has a unique posi-
tion at the interface of development, interpretation and imple-
mentation of regulations. The article shows that translation
processes are governed by transnational regulators e the IFRS
Foundation and, within the EU, the European Commission (EC) e
and that the translation work involves multiple, often changing,
constituents. In doing so, the article responds to the continued calls
to study the development and interpretation of accounting regu-
lations in their social context (Canning & O'Dwyer, 2013; Cooper &
Robson, 2006; Humphrey, Loft, & Woods, 2009; Suddaby, Cooper, &
Greenwood, 2007). It also responds to the call by Mennicken to
investigate ‘the networks of actors, instruments and the activities
that support … standardising agendas in local settings’
(Mennicken, 2008, p. 385).
Second, by shifting the focus on how translators and translation
reviewers address the practical issues of linguistic equivalence, the
current research makes a methodological contribution to the
literature on the translation of accounting regulations. The prob-
lematics of translation in the domain of accounting have previously
been addressed by conducting experimental studies (e.g., Doupnik
& Richter, 2003), comparing excerpts from translated accounting
standards with their English language counterparts (e.g., Dahlgren
& Nilsson, 2012; Nobes, 2006) or analysing English-language ac-
counting concepts in comparison to those in other languages (e.g.,
Evans, 2004; Kosmala-MacLullich, 2005). The current article ex-
amines the translation of IFRS as a social and institutional practice
or, more specifically, as institutional work required to establish and
maintain IFRS as a global, multilingual institution (Lawrence &
Suddaby, 2006). It provides an empirically grounded understand-
ing of the practical problems of linguistic equivalence, and how
they are addressed by translators and translation reviewers.
Accordingly, the article highlights the constructed and negotiated
nature of linguistic equivalence. This is a novel contribution to the
literature because the extant studies pertaining to the translation of
the IFRS (e.g., Dahlgren & Nilsson, 2012; Evans et al., 2015; Huerta,
Petrides, & Braun, 2013) tell us very little about the activities
involved in the creation of these translations.
The remainder of this article proceeds as follows. The following
section describes the theoretical background to the study. Section 3
serves as an introduction to the case of translating IFRS into the
Finnish language by describing the regulatory context in which the
translation of IFRS takes place. Section 4 provides an outline of the
case selection, research methods and empirical materials. Section 5
analyses how translations of IFRS are created and how the prob-
lems of linguistic equivalence are handled in practice by those
involved in the translation work. It also illuminates how the
translation of the IFRS is a contested area of expertise. The final
Section 6 presents a concluding discussion of the main arguments
and opportunities for future research.
3
See http://www.ifac.org/sites/default/files/publications/files/challenges-and-
successes-in.pdf.
4
See Thyssen (2011). Question for written answer to the European Commission
http://www.europarl.europa.eu/sides/getDoc.do?pubRef¼-//EP//TEXTþWQþP-
2011-008747þ0þDOCþXMLþV0//EN.
39
2. Literature analysis: translation
This section begins with a review of how prior accounting
literature has examined translation in the context of transnational
accounting regulation. It then offers a brief overview of how
scholars in linguistics and translation studies have conceptualised
translation. In particular, the notion of equivalence is introduced to
conceptualise the general problematic of translation. Further, the
concept of intertextuality is introduced for the analysis of how
translators and translation reviewers attempt to make sense of the
source text and search for equivalent target-language terms and
expressions. In order to shed light on the practical actions through
which linguistic equivalence is produced in the translation of the
IFRS, the paper draws on the concept of institutional work, which is
discussed in the fourth subsection. The final subsection offers a
summary.
2.1. Previous research on translation of accounting regulations
2.1.1. Translation as a barrier to transnational accounting
harmonisation
An emerging body of literature considers translation to be a
potential barrier to transnational accounting harmonisation
(Baskerville & Evans, 2011; Dahlgren & Nilsson, 2012; Doupnik &
Richter, 2003; Evans et al., 2015; Hellmann et al., 2010; Nobes,
2006, 2013; Sunder, 2011; Zeff, 2007). Some scholars point out
that translation might change the intended meaning of a regula-
tion, which, in turn, might hinder transnational harmonisation of
accounting practices (e.g., Dahlgren & Nilsson, 2012; Holthoff,
Hoos, & Weissenberger, 2015; Nobes, 2006, 2013; Sunder, 2011).
The following paragraphs introduce the different approaches ac-
counting scholars have taken to studying the translation of ac-
counting regulations. Thereafter, the next subsection describes how
the empirical topic of the current article complements these ap-
proaches and contributes to the wider literature on transnational
accounting regulation.
First, there is a tradition of conducting experimental research on
the translation and interpretation of expressions of probability and
uncertainty (such as reasonably possible, probable and virtually
certain) in accounting and auditing standards (e.g., Davidson &
Chrisman, 1993; Doupnik & Riccio, 2006; Doupnik & Richter,
2003, 2004; Huerta et al., 2013). Experimental studies follow the
positivist research tradition in that they apply quantitative
methods to examine whether speakers of one language interpret
uncertainty expressions differently than speakers of another lan-
guage. The findings from these studies suggest that the concepts
underlying the words differ between languages, which may have a
bearing on how adequately uncertainty expressions, and conse-
quently accounting and auditing standards, are translated into
other languages (Davidson & Chrisman, 1993; Doupnik & Riccio,
2006; Doupnik & Richter, 2003). Further, the translation of uncer-
tainty expressions (Huerta et al., 2013) and their interpretation
(Aharony & Dotan, 2004; Doupnik & Richter, 2003; Laswad & Mak,
1997; Simon, 2002) differ among individuals who are native
speakers of the same language. Based on the findings of experi-
mental studies, researchers have concluded that inconsistent
interpretation of uncertainty expressions by preparers of financial
statements with different native languages can lead to the incon-
sistent application of accounting standards, especially given that
accounting standards include such expressions in abundance (e.g.,
Doupnik & Riccio, 2006; Doupnik & Richter, 2003).
Although languages are embedded in cultures, experimental
studies have sought to disentangle the language effect from the
culture effect. These two effects have been conceptualised through
different theories. Both Davidson and Chrisman (1993), and
40 J. Kettunen / Accounting, Organizations and Society 56 (2017) 38e54
Doupnik and Richter (2003), conceptualised their studies on
different language groups with reference to the notion of linguistic
relativism and the SapireWhorf hypothesis, which suggests that
‘grammatical forms and categories provided by a language are
thought to affect the manner in which speakers of a given language
interpret the world’ (Doupnik & Richter, 2003, p. 19; see also; Sapir,
1949, pp. 207e14). The differences in the interpretation of uncer-
tainty expressions between cultural groups (in contrast to language
groups) have been hypothesised based on the HofstedeeGray
framework on how national culture affects accounting values
(Doupnik & Riccio, 2006; Doupnik & Richter, 2004; Gray, 1988;
Hofstede, 1980). Collectively, the studies on probability and un-
certainty expressions have provided insights into the interpretation
of a very specific category of words. These studies suggest that in
the context of accounting and auditing standards there are practical
issues around equivalence between languages. However, they do
not tell us much about the range of translation problems, or how
the translators or those who apply the standards attempt to address
these issues.
The second stream of research that has addressed the trans-
lation of accounting regulations comprises studies that analyse the
translated accounting terms in comparison to their English-
language counterparts. Applying qualitative analysis, such as
back-translation into English, these studies have often concen-
trated on terms representing vague concepts such as fair presen-
tation and a true and fair view (TFV) (Aisbitt & Nobes, 2001; Evans,
2003, 2004; Kirk, 2006; Kosmala-MacLullich, 2003, 2005; Nobes,
2009; Ordelheide, 1993; Walton, 1993; Zeff, 2007). Some studies
have argued that the translation of TFV into other languages has
been inconsistent (Nobes, 2009, 2013). For example, German
translations have rendered true and fair and fair identically, and
Danish and Swedish translations include only one signifier for true
and fair (Aisbitt & Nobes, 2001; Dahlgren & Nilsson, 2012; Nobes,
2009). Furthermore, when the examined target-language terms
are translated back into English, many of the literal back-
translations do not correspond to TFV (Nobes, 2009, 2013).
Despite the merits of studies on TFV, back-translation as a
method of analysis has certain limitations. When a target-language
term is back-translated into the source-language, the back-
translation can introduce new shifts in meaning. Therefore, com-
parisons of a source-language term, (such as TFV) and the back-
translated term (such as true and sufficient picture) can provide
only a limited insight into the quality of a translation. Although
back-translation may illustrate whether the translation from source
to target-language is a direct translation, it does not address the
issues around conceptual equivalence (Douglas & Craig, 2007).
Kosmala-MacLullich (2005) surveyed Polish accounting practi-
tioners and found a lack of consensus on the most appropriate
translation of TFV into Polish. She argued that this finding reflects
the local unfamiliarity with the concept which has been predomi-
nantly constructed in the British and American context, and raises
issues about the (un)transferability of concepts such as the TFV
across languages and cultures. Similarly, it has been argued that the
direct translation of other fundamental accounting terms, such as
gains and impaired, can be problematic because the closest equiv-
alent target-language terms are defined differently in the local
accounting regulations than in the IFRS (Dahlgren & Nilsson, 2012;
Evans et al., 2015; Huerta et al., 2013; Mourier, 2004; Zeff, 2007).
Third, recent studies point to overt mistakes or inaccuracies in
translations that alter the meaning of the IFRS (Dahlgren & Nilsson,
2012; Hellmann et al., 2010; Nobes, 2013; Sunder, 2011). For
instance, ‘unless the risk is not material’ has been replaced with
€ r va
‘såvida risken inte a €sentlig,’ which means, unless the risk is ma-
terial in the Swedish translation (Dahlgren & Nilsson, 2012, p. 49),
and need not has been translated into German as, weder noch, which
is a prohibition (Hellmann et al., 2010). These findings echo the
‘anecdotal stories of how Turkish or Japanese translations deviate
from the intent of the original,’ which, according to Sunder (2011,
pp. 301e302) are ‘difficult to evaluate in the absence of unanimity
behind bilingual authoritative voices.’ Furthermore, Nobes (2006)
discussed the Norwegian translation of International Accounting
Standard (IAS) 41, para. 34, which requires an unconditional gov-
ernment grant related to a biological asset to be recognised as in-
come when the grant becomes receivable. Receivable is expressed
as mottas in the Norwegian version, meaning received, an inter-
pretation that might postpone recognition of the given income and,
thus, not match the intention of the original standard. These studies
have argued that inaccurate translations can compromise the
comparability of the IFRS financial statements.
2.1.2. Translation work in the context of transnational accounting
regulation
Existing literature has not addressed the translation practices or
work involved in attempting to produce acceptable linguistic
equivalence in translated accounting and auditing standards in
detail. In relation to experimental studies on translation and those
examining inaccuracies or inconsistencies in translated regulations,
the current article's contribution is to focus on the translation work
and procedures through which transnational accounting regula-
tions are translated. As Toury, a translation theorist, noted, ‘there is
no real point in the product-oriented study [of translated texts]
without taking into account questions pertaining … to the strate-
gies governed by the norms of establishing a “proper” product’
(Toury, 1995, p. 13). Accordingly, the current study analyses how
translators and translation reviewers seek to render the meaning
by means of the target-language and make English-language ac-
counting concepts transferable to other languages. Thus, it does not
focus on (the inaccuracies or inconsistencies in) the products of
translation, and instead, steps back and provides detailed insights
into the institutional work required to reproduce the IFRS in other
languages.5
Studying the translation of the IFRS as social and institutional
practice contributes to the wider literature on accounting regula-
tion because translation lies at the interface between transnational
standard-setting and local implementation. If we wish to under-
stand how transnational accounting standards are created and then
translated into accounting practice in a multitude of settings, it is
important to map the whole chain of events from standard-setting
to interpretation and then to implementation of the standards.
While recent research attention has focused on the role of
globalisation and transnational institutions in accounting regula-
tion (e.g., Botzem, 2010; Crawford, Ferguson, Helliar, & Power,
2014; Humphrey et al., 2009; Samsonova-Taddei & Humphrey,
2015; Suddaby et al., 2007), the interactions and activities of the
dispersed network of actors that supports transnational regulation
agendas are still not well understood (Gillis et al., 2014; Suddaby
et al., 2007; however, see also; Barrett, Cooper, & Jamal, 2005;
Mennicken, 2008). The empirical focus of this article, the trans-
lation of IFRS, is one node of such institutional interactions within
the transnational regulatory arena. The International Accounting
Standards Board (IASB) -approved translations are co-produced by
representatives of the IFRS Foundation's staff, professional trans-
lators, accounting firms and other constituents approved by the
5
Some accounting scholars have briefly acknowledged that the IASC/IFRS
Foundation has developed and put in place a particular translation process for IFRS,
including a review by accounting experts (Baskerville & Evans, 2011; Dahlgren &
Nilsson, 2012; Evans et al., 2015; Huerta et al., 2013). Evans et al. (2015) referred
to a similar translation process for the International Standards of Auditing (ISA).
 J. Kettunen / Accounting, Organizations and Society 56 (2017) 38e54 41

standard-setter.6 The translation review committees are a point at rather than a formal correspondence is based upon “the principle of
which transnational standardisation, knowledge of local practices, equivalent effect” [Rieu & Phillips, 1954]’ (Nida, 1964, p. 159).
and English and target-language concept systems come together Formal correspondence and dynamic equivalence represent
(cf., Barrett et al., 2005; Mennicken, 2008). Examining how IFRS opposing approaches to what should be given priority in a trans-
translations are created offers further insight into the complex in- lation when perfect correspondence in form and meaning is absent;
teractions and practices that support transnational regulation. however, between these two approaches lie various gradations of
The following subsection gives a brief overview of how trans- acceptable translation, and preferences as to priority tend to vary
lation has been conceptualised in linguistics and translation over time and between professional fields (Nida, 1964; see also;
studies. As these conceptualisations highlight, translation usually Evans et al., 2015).
brings with it at least subtle changes in meaning and/or alters the To clarify, the studies on the translation of TFV presented in the
textual features of the source text. previous section can also be looked at through the lens of formal
correspondence versus dynamic equivalence. The studies implicitly
address the question of whether there is a formal correspondence
2.2. Translation and equivalence
between the term true and fair view and its translation. For instance,
in cases where the individual words true and fair are translated
Translation is a broad concept with multiple meanings. In the
with a single term into the target-language, a formal correspon-
academic fields of linguistics and translation studies, the notion of
dence is absent. Yet, all but one of these studies remain silent about
translation encompasses language and mainly refers to what
whether the target-language concept is interpreted and applied
Jakobson (1959/2000) termed ‘interlingual translation’ meaning
similarly to the notion of TFV in the source-language (cf. Kosmala-
translation from one language to another.7 More specifically, the
MacLullich, 2005). In other words, the dynamic equivalence be-
concept can be used in the sense of translation being (1) a process
tween TFV and its translations is not addressed.
(the act of translating) or (2) a product (translated text).
A translation oriented towards dynamic equivalence exemplifies
The academic discipline of translation studies employs several
a user-oriented approach to translation. Such a user-oriented or
approaches. Before the 1980s, research on translation (e.g., Catford,
functional approach to translation (Reiss, 1977/1989; Reiss &
1978; Jakobson, 1959/2000; Nida, 1964) mainly drew on linguistics
Vermeer, 1986; Nord, 1997) considers the function of the target
to provide its theoretical background, but since then translation
text as the most important criterion informing the translator's
studies as a field has developed, and now research into translation
decision. Within the functional approach, Reiss (1977/1989, pp.
is conducted from a variety of angles, if not multiple disciplines
113e114 in Munday, 2012) argued that the text, rather than a word
(Hatim & Munday, 2004). Rejecting a linguistic approach to trans-
or sentence, is the level at which equivalence is to be sought. Rather
lation, authors like Bassnett and Lefevere (1990) and Venuti (1998)
than fixating upon the source text, Reiss and Vermeer’s (1986)
encouraged researchers to shift the emphasis from the text to the
Skopos theory emphasises the target text and its purpose (i.e.,
‘broader issues of context, history and convention’ (Bassnett 1998,
skopos) as well as contextual factors. According to Evans et al.
p. 123), which they characterised as a cultural turn (Bassnett &
(2015) user-oriented or functionalist approaches to translation
Lefevere, 1990; Lefevere & Bassnett, 1998).
have gained ground within professional domains in recent decades,
Translation scholars have used the notion of equivalence to
albeit not in accounting.
describe the different forms of correspondence between the source
Overall, translation studies portray translation as a balancing act
text (or a part of it) and the target text (or a part of it) (e.g., Catford,
involving choices. Collectively, the different conceptualisations
1978; House, 2006; Jakobson, 1959/2000; Nida, 1964; Nida & Taber,
such as formal correspondence versus dynamic equivalencedor
2003). The equivalence problematic relates to the longstanding
Skopos theoryddirect attention to the tension between rendering
debate on the balance between form and content in translation;
the form or the message of the source text to the target-language.
translation is a balancing act. Equivalence is, however, a relative
They imply that translation introduces subtle changes in mean-
concept, and various types of equivalence have been con-
ing, and that it is relatively rare for there to be only one theoreti-
ceptualised. Within the linguistic paradigm of translation studies,
cally correct translation. Translation is instead directed by the
Nida (1964) distinguished between formal correspondence (or
preferences of the translator(s), or those who are in the position to
formal equivalence) and dynamic equivalence (see also, Nida &
set the principles that guide the translators’ work. Hence, like ac-
Taber, 2003) and that distinction remains influential (Hatim &
counting, translation is not a merely technical problem, and
Munday, 2004). According to Nida and Taber (2003, p. 201),
accordingly, translators and translation reviewers are in a position
formal correspondence is ‘a quality of translation in which the
of both responsibility and power (Evans et al., 2015). Therefore, the
features of the form of source text have been mechanically repro-
activities required to produce translations of IFRS, and other
duced in the receptor language’ (Nida & Taber, 2003, p. 201). In
transnational regulations, merit more consideration in research.
contrast, dynamic equivalence is a ‘quality of translation in which
the message of the original text has been so transported into the
2.3. Professional texts as outcomes of collaborative activity
receptor language that the response of the receptor is essentially
like that of the original receptors’ (Nida & Taber, 2003, p. 200). A
From a linguistic standpoint, accounting language is language for
‘translation which attempts to produce a dynamic equivalence
specific purposes (LSP) (Evans, 2010; Evans et al., 2015). The research
on LSP has described professional texts as ‘the result of collabora-
6
tive activities on the part of a number of professionals, who have
http://www.ifrs.org/Use-around-the-world/IFRS-translations/Pages/Role-of-
specific roles to play’ in text construction activities (Bhatia, 2004, p.
the-IFRS-Foundation-in-the-translation-process.aspx.
7
Accounting research (e.g., Chua & Taylor, 2008; Mennicken, 2008) has often 219). Indeed, both the original IFRS texts and their translations are
drawn on the concept of translation as it exists in the sociology of translation which outcomes of detailed processes involving collaborative activities,
originates in science and technology studies (e.g,. Callon, 1986; Latour, 1987, 1994). which has a bearing on the characteristics of the final texts.
In the Latourian sense (1994, p. 32), translation refers to the ‘displacement, drift, The inherent ambiguity of language and the difficulty to deter-
invention, mediation, creation of a new link that did not exist before and modifies
in part the two agents’. In this paper, the concept of translation is not used in the
mine meanings precisely pose challenges to translators, who are
Latourian but in a more traditional sense referring to translation from one language expected to retain the scope for interpretation unchanged in
into another. translation. Translation of an LSP may be particularly problematic,
42 J. Kettunen / Accounting, Organizations and Society 56 (2017) 38e54
partly because the translators need to distinguish meanings asso-
ciated with professional terminology from meanings in everyday
language. For instance, the terms conservatism or reserve may not
translate to the target-language in the accounting context in the
same way as they would in everyday language; nor do all ac-
counting terms refer unambiguously to well-specified referents,
and neither do accounting concepts overlap neatly in different
languages (e.g., Evans, 2004). Instead, their meanings are often
indeterminate, dependent on the context, and those meanings vary
between languages. The conceptual underpinnings of accounting
rely on the knowledge bases of other disciplines and traditions of
thought which adds another dimension to the general ambiguity of
language (Zambon & Zan, 2000). The IFRS are principles-based
accounting standards and, as such, are characterised by some
deliberate ambiguity to allow room for interpretation and appli-
cation of the standards in various business environments (Evans
et al., 2015).
Given that the texts produced in professional contexts are often
co-authored, they have a distinctly rich intertextual patterning
(Bhatia, 2004). I draw upon the notion of intertextuality in the
analysis of the translation of IFRS to emphasise the relational nature
of the textual meaning (see Bhatia, 2004; Kristeva, 1980) and to
describe the practices related to interpreting the source text and
searching for target-language terms. Intertextuality ‘refers to a
number of relationships that the text in question may have with
those which in some way have been used, referred to or exploited
either directly or indirectly in the construction of the text in
question’ (Bhatia 2004, p. 126). The analytical perspective of in-
tertextuality implies that the meaning is not seen to only reside in
the text in question. Instead, it is ‘seen as emerging from the re-
lations texts have with other texts’ (Solin, 2004, p. 267).
2.4. The translation of the IFRS as institutional work
The current research suggests that the notion of institutional
work offers a useful lens through which to examine the activities
required to produce and maintain multiple language versions of the
IFRS. The notion of institutional work makes it possible to focus on
the practical actions through which acceptable linguistic equiva-
lence, and accordingly, the IFRS as a global, translingual institution
is (re)produced in translation. Without the translation work, the
IFRS would remain inaccessible to accountants who do not read
English. The institutional work perspective is adopted in this paper
to draw attention to the constructed nature of the equivalence
between languages suggesting that translation ties together
threads of previously distinct concept systems, and to highlight the
collective efforts of local actors in the maintenance of the IFRS.
The study of institutional work investigates the interaction be-
tween actors and institutions (Lawrence & Suddaby, 2006;
Lawrence, Suddaby, & Leca, 2009). The practical actions through
which institutions are created, maintained and disrupted are cen-
tral to the research on institutional work (Lawrence et al., 2009).8
The perspective pays close attention to ‘the small worlds of insti-
tutional resistance and maintenance in which institutionalisation
and institutional change are enacted in the everyday getting by of
individuals and groups,’ acknowledging that a range of actors is
necessary to facilitate institutionalisation (Lawrence & Suddaby,
2006; Lawrence, Suddaby, & Leca, 2011, p. 57). The institutional
work perspective offers a way to address the activities undertaken
8
A few recent accounting studies have utilised the notion of institutional work to
investigate professional domain change (Suddaby, Saxton, & Gunz, 2015) and the
diffusion of management innovations (Chiwamit, Modell, & Yang, 2014; Hayne &
Free, 2014).
to create acceptable translations of newly issued standards and to
maintain the linguistic equivalence constructed in earlier trans-
lations. Significantly, the notion of institutional work also encom-
passes institutional maintenance, as a ‘supporting, repairing, and
recreating’ institutions (Lawrence & Suddaby, 2006, p. 230),
allowing the translation of IFRS to be examined as an ongoing and
recurrent activity rather than one-off event. As new standards are
issued and existing ones amended, the IFRS as a translingual
institution requires constant maintenance.
Following the practice tradition, Lawrence and Suddaby (2006)
view institutional work as intelligent, situated institutional action.
A practice orientation has been compared to process-oriented
studies: Studies of practice focus on ‘the internal life of the process’
(Brown & Duguid, 2000, p. 95 in; Lawrence & Suddaby, 2006, p. 218).
A ‘practice perspective highlights the creative and knowledgeable
work of actors which may or may not achieve its desired ends and
which interacts with existing social and technological structures’
(Lawrence & Suddaby, 2006, p. 219). Accordingly, the current study
examines the practical actions undertaken by translators and trans-
lation reviewers to render the IFRS into another language while
adhering to translation policies set by the standard-setters.
2.5. Summary
Research on the translation of accounting regulations has been
conducted in relation to (i) the problems of translating expressions
of uncertainty in accounting and auditing standards, (ii) the in-
compatibility of accounting concepts in different languages and (iii)
inaccuracies in some IFRS translations. In turn, linguistics and
translation studies have suggested that perfect equivalence be-
tween languages does not exist. Translators therefore need to bal-
ance the different linguistic characteristics of the target text, and
the ways in which they correspond to the source text. Furthermore,
translation unavoidably involves interpretation of the source text
and tends to lead to at least subtle changes in meaning. As trans-
lation is not merely a technical exercise, it warrants research within
the complex arena of transnational regulation. This paper attempts
to address the institutional work employed to produce language
translations of the IFRS. The following section introduces the
translation and language policies of the IFRS Foundation and the
European Commission (EC) concerning IFRS.
3. Introduction to the case: situating the translation of IFRS
in the regulatory context
The working language of the IASB is English, and the standard-
setter has stated that the approved version of any discussion
document or IFRS should be in English.9 The IASB also acknowl-
edges that the high-quality translation of the IFRS into other lan-
guages is imperative for the international distribution and use of its
standards.10 The IFRS Foundation has its own translation, adoption
and copyright policies, which includes a specific translation and
review process involving accounting experts.11 The IFRS Foundation
argues that in order to ‘ensure that IFRS remain uniform across all
languages and that translations are of the highest standards, the
IFRS Foundation maintains ultimate control over the translation
9
http://www.ifrs.org/Use-around-the-world/IFRS-translations/Pages/IFRS-
Translation-Review-Committees.aspx.
10
IASB, 2001. Changes and challenges. IASB Insight, Q1/Q2: 1 http://www.ifrs.org/
Archive/INSIGHT-journal/Q1-and-Q2-2008/Documents/INSIGHT_Q1Q208_lowres.
pdf.
11
http://www.ifrs.org/Use-around-the-world/IFRS-translations/Pages/Official-
translation-process-and-policies.aspx.
 J. Kettunen / Accounting, Organizations and Society 56 (2017) 38e54
process for IFRS.’12
However, within the EU, the Directorate General for Translation
(DGT) for the EC is in charge of the translation of IFRS because the
endorsed standards are enacted as EC regulations. The EC regula-
tions, including the IFRS as adopted by the EU, are equally legally
binding in all official EU languages (EC, 1958, 2002; see also; Nobes,
2013). In other words, the DGT currently provides the legally
binding translations of the endorsed IFRS.
The main difference between the translation policy of the IFRS
Foundation and that of the DGT for the EC is that while the IFRS
Foundation's translation policy requires a ‘review by a committee of
accounting experts who are native speakers with proven knowledge
and expertise in the area of the IFRSs’,13 the translation policies of the
DGT do not require any review by accounting experts.14 Conse-
quently, not all legally binding EU language translations are reviewed
by accounting experts and/or approved by the IFRS Foundation.
Fig. 1 depicts the steps for official translation of the IFRS as set
out by the IFRS Foundation. This study sheds light on the institu-
tional work employed in the construction of acceptable linguistic
equivalence in translation by providing insights into what happens
within and between the steps circled in the figure.
Each translation review committee (TRC) has a translator and a
coordinator,16 and TRCs are composed of representatives of the
large and mid-tier accounting firms, financial statement preparers,
academics, and specialists from industries such as the banking and
insurance sectors, all of whom are native speakers of the target-
language. Translations are further controlled by the IFRS Founda-
tion's translation coordinator who is based in London. The users of
financial statements are not, however, represented on the TRCs
which reflects the general tendency of the users not being actively
included in standard-setting (Durocher & Gendron, 2011;
Hopwood, 1994; Young, 2006). The IFRS Foundation has set the
objectives for the review process:
The purpose of the review process is to ensure the accuracy of
the translated text compared with the English original. The
purpose of the translation is not to interpret or explain the
standards, but merely to render the meaning of the English text
in another language. Consequently, Review Committee mem-
bers may not add, reduce, or alter in any way the substance and
the content of the standards and interpretations as approved by
the International Accounting Standards Board, although gram-
matical and syntax adaptations to improve the readability of the
text in the language in question are acceptable. (Terms of
Reference, IFRS Foundation).
This excerpt from the Terms of Reference (for TRCs containing
instructions set by the IFRS Foundation on composition of the TRCs,
workflow and principles of translation: unpublished) sets out
multiple goals: the concern with ensuring the accuracy of the
target-language text compared with the English original and a
prohibition against interpreting or explaining the text. With regard
to the literature on translation introduced in the previous section,
the goal of ensuring accuracy compared with the English source
text seems an ideal lacking a clear path to its realisation.
12
http://www.ifrs.org/Use-around-the-world/IFRS-translations/Pages/IFRS-
translations.aspx.
13
See http://www.ifrs.org/Use-around-the-world/IFRS-translations/Pages/
Official-translation-process-and-policies.aspx.
14
Verbal communication with the DGT translator.
15
Adapted from http://www.ifrs.org/Use-around-the-world/IFRS-translations/
Pages/Official-translation-process-and-policies.aspx.
16
The Finnish TRC differs slightly from that of other countries in terms of the
distribution of its work; The Finnish translator also coordinates the TRC.
43
Translation of the IFRS into Finnish started in 2002. The Finnish
TRC originally consisted of a translator and three reviewers. Today, it
comprises a translator and 19 reviewers including one representative
from the Accounting Regulatory Committee, one from the Ministry
of Employment and the Economy, and an observer from the DGT. The
Finnish translations, which the TRC produces in cooperation with the
DGT for publication in the Official Journal of the European Union, are
approved by both the EU and the IFRS Foundation.
4. The research setting, data and methods
A purposive sampling strategy was employed to select the case
(Miles & Huberman, 1994). In order to examine translation of the
transnational regulations in the realm of accounting, the IFRS were
selected for the following reasons: first, such translations are as
legally binding as the original English-language standards; second,
in the vast majority of countries where financial statements are
prepared in accordance with the IFRS, those preparing the financial
statements are not native speakers of English. Therefore, many of
them are likely to rely on the translated IFRS.
The analysis of the practical problems of linguistic equivalence
investigates how translation is conducted in practice. This article
explores and examines the case of the Finnish translation of IFRS as
an illustration (Siggelkow, 2007) of the translation of transnational
regulations. As indicated above, the activities and interactions
related to translation have received scant attention in the con-
textualist research even though translation has a unique position at
the interface between transnational standard-setting and local
implementation. Because it is the native language of the researcher
conducting the interviews, the Finnish language was selected to
avoid additional translation issues in the exchanges between the
researcher and the interviewees (Welch & Marchan-Piekkari, 2006;
see also,; Sunder, 2011). It also allowed the researcher to examine
the archival records in the target-language.
The Finnish language belongs to the Finno-Ugric language group
within the Uralic language family, in contrast to English and, for
example, French, German and Swedish, which are of the Indo-
European language family. Finnish vocabulary and syntax differ
substantially from those of English. For example, the Finnish lan-
guage does not differentiate between him and her, and there are
neither prepositions nor articles in Finnish. This can be illustrated
by a simple example: ‘in the house’ translates literally to Finnish
with one word, talossa, and the ending of the word indicates a
meaning that is similar to an English preposition. The specifying
meaning, which the definite article ‘the’ carries, is lost in trans-
lation. In addition to the languages having different roots, the
translation of IFRS between the language pair English-Finnish is
further complicated by the fact that the respective legal systems
have little in common. Accordingly, translation issues are to be
anticipated (Baskerville & Evans, 2011; Evans et al., 2015).
From the institutional viewpoint, translation into Finnish is
representative of translation within the EU in that all EU language
versions of the endorsed IFRS are equally legally binding. The
Finnish case is particularly interesting as the translations are also
created in accordance with the official translation process set out by
the IASB.17 In other words, the same Finnish translation is approved
by both the IASB and the EU. Examples of such IASB-approved
translations also exist for German, Japanese, and Russian trans-
lations but not those in Estonian, Italian or Swedish.18 An enquiry
17
http://www.ifrs.org/Use-around-the-world/Documents/Jurisdiction-profiles/
Finland-IFRS-Profile.pdf.
18
http://www.ifrs.org/Use-around-the-world/IFRS-translations/Pages/Available-
translations.aspx.
44 J. Kettunen / Accounting, Organizations and Society 56 (2017) 38e54
into translation into Finnish thus allows us to examine the creation
of the IASB-approved translations.
In order to understand the translation work and the interactions
between the translators and the accounting professionals review-
ing the translations, semi-structured interviews were carried out
with eleven Finnish TRC members including the TRC translator, and
one professional translator working for the DGT in Brussels. The
empirical part of the research project commenced in June 2010
with a meeting with the TRC translator in Finland. The data analysis
was interwoven with data collection to guide later interviews and
to select archival data for closer analysis (Miles & Huberman, 1994).
The initial three-hour interview with the key informant helped in
phrasing questions for the subsequent interviews. The TRC trans-
lator also provided documentary material including Terms of
Reference for TRCs [including instructions set by the IFRS Founda-
tion about the composition of the TRC, workflow and principles of
translation (unpublished)] and minutes from meetings where de-
cisions on terminology had been made.
The preliminary idea for the research was to explore the types of
linguistic problems inherent in translating accounting regulations,
given the references to such problems in the prior literature pub-
lished around the time of the transition to the IFRS in the EU (e.g.,
Doupnik & Richter, 2003; Evans, 2004; Nobes, 2006; Zeff, 2007). It
was soon discovered, however, that not only are the challenges of
translation are extensive, but that the translation processes are far
more complex and are more heavily influenced by the different
regulators than the prior academic literature and publicly available
materials had indicated. Specifically, the recurrent shifts between
the EU and IFRS Foundation in the governance of translation have
resulted in recurring changes in the translation practices and in the
constituents who participate in translation.19 Therefore, I decided
to broaden the investigation to include the work required to
address the practical problems of equivalence between languages.
The interviews yielded more than 16 h of material, of which 12
could be directly transcribed from audiotape. While the TRC
members were invited to participate in the interviews for this study
based on their roles on the committee, it is important to note that
the work of translation is not their principal duty as accounting
professionals. All the interviewees apart from the DGT translator
were senior accounting professionals, and the group included
advisory directors in accounting firms as well as senior accounting
experts from financial supervisory authorities (see also IFRS
Foundation, 2012, p. 1 for the composition of TRCs).
The data from the interviews were supplemented with a diverse
range of documents including excerpts from the translated stan-
dards. The archive data proved central to understanding the
translation work and conceptualising the more general problems of
translation through specific examples. It also helped to address the
recurrent methodological concern associated with the bias that
may arise from retrospective recall of the decision-making pro-
cesses that underpin the translation process (Toury, 1995).
Specifically, the archival data used for this study consist of nine
different types of written materials: (1) the guidelines for the
composition and workings of the translations review committee
issued by IFRS Foundation (Terms of Reference, unpublished), (2) a
list of new terms and their context plus the translator's suggestions
for alternative translations of the term, (3) records of TRC meetings,
most including terminology decisions, (4) records of TRC subgroup
meetings in which amendments to draft translations were deter-
mined, (5) draft translations, and TRC members' suggested
amendments to them, (6) translated IFRS, (7) the forewords to the
19
See http://www.europarl.europa.eu/sides/getDoc.do?pubRef¼-//EP//
TEXTþWQþP-2011-008747þ0þDOCþXMLþV0//EN.
actual standards, (8) printed email correspondence between
translators and TRC members and (9) training materials for EU
translators relating to common mistakes arising in the prior
translations of the IFRS.
In the first stage of analysis, I searched and examined publicly
available data on the translation policies of the EU and the IFRS
Foundation (e.g., the process chart of the official translation model
of the IFRS Foundation, Fig. 1), which form the background of the
construction of the translations (Bhatia, 2004).
In the second stage, I made notes on the translation problems
raised during the interviews and on the ways in which the TRC
members sought solutions to these problems. In addition, I noted
the interviewees' criticisms of translations that they had (not)
selected. Although the interviews were transcribed verbatim, I
listened to the recordings several times because transcriptions do
not capture non-verbal information (Bucholtz, 2007). By listening
to the recordings I was better able to infer the interviewees’ opin-
ions, such as whether they seemed to consider that a particular
translation problem was significant. Then, similar translation
problems identified by the interviewees were merged into broader
categories such as terminology, ambiguity and indeterminacy of
source text, and complexity and foreignness of the substance.
The third stage of analysis concentrated on seeking to identify
particular means adopted by the TRC members to interpret the
source text and how they participated in the creation of the
translation. In addition, I noted what the accounting experts said
about their role as reviewer and what they sought to accomplish
through the work of translation. Drawing on the notion of in-
tertextuality, I also paid attention to the ways in which other texts
were used to aid comprehension of the source text and to construct
the target text (Bhatia, 2004). I identified reoccurring phases in the
translations, and generated visualisations of the workflow of the
translation review committee. The visualisations and narrative
descriptions were then compared with the IFRS Foundation's offi-
cial process description (see Fig. 1) and other publicly available
material identified in the first stage.
In the fourth stage the findings from the interviews were
compared with documents copied from the archives of the TRC. The
archive material was especially helpful clarifying the practical
problems of constructing linguistic equivalence described by the
interviewees, and for analysing the suggestions and criticism
expressed during the terminology work. Accordingly, selected ex-
cerpts from the Finnish translations were compared with the
source text in order to infer a more detailed understanding of the
issues described by the interviewees. For example, some in-
terviewees indicated that the TRC had held lengthy discussions on
how to translate problematic terms, such as asset, measurement,
dealer market, broker market, observable and domestic partner.
Therefore, I compared the contexts in which these terms are used in
the source-language against the target-language standards, and the
ways in which the terms link to concept systems in the respective
languages so as to analyse why rendering these concepts
adequately in the target-language was problematic.
Furthermore, the archival materials were not only used to
illustrate but also to contrast the findings from the interviews.
Because the interviews were conducted over a longer period of
time, stages two, three and four were repeated to refine the analysis
each time that additional data was collected.
5. Constructing acceptable linguistic equivalence: the case of
translating the IFRS into Finnish
Maintaining the influence of the IFRS as a global institution
requires sustained institutional work, some of which takes place in
local settings and through micro-level activities. The institutional
 J. Kettunen / Accounting, Organizations and Society 56 (2017) 38e54
work of translation contributes to the maintenance of the IFRS as a
translingual institution and is essential if the standards are to have
global reach. The translation of the IFRS into the Finnish language,
which is the empirical focus of this section, connects the financial
reporting in Finland to the IFRS. It also enables the local accounting
profession to be part of ‘the new world’ of the IFRS, as an inter-
viewee put it. Without translation work, Finnish financial reporting
and the IFRS would have a considerably weaker connection. This
section also elaborates on the complexities of translation, illus-
trating that translation is more than a technical activity in the local
implementation of the global standards.
The section consists of three subsections. The first subsection fo-
cuses on the translation of terminology and shows that one-to-one
equivalence between a given English-language term and the corre-
sponding target-language term e as required by the IFRS Foundation's
translation approach e is often constructed during translation as
opposed to merely being a given. The efforts involved in the main-
tenance of one-to-one equivalence created in previous translations
and those made to select target-language terms capable of being
embedded in other texts are also examined in the first subsection.
The second subsection analyses how translators and the re-
viewers of translations deal with inherent ambiguity and uncertainty
as they seek to convey the meaning of the source text in the target-
language. It is argued that translation inevitably involves a degree of
interpretation of the source text, and that the TRC members infer its
meaning by drawing on their respective bodies of knowledge, by
referring to other texts invoked in reading the standard in question,
and through discussions between committee members.
The final subsection illustrates that the translation of the IFRS is
a contested area of expertise. It does so by investigating the
differing translation policies of the EU and the IFRS Foundation, and
analysing the accounting experts' and translators’ views on the role
of the expert review in the translation work. It looks into the
institutional dynamics surrounding the translation of the IFRS by
tracking the recurrent shifts in the organisation of the translation of
the IFRS in the EU.
5.1. Translation work on terminology
This subsection examines the activities undertaken by the TRC
to translate the IFRS concepts into the target-language, and the
practical problems of conceptual equivalence encountered therein.
The study will show that an existing word-for-word equivalence
between languages is not always available. In the absence of that
option, the formal correspondence of terminology is constructed
through committee activities that follow a process governed by the
standard-setter. Terminology work plays a central role in the IASB
translation approach: For each standard a list of key terms extrac-
ted by the IFRS Foundation is translated on a one-to-one basis, and
the translated terms are agreed upon in committee meetings.
In practice, more than one target-language term/word was often
suggested as an appropriate translation of a source-language term/
word: a situation typical of the absence of conceptual equivalence
Fig. 1. The IFRS Foundation translation process.15
45
between the source and target-language as described by Evans
(2004). The lack of conceptual equivalence between source and
target-language also meant that the TRC coined new terms to
describe source-language concepts, and a few terms were left un-
translated. Moreover, as languages are dynamic, the terms used in
the source text are occasionally changed even though the under-
lying concept does not change or vice versa, and in these cases it
must be decided whether to, and if so how to mirror these changes
in translation. These issues will be expanded upon below.
5.1.1. Addressing a lack of equivalence between the source-language
and target-language terminology.
Although the IFRS Foundation's translation policies generally
require that each source-language term is always translated using
the same target-language term across different standards, such
formal correspondence does not occur naturally and must be con-
structed during the translation process. The requirement for term-
for-term translation seems to be based on the assumption that the
two terms in these different languages would signify the same
referent in a text-external reality. However, several accounting
terms such as ‘income’ or ‘capital’ do not have an explicit referent
(Macintosh, Shearer, Thornton, & Welker, 2000), and in newly
developed accounting standards ‘things are fitted into the old cat-
egories, the categories are stretched and perhaps twisted and
themselves altered’ (Young, 2003, p. 621). Moreover, the underlying
concepts and their interconnections differ between languages. To
illustrate, the Finnish language does not have a term equivalent to
an asset in the singular.
Given that the IFRS include concepts previously absent from
local regulations, or the local accounting tradition more generally,
translators and translation reviewers seek out potential trans-
lations from a variety of other texts. A noteworthy example of such
intertextual practices (Bhatia, 2004) involved in translating termi-
nology is the search for a target-language equivalent to compre-
hensive income. The concept of comprehensive income was
unfamiliar to Finnish accounting thought, and the local regulations
did not contain any concept equivalent to it. Therefore, the trans-
lator specifically asked for the Finnish terms used to describe the
concept in the organisations where the TRC members work. One of
the TRC members noticed that cross listed companies filing with
the US Securities and Exchange Commission (which are few in
number) used the term laaja tulos in their Finnish language
financial statements. The term laaja tulos was later chosen by the
TRC from the four alternatives suggested by the translator as the
one most closely equating to comprehensive income.
Certain interviewees familiar with the US GAAP pointed out that
several new terms and expressions in the IFRS derive from the
standards set by the Financial Accounting Standards Board (FASB),
and they therefore considered those standards a relevant reference
point when inferring the meanings of the expressions newly
introduced into the IFRS. It seems likely that the intertextual ties
between the IFRS source text, and the US GAAP have also been
strengthened as a result of projects conducted jointly by the IASB
46 J. Kettunen / Accounting, Organizations and Society 56 (2017) 38e54
and the FASB (see e.g., Baudot, 2014; Pelger, 2016; cf. Bhatia, 2004).
When translating terms, the TRC engages in defining boundaries
of concepts and the extent to which the source-language and the
target-language concept overlap. To ensure the appropriateness of
target-language terms, the contexts in which the terms are used in
the respective languages were examined and compared. Doing so
involved drawing on various texts directly or indirectly depending
on the nature of the vocabulary used in the standard in question (cf.
Bhatia, 2004; Solin, 2004). Some of these surrounding texts were
referred to when reading the source text; others became reference
material through the use of a search engine, as some translation
reviewers described:
Interviewee D: The translator has always done some prepara-
tory work now and then by Googling to find out what contexts
the word appears in. So it's been a fairly typical procedure in
recent times, to use Google to find examples of the contexts the
word is used in. If somehow you're in the situation that some-
body suggests translating a term in a certain way, to see if it's
appropriate to use that term. So you must check whether the
word is used in any other contexts. To see if [the Finnish term]
will be misleading …
Interviewee E: The translator does most of the work when he or
she provides these suggestions [for Finnish terms] to be used …
Then you can read a text written by a professional in the field or
go to Google to find out on the internet where a certain kind [of
expression] is used. So you just try one way or another to find
out how the word is used elsewhere. But our translator goes
amazingly thoroughly through all the Finnish dictionaries and
these language questions to find out what sort of situation you
can use some [word] in … You have to try out how other people
understand it. So I've certainly quite widely [found out what my
colleagues think]. For example, in finance matters when there's
been [an issue finding Finnish equivalents to terms] … then I've
asked our finance experts who know better what sort of lan-
guage is used and what exactly something means. I've been
talking to them about the terms that were to be voted on.
In practice, the translator offers suggestions on the translated
terms, where applicable with information on their contexts in other
IFRS, or more broadly in the source and/or target-language.
Although the interviewee quoted above describes the search for
equivalent terms, or consulting colleagues, some interviewees said
that they did not get very involved in questions of terminology
prior to the committee meetings where the terms are discussed and
eventually voted upon. The interviewees did stress the importance
of these meetings in decision making on terminology. Having heard
other TRC members’ viewpoints, and their reasoning for selecting
or rejecting particular terms, the interviewees considered them-
selves able to form a more informed opinion, and to contribute to
the process of selecting appropriate target-language terms.
A particular difficulty noted by several interviewees e both
translators and accounting experts e was that not only are some of
the IFRS concepts foreign (a problem already described by Zeff,
2007) but the terms refer to things such as company forms or
financial instruments which are not part of routine business ac-
tivities in the jurisdiction of the target-language. Thus, problematic
absences encompass not only accounting terms but also the objects
whose accounting treatment is governed by the standards. Several
TRC members noted that in a small country like Finland, its capital
market has not been as subject to international influence as in other
territories, and has accordingly not developed to the same extent as
in the territories where the IFRS were developed. More than one
interviewee stated that the object or type of transaction under
scrutiny may not exist in the business environment of Finland:
Interviewee F: Sometimes the phenomenon did not yet exist
here; or it did but it was not talked about. Not at least in the
same way as the IFRS conceptualises it. […] And then there are
these English equivalents that weren't found. In the world of
financial instruments, if we have an option structure where
there is a cap, a floor, and a collar … well, then nobody would
translate them [literally] as katto (‘cap, ceiling, roof’) and lattia
(‘floor’) and kaulus (‘collar’) … Because it is so well-established
in the finance world, in practice, they are referred to using the
English terms, and it would feel like a naïve approach to
translate them […] So I think that not all the financial instru-
ment vocabulary can be translated.
Interviewee B: And then for example this ‘recoverable amount,’
which is not plain Finnish, so how do you say it to an ordinary
accountant who's preparing consolidated financial statements?
[…] The recoverable amount is not normal Finnish language,
which you would use when talking with your colleague […] So
in a way we invented new Finnish language and coined new
economic and accounting terminology […] So in that sense
there are terminology problems in that there are no words … or
there may be a word equivalent to ‘recoverable’ and a word
[equivalent] to ‘amount’ but when you combine the two words,
they don't make [a term] … no such [equivalent] word will be
found.
The above interview excerpts illustrate that one-to-one trans-
lation of terms is not without its limitations. The TRC members
considered that not every source-language concept could be
translated into comprehensible target-language terms. Therefore,
the above-mentioned source-language terms cap, floor and collar
are presented in parentheses in the Finnish language IFRS. Simi-
larly, although TRC members searched diligently for terms in a
variety of other texts, and contacted colleagues for advice in their
organisations, terms such as in-substance defeasance, sinking fund
and wash sale resisted translation and consequently were left
untranslated.
A few interviewees were critical of the comprehensibility of
certain terms that the TRC had opted to translate. An accounting
expert asserted that she would not be able to attach any meaning to
certain translated terms. ‘If someone talked to me about a miinu-
soptio [Finnish for out-of-money option] I wouldn't know what the
word means'. This provides further evidence to substantiate Zeff's
views (2007, p. 296) that ‘even if it is translated as accurately as
practicable into the language of the second country, the concept
may not be understood. The words may be understood, but the
concept may not be understood’.
Beyond the terminological issues examined above, the conno-
tations of professional terms may differ between languages. Young
(2003) argued with reference to the standards set by the FASB that
the rhetorical approaches evident in the standards employed by the
standard-setter serve to maintain the myth of accounting objec-
tivity. Accounting terms such as measurement connote, I would
suggest, the ideal of objectivity in accounting. When translating
terms into Finnish, certain TRC members rejected transferring some
of these unfamiliar connotations to the target-language. For
instance, a translation reviewer pointed out that measurement in
 J. Kettunen / Accounting, Organizations and Society 56 (2017) 38e54
accounting does not equate to objective measurement in other
contexts, and the Finnish language equivalent of the term mea-
surement does not traditionally denote valuation in the context of
accounting. The translation reviewer pondered the possibility of
translating measurement with the Finnish equivalent of valuation,
explaining that he had checked the term choice against the
German, French, Spanish, and Swedish translations of IAS 16
(although the translator had warned the reviewer against drawing
parallels between languages):
Although I do not know those languages that well, one can come
to a conclusion concerning the basis for the translation. Is it based
on valuation or measurement? It looks like only English-speaking
countries have taken this measurement road. Elsewhere they still
use the term valuation. I also dug into older literature in English.
Before the 1960s, the term measurement did not exist. In the fifties
and sixties there were a few articles and books discussing
whether valuation and measurement carried different meanings.
And in 1971, the committee of the Accounting Review made the
choice to use the term measurement.
Some of the interviewees described how they theorised certain
choices of term by referring to various texts or bodies of knowledge,
as illustrated by the above quotation. Having discussed the alter-
native target-language terms in a meeting, the TRC translated the
term measurement using the target-language equivalent of valua-
tion, a term traditionally used in the accounting contexts in the
Finnish language. Similarly, the TRC considered what sort of
observation the observable in observable prices or observable
market designates and to what extent it might be captured by
alternative target-language words.
5.1.2. Maintaining a consistent translation
The IFRS Foundation's translation policy requires that each
source-language term is always translated with the same target-
language term (IFRS Foundation, Terms of Reference, unpublished),
and any exception to this rule must be justified. According to the
translators interviewed, the Directorate General for Translation has
a similar requirement. The practice of ensuring the consistent
translation of terms within and across standards may be an
important form of institutional work for the maintenance of the
IFRS as a translingual institution. An outcome of the term-for-term
correspondence may be that issues around conceptual equivalence
remain unproblematised in situations where different language
versions of an IFRS are read in parallel.20
The interviewees reported the TRC worked to maintain consis-
tent terminology across different standards, and often referred to
the translations of previous standards when a new standard or an
amendment was being translated. Additionally, the DGT translators
might suggest aligning the terminology to a certain extent with
other EU texts. As a consequence, there is a strong intertextual
patterning between the translations of different IFRS as the choice
of vocabulary and phrasing for the translations of any newly issued
or amended IFRS is strongly affected by the way earlier standards
have been translated (cf. Bhatia, 2004).
In addition to collaborative activities between TRC members,
maintaining consistency across different IFRS was facilitated
20
To give an example of the opposite, translations of the true and fair view (TFV)
into different languages have been problematized in several studies on the grounds
that the term has been translated inconsistently (Dahlgren & Nilsson, 2012; Nobes,
2009, 2013). Furthermore, an interviewee indicated that it is not uncommon that
she or her colleagues read at their daily work the original IFRS and the translation
in parallel to interpret a certain requirement.
47
technologically through the use of translation memory21 software
and the list of translated terms. Despite the technological aids,
maintaining terminological consistency requires effort and judge-
ment. The translation reviewers frequently underlined the signifi-
cant role of the TRC translator in maintaining the consistency of
terminology across standards. They pointed out that the translator
had the most knowledge of the basis for the decisions made in
translating previous standards. However, previous translations were
not simply mechanically reused in subsequent standards due to the
context-dependent meaning of words in both languages. Difficulties
arose partly because target-language equivalents must be chosen
based on their perceived appropriateness in the context of the
standard where they first appeared. When source-language terms
reappear in subsequent standards in different contexts, or as a part of
another word combination, the previously selected target-language
equivalents might not carry a meaning that is appropriate in the
new context.
A TRC member described the difficulties arising when funda-
mental concepts are used in the source text in a manner that ap-
pears inconsistent to a Finnish reader.
Interviewee D: And then, relating to the terminology, it has been
clear to us in Finland what expenditure is (in Finnish, meno),
what expense is (kulu), and what a cost (kustannus) is, but in
English it is much less well-defined because the term cost en-
compasses costs in cost accounting, but it is also meno, expen-
diture, in other contexts as meno would be defined in Finnish.
When the same term has varying meanings in English, it is
utopian to think that some absolute consistency could be
attained. There have been some differences in usage and prob-
ably will be because of these traditions …
This is another illustration of why the term-for-term requirement
is problematic in translation when the meaning depends on the
context. Abstract concepts often cover a certain semantic field rather
than referring solely to a well-specified referent. Therefore, the TRC
strives to determine whether the meaning of a previously selected
target-language term overlaps with its source-language counterpart
in a given context, thus, implicitly defining the boundaries of the
concepts. Another example of conceptual differences is the trans-
lation of the terms depreciation and amortisation. The English lan-
guage conceptualises the gradual decreasing usefulness of assets by
writing off costs in a process called depreciation for tangible fixed
assets and amortisation for goodwill and other intangible assets; In
contrast, the Finnish language uses the same term, poisto, for both.22
The English language also uses the term amortisation, in the context
of repayment of a debt by a borrower, while the Finnish language
uses a different term, takaisinmaksu, as poisto is strictly an account-
ing term. The Finnish TRC chose, however, to translate both amor-
tisation and depreciation of assets using same term, subsequently
justifying the choice in the foreword to the translated bound volume
of the standards.
5.1.3. Translation problems arising from changes in source-
language terminology
It is recognised that language is dynamic, and accounting
21
A translation memory is a database that consists of text segments in the source-
language and their previous translations to the target-language. Translation
memory software suggests possible translations of segments based on prior
translations, and shows the technical level of equivalence between a given source-
language segment and the previously translated text segment in percentage form.
22
According to Dahlgren and Nilsson (2012), similar difficulties arose with the
translation of ‘depreciation’ and ‘amortisation’ into Swedish.
48 J. Kettunen / Accounting, Organizations and Society 56 (2017) 38e54

terminology changes over time (Evans, 2010; Mills, 1989). However, as in forthcoming financial statements or media articles.
the current research establishes that the terms used in the IFRS and
the underlying concepts do not always change simultaneously. The
5.2. Coping with ambiguity and uncertainty in the meaning
shifting meaning of a term, or changes to the term when its referent
has not changed, combined with the word-for-word requirement in
The ambiguity and indeterminacy of language have attracted
particular, creates issues for those involved with the translation. In
considerable attention in linguistics and legal theory (e.g., Cao,
such circumstances the translators and translation reviewers
2007; Joseph, 1995). It is argued in this subsection that these
pondered whether and when the translation should reflect termi-
inherent properties of language pose considerable practical chal-
nological changes in the source text as compared to earlier standards.
lenges to the IFRS translators and translation reviewers. A certain
An event that was often brought up in the interviews was the
degree of interpretation appears to be an unavoidable aspect of the
decision-making process regarding the translation of the statement
translation of IFRS, although the IFRS Foundation maintains that
of financial position when this term replaced the balance sheet in IAS
the ‘purpose of the translation is not to interpret or explain the
1 in 2007. Before the terminology meeting, all but one of the TRC
standards’ (IFRS Foundation, Terms of Reference, unpublished). This
members signed on to a suggestion to adopt a more literal trans-
subsection investigates how translators and translation reviewers
lation of the statement of financial position. But during the meeting,
deal with the ambiguity and uncertainty around intended mean-
one member convinced the rest of the committee that the Finnish
ings within the source text. Although the translation process
term could not be changed. Interviewees recalled her saying, for
established by the IFRS Foundation highlights the formal corre-
example, that ‘the balance sheet [in Finnish tase] is the balance
spondence of terminology, the translators and translation re-
sheet and it can't be anything else’.23 This decision making by
viewers underlined the efforts made to render the meaning(s) of
consensus also illustrates that meetings significantly affect the
the source text appropriately in the translation. Rendering the
decisions made by the TRC. An interviewee remarked that ‘luckily’
intended meaning beyond individual words is undoubtedly a
they did not change the term because ‘doing so could have
desirable goal. Furthermore, the constituent perceptions of the
confused financial statement users and the media.’ Indeed, intro-
extent to which that goal is being achieved will likely have an
ducing a new target-language term would have created new
impact on the stability of the IFRS as a translingual institution.
terminological differences between the Finnish language IFRS and
For each standard, a subgroup of TRC members was chosen to
local accounting regulations.
review the draft translation to ensure that the meaning was not
The decision not to amend the Finnish term corresponding to the
altered in translation (see Fig. 2). A TRC member described the
balance sheet may be interpreted as reflecting a concern with ensuring
difficulty of interpreting and rendering the meanings of longer text
the success of translation in the form of forthcoming intertextuality.
segments in translations as follows:
The familiar and concise target-language term, tase, continues to be
embedded in other texts, including financial statements. When the Interviewee G: One of the main problems in the translation of
target-language IFRS terms are used in other contexts by the pre- IFRS is not terminology, which is also a difficult issue, but un-
parers of financial statements and the media, among others, these derstanding what something means. A single sentence in which
constituents implicitly preserve and reproduce the one-to-one there is no foreign word can be discussed in the group for a long
equivalence constructed by the committee. This serves to maintain time. Sometimes the text can be understood in several ways.
the equivalence of terminology constructed in the translation. Also, native speakers who we consult can have difficulties [with
In summary, this section has examined the practical problems of understanding the text].
linguistic equivalence, and the activities that trigger the creation and
maintenance of a formal correspondence of terminology between
Another TRC member emphasised that ‘The role of the reviewer
languages. The lack of a pre-existing conceptual equivalence be-
requires that you should see not only if it looks fluent, but also that
tween the languages is apparent, in that often several different
the idea corresponds to what has been said in the original text.’
target-language terms were suggested as translations for one source-
These statements show that translators and translation reviewers
language term. However, the TRC coined new terms to describe IFRS
find it problematic to make sense of the source text and transfer the
concepts, and a few source-language terms were left untranslated.
meaning unchanged to the target text. They also illustrate that the
It was also argued in this section that the consistent translation of
interviewees strive to convey the meanings of longer segments of
terms appears to be an important form of the institutional work in
text, rather than the individual words, into the target-language.
the maintenance of the target-language IFRS. As a result, there is a
While research has acknowledged that translators should not try
strong intertextual patterning across different standards in the
to resolve ambiguities but aim to capture them in the translation
translated IFRS. The terms used in the IFRS and the underlying
(Evans et al., 2015), ambiguity cannot always be maintained in
concepts do not, however, always change simultaneously, which,
translation, as some of the interviewees pointed out.
combined with the term-for-term requirement, causes issues for
The data informing the current study indicate that uncertainty
those involved in translation. Due to the context-dependent mean-
about the meaning of the source text arises from two main sources.
ing of words, the TRC engages in defining the boundaries of concepts
First, with regard to how the text is structured verbally (including
by deciding whether their meanings overlap adequately in a
syntax), both the translators and accounting experts noted that it is
particular context. Occasionally, the translation reviewers also
not always clear which word or phrase a particular (relative) pro-
focused on selecting target-language terms capable of being
noun refers to (i.e., what the antecedent of the subordinate clause
embedded in other texts, thus reproducing the constructed linguistic
is). The following excerpts briefly illustrate this problem of gram-
equivalence intertextually in texts outside the actual standards, such
matical ambiguity, in particular that of ambiguous cross-
references:
When management is aware, in making its assessment, of ma-
23
Similarly, the term Bilanz was retained in the German translation (Evans et al., terial uncertainties relating to events and conditions that [italics
2015). ‘This is not an oversight e the possibility of using alternative translations was
debated during the translation of the 2007 revision of IAS1, but the reviewer chose
added] may cast significant doubt upon the entity's ability to
to retain the term dBilanz. We are grateful to Robin Bonthrone for pointing this continue as a going concern, the entity shall disclose those un-
out.’ (Evans et al., 2015, p. 17.). certainties (IFRS., 2008; IAS 1.25).
 J. Kettunen / Accounting, Organizations and Society 56 (2017) 38e54
Fig. 2. The translation process based on the data of the current study.
… eliminate all deferred losses and gains arising on derivatives
that [italics added] were reported in accordance with previous
GAAP as if they were assets or liabilities (IFRS., 2008; IFRS 1.B4).
In the first excerpt above, the term that could refer to either
uncertainties or events and conditions. In the latter excerpt, that
could refer to either losses and gains or derivatives. Ensuring that the
translation correctly captures the intention in these sentences, if a
dual reference was intended, is not possible in Finnish. Theoreti-
cally, a translation carrying either of the two possible meanings is a
correct translation. While a certain degree of indeterminacy may
sometimes be intentional in order to allow the standard to be
applicable in different business environments, the referential am-
biguity in the above excerpts is scarcely likely to be intentional on
the part of the standard-setter. For these reasons, TRC members
needed to decide which meaning to render in a translation. If the
translator is not an accounting expert, it may be even more difficult
to infer the referential relationships (and thus the standard-setter's
intended meaning) in each context. As the DGT translator noted:
Referential ambiguity brings forth [a problem], because none of
us [in the translation team at DGT] is an accounting expert. And
then if there is a short elliptical sentence, we may invert the
referential relationship when translating.
According to more than one translation reviewer, however, ac-
counting experts can also struggle with sentences in which the
grammatical structure permits of two interpretations, although
accounting experts can draw on the context to resolve such
ambiguities.
Second, issues with comprehending the original standards, and
uncertainty about the standard-setter's intentions, were said to
derive from the general complexity or foreignness of the substance
of the regulation. One of the accounting experts interviewed re-
flected on the need for interpretation and the difficulty of
comprehension as follows:
49
Interviewee C: [the] IFRS can be difficult, which is not only a
linguistic matter. It's just that they are so difficult to compre-
hend that also we as professionals can have [difficulty under-
standing] what IASB wants, if it's not totally clear … and they are
forthcoming norms … and they have not been in use yet, so it's
not even known … what the IASB is seeking …
To some extent, making sense of the source text may become a
negotiation about another negotiation, in that the TRC members
negotiate the content produced by negotiations conducted at the
IASB. It should be noted, however, that the word negotiation is not
used here in the sense that TRC members would have sought to
create translations in a self-interested manner. Instead, the
differing professional backgrounds of TRC members seemed to in-
fluence which issues they considered in the course of the trans-
lation. It is worth noting that the translation reviewers are able
define the bodies of knowledge they deem relevant in termino-
logical decision making (cf. Pelger, 2016). For example, the aca-
demic (whose research interests include measurement theories)
presented viewpoints on the meanings of terms such as observable,
and measure/measurement versus valuate/valuation and on their
conceptual equivalence with target-language terms in various
contexts. This again illustrates that the meaning is inferred in
relation to other texts or bodies of knowledge invoked during the
reading and translation process (Bhatia, 2004).
Furthermore, TRC members reported they found translation
problematic because ‘standards are by nature a compromise,’ and
because ‘the IFRS are [written in] neither American nor British
English’ or ‘some standards have been created in a rush.’ Indeed,
the conceptual roots of standards are diverse and the standard-
setters have tried not to bind the terms tightly to any geograph-
ical location. One could also say that the perceived difficulty of
translating a text is not independent of who is translating it. An
interviewee noted that TRC is heterogeneous and they ‘are in
contact with the standards in different ways,’ also saying that:
[The relative difficulty of comprehension] is related to our
differing backgrounds. I follow all the time what happens with
50 J. Kettunen / Accounting, Organizations and Society 56 (2017) 38e54
the IFRS, beginning from when they are being developed. So I
have some kind of understanding of what they want to say.
Then, what is the final English-language version of the standard
… I know … or I think that I know, what they mean by it. But if
you haven't followed it … or read the basis for conclusions …
and if you really haven't followed it from the beginning of the
project, when they made some preliminary decisions and when
they changed it, and those bases and discussions … then it is
certainly more difficult.
The meaning of the original text to be conveyed in the target-
language was not only inferred from the final text of any standard
per se. Rather, the meanings were constructed intersubjectively
and intertextually based on the TRC members' interpretations of
the standard-setters’ intended meanings. Those intended meanings
in turn tend to be formed in the course of the standard-setting
process and with reference to other texts such as similar stan-
dards or the Conceptual Framework. In particular, the goal of dy-
namic equivalence cannot be achieved in translation, if it can be
achieved at all, only on the basis of the current text as the meanings
of sentences in the present standards are the result of developing
and revising standards since the formation of the International
Accounting Standards Committee (IASC).
In summary, the TRC members make efforts to infer the
meaning in the source text as intended by the standard-setter and
to convey it unchanged into the target-language. The institutional
work aimed at capturing the meaning in translation takes different
forms ranging from the grammatical analysis of sentences to
actively following standard-setting due process from its early
stages. In other words, the hybrid expertise possessed by the
committee, and their long term commitment to the standards
makes it possible to approach instances of ambiguity or uncertainty
in the meaning simultaneously from various angles. Ambiguity and
indeterminacy are, however, inherent properties of language and,
as such, stand as challenges to the work of translation.
5.3. Translation of the IFRS as a contested area of expertise and
regulatory influence
This subsection addresses the institutional dynamics sur-
rounding the translation of the IFRS as it continues to be an un-
settled area marked by a range of policies and regulatory
influences. More specifically, responsibility for translating the EU-
endorsed IFRS has shifted between the DGT and the local TRCs set
up by the IASC/IFRS Foundation. The differing translation policies of
the DGT and the IFRS Foundation and their impact on the degree of
involvement of translators and accounting professionals in the
translation process, were briefly described in Section 3. These
policies will be discussed further in this section. The IFRS Founda-
tion describes its approach to authorising translations in its trans-
lation, adoption and copyright (TAC) policy as follows24
The Foundation owns the worldwide copyright to the IFRSs in all
languages and therefore owns the exclusive right to reproduce,
or authorise others to reproduce or translate, IFRSs. … The
quality of the translation is protected e as copyright owner, the
Foundation sets the translation process, and decides who carries
out the translation, ensuring that there is a single, high-quality
translation in each language … as copyright owner, the Foun-
dation decides who to license to translate. Licences are non-
exclusive but the Foundation supports its licensees’ shared
24
http://www.ifrs.org/Use-around-the-world/Adoption-and-copyright/
Documents/Translation-Adoption-Copyright-Policy-August-2013.pdf.
commercial interests, ensuring that a single, authorised trans-
lation is sustainable. Income from authorised sales of official
translated publications helps the Foundation to cover the
translation costs.
Significantly, the TAC policy declares that the IFRS Foundation
has sole responsibility for establishing the translation process; a
policy intended to protect the quality of translations. Overall, the
IFRS Foundation emphasises that translation into different lan-
guages must adhere to a uniform, well-specified process. The
practices of translating into different languages are, however, more
varied. Despite references by the IFRS Foundation to a uniform
translation process and ensuring a single high-quality translation in
each language, the IFRS Foundation does not have direct control
over the translations into several EU languages. Likewise, multiple
translations exist for languages such as French and Portuguese, one
translation undertaken by the DGT and another by a local profes-
sional association outside the EU.
In sum, the translations created by the DGT, as well as those by
the TRCs affiliated with the IFRS Foundation, are not always rec-
ognised by the IFRS Foundation or the EC respectively.25 The
translations of the EU-endorsed IFRS are not necessarily reviewed
by accounting experts. In contrast, the IFRS Foundation only ap-
proves the translations undertaken in accordance with its trans-
lation process including an expert review.
5.3.1. Shifting sites of translation
The trajectories of organising the translation work reflect the
dynamics arising from the duality of the IFRS that originated as a
form of soft regulation, but are now legally binding regulations in the
EU. The first translations of the then IAS were created by professional
organisations, such as the former IASC member bodies before the
widespread adoption of IFRS by the EU. The then IASC acknowledged
the need for an official translation process when it undertook the
German translation in 1997 (IASB., 2001). For the EU, the IASC
Foundation (the predecessor of the IFRS Foundation) was contracted
by the EC to provide official translations of the endorsed IFRS in
cooperation with the local TRCs during various periods between
2003 and 2009. From October 2005 to June 2007, the DGT provided
translations without the participation of the IASC Foundation. The
translation contract between the EC and the IASC Foundation was
terminated at the request of the latter in June 2009, and the DGT has
been responsible for the EU translations since then.26
The case of the Finnish language translation examined in this
paper appears to be an exception to the current translation prac-
tices of the EU-endorsed IFRS. Despite the recurring organisational
changes described above, all the IFRS have been translated into
Finnish by the same translator and reviewed by accounting experts,
except for the period 2005e2007 and a short period in 2009e2010,
when the DGT alone was in charge of the translation. Currently, the
Finnish translations are approved by the IFRS Foundation and
endorsed as EC regulations. In practice, the former Finnish TRC has
been supplying one of the DGT's subcontractors with translations
25
In an answer to a parliamentary question, Ms Vassiliou on behalf of the EC
stated as follows: “As for the uniform application of standards within the European
Union, it is the version published in the Official Journal of the European Union that is
binding, regardless of the possible existence of other translations.” (http://www.
europarl.europa.eu/sides/getAllAnswers.do?reference¼P-2011-
008747&language¼EN). Not all EU languages are included in the list of available
translations by the IFRS Foundation. (http://www.ifrs.org/Use-around-the-world/
IFRS-translations/Pages/Available-translations.aspx).
26
See http://www.europarl.europa.eu/sides/getAllAnswers.do?reference¼P-
2011-008747&language¼EN and http://www.khtmedia.fi/julkaisut/kategoriat/ifrs/
ifrs-standardit-2013.
 J. Kettunen / Accounting, Organizations and Society 56 (2017) 38e54
that have been further reviewed by the DGT.27 The present
involvement of the former TRC contradicts the understanding
expressed in prior literature that the DGT ‘has been preparing all
translations of endorsed IFRS since June 2009, without input from
the IFRS Foundation’ (Evans et al., 2015, p. 4; see also; Somssich
et al., 2012). While the issues affecting translations into other lan-
guages are beyond the scope of this paper, it is worth noting here
that DGT translators may refer to the translations approved by the
IFRS Foundation as a basis for creating the EU-endorsed trans-
lations if such translations exist.28
5.3.2. Translation of the IFRS as a contested area of expertise
The site of translation has an impact on who is entitled to carry
out the translation, and on the practices and procedures through
which the translations are created. As Cooper and Robson (2006,
415) stated, the ‘institution and location where regulation takes
place affect both the outcome and the legitimacy of the rules and
practices produced.’ The translation of the IFRS appears as a con-
tested area of expertise, where the relative involvement of occu-
pational groups varies between languages and jurisdictions, and
the appropriateness of translations is ultimately questioned. Simi-
larly, the interviewees in this study called into question certain
target-language equivalents chosen by translators (as opposed to
accounting experts), and a few expressed disbelief that translators
alone could render the IFRS adequately into other languages.
Translation at the DGT differs from the translation process
created by the IFRS Foundation in a number of aspects. The EU-
endorsed IFRS are either translated by the in-house translators of
the DGT or outsourced to contractors and subjected to the DGT's
internal quality control. Because the DGT is a public organisation
and part of the EC, the DGT translators tend to primarily rely on
experts in national ministries of finance for advice rather than
those working in the private sector,29 which is in line with the
European tradition of the State being involved in the regulatory
processes (see Chiapello & Medjad, 2009; Crawford et al., 2014). For
some languages, however, the DGT or its subcontractors maintain
contact with national accounting associations or the TRCs associ-
ated with the IFRS Foundation. Regarding the translation process,
the DGT translators are not expected to follow the sequence of
activities required by the IFRS Foundation.
As the above description of the shifting sites and the variations
between languages in constituents and practices illustrates trans-
lation of the IFRS remains an unsettled area of expertise. Tradi-
tionally, the translation of EC regulations has been undertaken and
monitored solely by translators. The domain of accounting has
however expanded into translation work through the involvement
of the profession in the setting of standards and their claim to
possess the required technical expertise. Nevertheless, the contest
over who possesses legitimate expertise has not taken the form of
an overt struggle between accountants and translators. In fact, the
IASB approach to translation relies on hybrid expertise and the co-
production of translations by accountants and translators.
While the interviews conducted for this study focused on the
practical problems of linguistic equivalence and translation into
Finnish within the framework set by the IASB, the interviewees'
perceptions of and concerns about the quality of the translations
not reviewed by accounting experts emerged in several interviews.
Many of the interviewees (from both the translator and accountant
27
It is noted on the IASB webpage that a Finnish translation of the 2014 Blue Book
is available. However, no reference is made to the existence of other DGT
translations.
28
Verbal communications with an anonymous DGT translator, June 2015.
29
Verbal communications with three anonymous DGT translators, June 2015.
51
groups) emphasised that they considered a knowledge of ac-
counting, and specifically expertise in the IFRS, to be a prerequisite
for the translation work. Similarly, some of the accounting experts
interviewed criticised the translators' approach towards devising
Finnish terms intended to correspond with the detailed profes-
sional terminology in English. An interviewee went on to say that
‘the translators are taught to use such voluptuous pure Finnish
which no [accounting] expert would use even in her worst night-
mare.’ The interviewee was opposed to what she described as
translations that were ‘too Finnish,’ and was concerned that they
might not be used in practice e a concern that a few other re-
viewers also raised.
On the practical side, the discontinuities resulting from the
contractual issues between the EU and the IFRS Foundation influ-
enced the production of the subsequent translations. As those
involved in the translation process had changed, those charged
with conducting the later translation abandoned the idea of using
some of the previous translations due to perceived mistakes or
inaccuracies. An interviewee described the situation as follows:
We have that kind of problem that when the standards were
translated by freelancers … those translation memories [data-
base content] … and also the versions published in the Official
Journal at that time were bad … and were corrected for the
volume by [the name of a publisher] … which means that we
can't use all those previous texts … which as such are supposed
to be official regulations. As a matter of fact, we have been able
to create the kind of translation memory that is reliable enough
[only going back to translations since] 2008.
The preceding quotation illustrates that the later translators
were hesitant about complying with terms chosen by their pre-
decessors. Although the TRC normally worked to maintain the one-
to-one correspondence of terminology that they had established in
previous translations, certain translations by subcontractors were
considered to be inadequate as reference material in the creation of
new translations by both later translators and IFRS experts.
The above narrative on the recurrent shifts between the DGT and
the TRCs complement recent studies on the EU/IASB dynamic (e.g.
Chiapello & Medjad, 2009). Prior analysis, such as that of Bengtsson
(2011) and Crawford et al. (2014), focused on the overt struggles
over the control of accounting standards within the EU, whereas
this section has drawn attention to the somewhat more mundane
(yet important) activities supporting transnational regulation. Spe-
cifically, it has highlighted the recurrent shifts in the organisation of
translation work, and the discrepancies between the translation
policies of the EU and the IFRS Foundation. This subsection has also
shown that the institutions and locations of translation are inter-
twined with the professions entitled to carry out the translation of
the IFRS. As was discussed in this subsection, the IFRS Foundation's
translation policy delegates a more integral role in translation work
to the accounting profession than does the translation policy of the
EU. Notably, the difficulties of the translating the IFRS into the EU
languages have not dissipated, despite the fact that the IFRS were
adopted by the EU more than a decade ago.
Despite changes in how the translation of the IFRS is organised
at the transnational level, the particular case of the Finnish lan-
guage translation studied in this paper does not stand out as an
illustration of a struggle between professions. Rather, the DGT
translators and the former members of the Finnish TRC continue
their collaboration (albeit under an amended arrangement) to
maintain what both parties consider good translation practice, and
thereby to maintain the legitimacy of the translations. In other
words, the local actors resisted the institutional change taking place
at the transnational level, specifically by continuing the provision of
52 J. Kettunen / Accounting, Organizations and Society 56 (2017) 38e54
the IASB-approved translations of the EU-endorsed IFRS. This ac-
tion contributed to preserving the existing translation process by
retaining an expert review. The continued expert review in the case
of Finnish language translation appears to be an exception to cur-
rent practices of translating the IFRS into the EU languages, and it is
important to note that studying the trajectory of the organisation of
the translation work for other languages such as French, German or
Swedish would have yielded a quite different account.
6. Discussion and conclusions
This paper has shed light on the practical problems of linguistic
equivalence by enquiring into the empirical topic of the translation
of the IFRS into the Finnish language. Drawing inspiration from the
notion of institutional work (Lawrence & Suddaby, 2006), the
present study has explored the detailed practical activities and
institutional interactions aimed at creating and maintaining an
acceptable linguistic equivalence in the course of translation, and
thereby maintaining the IFRS as a translingual institution. Indeed,
the institutional work of translation is essential for the global reach
of the IFRS. It has been shown in the paper that the formal, one-to-
one equivalence between terms in the English-language IFRS and
their translations is often constructed during the translation com-
mittee processes, as opposed to simply being apparent. Similarly,
the translators and translation reviewers strive to maintain the
previously established one-to-one equivalence when translating
the current standard. Efforts are also made to select target-
language terms that it is assumed will embed well into other
texts, ensuring the long-term sustainability of the translations.
Overall, the translation of the IFRS is an ongoing activity which
involves negotiation and the balancing of case-specific criteria that
the committee members present and must reach an accord on.
The present study contributes to two related bodies of ac-
counting literature. First, it contributes to the literature on the
translation of accounting and auditing standards by examining the
procedures through which different regulators, translators and
accounting experts reviewing the translations address the practical
problems of linguistic equivalence. This is a novel contribution to
the literature because extant studies have examined the outcomes
of translation activities, that is, translated text segments and terms,
focusing primarily on inaccuracies, translation errors (e.g.,
Dahlgren & Nilsson, 2012; Nobes, 2006; Sunder, 2011), or differ-
ences in the interpretation of uncertainty expressions (e.g.,
Aharony & Dotan, 2004; Doupnik & Richter, 2003). Several ac-
counting studies have suggested that translation is inherently
problematic (Baskerville & Evans, 2011; Evans, 2004; Evans et al.,
2015; Zeff, 2007). When compared with prior research, the pre-
sent study clearly elaborates on the complexities of translation and
contributes a focus on translation work in its social and institu-
tional context.
While prior accounting research on translation recognises that
accounting concepts may not neatly overlap in different languages,
the findings from the study of Finnish language translation illus-
trate that issues arise in attempting to render the meanings of text
segments (as opposed to words) into another language. In partic-
ular, inferring the meaning in the source text was considered
problematic because of the occasional foreignness of the substance
and the inherent ambiguity and indeterminacy of language. In
other words, problems of translation do not only arise from the
absence of the means to render a specific meaning in the target-
language (which is also a common problem) but from the need to
make sense of the source text. Following the standard-setting
process from its early stages contributes greatly to making sense
of the final text in the IFRS. Overall, the meaning was not merely
carried by the text under review itself but the TRC members drew
on other texts and their professional knowledge bases to infer the
meaning of the source text to be rendered in the target-language.
Second, with regard to the scholarship on transnational ac-
counting regulation and standard-setting, the study of the trans-
lation of regulatory texts furthers our understanding of the
institutional interactions and activities within the transnational
financial regulatory arena. The translation of the IFRS has a unique
position at the interface of setting and implementing standards,
and it also lies in the middle ground of activities supporting regu-
lation governed by the IFRS Foundation and those by the EC. Apart
from transnational standard-setters and regulators, the translation
review committees are a node of interaction between representa-
tives of large and mid-tier audit firms, industry specialists, the
preparers of financial statements, and other constituents approved
by the standard-setter. Importantly, the IASB-approved translations
of the IFRS are not created by a single translator but are the
outcome of collaborative activity between translators and selected
constituents following a specific process. Despite the detailed
process set up by the standard-setter, however, the committee
members have the responsibility for the translation and it is they
who ultimately decide what aspects are taken into consideration in
translation.
While the present study has focused on institutional work car-
ried out by translators and translation reviewers to maintain the
IFRS as a translingual institution, the standard-setter also engages
in institutional work to maintain the status of the translations of
the IFRS. For instance, the IFRS Foundation argues that ‘multiple
translations of IFRSs into the same language would endanger
comparability, transparency and the long-term sustainability of
high-quality IFRS translations.’30 Equally importantly, multiple
translations might serve to problematise the ideal of linguistic
equivalence, and in particular the one-to-one equivalence of ter-
minology between languages, which the IFRS Foundation's trans-
lation process relies on.
The developing and shifting nature of the organisation of the
translations of the IFRS into EU languages aligns with other analyses
of the dynamics associated with the transnational financial regula-
tory structures (see e.g. Humphrey et al., 2009). The organisation of
the translation of the IFRS is not a static phenomenon but one that
shifts and changes. The roles of the transnational agencies, and
accordingly, the actual constituents undertaking translation have
changed recurrently, often abandoning or reintroducing the IFRS
Foundations’ translation approach. For the moment, the end of the
contract between the EU and the IFRS Foundation appears to have
prompted discontinuing the expert review of several EU language
translations. Interestingly, large and mid-tier audit firms, or other
private-sector constituents are no longer represented in the repro-
duction of the non-English-language versions of the endorsed IFRS.
This observation on the present situation contrasts with the findings
of several studies pointing to the large audit firms occupying more
terrain in transnational accounting regulation and related activities
(e.g., Cooper & Robson, 2006; Humphrey et al., 2009).
The interviews and archival materials this paper draws upon are
restricted to the case of Finnish language translation, and so the
paper does not address the intra- or inter-organisational negotia-
tions and decision making of the IFRS Foundation and the EU on the
organisation of translation work. Similarly, although it is stated that
‘The IASB is keen to avoid issues problematic to translation in these
documents’ (IFRS Foundation, 2010, p. 2), we do not know whether
and how the standard-setters or the staff attempt to address issues of
translation during the standard-setting process. While this paper has
30
http://www.ifrs.org/Use-around-the-world/IFRS-translations/Pages/IFRS-
translations.aspx.
 J. Kettunen / Accounting, Organizations and Society 56 (2017) 38e54
examined the organisation of translation for an EU language, the IFRS
are translated in a multitude of jurisdictions beyond the EU. The
institutional and linguistic contexts where translation activities take
place are diverse and might differ substantially from those examined
in the present paper. This issue could be a matter of future research.
In sum, there are clear opportunities for research to understand
better how accountants and auditors deal with the issues of
multilingualism and translation. It would be useful to gain insights
into various language and translation related practices in the realm
of financial reporting, including the interpretation of transnational
standards and regulations by accountants and auditors who are not
native English speakers, and the preparation of financial statements
in multiple languages. Despite the growing body of literature on
translation as a potential barrier to transnational accounting har-
monisation we have very little insight into the views of those who
prepare, use and audit financial statements on the approaches they
adopt to address the practical problems involved in handling lin-
guistic equivalence in the context of their daily work, or into the
negotiations on the degree of discretion possible with specific
translations. This again, as suggested by Cooper and Robson (2006),
highlights the sustained importance of conducting research on the
broader topic of the interpretation and implementation of rules.
Acknowledgements
I would like to thank the editors of the special issue and the two
anonymous reviewers for their helpful comments. I also benefitted
from the comments made by Kimmo Alajoutsija €rvi, Lisa Baudot,
Rachel Baskerville, Michael Bourne, Lisa Evans, Marko J€ €a
arvenpa €,
Tanja Leppa €aho, Hannele M€ €, Toni Ma
akela €tto
€ , Jukka Pellinen, Antti
Rautiainen, Bob Scapens, Marjo Siltaoja, Aila Virtanen, and the
participants at the AOS workshop at the LSE (2011), Critical Per-
spectives on Accounting Conference in Clearwater (2011), IPA
Emerging Scholars' Colloquium in Cardiff (2012) and EAA's Doctoral
Colloquium in Saclay, Paris (2013) on the earlier drafts of this paper.
I also want thank the interviewees who participated in this study
for their time and insights. Finally, I gratefully acknowledge the
financial support received from the Finnish Foundation for Eco-
nomic Education.
Appendix 1
Abbreviations
EC The European Commission
EU The European Union
DGT The Directorate-General for Translation
FASB The Financial Accounting Standards Board
IASB The International Accounting Standards Board
IASC The International Accounting Standards Committee
IFRS The International Financial Reporting Standards
IAS The International Accounting Standards
ISA The International Standards on Auditing
LSP Language for Specific Purposes
TRC Translations Review Committee
References
Aharony, J., & Dotan, A. (2004). A comparative analysis of auditor, manager and
financial analyst interpretations of SFAS 5 disclosure guidelines. Journal of
Business Finance and Accounting, 31(3&4), 475e504.
Aisbitt, S., & Nobes, C. (2001). The true and fair requirement and its recent national
implementations. Accounting and Business Research, 21(2), 83e90.
Barrett, M., Cooper, D. J., & Jamal, K. (2005). Globalization and the coordinating of
work in multinational audits. Accounting, Organizations and Society, 30(1), 1e24.
53
Baskerville, R., & Evans, L. (2011). The darkening glass: Issues for the translation of
IFRS. Edinburgh: The Institute of Chartered Accountants of Scotland.
Bassnett, S.. (1998). The translation turn in cultural studies. In Bassnett, S. & Lefe-
vere, A. (eds.), Constructing cultures (pp. 1e11). Essays on literary translation.
Clevedon: Multilingual Matters.
Bassnett, S., & Lefevere, A. (1990). Translation, history and culture. London and New
York: Pinter.
Baudot, L. (2014). GAAP convergence or convergence gap: Unfolding ten years of
accounting change. Accounting, Auditing and Accountability Journal, 27(6),
956e994.
Bengtsson, E. (2011). Repoliticalization of accounting standard setting e the IASB,
the EU and global financial crisis. Critical Perspectives on Accounting.
Bhatia, V. (2004). Worlds of written discourse: A genre-based view. London:
Continuum.
Botzem, S. (2010). The politics of accounting regulation. Organizing transnational
standard setting in financial reporting. Cheltenham : Edward Elgar.
Botzem, S., & Dobusch, L. (2012). Standardization cycles: A process perspective on
the formation and diffusion of transnational standards. Organization Studies,
33(5e6), 763e790.
Brown, J. S., & Duguid, P. (2000). Balancing Act: How to capture knowledge without
killing it. Harvard Business Review, 78(3), 73e80.
Bucholtz, M. (2007). Variation in transcription. Discourse Studies, 9(6), 784e808.
Callon, M. (1986). Some elements of a sociology of translation: Domestication of the
scallops and the fishermen of st Brieuc Bay. In J. Law (Ed.), Power, action and
belief: A new sociology of knowledge? (pp. 196e223). London: Routledge & Kegan
Paul.
Canning, M., & O'Dwyer, B. (2013). The dynamics of a regulatory space realignment:
Strategic responses in a local context. Accounting, Organizations and Society,
3(38), 169e194.
Cao, D. (2007). Inter-lingual uncertainty in bilingual and multilingual law. Journal of
Pragmatics, 1(39), 69e83.
Catford, J. C. (1978). A linguistic theory of translation: An essay in applied linguistics.
Oxford, England: Oxford University Press.
Chiapello, E., & Medjad, K. (2009). An unprecedented privatisation of mandatory
standard-setting: The case of European accounting policy. Critical Perspectives
on Accounting, 4(20), 448e468.
Chiwamit, P., Modell, S., & Yang, C. L. (2014). The societal relevance of management
accounting innovations: Economic value added and institutional work in the
fields of Chinese and Thai state-owned enterprises. Accounting and Business
Research, 44(2), 144e180.
Chua, W. F., & Taylor, S. (2008). The rise and rise of IFRS: An examination of IFRS
diffusion. Journal of Accounting and Public Policy, 27(6), 462e473.
Cooper, D., & Robson, K. (2006). Accounting, professions and regulation: Locating
the sites of professionalization. Accounting, Organizations and Society, 31(4e5),
415e444.
Crawford, L., Ferguson, J., Helliar, C., & Power, D. (2014). Control over accounting
standards within the European Union: The political controversy surrounding
the adoption of IFRS 8. Critical Perspectives on Accounting, 25(4e5), 304e318.
Dahlgren, J., & Nilsson, S.-A. (2012). Can translations achieve comparability? The
case of translating IFRSs into Swedish. Accounting in Europe, 9(1), 39e59.
Davidson, R. A., & Chrisman, H. H. (1993). Interlinguistic comparison of interna-
tional accounting standards: The case of uncertainty expressions. The Interna-
tional Journal of Accounting, 28(1), 1e16.
Djelic, M.-L., & Sahlin-Anderssen, K. (2006). Transnational governance: Institutional
dynamics of regulation. Cambridge: Cambridge University Press.
Douglas, S., & Craig, C. (2007). Collaborative and iterative Translation: An alternative
approach to back translation. Journal Of International Marketing, 15(1), 30e43.
Doupnik, T., & Riccio, E. (2006). The influence of conservatism and secrecy on the
interpretation of verbal probability expressions in the Anglo and Latin cultural
areas. The International Journal of Accounting, 41(3), 237e261.
Doupnik, T., & Richter, M. (2003). Interpretation of uncertainty expressions: A cross-
national study. Accounting, Organizations and Society, 28(1), 15e35.
Doupnik, T., & Richter, M. (2004). Impact of culture on the interpretation of “in-
context” verbal probability expressions. Journal of International Accounting
Research, 3(1), 1e20.
Durocher, S., & Gendron, Y. (2011). IFRS: On the docility of sophisticated users in pre-
serving the ideal of Comparability. European Accounting Review., 20(2), 233e262.
EC. (1958). Language regulation of EC (1/1958).
EC. (2002). Regulation (EC) No 1606/2002 of the European parliament and of the
council of 19 July 2002 on the application of international accounting standards.
Erb, C., & Pelger, C. (2015). “Twisting words”? A study of the construction and
reconstruction of reliability in financial reporting standard-setting. Accounting,
Organizations and Society, 40, 13e40.
Evans, L. (2003). The true and fair view and the fair presentation override of IAS 1.
Accounting and Business Research, 33(4), 311e325.
Evans, L. (2004). Language, translation and the problem of international accounting
communication. Accounting, Auditing and Accountability Journal, 17(2), 210e248.
Evans, L. (2010). Observations on the changing language of accounting. Accounting
History, 15(4), 439e462.
Evans, L., Baskerville, R., & Nara, K. (2015). Colliding Worlds: Issues relating to
language translation in accounting and some lessons from other disciplines.
Abacus, 51(1), 1e36.
Gillis, P., Petty, R., & Suddaby, R. (2014). The transnational regulation of accounting:
Insights, gaps and an agenda for future research. Accounting, Auditing and
Accountability Journal, 26(6), 894e902.
54 J. Kettunen / Accounting, Organizations and Society 56 (2017) 38e54
Gray, S. (1988). Towards the theory of the cultural influence on the development of
accounting systems internationally. Abacus, 24(1), 1e15.
Gro€jer, J. (2001). Intangibles and accounting classifications: In search of a classifi-
cation strategy. Accounting, Organizations and Society, 26(7e8), 695e713.
Hatim, B., & Munday, J. (2004). Translation. An advanced resource book. London and
New York: Routledge.
Hayne, C., & Free, C. (2014). Hybridized professional groups and institutional work:
COSO and the rise of enterprise risk management. Accounting, Organizations and
Society, 39(5), 309e333.
Hellmann, A., Perera, H., & Patel, C. (2010). Contextual issues of the convergence of
the International Financial Reporting Standards: The case of Germany. Advances
in Accounting, 26(1), 108e116.
Hines, R. (1988). Financial accounting: In communicating reality we construct the
reality. Accounting, Organizations and Society, 13(3), 251e261.
Hofstede, G. (1980). Culture's consequences: International differences in work-related
values. Beverly Hills, CA: Sage Publications.
Holthoff, G., Hoos, F., & Weissenberger, B. (2015). Are we lost in Translation? The
impact of using translated IFRS on decision-making. Accounting in Europe, 12(1),
107e125.
Hopwood, A. (1994). Some reflections on 'The harmonizing of accounting within
the EU'. European Accounting Review, 2(3), 241e253.
House, J. (2006). Text and context in translation. Journal of Pragmatics, 38(3),
338e358.
Huerta, E., Petrides, Y., & Braun, G. (2013). Translation of IFRS: Language as a barrier
to comparability. Research in Accounting Regulation, 25(1), 1e12.
Humphrey, C., Loft, A., & Woods, M. (2009). The global audit profession and the
international financial architecture: Understanding regulatory relationships at a
time of financial crisis. Accounting, Organizations and Society, 34(6e7), 810e825.
IASB. (2001). IASB Insight, October 2001. International Accounting Standards
Committee.
IFRS. (2008). International financial reporting standards.
Jakobson, R. (1959/2000). On linguistic aspects of translation in Brower, R.
(Reprinted in Venuti, L.. In On translation (pp. 232e239) Cambridge MA: Har-
ward university press Accessed online http://www.stanford.edu/~eckert/PDF/
jakobson.pdf)
Joseph, J. E. (1995). In M. Morris (Ed.), Interdermicy, translation and the law in trans-
lation and the law. Amsterdam/Philadephia: John Benjamins Publishing Company.
Kirk, N. (2006). Perceptions of the true and fair view concept: An empirical
investigation. Abacus, 42(2), 205e235.
Kosmala-MacLullich, K. (2003). The true and fair view construct in the context of
the Polish transition economy: Some local insights. European Accounting Review,
12(3), 465e487.
Kosmala-MacLullich, K. (2005). True and fair view or rzetelny i jasny obraz1? A
survey of polish practitioners. European Accounting Review, 3(14), 579e602.
Kristeva, J. (1980). Desire in language: A semiotic approach to literature and art. New
York: Columbia University Press.
Laswad, F., & Mak, Y. (1997). Interpretations of probability expressions by New
Zeeland standards setters. Accounting Horizons, 11(4), 16e23.
Latour, B. (1987). Science in action: How to follow scientists and engineers through
society. Cambridge, MA: Harvard University Press.
Latour, B. (1994). On technical mediation, philosophy, sociology, genealogy. Com-
mon Knowledge, 3(2), 29e64.
Lawrence, T., & Suddaby, R. (2006). Institutions and institutional work. In R. Clegg,
S. Hardy, T. Lawrence, & W. Nord (Eds.), Sage handbook of organization studies
(2nd ed., pp. 2015e2254). London: Sage.
Lawrence, T., Suddaby, R., & Leca, B. (2009). In Lawrence, Suddaby, & Leca (Eds.),
Introduction: Theorizing and studying institutional work in institutional Work:
Actors and agency in institutional studies of organizations. Cambridge: Cambridge
University Press.
Lawrence, T., Suddaby, R., & Leca, B. (2011). Institutional Work: Refocusing insti-
tutional studies of organization. Journal of Management Inquiry, 20(1), 52e58.
Lefevere, A., & Bassnett, S. (1998). Introduction. Where are we in translation
studies?. Essays on literary translation In S. Bassnett, & A. Lefevere (Eds.),
Constructing cultures (pp. 1e11). Clevedon: Multilingual Matters.
Macintosh, N., Shearer, T., Thornton, D., & Welker, M. (2000). Accounting as simu-
lacrum and hyperreality: Perspectives on income and capital. Accounting, Or-
ganizations and Society, 25(1), 13e50.
Mennicken, A. (2008). Connecting worlds: The translation of international auditing
standards into post-Soviet audit practice. Accounting, Organizations and Society,
33(4e5), 384e414.
Miles, M., & Huberman, M. (1994). Qualitative data analysis: An expanded sourcebook
(2nd ed.). Thousand Oaks: Sage.
Mills, P. (1989). Words and the study of accounting history. Accounting, Auditing and
Accountability Journal, 2(1), 21e35.
Mourier, L. (2004). Communicating financial reporting across continents. Hermes.
Journal of Linguistics, 32, 145e165.
Munday, J. (2012). Introducing translation studies. London and New York: Routledge.
Nida, E. (1964). Toward a science of translating: With special reference to principles
and procedures involved in Bible translating. Leiden, Netherlands.
Nida, E., & Taber, C. (2003). The theory and practice of translation. Leiden: Brill.
Nobes, C. (2006). The survival of international differences under IFRS: Towards a
research agenda. Accounting and Business Research, 36(3), 233e245.
Nobes, C. (2009). The importance of being fair: An analysis of IFRS regulation and
practice e a comment. Accounting and Business Research, 39(4), 415e427.
Nobes, C. (2013). The continued survival of international differences under IFRS.
Accounting and Business Research, 43(2), 83e111.
Nord, C. (1997). Translating as a purposeful activity. Functionalist approaches
explained. Manchester: St. Jerome.
Ordelheide, D. (1993). True and fair view. A European and a German perspective.
European Accounting Review, 1, 81e90.
Pelger, C. (2016). Practices of standard-setting e an analysis of the IASB's and FASB's
process of identifying the objective of financial reporting. Accounting, Organi-
zations and Society, 50, 51e73.
Reiss, K. (1977/1989). Text types, translation types and translation assessment. In
A. Chesterman (Ed.), Readings in translation theory (pp. 75e95). Helsinki: Finn
Lectura.
Reiss, K., & Vermeer, H. J. (1986). Mita€ ka
€a€nta
€minen on: Teoriaa ja ka
€yta
€nto
€a€. Helsinki.
Finland: Gaudeamus.
Rieu, E., & Phillips, J. (1954). Translating the gospels. Concordia Theological Monthly,
25, 754e765.
Samsonova-Taddei, A., & Humphrey, C. (2015). Risk and the construction of a Eu-
ropean audit policy agenda: The case of auditor liability. Accounting, Organi-
zations and Society, 41(1), 55e72.
Sapir, E. (1949). The status of linguistics as a science. In D. G. Mandelbaum (Ed.),
Selected writings of Edward Sapir in language, culture, and personality. Berkeley/
Los Angeles, Washington DC: University of California Press, 2.
Siggelkow, N. (2007). Persuasion with case studies. Academy of Management Journal,
50, 20e24.
Simon, J. (2002). Interpretation of probability expression by financial directors and
auditors of UK companies. The European Accounting Review, 11(3), 601e629.
Solin, A. (2004). Intertextuality as mediation: On the analysis of intertextual re-
lations in public discourse. Text - Interdisciplinary Journal for the Study of
Discourse, 24(2), 267e296.
Somssich, R., Sonnevend, P., Barret, P., Zakekas, F., Szabados, T., & Grosu, M. (2012).
Studies on translation and multilingualism. Language and Translation in Inter-
national Law and EU Law. European Commission. Accessed at http://ec.europa.
eu/dgs/translation/publications/studies/language_translation_international_
eu_law_en.pdf.
Suddaby, R., Cooper, D., & Greenwood, R. (2007). Transnational regulation of pro-
fessional services: Governance dynamics of field level organizational change.
Accounting, Organizations and Society, 32(4e5), 3333e3362.
Suddaby, R., Saxton, G., & Gunz, S. (2015). Twittering change: The institutional work
of domain change in accounting expertise. Accounting, Organizations and Soci-
ety, 45(5), 52e68.
Sunder, S. (2011). Ifrs monopoly: The pied piper of financial reporting. Accounting
and Business Research, 41(3), 291e306.
Toury, G. (1995). Descriptive translation studies and beyond. Amsterdam/Philadel-
phia: John Benjamins.
Venuti, L. (1998). The scandals of translation: Towards an ethics of difference. London;
New York: Routledge.
Walton, P. (1993). Introduction: The true and fair view in British accounting. Eu-
ropean Accounting Review, 2(1), 49e58.
Welch, C., & Marchan-Piekkari, R. (2006). Crossing language boundaries: Qualitative
interviewing in international business. Management International Review, 46(4),
417e437.
Wong, P. H. Y. (2004). Challenges and successes in implementing international stan-
dards: Achieving convergence to IFRSs and ISAs. New York: International Feder-
ation of Accountants. http://www.ifac.org/sites/default/files/publications/files/
challenges-and-successes-in.pdf.
Young, J. (2003). Constructing, persuading and silencing: The rhetoric of accounting
standards. Accounting, Organizations and Society, 28(6), 621e638.
Young, J. (2006). Making up users. Accounting, Organizations and Society, 31(6),
579e600.
Zambon, S., & Zan, L. (2000). Accounting relativism: The unstable relationship be-
tween income measurement and theories of the firm. Accounting, Organizations
and Society, 25(8), 799e822.
Zeff, S. (2007). Some obstacles to global financial reporting comparability and
convergence at a high level of quality. The British Accounting Review, 39, 290e302.
Internet References (all accessed 31.5, 2016)
IFRS Foundation: IFRS Foundation and its translation policies http://www.ifrs.org/
Useþaroundþtheþworld/IFRSþtranslations/IFRSþtranslations.htm.
IFRS Foundation: Available translations. http://www.ifrs.org/Use-around-the-
world/IFRS-translations/Pages/Available-translations.aspx.
IFRS Foundation: Official translation process and policies. http://www.ifrs.org/Use-
around-the-world/IFRS-translations/Pages/Official-translation-process-and-
policies.aspx.
IFRS Foundation. (2010). Translations update, 14.10.2010 http://www.ifrs.org/
Documents/files/TransUpdateJul10.pdf.
IFRS Foundation. (2012). Translations update. October 2012 http://media.ifrs.org/
2012/Translations/TransUpdateOct2012.pdf.
Thyssen, M. (2011). Question for written answer to the Commission. http://www.
europarl.europa.eu/sides/getDoc.do?pubRef¼-//EP//TEXTþWQþP-2011-
008747þ0þDOCþXMLþV0//EN.
Vassilou. (2011). Answer given on behalf of the Commission. http://www.europarl.
europa.eu/sides/getAllAnswers.do?reference¼P-2011-008747&language¼EN.
 Accounting, Organizations and Society 56 (2017) 55e67

Contents lists available at ScienceDirect

Accounting, Organizations and Society

journal homepage: www.elsevier.com/locate/aos

Representing the market perspective: Fair value measurement for

non-financial assets
Richard Barker a, *, Sebastian Schulte b
a
Saïd Business School, Oxford University, United Kingdom
b
Judge Business School, Cambridge University, United Kingdom

a r t i c l e i n f o a b s t r a c t

Article history: Fair value measurement (FVM) in IFRS calls for a market-oriented representation of economic ‘reality’,
Available online 16 January 2015 whereby the values attributed to rights (assets) and obligations (liabilities) are in principle determined
from the perspective of the ‘market participant’ rather than that of the reporting entity. We argue,
however, based upon Searle’s analysis of institutional reality, that such rights and obligations exist and
are knowable only under certain conditions, that when those conditions hold FVM is not distinctive, and
that when they do not hold the requirements of FVM are wishful and incoherent. Based upon this
analysis, and using case study data, we explore how FVM is applied in practice to non-financial assets.
We find, for a predominance of core operating assets, that fair value is unknowable, because of the
absence of the institutional reality on which the FVM idea implicitly depends. In these cases, actors’
representations of fair value were found to be expedient, unstable and ultimately in direct contradiction
of the market participant’s perspective that is ‘wished-for’ in IFRS.
© 2016 Published by Elsevier Ltd.

Introduction Instead of accounting representations being neutral, Davis et al.

At the heart of the method of financial accounting is the balance
sheet, which is a representation of economic ‘reality’, in the form of
a summary of the rights (assets) and obligations (liabilities) of a
reporting entity. That representation, undertaken in accordance
with prevailing accounting standards, can be viewed as a process of
translating activities and events into financial metrics (Robson,
1991) and, thereby, as a mechanism for enabling economic
discourse with respect to those activities and events (Burchell,
Clubb, Hopwood, Hughes, & Nahapiet, 1980; Hopwood, 1987).
The accounting process is not neutral in this regard, however, and
what counts as an accounting representation changes over time
(Burchell, Hopwood, & Clubb, 1985; Davis, Menon, & Morgan, 1982;
Hines, 1988; Miller, 1998). The map-making metaphor proposed by
standard-setters themselves is therefore rejected by Hines (1991),
while Young (1994) describes accounting problems as being con-
structed as opposed to simply ‘being there,’ and Young and
Williams (2010) identify the value judgements that are unavoid-
ably implicit in identifying and classifying amounts required to be
recognised in the financial statements.
* Corresponding author.
(1982) and Morgan (1988) view them instead as partial, being
determined by reference to the defining attributes of a chosen
metaphor. So, for example, representation according to an ‘ac-
counting as history’ metaphor is concerned with ‘providing a
faithful record of the transactions of an enterprise’ (Morgan, 1988).
The image of accounting here is one of serving the function of social
memory (Basu, Kirk, & Waymire, 2009) and of providing an account
of the stewardship of resources (Ijiri, 1983; Murphy, O’Connell, &
O’hOgartaigh, 2013). This leads to an emphasis being placed on
information that is perceived to be reliable and verifiable, that can
be ‘counted on’ to underpin consensus in preparation and use (Ijiri
& Jaedicke, 1966). These are informational characteristics typically
associated with historical cost accounting (Ijiri, 1983). An alterna-
tive metaphor, also proposed by Morgan (1988), is ‘accounting as
economics’, which is ‘the view that accounting should try to mirror
current economic realities and reflect basic economic principles.’
While the development of economic thought can be said to have
itself built upon a metaphor of accounting practice (Klamer &
McCloskey, 1992), ‘accounting as economics’ reverses that causa-
tion. The image here is one of economic theory guiding accounting
practice, with accounting information seeking to capture current
market prices and to inform current, economic decision-making.
Varying in emphasis, for example from seeking to determine

http://dx.doi.org/10.1016/j.aos.2014.12.004
0361-3682/© 2016 Published by Elsevier Ltd.
56 R. Barker, S. Schulte / Accounting, Organizations and Society 56 (2017) 55e67
profit as a measure of economic performance (e.g. Edwards & Bell,
1961; Sterling, 1970) to simply viewing accounting information as
an input into investment decision-making (Beaver, 1989), a com-
mon theme here, in contrast with ‘accounting as history’, is an
appeal to the underlying discipline of economics, and so to the
importance of market signals and expected cash flows in the
determination of accounting practice.
In this paper, we interpret the introduction of fair value mea-
surement (FVM) in IFRS as a shift in metaphorical construct, from
the historical to the economic, and so also as a shift in the ac-
counting representation of economic ‘reality’. Our aim is to explore
the nature of this change, and to make a research contribution that
is both theoretical and empirical.
The paper is structured as follows. In Section ‘Accounting as
economics: the fair value idea’, we position fair value in IFRS 13 as a
transforming idea, as an extension in accounting of an underlying
logic of financial economics, which introduces a concept of ‘the
market’ to replace a more traditional, transaction-based perspec-
tive of accountability and stewardship (Morgan, 1988; Power, 2010;
Ravenscroft & Williams, 2009). Explicitly shifting the emphasis in
accounting practice, from what the IASB terms the ‘reliable’ mea-
surement of carrying amounts in the balance sheet, to the IASB’s
more recent conception of the ‘faithful representation’ of what the
current values of assets might hypothetically be, FVM points to-
wards change in the accounting representation of economic ‘re-
ality’. This idea of fair value in IFRS 13 contains several implicit
assumptions, both about the ontological nature of the ‘reality’ that
is being represented in accounting, and also about epistemological
claims that can be made with respect to that ‘reality’. In Section
‘Searle’s analysis of the nature of institutional reality’ we draw upon
Searle’s analysis of social reality (Searle, 1995, 2010), in order to
provide a theoretical basis from which these implicit assumptions
in IFRS 13 can be made explicit, and thereby better understood. Of
central importance here is Searle’s notion of an ‘institutional fact’,
which is something that can be known objectively about an insti-
tution. In turn, and as will be explored in greater depth in Section
‘Searle’s analysis of the nature of institutional reality’, an institution
in Searle’s analysis is an agreed-upon means for creating rights and
obligations among agents (Searle, 2005). Financial reporting, being
concerned with economic rights and obligations, can therefore be
seen as a system for representing institutional facts. This notion is
applied in Section ‘FVM and the representation of institutional
facts’, which brings together the previous two sections, applying
theory in Searle to the analysis of FVM in IFRS 13. We argue that the
process of representing fair values in financial statements does not
in itself involve the creation of institutional facts. Rather, the
preparation of financial statements involves either the reporting of
institutional facts already in existence, or else the creation of data
that cannot themselves constitute new institutional facts. We show
that the second of these two possibilities undermines the gener-
ality of the fair value idea that is presupposed in IFRS 13, because
there exists neither an institutional reality to be represented nor
the possibility that such a reality can be ‘known’. We also show that
this undermines the IASB’s conceptual shift from the notion of
‘reliability’ to that of ‘faithful representation’, because that shift is
redundant when institutional facts are already in existence, while it
is incoherent when they do not already exist. In Sections ‘Field-
work’ and ‘Case study evidence’, we turn to the empirical compo-
nents of our paper. Section ‘Fieldwork’ summarises our research
method, which employs case study evidence, from companies in
Germany, Switzerland and the UK, to explore how FVM is inter-
preted and applied in practice. Section ‘Case study evidence’ then
sets out the case study evidence from our fieldwork. We find that,
faced with the conundrum that there is an absence of institutional
facts, yet also a requirement to implement IFRS 13, preparers of
financial accounts sought to represent fair value in one or more of
the following ways: transferring the problem elsewhere, narrowing
the problem to make it more tractable, or finding an expedient
solution by subverting the requirements of IFRS 13. We suggest that
these practices led to a varied and inherently unstable accounting
representation of fair value, at odds with the fair value idea that is
wished-for in IFRS 13. In Section ‘Conclusion’ we bring together
theory and evidence, drawing out and discussing the main findings
and implications of our research.
Accounting as economics: the fair value idea
Fair value has emerged in recent years as the preferred mea-
surement model of the IASB. As described by a former IASB board
member, ‘fair value meets the conceptual framework criteria better
than other measurement bases considered’ (Barth, 2007). Along
similar lines, another former board member concluded that ‘fair
value is here to stay ... conceptual support for fair value is
demonstrable’ (McGregor, 2007). Fair value has duly appeared in
accounting standards introduced by the IASB, such as IFRS 2, IFRS 3
and IFRS 9, as well as in revisions to earlier standards, such as IAS 16
and IAS 36 (IASB, 2013), and most recently in proposed revisions to
the IASB’s Conceptual Framework.1 While the 2008 credit crisis
somewhat tarnished the status of FVM, any serious alternative to
fair value in IFRS has been conspicuous by its absence. Indeed, IFRS
13 is unique in being the only official IASB pronouncement with the
explicit purpose of setting out a theory and practice of
measurement.
There are two respects in which the adoption of FVM implies a
transition from ‘accounting as history’ to ‘accounting as economics’.
First, in contrast with a more traditional, transaction-based
perspective, FVM is explicitly market-oriented, whereby the
perspective of the market is proposed as the foundation for ac-
counting representation. Second, IFRS 13 insists on the generality of
a market perspective, whereby the objective to report market
values holds whether or not markets themselves exist, which leads
to what Power (2010) terms a ‘transformation of reliability’. In this
section of the paper, we set out and discuss these two defining
aspects of the fair value idea.
At the heart of IFRS 13 is a distinctive concept of ‘the market’.
This is consistent with the ‘accounting as economics’ metaphor,
with its emphasis on current market prices as a guide to economic
decision-making. The centrality of the market concept is indicated
in the stated objective of FVM in IFRS 13, which is ‘to estimate the
price at which an orderly transaction to sell the asset or to transfer
the liability would take place between market participants at the
measurement date under current market conditions’ (para 2; italics
in the original). Implicit in the characterisation of value as ‘fair’ is
that it represents the outcome of an arm’s length agreed exchange,
of an ‘orderly transaction ... between market participants ... under
current market conditions.’
In the following extract from its conceptual framework, the FASB
summarises the ideological appeal of this market orientation (FASB,
2000, para. 26):2
Among their many functions, markets are systems that transmit
information in the form of prices. Marketplace participants attri-
bute prices to assets and, in doing so, distinguish the risks and
rewards of one asset from those of another ... An observed market
price encompasses the consensus view of all marketplace
1
It should be noted that the incorporation of fair value in proposed revisions to
the IASB’s Conceptual Framework is in the context of a mixed measurement model,
as opposed to a ‘full fair value’ approach.
 R. Barker, S. Schulte / Accounting, Organizations and Society 56 (2017) 55e67 57

participants about an asset or liability’s utility, future cash flows, market participant that holds the asset’ (para. 21), even though ‘an
the uncertainties surrounding those cash flows, and the amount entity need not identify specific market participants’ (para. 23) and
that marketplace participants demand for bearing those that it ‘need not undertake exhaustive efforts to obtain information
uncertainties. about market participant assumptions’ (para. 89).3 As Bougen and
Young (2011) describe, there is a sort of ‘longing’ in this require-
This is an abstract, instinctive appeal, inspired not least by
ment to represent an arm’s length transaction between hypothet-
Hayek, to the informational system of ‘the market’ as the desired
ical market participants. There is the wish that an imagined reality
basis for accounting representation (Hayek, 1945; Slater & Tonkiss,
might itself become real, assuming the role that Baudrillard (1983)
2001, chap. 2). The cultural authority of financial economics comes
defines as a ‘hyperreality’ (Macintosh, Shearer, Thornton, & Welker,
through loud and clear, in the form of an endorsement of, first, the
2000). We have here an idea, which is initially born out of econo-
claimed superiority of ‘the market’ as the revealer of economic
mists’ theorising about revealed preferences in markets, which is
preferences and, second, the ‘textbook’ characterisation of the
then disembodied from the function of markets in practice and
valuation methods that underpin market-oriented calculative
carried into the realm of hypothetical markets, and which is then
practice (Power, 2010; Whitley, 1986).
finally called upon to act performatively in creating accounting
An implication of this economic logic is that there is no
representations in its own image. Indeed, to the extent that eco-
distinction in IFRS 13 between different ‘types’ of fair value, but
nomic theorising is itself historically grounded in the logic and
instead a single, unifying measurement objective. Stated differ-
practice of accounting, we have what Klamer and McCloskey (1992)
ently, the concept of an equilibrium, market-clearing price is an
describe as accounting ‘eating its own tail’, where theory abstracted
inexorable and unique outcome of economic logic; no alternative
from practice is taken as a guide to enacting practice.
ontology is conceptually permissible. In this paper, we call this the
IFRS 13 therefore maintains an insistence on the possibility of a
‘market ontology’. The market-orientation in IFRS 13 is therefore
clear distinction between the valuation perspective of the reporting
not partial but is instead pervasive and general. Yet this conceptual
entity and that of the market participant.4 This distinction is cen-
positioning raises a practical question, namely how should fair
tral, because without it the conceptual purity of the adopted eco-
value be determined in cases where market prices are not readily
nomic metaphor could not itself be maintained. The Introduction to
available? It is in the answer to this question that IFRS 13 introduces
IFRS 13 states this position forcefully: ‘fair value is a market-based
the second aspect of the idea of FVM, namely the insistence on
measurement, not an entity-specific measurement ... an entity uses
generality.
assumptions that market participants would use when pricing the
IFRS 13 explicitly acknowledges that different types of data are
asset ... an entity’s intention to hold an asset ... is not relevant when
available for the purposes of an accounting representation of fair
measuring fair value’ (para. IN9). We see here both aspects of the
value. In other words, while (in the sense described above) there is
idea of FVM: it is market-based and it can (indeed, it must) be
a single ontology e the market ontology as a universal concept of
applied universally by means of adopting the market participant’s
fair value e there is also acknowledgement of epistemological di-
perspective.
versity, of different types of data with which to make epistemo-
Power (2010) draws attention to the ‘transformation of reli-
logical claims with respect to the representation of fair value.
ability’ that characterises the transition being described here from
Specifically, ‘Level 1’ data are quoted prices in active markets, while
an historical, transaction-based accounting model to an economics-
‘Level 2’ data are observable inputs other than Level 1 data, and
based model, grounded in the expected cash flows signalled by
‘Level 3’ data are unobservable, meaning that they are the pre-
market prices. Citing Barth (2007) and others, Power argues that
parer’s own assumptions about ‘the assumptions that market par-
‘the very idea of reliability is being reconstructed … deep down the
ticipants would use when pricing the asset’ (IFRS 13, para 87).
fair value debate seems to hinge on fundamentally different con-
It is particularly at Level 3 that the second aspect of the fair value
ceptions of the basis of reliability in accounting.’ In this respect, the
idea becomes evident. Given the IASB’s insistence on the univer-
transformation in IFRS 13 is consistent with a broader re-
sality of FVM, even at Level 3 where ‘data are unobservable,’ IFRS 13
positioning in IFRS, which is signalled by the 2010 revision to the
require preparers of financial statements to hypothesise the exis-
IASB’s Conceptual Framework, in which the IASB replaced its
tence of markets, such that FVM simulates the price that would
concept of ‘reliability’ with an alternative concept termed ‘faithful
exist if the market existed (Bougen & Young, 2011; Bromwich,
representation’. While these two concepts are in several respects
2007; Laux & Leuz, 2009; Lennard, 2007; Walton, 2007). The
similar to one another, a defining difference is in the role played by
‘problem’ that Level 3 falls outside the traditional, transactional
the notion of ’verification’.5 While both concepts carry the repre-
scope of accounting is thereby ‘solved’ by extending the reach of
sentational ontology, rejected by Hines (1991), that ‘faithful rep-
accounting technique (Williams, 2002). While there are no eco-
resentation is the depiction in financial reports of the economic
nomic preferences revealed in actual market transactions, ac-
phenomena they purport to represent,’ it is only the concept of
counting representation must proceed as if there were. The novel
reliability that is defined to embody the epistemological condition
mechanism invoked to facilitate this move is that of the ‘market
of verifiability. Hence, when, in 2010, the IASB introduced ‘faithful
participant’. Employed rhetorically in much the same way as the
representation’ to replace ‘reliability’, it argued that ‘including
‘made-up user’ (Young, 2006), the market participant is the hypo-
verifiability as an aspect of faithful representation could result in
thetical actor to whom IFRS 13 appeals in the determination of fair
value. IFRS 13 calls upon preparers of financial statements to as-
sume that hypothetical market transactions are at arm’s length,
unforced and reasonably well-informed (BC55-59), and that an 3
Illustrative Example 1 in IFRS 13 does, however, illustrate the desirability of
asset (in its current state) is being valued as if it had been trans- greater information with respect to a realistic identification of a market participant.
4
ferred to the market participant for use in his or her own (hypo- Indeed, and in a move that seems to belie the (false) dichotomy between an
external ‘reality’ and an epistemologically objective representation of that ‘reality’
thetical) business. Specifically, IFRS 13 states that ‘fair value (Hines, 1991), IFRS 13 even allows an entity to assume that its own use of an asset
measurement shall ... (be) considered from the perspective of a approximates that of a market participant.
5
BC26 of the IASB’s Conceptual Framework states that ‘verifiability means that
different knowledgeable and independent observers could reach consensus,
although not necessarily complete agreement, that a particular depiction is a
2
Note that IFRS 13 was authored jointly by the IASB and the FASB. faithful representation’ (IASB, 2013).
58 R. Barker, S. Schulte / Accounting, Organizations and Society 56 (2017) 55e67
excluding information that is not readily verifiable ... (which) would
make the financial reports much less useful … (and so verifiability
is) desirable but not necessarily required’ (IASB, 2010; BC3.36).
Presumably for emphasis, the point is later repeated: ‘faithful
representation does not mean accurate in all respects ... for
example, an estimate of an unobservable price or value cannot be
determined to be accurate or inaccurate’ IASB, 2010; QC15).
There is an important departure here between ‘accounting as
economics’ in the fair value idea and its manifestation in earlier
attempts at current value accounting, notably Edwards and Bell
(1961), Chambers (1966) and Sterling (1970). While the ideas of
fair value and of current value share a commitment to current
market prices as a basis for ‘information-usefulness’, and while
they can both be said to differ from ‘accounting as history’ in this
regard, the current value idea does not share the readiness of the
fair value idea to downplay the importance of verifiability. Indeed,
such an approach is rejected as antithetical to the very purpose of
accounting with, for example, Edwards and Bell (1961) and Barton
(1974) making a critical distinction between subjective cash flow
forecasts and confirmatory accounting information, Chambers
(1965) stressing the discipline of measurability, and Sterling
(1970) seeking to ground accounting representation in economic
opportunities that are realisable as opposed to hypothetical. Fair
value therefore stands apart, as a distinctively idealistic and sub-
jective interpretation of the ‘accounting as economics’ metaphor.
There are two points being made here. The first is to reinforce
Power’s (2010) observation that in the revisions to the Conceptual
Framework, as in the presumed universality of fair value in IFRS 13,
there is an underlying transformation of the concept of ‘reliability’
in IFRS, whereby faithful representation, or equivalently the uni-
versality of the market ontology, is deemed to be possible even
when epistemological claims cannot be verified. The second point
is that this transformation makes ontological and epistemological
assumptions, and that these need to be unpacked in order that the
nature of the transformation can be better understood. That
unpacking is the subject of the next two sections of the paper. We
will start this process in Section ‘Searle’s analysis of the nature of
institutional reality’, by introducing Searle’s theory of institutional
reality, and in Section ‘FVM and the representation of institutional
facts’ we will then bring Searle’s theory to apply to the discussion of
FVM above. In turn, this will lay the foundation for the empirical
component of the paper, in Sections ‘Fieldwork’ and ‘Case study
evidence’.
Searle’s analysis of the nature of institutional reality
We have argued so far that IFRS 13’s introduction of FVM implies
transformation in accounting representation, and that there is an
implicit ontology and epistemology to be understood here. In this
section, we start to explore this issue by setting out Searle’s theory
of institutional reality (Searle, 1995, 2010; see also Shapiro, 1997;
Mouck, 2004; McKernan, 2007, for discussion of Searle in an ac-
counting context). We will draw a conclusion that is broadly similar
to that in Berger and Luckmann (1966, p78), that ‘the objectivity of
the institutional world … is a humanly produced, constructed ob-
jectivity.’ In contrast with Berger and Luckmann, however, and as
will be evident especially in Section ‘FVM and the representation of
institutional facts’, Searle’s theory is particularly pertinent to this
paper because it enables us to explore the philosophical founda-
tions of this institutional reality, as opposed to being interested
primarily in how our sense of that reality is constructed (Elder-Vass,
2013).
We restrict our discussion to those basic elements of Searle’s
theory that contribute directly to the paper. These elements, which
will be described below, are: first, intentionality and the directions
of fit of intentional states; second, collective intentionality and
conditions of satisfaction; third, status function declarations; and,
fourth, the ontological and epistemological status of institutional
facts.
Intentionality is ‘that capacity of the mind by which it is directed
at, or about, objects and states of affairs in the world … intentional
states are always about, or refer to, something’ (Searle, 2010, p25).
Each intentional state is of a given type (e.g. belief, fear, desire) and
with a specific content, for example the propositional content that
‘the sun is shining.’ Different intentional states can be said to have
different directions of fit (Searle, 2010, p27). If the intentional state
is a belief, then it is said to have a ‘mind-to-world direction of fit’,
meaning that the belief can be said to be true or false: either the sun
is shining, or it is not, hence the condition of satisfaction of the
intentional state is whether the sun actually is shining. If the
intentional state is a desire, then it is said to have a ‘world-to-mind
direction of fit’, meaning that the desire can be either satisfied or
frustrated: either the sun will come out and satisfy the desire, or it
will not, hence the condition of satisfaction of the intentional state
is whether or not the sun will come out.6
Central to Searle’s interpretation of the social is the notion of
collective intentionality, which is a precondition of cooperative
behaviour, whereby the intentional state of an individual is shared
with those of other members of a group. Collective intentionality
can be surfaced by asking the question ‘what exactly is the collec-
tive trying to do?’ (Searle, 2010, p55). An example in Searle (1990) is
that of a group of people sitting outside, who each individually run
for shelter when it starts to rain. He contrasts this with the same
group performing the same actions, except this time in the context
of acting out a scene in a play where the actors run for shelter. In the
former case, there is intentionality for each individual in the form “I
intend to do x” while, in the latter case, there is a collective
intention in the form “we intend to do y.” There is likewise col-
lective intentionality in playing team sports, because the partici-
pants are consciously engaged in social behaviour, where what “I”
am doing is a part of what “we” are doing.
What the collective is trying to do is a question of causality, of a
social outcome that the collective is seeking to bring about. Searle
gives the example of a line of stones on the ground, which inter-
ested parties agree defines a boundary between properties. What is
happening is that the stones are assigned the function of being a
boundary, which causes it to be so, even though this function is
observer-relative and is not an intrinsic property of the stones
themselves. This function is created by virtue of collective inten-
tionality and sustained by means of collective recognition: the
stones act as a boundary because, and only because, all parties
agree that this is the case. The mechanism is that functions are
created by speech acts that are constitutive rules, which make
something the case by representing it as being the case. Searle
terms these speech acts ‘status function declarations’, which take
the logical form ‘X (object) counts as Y (function) in C (context).’
Hence, on the basis of collective intentionality with respect to the
stones, the statement ‘we declare that these stones (X, object)
count as a boundary (Y, function) in the relationships among
interested parties (C, context)’ is constitutive of the institutional
reality that a boundary has been created, the subsequent existence
of which depends entirely on its collective recognition. Likewise,
pieces of paper function as bank notes (or money) because we
intend them to do so, and the mechanism is that collective
6
It might be noted that there is a certain ontology being assumed in these ex-
amples, namely that ontologically objective existence is possible. As will later be
described, however, Searle’s analysis is consistent with viewing all forms of ac-
counting as having an ontologically subjective mode of existence.
 R. Barker, S. Schulte / Accounting, Organizations and Society 56 (2017) 55e67
intentionality confers a function on the piece of paper that creates a
new institutional reality. In general, an institution is a system of
constitutive rules, and such a system automatically creates the
possibility of institutional facts. In this way, Searle (2010, p13) ar-
gues that ‘all human institutional reality is created and maintained
in existence by status function declarations.’
Four important points should be noted here. First is that the
creation and maintenance of institutional facts requires collective
intentionality, status function declarations and collective accep-
tance; institutional facts do not otherwise exist. Second is that
functions in general serve a purpose, and that, in the case of status
function declarations, the primary purpose is to create and regulate
power relationships, for example rights and obligations with
respect to a property boundary, a monetary amount or a contract.
The reason for the constitutive rule that creates the boundary is to
determine and regulate the power relationships among interested
parties; otherwise, the function would serve no purpose. Searle
uses the term ‘deontic power’ in this context, where rights and
obligations are, respectively, positive and negative deontic powers.
The third point to note is that status function declarations need not
be bestowed upon objects, as for example stones in the example
above. They can instead be created ‘out of thin air’, adopting what
Smith (2003) terms ‘freestanding Y terms.’ An example here is
electronic money, where a status function is assigned, and collec-
tively recognised, without being bestowed on any object. Yet Searle
notes that ‘the freestanding Y terms always bottom out in actual
human beings who have the powers in question because they are
represented as having them ... you don’t need to have a physical
realisation to have money or a corporation ... but you do have to
have owners of money and officers and shareholders of a corpo-
ration’ (Searle, 2010, 108e9). Finally, the fourth point to note is that
while all institutional reality created by means of status function
declarations is ontologically subjective, because it does not exist
independently of people, this does not preclude the possibility that
it is epistemologically objective, and thereby permits observer-
independent ascertainment of truth or falsity. While money is an
institution, and not an intrinsic property of pieces of paper, it is
epistemologically objective (an institutional fact) that, for example,
the closing share price of Pearson plc on 17 February 2014 was
£11.17. In short, we can make epistemologically objective claims
with respect to ontologically subjective social realities.
In brief summary, Searle’s account of institutional reality re-
quires that collective intentionality is enacted by means of status
function declarations, enabled by collective intentionality and
maintained by collective acceptance, which create (ontologically
subjective) institutions and thereby (epistemologically objective)
institutional facts.
FVM and the representation of institutional facts
Applying Searle’s theory, we can now start to explore the
ontological and epistemological foundations of FVM. We first note
that, in general, accounting represents an economic ‘reality’ that
does not exist independently of human experience. For example,
while something such as oil can be said to have an ontologically
objective existence, it is only by social consensus that it can be said
to exist as an ‘asset’ with an attribute such as ‘market value’; the oil
price that we record in the accounts is therefore ontologically
subjective. So, too, all accounting representations are ontologically
subjective, whether assets, liabilities or equity, because they do not
exist independently of human experience. As Searle argues, how-
ever, ontological subjectivity is not inconsistent with epistemo-
logical objectivity. For example, while the existence of an equity
claim depends subjectively upon human experience, the share
price at which the claim trades at any specific point in time is an
59
institutional fact that can be known.
What, then, of fair value? Consider first Level 1, which is defined
in terms of observable market prices, and which therefore invites
us to reflect upon the social and institutional reality of existing
markets. In the context of IFRS 13, while the general, conceptual
appeal to ‘the market’ is itself detached from the materiality of
markets in practice, any accounting representation that uses Level 1
data to enact IFRS 13 must relate to the complexities of a practical
setting. To illustrate, Callon (1998), MacKenzie (2009) and others
argue that, for market transactions to be possible in practice, and so
for markets to exist, there is a need for ‘framing’. This requires that
items being traded are capable of ‘disentanglement’, whereby
transfers of property rights between calculative agents require
(financial) measurability, separability and so transferability. In turn,
this depends upon the existence of a legal or regulatory framework,
standards of conduct and calculation, and some form of physical
structure or process that constitutes the marketplace itself. Roberts
and Jones (2009), Vollmer, Mennicken, and Preda (2009) and others
emphasise the nature of the market as a network, whereby market
transactions and the environmental contexts for those transactions
are co-determined. In short, we have here a complex, subjective
social ontology. Yet Searle’s analysis suggests that we also have a
product of this network that is a simple, epistemologically objec-
tive, institutional fact which (like money) conveys a deontology of
economic rights and obligations; namely, the market price. This
institutional ‘facticity’ is important from an accounting perspective,
because an accounting representation using Level I data can be
viewed simply as a recording of an institutional fact, as a ‘mapping’
from a price recorded in the market to the same price reported in
the financial statements. While such a mapping may require the
IASB to act as a regulative rule-maker, indicating how the fair values
should be presented in the financial statements, there is no
constitutive rule here that would create a new institutional reality,
other than in the trivial sense that ‘the observed market price (X)
counts as the fair value (Y) in the context of financial reporting (C).’
To borrow the language of intentionality, there is a mind-to-world
direction of fit, where the IASB’s concept of verifiability is the
mechanism by which the reported institutional fact can be said to
be true or false. Moreover, there is no practical difference in this
case between the concept of ‘reliability’ and that of ‘faithful rep-
resentation’. This is because it is the market participant’s perspec-
tive itself that is being verified, such that establishing ‘reliability’ in
the IASB’s sense of the term implies making an epistemologically
objective claim to the market ontology, which has the same
meaning as the IASB’s concept of ‘faithful representation’.
This position stands in sharp contrast with Level 3 fair values,
where there is nothing to be observed and therefore epistemo-
logical subjectivity in the making-up of a market participant’s
perspective. While, at Level I, there is collective intentionality
among market participants (of the form ‘we are trading on a mar-
ket’), as well as a status function declaration with respect to market
prices (of the form ‘market prices count as economic claims’), at
Level 3 there cannot in principle be either collective intentionality
or status function declaration, because market participants do not
exist. In the sense described by Searle, therefore, institutional facts
do not exist at Level 3, and so they cannot be reported.
An implication of this absence of institutional facts is that there
is logical incoherence in the market ontology maintained in IFRS 13
and, more generally, in the IASB’s notion of ‘faithful representation’.
This is because the existence of institutional facts simultaneously
implies a subjective social ontology and an objective epistemo-
logical claim with respect to that ontology: there exists something
to be known, as well as an objective claim that it is known. At Level
3, however, IFRS 13 requires that the individual preparer of accounts
represents a hypothetical ontology. In contrast with Level I, this
60 R. Barker, S. Schulte / Accounting, Organizations and Society 56 (2017) 55e67
ontology has a world-to-mind direction of fit. It cannot in principle
be ‘known’ but is instead a ‘wished-for’ social outcome, whereby
the preparer is required to adopt the perspective of the market
participant in a neoclassical economic world of the IASB’s imagi-
nation. Hence, the market ontology at Level 3 is not observable as a
social outcome, in the manner of market prices at Level 1, but
instead it is inaccessible to the preparer and does not permit the
making of an epistemologically objective claim. Alternatively
stated, the absence of verifiability (and so reliability) arises jointly
with the absence of the possibility of faithful representation, for the
simple reason that there is no institutional fact to be represented in
the first place. The ontological status of FVM at Level 3 is therefore
fundamentally different from that at Level 1. While both are
ontologically subjective, the epistemological objectivity at Level I is
consistent with the single ‘market ontology’ described above, while
the epistemological subjectivity at Level 3 renders implausible the
preparer’s ability to represent this single ontology, thereby
rendering empty any claim to faithful representation. Moreover,
while the concept of faithful representation is meaningful in the
presence of institutional facts at Level I, it is also in this context
redundant because if the institutional facts have been verified, then
they are necessarily representationally faithful. In short, the IASB’s
notion of ‘faithful representation’ is meaningless at Level 3 and
redundant at Level 1.
Overall, we have neither a fully representational nor a fully
constructionist view of accounting, but instead a hybrid of the two.
In contrast with Hines (1988), communication does not construct
reality at Level 1, but instead an existing institutional reality is re-
ported. At Level 3, there is a form of reality construction in that fair
values are imagined and reported in a way that does not correspond
with an existing institutional reality. Yet, in Searle’s sense, we do
not have the creation of new institutional facts in this case either,
because there is no collective intentionality in generating these fair
values. This is not to deny that the subsequent usage of these re-
ported data might not itself lead to the creation of new institutional
facts. This is possible because reported fair values are observable
and can form a basis for collective agreement and action. We are,
however, making a distinction here, which is otherwise conflated in
the literature, between the accounting representation itself and the
subsequent (social) use of accounting representations, and we are
arguing that the accounting representation itself does not create
new institutional facts, either at Level 1 or at Level 3. Accounting
data are first produced and then they are used; the former process
e the accounting representation itself e does not lead to the cre-
ation of institutional facts.
In the empirical sections of our paper, to which we now turn, we
present evidence concerning how the accounting representation
itself is made. We have argued that IFRS 13 presupposes a market
ontology that cannot in principle be a general basis for accounting
representation. Yet the implementation of IFRS 13 by companies is
nevertheless not optional. Our evidence therefore concerns how
preparers of accounts enact the (unrealistic) market ontology of
IFRS 13 in making accounting representations of fair value.
Fieldwork
Our fieldwork was designed to explore the implementation in
practice of IFRS 13, in particular in what ways the representation of
fair value constitutes either the reporting of pre-existing institu-
tional facts or the ‘making-up’ of a market perspective and, if the
latter, what characterises the process of accounting in the absence
of the creation of new institutional facts.
We adopt a case study-based research design, investigating the
fair value determination practice of publicly listed companies in the
pharmaceutical (P), electricity utility (E), telecom (T) and general
industrial (I) sectors, in Germany, Switzerland and the UK, all of
which applied IFRS as primary accounting standards. We employ
the case study method (Cooper & Morgan, 2008) in seeking to go
beyond the abstraction and assumption of the ‘accounting as eco-
nomics’ metaphor adopted in IFRS 13, and to explore the enactment
in practice of IFRS 13’s distinctive approach to ‘the market’ as the
basis of accounting representation.7 Given the open-ended and
exploratory nature of our study, we sought to select case studies
based upon an information-oriented selection. In contrast with a
random selection, which aims to avoid systematic bias in the
sample, an information-oriented selection acknowledges that a
typical or average case is often not the richest in information
(Flyvbjerg, 2006) and cases are instead selected on the basis of
expectations about their information content (Eisenhardt, 1989;
Ragin & Becker, 1992; Roesch, 1978), including extreme cases in
which the process of interest is ‘transparently observable’
(Pettigrew, 1995).
In seeking this richness of information, our selection of case
studies rested upon potential variety, scope and challenge of FVM
application in case settings. We took company size as a proxy for
complexity and diversity, with larger companies being potentially
more diverse by business and geography. Specifically, we aimed to
select listed companies that were the constituents of the national
main stock index, namely the DAX30 in Germany, the SMI25 in
Switzerland and the FTSE100 in the UK. Our choice of countries
increased the richness of information by drawing upon different
accounting traditions (Nobes, 1983).8 We also looked for diversity
across industries and for complex, economically significant non-
financial assets within industries. We excluded financial assets in
part to avoid scoping the fieldwork too broadly, but primarily
because of the information-oriented selection described above,
according to which we expected non-financial assets to provide
rich and diverse data, as the applicability of FVM can be viewed, a
priori, as challenging, given a likely absence of markets, insepara-
bility of assets and value generated by assets in use rather than in
exchange. For pharmaceuticals, while production is not a capital-
intensive process, the centrality of intellectual capital generates
complex and unique intangible assets such as patents, licences and
other subclasses of long-lived intangible assets. For electricity
utility companies, specialised assets are employed across a range of
oil, natural gas, coal, nuclear power, hydro power, renewable fuel,
wind turbine and photovoltaic technologies, with additional
complexity resulting from the interaction between different assets
in a network and because of the effects of regulatory control. Tel-
ecom companies also exhibit complex asset interactions and reg-
ulatory processes, while also combining substantial tangible assets,
in the form of telecom networks, with intangible assets, notably
7
Given the importance of FVM, there is surprisingly little empirical research that
is directly relevant to this paper. Most research addressing FVM adopts a market-
based, ‘archival-empirical’ methodology (Landsman, 2007), which has a ‘black
box’ character which bypasses the main themes of this paper. While this suggests
an important role for alternative research methods, such studies are few and far
between. An experimental method is employed in Gaynor, McDaniel, and Yohn
(2011) to evaluate users’ understanding of credit risk changes, for liabilities car-
ried at fair value. Gwilliam and Jackson (2008) employed a case study approach to
explore the manner in which Enron arrived at fair values, including identification of
the ease with which the company used special purpose entities to ‘monetize’
physical assets and so to achieve fair value accounting. The informational richness
of the case study approach is further illustrated by Benston (2006), who analysed
the escalating use of fair values by Enron, identifying the sources of the escalation,
the process by which (in particular) Level 3 fair values came increasingly to be used,
and the consequences for the financial health of the company.
8
Barrett, Cooper, and Jamal (2005) find variation even in standardised audit
procedures conducted by multi-national audit firms, and such variation might be
expected to even more pronounced across reporting entities from different
geographical and cultural settings.
 R. Barker, S. Schulte / Accounting, Organizations and Society 56 (2017) 55e67 61

Table 1
Case study participants.

Market capitalisationa Turnovera IFRS US-GAAP

Panel A: company size and accounting standards

T1 Top 5 Top 5 Yes Yesb
T2 Top 5 Top 5 Yes Yesb,c
T3 Top 10 Top 10 Yes Yesb
P1 Top 5 Top 5 Yes Yesb
P2 Top 20 Top 10 Yes No
P3 Top 5 Top 5 Yes Yesc
I1 Top 10 Top 10 Yes No
I2 Top 5 Top 5 Yes No
I3 Top 30 Top 50 Yes No
E1 Top 20 Top 20 Yes Yes
E2 Top 5 Top 5 Yes Yesb,c
E3 Top 50 Top 30 Yes Yes

Intangibles Goodwill Land and buildings Plant and equipment

Panel B: composition of long-lived, non-financial assets (% of total asset value)

Telecom 3.6e28.5 2.2e26.7 2.9e7.8 25.3e56.0
Avg. 12.2 Avg. 15.0 Avg. 5.3 Avg. 36.9
Pharmaceutical 2.6e15.5 3.0e15.7 7.1e10.9 9.0e11.6
Avg. 10.3 Avg. 9.7 Avg. 8.9 Avg. 10.4
General industrial 4.0e7.0 7.1e12.7 4.4e14.7 8.3e48.7
Avg. 4.4 Avg. 10.3 Avg. 10.0 Avg. 22.2
Electricity utility 0.9e4.3 1.7e11.8 1.7e4.2 17.9e28.9
Avg. 2.7 Avg. 5.6 Avg. 3.3 Avg. 23.1

T ¼ telecom; P ¼ pharmaceutical; I ¼ general industrial; E ¼ electricity utility.

a
In comparison with Western European peer companies.
b
Listed on the NYSE, involving reconciliation to US-GAAP and provision of SEC Form 20F.
c
Experience of change from US-GAAP to IFRS in current or previous year.

wireless licences and consumer-oriented brands. Finally, the gen-
eral industrials sector was considered to offer ‘textbook’ examples
of capital-intensive business, with property, plant and equipment
likely to be specialised in nature.
A selection frame was created by using the ‘Industry Classifi-
cation Benchmark’ (ICB), which comprised 28 pharmaceutical, 31
electricity utility, 18 telecom and 106 industrial companies in
Germany, Switzerland and the UK.9 While the recruitment of par-
ticipants proved to be challenging, we were successful in gaining
access to eleven companies, one short of our target but including
several large, global companies.10 As reported in Table 1, which
provides a summary of case study participants, the market capi-
talisation and turnover of P1, P3, E2, T1, T2, and I2 are all in the top 5
amongst their Western-European peers. They are joined by T3 and
E3 as being in the top 3 national peer group.
The core parts of the case studies were semi-structured in-
terviews, which enabled a detailed investigation of complex,
entity-specific and asset-specific issues. The interview questions
were adapted from case to case to account for the varying struc-
tural, national and regulatory settings of the case companies and, in
order to increase contextual understanding, a considerable amount
of additional data and sources were also used in each case, such as
annual accounts, 20-F forms, interim reports, management reports
(also of main competitors) and publicly available documents or
presentations dealing with the implementation or adoption of IFRS.
9
The telecom subset comprises companies classified as ICB 6535 (Fixed Line
Telecommunications) and 6575 (Mobile Telecommunications), the pharmaceutical
subset comprises companies classified as ICB 4577 (Pharmaceuticals), the industrial
subset comprises companies classified as ICB 2757 (Industrial Machinery), 2753
(Commercial Vehicles & Trucks), 2727 (Diversified Industrials) and 2713 (Aero-
space). The electricity utility subset comprises companies classified as ICB 7535
(Electricity) and 7575 (Multiutility).
10
For the 12th company, an interview was agreed upon, but then postponed
several times. Eventually, the interviewee withdrew willingness to participate due
to time constraints.
We sought input only from senior members of the group financial
reporting department, and only in very large companies where
there was resource dedicated to high-level technical accounting
expertise. This need for depth of knowledge and experience of FVM
among interviewees arose simply to ensure, as far as possible, high-
quality, well-informed evidence. Interviewees were asked about
their actual IFRS experience with FVM and also about the applica-
bility of IFRS 13 in the specific contexts of assets in their busi-
nesses.11 Interviews lasted from 45 min to 4 h, and were conducted
in either English or German. In all but one case, the interviews were
conducted face-to-face, in Germany, Switzerland or the UK, be-
tween March 2006 and February 2007. The number of interviewees
in each interview ranged from 1 to 3, and in total 19 interviewees
contributed to the research.
Case study evidence
The discussion in Section ‘FVM and the representation on
institutional facts’ concluded that fair values are either reported as
pre-existing institutional facts, or else they are ‘made-up’ repre-
sentations, the preparation of which does not involve the creation
of new institutional facts. In presenting our case study evidence in
this section, we first highlight the importance of the second of
these two categories. We found that there was typically an absence
of Level 1 data for non-financial assets, as well as perceived epis-
temological limitations of Level 2 data with respect to the market
ontology of IFRS 13. The reporting of pre-existing institutional facts
was therefore largely precluded.
Notwithstanding these difficulties with respect to the absence of
institutional facts, there was nevertheless a requirement to
11
IFRS 13 was preceded by an Exposure Draft and a Discussion Paper. As IFRS 13 is
the current document, and as it does not differ significantly from the earlier doc-
uments, it is IFRS 13 that is referred to in this paper. At the time of the interviews,
however, the current document was the DP (and, under US GAAP, SFAS 157).
62 R. Barker, S. Schulte / Accounting, Organizations and Society 56 (2017) 55e67
implement IFRS 13, such that some form of accounting represen-
tation of fair value was unavoidable. In the second part of this
section of the paper we explore how preparers responded to this
dilemma. We found that representations of fair value were made in
one or more of the following ways: transferring the problem else-
where, narrowing the problem to make it more tractable, or finding
an expedient solution by subverting the requirements of IFRS 13.
In the final part of this section of the paper, we suggest evidence
that the above practice in applying IFRS 13 led to a varied and
inherently unstable representation of economic ‘reality’, as agents
noted the subjectivity in their own representations and considered
the incentives and interpretations of others.
An absence of institutional facts
A striking feature of the case studies was how limited in practice
were the availability of Level 1 data, especially for the core assets of
the case companies; the simple case of representing fair values by
reporting pre-existing institutional facts was largely irrelevant. This
arose is part because the cash-generating capacity of individual
assets often could not be separated from that of the business as a
whole, which led naturally to consideration of larger units of ac-
count, thereby drawing away from the transactional foundation of
‘accounting as history’. In the electricity utility business, for
example, E1 claimed that the core assets were ‘one big and insep-
arable unit,’ while a telecom interviewee noted that ‘the end
customer normally does not realise whether he is currently using a
3G or a 2G frequency band, nor does he pay a different price’ (T3),
implying that a market-based perspective did not permit cash flows
to be attributed to separable assets.
At the level of individual assets, also, it was widely claimed that
Level 1 data were unavailable. There were some limited exceptions,
yet these commonly applied (in impairment tests and purchase
allocations) only for assets that represented a modest component of
the balance sheet, such as cars, trucks and buildings. More gener-
ally, the nature of core assets for many of the companies precluded
the existence of observable market prices. An illustration, offered
by P3, was that the company had never been approached by a
competitor offering to buy one of its key brands, which were
described as ‘unique by their very nature;’ in contrast, ‘dossiers’ for
off-patent (generic) drugs, covering all product characteristics and
aspects of manufacturing, were frequently traded, yet at values not
reaching the materiality threshold in P3’s financial statements. The
uniqueness of core non-financial assets also applied for the most
part to the fixed and mobile telephony network assets in the tele-
com group of case study companies, the production, transmission
and distribution system in the electricity utility group and, to a
certain extent, the production plants and facilities of the industrial
case companies. The absence of market prices was therefore a
significant challenge, not a marginal one. There was nothing like
the continuous trading associated with an active market. As one
interviewee described:
‘There is neither a market for nuclear power plants nor for our
regional electricity transmission lines ... We own huge infrastruc-
ture assets which simply never change hands at all ... in terms of
Level 1, there is no data at all, nothing.’
E1
The difficulties described here were not restricted to the
absence of Level 1 data. They were also evident in the context of
Level 2 data, where the problem was not so much that data were
unavailable, but rather that (in substance) case study participants
challenged their epistemological status with respect to IFRS 13’s
market ontology. An example was provided by T2 which, for the
purposes of an annual goodwill impairment test, had been sought
to determine the fair value of a major operation in a foreign
country. Unlike in previous years, the past year had yielded Level 2
data, in the form of the price offered in a takeover bid by a global
competitor of another company in the industry. The interviewees
struggled, however, to make use of this market information, the
main reason being the perceived incomparability between their
own operation and the competitor’s operation. In particular, the
size of both businesses, the competitive environments in both
countries, the customer behaviour and a number of other value-
affecting factors were said to differ too much. One of the in-
terviewees at T2 summarised as follows:12
Let us assume we want to value our British CGU, but we only know
about a comparable transaction in Denmark; that would not help
us. There are too many differences to warrant an adjustment
approach. We cannot just say that the Danish business is equiva-
lent to the British one. This would be a desperate attempt to use a
mark-to-market approach.
T2
A similar example arose in the case of wireless licences, which
were not regularly traded yet where a recent auction in the US
(FCC-auction 66) was characterised as a ‘stroke of luck for mark-to-
market FVM for non-financial assets’ (T2).13 The possibility of using
the results from auction 66 to determine the fair value of the
existing portfolio of FCC-licences became heavily discussed in the
telecom industry. As reported in Table 2, however, alternative in-
terpretations of the market price were possible, with four ‘official’
alternatives being offered by the FCC, and these alternative
methods gave rise to significant variation in prices, both at a point
in time for any given auction, and also for the rate of price increase
or decrease from one auction to another.
T2 illustrated the effects of this variation in price for a licence
acquired in the Washington DC area, for which the range of po-
tential fair values lay between $29.9 million and $179 million, ac-
cording to whether valuation was undertaken at a portfolio or
individual licence level, whether adjustment was made for varia-
tion in bandwidth, and whether or not auction revenue was
weighted by the population served by the license.
In short, these telecom examples illustrate that while (episte-
mologically objective) market prices made possible the represen-
tation of fair value based upon some sort of market perspective,
they did not make accessible the market ontology demanded by
IFRS 13, being the market participants’ perspective specific to the
unique assets of the company. A similar frustration was evident in
the selection of discount rates, where market data were available in
abundance, yet where this resulted in indeterminacy, because none
of the multiplicity of alternative metrics had any greater episte-
mological claim than any other. It was stated, for example, that
because Bloomberg provides betas based on different markets, it
was possible to ‘get pretty much any value within a certain area by
simply changing the parameters on the Bloomberg engine’ (T3). In
addition, further subjectivity was caused by the choice of peer
companies, and several interviewees emphasised difficulties in
finding companies that operated in the same industry, with com-
parable operational and financial risk structures.
12
A CGU is a cash-generating unit.
13
The FCC is the Federal Communications Commission, which regulates the US
telecom industry.
 R. Barker, S. Schulte / Accounting, Organizations and Society 56 (2017) 55e67 63

Table 2
Estimates of historical prices in US wireless telecom licence auctions.

FCC auction Pricing method 1 Pricing method 2 Pricing method 3 Pricing method 4
(prices in $
p.c.)

Price equals: Total auction revenue Price equals: The average of revenue per licence Price equals: As method one, Price equals: As method two,
divided by total auction population divided by population of respective licence area adjusted for licence bandwidth adjusted for licence bandwidth
Auction 4 15.54 13.23 0.52 0.44
Auction 5 39.88 23.33 1.33 0.78
Auction 10 58.28 46.64 1.94 1.55
Auction 35 42.37 20.25 4.03 1.96
Auction 58 9.89 4.62 0.96 0.46
Auction 66 7.99 3.25 0.43 0.18

Faced with this absence of institutional facts, case study in-
terviewees rejected the notion that ‘making-up’ the perspective of
the hypothetical market participant was a meaningful alternative
route to the representation of fair value. In other words, there was a
rejection of the generality of the market participants’ perspective,
which is presumed in IFRS 13 and which was described in Section
‘Accounting as economics: the fair value idea’ as one of the two
defining ideas of fair value. To illustrate, P1 claimed to be unable to
distinguish between the (entity specific) value-in-use and the
(market-based) fair value of a brand, asserting that ‘it is impossible
to create a market where no market exists.’ Likewise, I2 explicitly
questioned ‘the possibility to reflect the perspective of a market
participant,’ while E2 saw ‘huge difficulties in distinguishing be-
tween market and entity specific values.’ Such a sentiment was
commonplace in the case studies. P1 discussed at some length the
conceptual possibility of FVM’s market orientation, yet concluded
with the stark observation that ‘we simply do not know the
external perspective.’ E1 summarised the conundrum as follows.
who decided what should count as fair value. This case is inter-
esting because it involved a switch from US GAAP to IFRS, which
made it possible to observe alternative accounting representations,
the difference being that discount rates were based either on
average data from three investment banks (CSFB, Citigroup and
Morgan Stanley) or on data from Bloomberg. As shown in Panel A in
Table 3, the unlevered betas of six competitors was averaged over
three annual periods, and the resulting beta of 1.15 was adjusted to
the actual debt-to-equity ratio of the particular CGU and used for
determining all FVMs in the operating segment in question. Panel B
Table 3
Illustrative example of subjective selection of discount rates.
Company Country Year Average
1/05e1/06 1/04e1/05 1/03e1/04

Panel A: peer group-based calculation of b

In theory, I could go to a third party and ask them to perform a Arcelor Canada 1.05 0.95 0.67 0.89
valuation, but how could a third party do it, when they are not Corus Group UK 1.72 1.50 n.a. 1.61
considering buying and operating the asset that I want to value? Rautaruukki Finland 1.60 0.45 0.31 0.79
Salzgitter Germany 0.82 0.34 n.a. 0.58
E1 SSAB Sweden 1.03 0.71 0.53 0.76
US Steel USA 2.92 2.78 1.18 2.29

Average 1.15
This rejection of the generality of the fair value idea is consistent
with the discussion in Section ‘FVM and the representation on Segment A B C D E F G H I J K
institutional facts’, where it is argued that preparers of accounts
Panel B: US-GAAP vs. IFRS assumptions
cannot in principle ‘make-up’ institutional facts that would
CSFB/Citigroup/ 0.77 0.58 0.77 0.28 0.44 0.60 0.46 0.83 0.49 0.55 0.53
constitute IFRS 13’s market ontology. In spite of this, however, the Morgan Stanley
case study companies were nevertheless faced with the practical (for US-GAAP)
requirement to implement IFRS 13 and to represent fair values, Bloomberg (for IFRS) 1.15 0.80 1.01 0.56 0.92 0.79 0.68 1.04 1.04 0.77 0.67
notwithstanding the inherent conceptual ambiguity underlying
Scenario bim rf rm  rf Fair value ‘Buffer’
that representation. Our evidence suggested that this was under-
taken in one or more of three ways: transferring the problem abs. (V) D abs. (V) D
elsewhere, narrowing the problem to make it more tractable, or Panel C: effects on the fair value of a CGU
finding an expedient solution by subverting the requirements of IFRS “base case” 1.15 0.0340 0.050 5,080,280 n.a. 926.554 n.a.
IFRS 13. These are discussed below. Scenario A 0.77 0.0450 0.050 6,198,180 22.0 2,044,454 121.0
Scenario B 1.15 0.0481 0.050 5,089,543 0.2 935,817 1.0
Scenario C 1.15 0.0450 0.047 5,253,782 3.4 1100.056 18.7
Representing fair value Scenario D 0.77 0.0481 0.047 6,383,492 25.6 2229.766 141.0

This example illustrates an impairment test for a cash-generating unit (CGU),

One way for the case companies to represent fair value was
simply to allow another organisation to take responsibility for it.
For example, E2 described mandating Standard and Poors to
determine fair value purchase price allocations, for the purposes of
accounting for business combinations under IFRS 3. Another
approach, described by P2, was the ‘outsourcing’ of WACC calcu-
lations to Bloomberg, which was motivated by the desire to avoid
‘endless and useless discussions’ with the auditors. In a different
context, I2 described in some detail a case where it was the auditor
comprising assets with an overall carrying amount of V4.15 million, including V0.89
million of goodwill. With the parameters used for IFRS, the fair value of the CGU
amounted to V5.08 million. Thus, the ‘buffer’ of the CGU was V0.93 million, which
first had to be used up before goodwill would need writing down. As reported under
scenario A, the effect of using the beta according to the data provided by CSFB/
Citigroup/Morgan Stanley (b ¼ 0.77) rather than that from Bloomberg (b ¼ 1.15)
would have been to increase the CGU’s fair value by 22%. Accordingly, the goodwill
impairment buffer would have increased by 121% to more than V 2.0 million. The
rather slight changes in rf and rm  rf have less strong effects on fair value (scenarios
B and C), but changing all parameters together would have increased the goodwill
buffer by 141% (scenario D).
64 R. Barker, S. Schulte / Accounting, Organizations and Society 56 (2017) 55e67
in Table 3 provides a side-by-side comparison of the betas ac-
cording to the two different methods for all fifteen of the company’s
operating segments. The differences are significant, with Bloom-
berg betas being substantially larger. Panel C, which concerns the
application of FVM to a CGU, and which was discussed in the
interview, illustrates that the effects of these differences amounted
to an increase of significantly greater than 100% in the goodwill
buffer.
While both sets of parameters were market-based and could
therefore claim to represent the market participant’s perspective, it
was the auditor that ‘negated suddenly’ the selection of one
method over the other, thereby arbitrarily determining which ac-
counting representation of fair value would be made. This arbi-
trariness was evident also in the interview with P2, where it was
argued that: ‘To speak quite frankly, when auditing a complex fair
value, no auditor can prove that a 5% revenue growth rate is incorrect
nor can they say it has to be 4.5 or 5.5% ... we are able to influence the
result dramatically just by minimally changing one of the more
important, and less auditable input values.’ E2 noted the following
superficial concern, which was restricted to only the appearance of
epistemological objectivity: ‘It is not possible to adjust our plans
and assumptions every year, because the auditors will not tolerate
such changes, and we as a company would appear unreliable.’
In this context of being faced with the possibility of extensive
and ultimately arbitrary discussions with its auditors, P1 adopted
an approach of simplifying the information provided to the audi-
tors, thereby narrowing the scope for discussion and the potential
for disagreement. P1 described that its own financial control
department routinely modelled alternative scenarios for future
revenues from major drug patents, yet only a subset of this infor-
mation was discussed with the auditors in the determination of
how to represent the fair values of these intangibles:
‘We find it hard to argue with our auditors for every small
parameter. The more complex valuation gets, the more flexibility
we have, and the more discussions we have with the auditors. This
is why we come up with a single revenue stream and avoid
scenario-based valuation models, where we have to argue about
the probabilities and the particular cash flow streams.’
P1
This ‘management’ of the process of representing fair values,
given the absence of any epistemologically objective claim to the
market ontology demanded by IFRS 13, came down to finding
whatever alternative representation would most readily satisfy the
auditors. A striking consequence of this approach was to represent
an ontology that is explicitly inconsistent with IFRS 13, namely the
reporting entity’s own perspective, as opposed to the perspective of
the market. This subverting of IFRS 13 was evident throughout the
case studies. For example, P3 claimed to ‘stick with the principle of
using the same assumptions for fair value as we use for our con-
trolling purposes, as long as the auditors approve.’ In much the
same spirit, I2 defended the use of entity-specific cash flows as
follows: ‘to justify our practice to the auditors, we help ourselves
with the thought that we do not see any reasons as to why an
external would apply different assumptions and methods from an
internal.’ This was a general theme. I1 claimed to proceed ‘by
simply assuming a market participant would both operate the plant
and the portfolio in the same way as the reporting entity does and
have the same supply conditions.’ E2 claimed to ‘see huge diffi-
culties in distinguishing between market- specific and entity-
specific values. Hence, we think there is always the presumption
that the market’s view is equivalent to our view.’ E1 argued as
follows:
With the knowledge I have, how can I pretend I am a typical market
participant? I am not, so I will include entity-specific elements
automatically.
E1
A further illustrative example arose in the context of impair-
ment testing where, for each of the six case study companies which
apply IFRS and US-GAAP ‘side-by-side’, or which recently have
switched from US-GAAP to IFRS, there was claimed to be hardly any
difference between fair value under US-GAAP and value-in-use
under IFRS.14 In other words, the fall-back position is the
perspective of the entity, rather than that of the market.
Overall, the evidence here is not only that fair values do not
represent the market ontology in IFRS 13, but that they also actually
subvert the fair value idea by instead seeking to represent the
ontology that IFRS 13 explicitly rules out. In turn, and as we discuss
next, we found evidence to suggest a sort of irony in the application
of IFRS 13, whereby the practices above, which might be described
as stabilising the internal preparation of fair values, were associated
with a varied and inherently unstable accounting representation.
An unstable representation
We have so far described a process that does not lead to the
creation of new institutional facts. The underlying problem is the
absence of epistemologically objective claims that could serve as
the basis for collective intentionality with respect to IFRS 13’s
market ontology. In turn, a consequence of this absence of under-
lying unity of purpose was that different practices emerged. There
was, for example, significant variation across the case companies in
the design of DCF models, even when the models were being
applied to similar assets. In the telecom group, one case company
generally used detailed cash flows for 10 years, whilst an inter-
viewee in an industrial case company insisted on the impossibility
of making detailed predictions for longer than five years, ‘because
of the rapidly changing characteristics of the industry’ (I3). Similar
observations were made within the pharmaceutical group, where
the forecasting horizon ranged between three and ten years. In
another variation, P2 described its process as follows:
We use a standard valuation model, independent of whether it is a
research project or a patent which is to be valued. Generally, we
apply a unified procedure: for the first five years, numbers are as
exact as possible; for the next five years, we use a rough estimate;
and for another five years, we work by ‘rule of thumb’.
P2
A number of interviewees stated they were uncertain to what
extent future cash flows should include future developments like
technological improvements and expansion to new or different
markets. Interviewees perceived conflict between accounting in an
‘as of today’ status and a valuation that should primarily be based
on the perspective of the market. As one interviewee stated, this
issue is conceptually less problematic for value-in-use, where the
measurement objective clearly excludes ‘any estimated future cash
inflows or outflows expected to arise from future restructurings or
14
In contrast with US GAAP, recoverable amount under IAS 36 is based upon the
higher of (entity-specific) value-in-use and fair value less costs to sell. If there were
differences with fair value under US GAAP, they were caused by the ‘cost to sell’
component, which in this context is conceptually the only difference between fair
value usage in IFRS and in US GAAP (arising in cases where IFRS fair value less cost
to sell is greater than value-in-use but less than the carrying amount).
 R. Barker, S. Schulte / Accounting, Organizations and Society 56 (2017) 55e67
from improving or enhancing the asset’s performance’ (IAS 36, par.
33). In contrast, the struggle to conceptualise the market partici-
pant’s perspective seemed to lead to speculative thinking. E1
perceived that the ‘market valuation always tends to include a sort
of hope value, whilst the business plan generally does not, as it is
more based on what one knows.’ The term ‘hope value’ describes an
increase in value which is produced by the belief that there is a
chance that the economic, technological or political situation im-
proves significantly, so that the expected cash flows turn out to be
higher than expected in the business plan. E1 argued that this hope
value should theoretically be taken into account in the way that
cash flows are forecasted, if the market’s view is implemented.
There are two things happening here. The first is the trans-
formation from ‘accounting as history’ to ‘accounting as eco-
nomics’, as the representation of fair value served to draw
accounting away from its traditional transactions-basis and to-
wards speculation around future cash flows. P2, for example,
referred to a ‘turning away from fact-based accounting numbers,’
which was perceived to be a major change from prior experience.
Second, however, the ‘accounting as economics’ being practised
was not the single market ontology of IFRS 13 but instead a range of
subjective ontologies, as each reporting entity sought to represent
fair value in its own way. This is in direct conflict with the fair value
idea under IFRS 13, whereby reporting entities are in principle
required to not represent their own views (which could reasonably
be idiosyncratic) but instead the collective view of the market
(which IFRS 13 characterises as external and objective).
A consequence of this variety in practice, and of the absence of
social consensus lying behind the variety, was a general sense of
unease about where the process of FVM implementation might
lead. The concern here lay beyond the preparers’ capacity to sta-
bilise and control their own processes of preparing fair values.
Rather, the concern followed from recognising that all entities were
building castles on the sand. The anxiety was expressed, for
example, that some companies would ‘use the areas of judgement
more than others’ (I2), and that this might ‘have a significant
impact on the accounts and consequently on the competition at the
capital markets’ (P1). The concern here goes beyond whether it was
possible to represent the market ontology of fair value. Instead, the
concern is one of uncertainty with respect to the inherently arbi-
trary and unstable nature of the representation of fair value
required by IFRS 13, and of the potential for others to enact those
requirements differently. In a similar manner, the uncertain
response of investors was also a source of concern. E2 asserted that
‘financial statement users do not understand how the fair values
are calculated,’ in comparison with cost-based accounting mea-
sures, which E3 described as ‘easily comprehensible, based on
historical cost, expected useful life and asset age.’ And if these were
data that investors did not understand, then this gave rise to a
related concern, over how investors would respond to financial
statements prepared on the basis of FVM. As I2 summarised in the
context of the income statement effect of FVM, there was a
perceived risk of ‘a dangerous and uncontrollable factor for the
financial statements.’
Conclusion
We have argued in this paper that the process of representing
fair values in financial statements involves either the reporting of
institutional facts already in existence or, in cases where IFRS 13’s
market ontology does not exist and so cannot be represented
objectively, the creation of data that do not themselves constitute
new institutional facts. We find empirically that the latter of these
two possibilities is dominant, and that preparers of financial
statements therefore find ways to ‘work around’ the ‘requirement’
65
in IFRS that reported fair values should represent the unknowable.
The main theoretical implication of our paper rests upon the
observability of the market ontology in FVM, which in turn rests
upon the existence of an underlying institutional reality. Searle’s
analysis is that collective intentionality is required to create
constitutive rules (status declaration functions) which, if generally
accepted, bestow rights and obligations. While ontologically sub-
jective, these rights and obligations constitute the institutional
reality that financial reporting seeks to represent in the form of
assets and liabilities. Such a reality is in principle knowable. That is
to say, whether or not there exist (epistemologically objective)
institutional facts, there does exist an institutional reality. If, how-
ever, in the cases where FVM’s market participant is hypothetical,
and so there does not exist an institutional reality, then of course
that ‘reality’ cannot be known. While IFRS 13 does explicitly
address epistemological variation in its Levels 1e3 classification,
such variation is in principle irrelevant if there does not exist an
institutional reality to which an epistemological claim can be made.
What IFRS 13 fails to address, therefore, is ontological variation.
Instead, it makes a commitment to a single, market ontology, both
in cases where there does exist an institutional reality, and yet also
in cases where that ‘reality’ can only be wished-for.
An implication of the above for accounting practice, except at
Level 1, is that representations of fair value cannot be from the
market participant’s perspective e as they are purported to be
under IFRS e and that a set of demands is thereby placed upon
actors to find ways to represent that which cannot be represented.
A further implication is that an aggregate amount represented as
‘fair value’ can be the sum of the values of both institutional facts
and non-institutional facts, giving that amount no obvious
meaning.
Our evidence is that actors’ responses to these demands can be
interpreted as going partway towards the creation of institutional
facts. There is collective intentionality on the part of preparers,
third party experts and auditors, in the acts of outsourcing fair
value determination to third party experts, finding ways to reach
agreement with auditors on acceptable representations of fair
value, and discussing within reporting entities how to apply IFRS
13. This collective intentionality takes the form ‘we are seeking to
represent fair value.’ And having reached agreement, there is then
the imposition of a status declaration function, of the form ‘these
reported amounts (X) count as fair values (Y) in the context of
financial reporting (C).’ Critically, however, the creation of institu-
tional facts requires a further step, beyond collective intentionality
and status function declaration, which is that there must also be
collective acceptance. In one sense, this step cannot be taken in
practice for FVM outside Level 1, because the possibility of collec-
tive acceptance is denied by the absence of an institutional reality,
by the reported fair values being a ‘representation’ of non-existent
rights and obligations. In another sense, however, the accounting
representation might, so to speak, substitute for the real thing. In
general in financial reporting, the users of financial statements
observe the accounting data, and not the underlying rights and
obligations that those data purport to represent. The possibility
exists, therefore, that there might be collective acceptance of re-
ported values, and that these might be accepted as institutional
facts, notwithstanding that they do not actually represent any un-
derlying institutional reality and also that (as our evidence sug-
gests) they might actually represent the perspective of the preparer
instead of the wished-for market ontology of IFRS 13. While this
possibility may exist, however, our fieldwork suggested reasons to
question its role in practice. We found that, ‘under the surface’ of
preparers’ attempts to stabilise accounting representations of fair
value, there was variation and inherent instability in practice in
determining reported amounts, as well as a degree of anxiety over
66 R. Barker, S. Schulte / Accounting, Organizations and Society 56 (2017) 55e67
how those amounts might be interpreted and used. The implication
is that the accounting might not ‘substitute for the real thing,’ and
that some form of instability is implied by detaching representa-
tions of fair value from an anchor of institutional reality. Our paper
therefore leads to the question of the status of reported fair values
that are neither institutional facts nor are likely to nevertheless be
accepted as such.
Our evidence does not, however, extend far enough to explore
this question further, because it is restricted to the preparation of
financial statements. While, in common with studies such as
Burchell et al. (1985) and Hopwood (1987), we have explored the
calculative role of accounting in reframing economic space, a
contrast with these studies is that we have not also explored
whether and in what ways such reframing contributes to consti-
tuting new social and institutional reality, for example by means of
different nexuses of practices and material arrangements. Our pa-
per therefore presents only part of a story. We have not fully
explored how, in the sense described by Schatzki (2005), the
reconstruction of social reality is enacted through the development
of new practices and material arrangements, as preparers find new
ways of ‘knowing’ and of following different rules towards different
ends. The opportunity is to explore more broadly the context
within which the social is reconstructed (Schatzki (2005). A wider
story would also bring into play what Miller and Power (2013)
describe as an accounting complex, involving four key roles of ac-
counting in territorialising (the recursive construction of calculable
spaces), mediating (linking actors and arenas), adjudicating (eval-
uating performance) and subjectivising (control or regulation).
While our paper touches upon each of these roles, it does so only
incompletely, as each also has social and institutional dimensions
that lie beyond the scope of our analysis of the fair value idea in
IFRS and of fieldwork that includes only preparers of financial
statements. Most obviously, the representation of fair values cre-
ates new, epistemologically objective data which, while not in
themselves institutional facts, can form the basis for collective
intentionality and so the creation of new institutional facts that
confer deontic powers. A simple example would be that fair values
represented using Level 3 data could become embedded in
contractual agreements, thereby conferring the deontic powers of
economic rights or obligations. While our paper has explored how
practitioners seek to represent fair values, a further level of analysis
lies in consideration of how those fair values then further influence
behaviours, within and outside reporting entities. A related, alter-
native approach would be to consider why preparers and auditors
are willing to accept the requirement to represent the hypothetical
market ontology in the first place, and why there is not resistance
but instead acquiescence with respect to the constitutive rule-
making authority of the IASB. In this context, IFRS itself can be
understood as an institutional fact. There is collective intentionality
on the part of the IASB and its stakeholders (‘we are setting ac-
counting standards’), there is a status declaration function (‘IFRS (x)
counts as the rules for accounting representations (y) in the context
of financial reporting (c)’) and the evidence above is that there is
also collective acceptance (by preparers and auditors) of the insti-
tutional reality so created, of IFRS itself. Yet, while the relevant
actors agree that IFRS has binding normative force, and while there
is epistemological objectivity (institutional facticity) in the wording
of the accounting standards themselves, such standards remain
ontologically subjective, being the product of human negotiation
and agreement and having no reality outside of this collective
agreement process. This subjectivity of foundation, together with
the instability of outcome described above, suggests fertile ground
for further change in what counts as an accounting representation.
This further consideration of the accounting complex links our
study with others in the literature. For example, as our fieldwork
did not include auditors, an obvious companion to our study would
be an exploration of the representation of fair value from the au-
ditors’ perspective, contributing to the literature on the role of the
audit profession in the development of financial reporting practice
(Botzem & Quack, 2009; Cooper & Robson, 2006; Power, 2003). The
role of the auditor is particularly interesting in the light of the
inherently unauditable nature of the (hypothetical) market
ontology in IFRS 13 (Bayou, Reinstein, & Williams, 2011) with im-
plications for social acceptance and for the strategies that auditors
adopt in order to represent themselves as relevant experts in an
unknowable domain (Power, 1997). There are likewise implications
for the accountability of management because, notwithstanding
the hypothetical nature of the market ontology, newly reported
amounts are likely to create demands for new behaviours, to
explain and respond to the entity’s revised representation of itself
(Williams, 2002; Young & Williams, 2010). A further issue, also
related to the implausibility of the market ontology, is the inherent
instability of the attempt to represent fair value, which draws
parallels with other attempts in analogous domains, such as ac-
counting for brands (Napier & Power, 1992), inflation (Robson,
1994), value-added (Burchell et al., 1985) and liabilities (Morley,
2014). The common theme here is for further research to go
beyond how fair value is represented and to consider also how it is
itself constitutive, as the representation of fair values becomes part
of what Callon (1998) describes as the ‘shared culture, rules, pro-
cedures, routines or conventions’ that influence the calculative
frame of behaviour in representing economic activity (Carruthers,
1995; Hines, 1991; Vollmer et al., 2009).
Acknowledgments
The authors are very grateful to Robert Monks and to the
Cambridge European Trust, for providing financial support for this
research project. The authors also gratefully acknowledge valuable
comments from the Reviewers and from Andrew Brown, Chris
Chapman, Christopher Napier, Mike Power, Keith Robson, Peter
Walton, Geoff Whittington and Joni Young.
References
Barrett, M., Cooper, D., & Jamal, K. (2005). Globalization and the coordinating of
work in multinational audits. Accounting, Organizations and Society, 30(1), 1e24.
Barth, M. E. (2007). Standard-setters, measurement issues, and the relevance of
research. Accounting and Business Research, 37(3), 7e15.
Barton, A. D. (1974). Expectations and achievements in income theory. The Ac-
counting Review, 49(4), 664e681.
Basu, S., Kirk, M., & Waymire, G. (2009). Memory, transaction records, and The
Wealth of Nations. Accounting, Organizations and Society, 34(8), 895e917.
Baudrillard, J. (1983). Simulations. New York: Semiotext(e).
Bayou, M., Reinstein, D., & Williams, P. (2011). To tell the truth: A discussion of
issues concerning truth and ethics in accounting. Accounting, Organizations and
Society, 36, 109e124.
Beaver, W. (1989). Financial reporting: An accounting revolution. Englewood Cliffs,
NJ: Prentice-Hall.
Benston, G. (2006). Fair-value accounting: A cautionary tale from Enron. Journal of
Accounting and Public Policy, 25, 465e484.
Berger, P., & Luckmann, T. (1966). The social construction of reality: A treatise in the
sociology of knowledge. Garden City, NY: Anchor Books.
Botzem, S., & Quack, S. (2009). (No) limits to Anglo-American accounting? Recon-
structing the history of the International Accounting Standards Committee: A
review article. Accounting, Organizations and Society, 34, 988e998.
Bougen, P. D., & Young, J. J. (2011). Fair value accounting: Simulacra and simulation.
Critical Perspectives on Accounting.
Bromwich, M. (2007). Fair values: Imaginary prices and mystical markets e A
clarificatory review. In P. Walton (Ed.), The Routledge companion to fair value and
financial reporting (pp. 46e68). Routledge.
Burchell, S., Clubb, C., Hopwood, A., Hughes, S., & Nahapiet, J. (1980). The roles of
accounting in organisations and society. Accounting, Organizations and Society,
5, 5e27.
Burchell, S., Hopwood, A., & Clubb, C. (1985). Accounting in its social context: To-
wards a history of value added in the UK. Accounting, Organizations and Society,
10(4), 381e413.
 R. Barker, S. Schulte / Accounting, Organizations and Society 56 (2017) 55e67
Callon, M. (Ed.). (1998). The laws of the markets. London: Blackwell.
Carruthers, B. (1995). Accounting, ambiguity, and the new institutionalism. Ac-
counting, Organizations and Society, 20(4), 313e328.
Chambers, R. (1965). Measurement in accounting. Journal of Accounting Research,
3(1), 32e62.
Chambers, R. (1966). Accounting, evaluation and economic behavior. New Jersey:
Prentice-Hall.
Cooper, D., & Morgan, W. (2008). Case study research in accounting. Accounting
Horizons, 22(2), 159e178.
Cooper, D., & Robson, K. (2006). Accounting, professions and regulation: Locating
the sites of professionalisation. Accounting, Organizations and Society, 31(4/5),
415e444.
Davis, S., Menon, K., & Morgan, G. (1982). The images that have shaped accounting
theory. Accounting, Organizations and Society, 7(4), 307e318.
Edwards, E. O., & Bell, P. W. (1961). The theory and measurement of business income.
University of California Press.
Eisenhardt, K. (1989). Building theories from case study research. Academy of
Management Review, 14(4), 532e551.
Elder-Vass, D. (2013). The reality of social construction. Cambridge: Cambridge
University Press.
FASB (2000). Status of concepts statement no. 7, using cash flow information and
present value in accounting measurements. Norwalk, CT: FASB.
Flyvbjerg, B. (2006). Five misunderstandings about case-study research. Qualitative
Inquiry, 12(2), 219e245.
Gaynor, L., McDaniel, L., & Yohn, T. (2011). Fair value accounting for liabilities: The
role of disclosure in unraveling the counterintuitive income statement effects
from credit risk changes. Accounting, Organizations and Society, 36, 125e134.
Gwilliam, D., & Jackson, R. (2008). Fair value in financial reporting: Problems and
pitfalls in practice: A case study analysis of the use of fair valuation at enron.
Accounting Forum, 32, 240e259.
Hayek, F. (1945). The use of knowledge in society. American Economic Review, 35(4),
519e530.
Hines, R. (1988). Financial accounting: In communicating reality, we construct re-
ality. Accounting, Organizations and Society, 13(3), 251e261.
Hines, R. (1991). The FASB’s conceptual framework, financial accounting and the
maintenance of the social world. Accounting, Organizations and Society, 16(4),
313e331.
Hopwood, A. (1987). The archaeology of accounting systems. Accounting, Organi-
zations and Society, 12, 207e234.
IASB. (2010). The conceptual framework for financial reporting. London: IASB.
IASB. (2013). International financial reporting standards. London: IASB.
Ijiri, Y. (1983). On the accountability-base conceptual framework of accounting.
Journal of Accounting and Public Policy, 2, 75e81.
Ijiri, Y., & Jaedicke, R. (1966). Reliability and objectivity of accounting measure-
ments. The Accounting Review, 474e483.
Klamer, A., & McCloskey, D. (1992). Accounting as the master metaphor of eco-
nomics. European Accounting Review, 1(1), 145e160.
Landsman, W. (2007). Is fair value accounting information relevant and reliable?
Evidence from capital market research. Accounting and Business Research, 37(3),
19e30.
Laux, C., & Leuz, C. (2009). The crisis of fair value accounting: Making sense of the
recent debate. Accounting, Organizations and Society, 34(6/7), 826e834.
Lennard, A. (2007). Between a rock and a hard place. In P. Walton (Ed.), The Rout-
ledge companion to fair value and financial reporting (pp. 115e122). Routledge.
Macintosh, N., Shearer, T., Thornton, D., & Welker, M. (2000). Accounting as simu-
lacrum and hyperreality: Perspectives on income and capital. Accounting, Or-
ganizations and Society, 25(1), 13e50.
MacKenzie, D. (2009). Material markets. Oxford: Oxford University Press.
McGregor, W. (2007). Whither fair value accounting? In P. Walton (Ed.), The Rout-
ledge companion to fair value and financial reporting (pp. 103e114). Routledge.
McKernan, J. (2007). Objectivity in accounting. Accounting, Organizations and Soci-
ety, 32, 155e180.
Miller, P. (1998). The margins of accounting. European Accounting Review, 7,
605e621.
Miller, P., & Power, M. (2013). Accounting, organizing and economizing: Connecting
accounting research and organization theory. The Academy of Management
Annals, 7(1), 557e605.
Morgan, G. (1988). Accounting as reality construction: Towards a new epistemology
for accounting practice. Accounting, Organizations and Society, 13(5), 477e485.
67
Morley, J. (2014). Where culture and lobbying collide: The case of the IASB Liabil-
ities Project. Working paper: LSE.
Mouck, T. (2004). Institutional reality, financial reporting and the rules of the game.
Accounting, Organizations and Society, 29(5/6), 525e541.
Murphy, T., O’Connell, V., & O’hOgartaigh, C. (2013). Discourses surrounding the
evolution of the IASB/FASB Conceptual Framework: What they reveal about the
“living law” of accounting. Accounting, Organizations and Society, 38, 72e91.
Napier, C., & Power, M. (1992). Professional research, lobbying and intangibles:
Review essay. Accounting and Business Research, 23(89), 85e95.
Nobes, C. (1983). A judgmental international classification of financial reporting
practices. Journal of Business Finance and Accounting, 10(1), 1e19.
Pettigrew, A. (1995). Longitudinal field research on change e Theory and practice. In
G. Huber, & A. Van de Ven (Eds.), Longitudinal field research methods: Studying
processes of organizational change. Thousand Oaks, CA: Sage Publishers.
Power, M. (1997). Expertise and the construction of relevance: Accountants and
environmental audit. Accounting, Organizations and Society, 22(2), 123e146.
Power, M. (2003). Auditing and the production of legitimacy. Accounting, Organi-
zations and Society, 28, 379e394.
Power, M. (2010). Fair value accounting, financial economics and the transformation
of reliability. Accounting and Business Research, 40(3), 197e210.
Ragin, C., & Becker, H. (Eds.). (1992). What is a case? Exploring the foundations of
social inquiry. Cambridge: Cambridge University Press.
Ravenscroft, S., & Williams, P. (2009). Making imaginary worlds real: The case of
expensing employee stock options. Accounting, Organizations and Society, 34(6/
7), 525e541.
Roberts, J., & Jones, M. (2009). Accounting for self-interest in the credit crisis. Ac-
counting, Organizations and Society, 34, 856e867.
Robson, K. (1991). On the arenas of accounting change: The process of translation.
Accounting, Organizations and Society, 16(5/6), 547e570.
Robson, K. (1994). Inflation accounting and action at a distance: The Sandilands
episode. Accounting, Organizations and Society, 19(1), 45e82.
Roesch, E. (1978). Principles of categorization. In E. Roesch, & B. B. Lloyd (Eds.),
Cognition and categorization (pp. 27e48). Hillsdale: Lawrence Erlbaum
Associates.
Schatzki, T. R. (2005). Peripheral vision the sites of organizations. Organization
Studies, 26(3), 465e484.
Searle, J. (1995). The construction of social reality. London: Penguin.
Searle, J. (2005). What is an institution? Journal of Institutional Economics, 1(1),
1e22.
Searle, J. (2010). Making the social world. Oxford: Oxford University Press.
Searle, J. (1990). Collective intentions and actions. In P. Cohen, J. Morgan, &
M. E. Pollack (Eds.), Intentions in communication. Cambridge, Mass: Bradford
Books, MIT Press.
Shapiro, B. (1997). Objectivity, relativism, and truth in external financial reporting:
What’s really at stake in the disputes? Accounting, Organizations and Society,
22(2), 165e185.
Slater, D., & Tonkiss, F. (2001). Market society: Markets and modern social theory.
Cambridge: Polity Press.
Smith, B. (2003). John Searle: From speech acts to social reality. In B. Smith (Ed.),
John Searle. Cambridge: Cambridge University Press.
Sterling, R. (1970). Theory of the measurement of enterprise income. Lawrence, KA:
University of Kansas Press.
Vollmer, H., Mennicken, A., & Preda, A. (2009). Tracking the numbers: Across ac-
counting and finance, organizations and markets. Accounting, Organizations and
Society, 34(5), 619e637.
Walton (Ed.). (2007). The Routledge companion to fair value and financial reporting.
Routledge.
Whitley, R. (1986). The transformation of business finance into financial economics:
The roles of academic expansion and changes in US capital markets. Accounting,
Organizations and Society, 11, 171e192.
Williams, P. F. (2002). Accounting and the moral order: Justice, accounting, and
legitimate moral authority. Accounting and the Public Interest, 2, 1e21.
Young, J. (1994). Outlining regulatory space: Agenda issues and the FASB. Ac-
counting, Organizations and Society, 19(1), 83e109.
Young, J. (2006). Making up users. Accounting, Organizations and Society, 31,
579e600.
Young, J., & Williams, P. F. (2010). Sorting and comparing: Standard-setting and
“ethical” categories. Critical Perspectives on Accounting, 21, 509e521.
 Accounting, Organizations and Society 56 (2017) 68e83

Contents lists available at ScienceDirect

Accounting, Organizations and Society

journal homepage: www.elsevier.com/locate/aos

Relative reliability and the recognisable firm: Calculating goodwill

impairment value
Jari Huikku a, Jan Mouritsen b, *, Hanna Silvola a
a
Aalto University School of Business, Department of Accounting, Runeberginkatu 22-24, 00100 Helsinki, Finland
b
Copenhagen Business School, Department of Operations Management, Solbjerg Plads 3, B.5.10, 2000 Frederiksberg, Denmark

a r t i c l e i n f o a b s t r a c t

Article history: This paper complements financial accounting research by a qualitative study of financial accounting
Received 30 September 2012 practices. Its object is goodwill impairment tests (IAS 36) under the influence of International Financial
Received in revised form Reporting Standards, which it uses to illustrate how financial accounting is produced. The aim is to
5 February 2016
investigate how accounting standards are translated into accounting practices, and to investigate how
Accepted 31 March 2016
Available online 21 April 2016
this is reliable. Drawing on actor network theory, the paper proposes calculative practices to be a net-
worked and distributed affair. The study has two main contributions. Firstly, it shows that in the case of
goodwill impairment tests, financial accounting is a process of finding, qualifying, stabilizing and
calculating traces that often have to be found beyond the company infrastructure of sheets of accounts
and the financial ledger. Secondly, it shows that these traces increase reliability when they are recog-
nisable and impersonal. No single person is responsible for the financial calculation and the traces used
assume that a firm cannot systematically outperform the broader economy or the history of the firm. It
also helps to increase reliability if institutional roles such as auditors and valuation experts tolerate the
calculation. Reliability increase when traces and supporting institutional actors that take part in the
calculation are at a distance. Because of this production process, readers of financial statements face the
following paradox: the things they see are less associated with specific entrepreneurial activities in the
firm and more with normalised trends inside and outside the firm. Seeing the firm requires them to look
at its past, at negotiated budgets, at its competitors, at industrial outlook, and at the statistical bureaus
that compile information on the economic development of industries and countries; they may also have
to listen to valuation experts and auditors. Seeing the value of a firm requires actors to look elsewhere.
© 2016 Elsevier Ltd. All rights reserved.

1. Introduction accounting communicates the world. We observe a construction

“That's the paradox. That's where we walk a very thin line. We
communicate reality: that is the myth; that is what people believe.
It is even what most of us believe. And, in a sense, we do
communicate reality. There is something there: bricks and people
and so on. And the organisation can, say, be ‘doing well’, or ‘doing
badly’, in whatever sense you take that to mean. And it is our job to
convey it. But what is 'the full picture'? There is no full picture. We
make the picture. That is what gives us our power: people think and
act on the basis of that picture! Do you see? Are you beginning to
see?” (Hines, 1988, p, 265).
Ruth Hines' (1988) famous fable about financial accounting asks
of us to contemplate what it is that we see when financial
* Corresponding author.
E-mail addresses: jari.huikku@aalto.fi (J. Huikku), jm.om@cbs.dk (J. Mouritsen),
hanna.silvola@aalto.fi (H. Silvola).
and see less than a full picture, she says: “There is no full picture”
(ibid., p. 265). So, which picture does financial accounting make us
see? More specifically, when readers of financial statements
observe a calculation of goodwill impairment based on net present
value, what do they see? To answer this general question, it is
necessary to study how financial accountants produce financial
statements. While there is a discernible body of market-based
research designed to test the effects of financial accounting
choices, e.g. in relation to fair value accounting (Laux & Leuz, 2009),
empirical research about the production of accounting is largely
absent (Durocher & Gendron, 2011; Hopwood, 2000; Young, 2006).1
1
There are important studies of auditing practices (for an early review of audit
practices, see Power (2003), of auditing firms (Anderson-Gough, Grey, & Robson,
2001; Cooper, Greenwood, Hinings, & Brown, 1998; Gendron et al., 2007;
Kornberger, Justesen, & Mouritsen, 2011; Suddaby, Cooper, & Greenwood, 2007),
and of audit-committee practices (Gendron & Be
dard, 2006; Gendron et al., 2004).

http://dx.doi.org/10.1016/j.aos.2016.03.005
0361-3682/© 2016 Elsevier Ltd. All rights reserved.
 J. Huikku et al. / Accounting, Organizations and Society 56 (2017) 68e83
This generally motivates the paper's interest in translations from
financial standards into financial accounting practices, which are
critical in order to understand what financial accounting makes
visible (Robson & Young, 2009).
The paper has two main aims. Firstly, it seeks to explore trans-
lations between financial accounting standards and financial ac-
counting practices. As a construction, financial accounting is often
presented as easily mouldable because it is mathematical (Vollmer,
2003, 2007) and easy for managers to manipulate by changing the
calculation to undertake earnings management (Macintosh, 2006,
2009; Ramanna, 2008). When understood as this type of construc-
tion, accounting is in the hands of the few who can design it to suit
their interests. However, there may be a limit with regard to how far
this can go because the more personal financial accounting is the less
reliable it will be and then it will not engender trust and comfort
(Pentland, 1993; Power, 1995, 1996, 1997, 2003). It is important
therefore to investigate whether and how a financial accounting
construction is different from a personal statement. The second aim
is to explore what readers of financial statements see when financial
standards are translated into practices. Accounting standards delimit
the financial accounting object in principle, but they do not specify
the empirical demarcations that locate the standard in practices of
financial accounting (Lezaun, 2006). Financial accounting un-
derstands the economic world from the classifications produced by
sheets of accounts and the general ledger. They organise transactions
and records which are the remaining simple traces from complex
economic selling, purchasing and production events. The records in
financial accounting database are typically understood as traces of
past events. However, International Financial Reporting Standards
(IFRS) pose the challenge that financial accounting increasingly is
tasked to engage with the future. Traces therefore have to be in-
dications of the future and these traces may not intuitively be part of
the set of historical records found in the financial accounting data-
base. It is therefore not clear what it is readers of financial accounting
can see when they observe financial accounting.
To achieve these aims, the paper investigates how goodwill
impairment is produced. This is a critical case for IFRS because
goodwill is a level three asset that requires being tested for
impairment by means of models. It has no market value per se
(Bougen & Young, 2012; Macintosh, Shearer, Thornton, & Welker,
2000). Goodwill is difficult for two reasons. First it is a residual
value and has no associated discernible and separable asset; and
second it is about the future. It is a critical incidence for IFRS. If
goodwill accounting is reliable e in the sense of being able to be
relied upon e this may also be so for other IFRS based valuations.
Accordingly, the specific research questions are: how do financial
accounting practices produce goodwill impairment value, and how
is the financial accounting calculation reliable?
Drawing broadly on actor network theory (e.g. Latour, 1989, 1987,
2005), the study examines the practices of calculation as a distributed
network.2 According to this approach, the preparer is not a mind or
brain that more or less liberally interprets accounting and changes it
to suit individuals' interpretations and strategies. Instead, financial
accountants are a part of a wider set of actors including both human
actors and non-human actants who in their own ways influence the
2 properties of the firm disappear from the calculation and what
Prior research on goodwill accounting has addressed goodwill impairment
testing using quantitative methods. This research suggests that impairment testing
procedures help opportunistic management discretion in relation to the timing and
magnitude of goodwill write-offs (Beatty & Weber, 2006; Massoud & Raiborn,
2003; Ramanna, 2008; Ramanna & Watts, 2012; Wines, Dagwell, & Windsor,
2007). New CEOs may use goodwill write-offs to clean the books (Masters-Stout,
Costigan, & Lovata, 2008), and managers may engage in big bath earnings man-
agement and write goodwill off when earnings are already depressed (Jordan,
Clark, & Vann, 2007).
69
preparation of financial statements. Financial accountants may find
themselves in a centre of calculation which is obligated to develop
financial statements, but they cannot do this only by themselves. At
least, as a centre of calculation, the financial accounting office requires
records to calculate on. These records are typically traces of activity
that has happened elsewhere in time and space. The financial ac-
counting office cannot calculate if it does not have traces that enable it
to translate the financial accounting standard. The financial ac-
counting database is a “large star-shaped web of mediators” (Latour,
2005, p. 217) which allows things to flow into and out of the finan-
cial accounting office: traces flow in and financial statements flow
out. As Latour (2005) says, any actor such as a financial accounting
office is made to exist by many relations and entities. Therefore, the
financial accounting office's efforts to develop financial statements
are mediated by non-human actants (e.g. traces in the form of re-
cords) and human actors (e.g. auditors) that together negotiate what
the financial statement is about. Through this approach the preparer
is a network more than a single person or mind.
The empirical analysis is based on Finnish data. Finland is a critical
case for analysing effects of IFRS on financial accounting practices
because IFRS were a radical step for Finnish preparers (Nobes, 2013).
Not only did the regulation change from a classical continental Eu-
ropean conservative focus to an IFRS fair values approach almost
overnight (Erb & Pelger, 2015; Power, 2010), it also made IFRS
regulation to be Finnish regulation with no adaptation (Kettunen,
2014). No preparer could be expected to have expertise.3 Drawing
on interviews with 55 financial accountants, auditors, financial ad-
visors, the financial supervisory authority, financial analysts, in-
vestors, creditors, media and practice-influencing academics with a
focus on their experiences working with goodwill calculations.
The study has two main contributions. As a study of financial
accounting in action, it shows firstly that as practice, preparers of
financial statements are busy finding, qualifying, stabilizing and
calculating traces typically found outside the financial accounting
database. The study shows that the traces that are favoured by
preparers construct a financial statement, which when observed by
readers make them see away from the specifics of the firm.
Secondly, the urge to see away from the firm is an effect of pre-
parers' understanding of reliability. It appears that traces produced
by external statistical bureaus, external advisors and consultants are
preferred to internal ones; internal traces that are negotiated such as
budgets or used for several purposes are preferred to individual and
singular ones. Individual traces proposed by entrepreneurial man-
agers are not trusted. This matters because traces are then under-
stood to represent an impersonal “view from nowhere” (Nagel,
1986; Porter, 1992, 1994b). The reliability of the accumulation of
traces is helped by many people tolerating it; people who occupy
institutionalised positions or roles such as auditors and experts are
stronger than financial accountants and managers.
These characteristics make the calculation of goodwill impair-
ment recognisable, realistic and un-surprising. This practice is not as
much concerned with seeing the economics of the particular entre-
preneurial activities of the firm as may be the ambition of IFRS (Barth,
2007). Instead, drawing on country and industry averages, on his-
torical growth-rates, and on negotiated budgets, the calculation is
more average to the firm and the economy than might be expected
(see e.g. Ramanna & Watts, 2012). To some extent, the specific
3
This makes Finland a critical case for the analysis of the implications of the
change of accounting regulation. It is likely that the case of Finland will be a more
systematic experiment of the effects of adoption of goodwill impairment testing
than Anglo-Saxon countries (Mennicken & Millo, 2012; Nobes, 2013). A few Finnish
firms already had a little exposure to goodwill accounting having applied US-GAAP.
70 J. Huikku et al. / Accounting, Organizations and Society 56 (2017) 68e83
readers of financial statements see when they observe a financial
statement is outside the firm either in time (as in historical growth
rates) or in space (statistical offices predict macro growth-rates). To
observe the firm through financial statements, readers see elsewhere.
The remainder of the paper is organised as follows. The next
section discusses a set of theoretical resources that make it possible
to study calculation as a process and which draws on actor network
theory. Thereafter, the method section is outlined, describing how
this qualitative study of financial accounting practices was con-
ducted. The empirical section provides evidence of the process of
calculating goodwill impairment value. The discussion makes clear
the properties of the calculative practice that produces a goodwill
impairment value. Finally, conclusions are provided.
2. Understanding financial accounting as practice
Prior studies of financial accounting have focused more on its
institutional dimension than on preparers' financial accounting
practices (Gendron, Cooper, & Townley, 2007; Hines, 1991; Robson,
1991, Robson & Young, 2009; Smith-Lacroix, Durocher, & Gendron,
2012; Suddaby, Gendron, & Lam, 2009). This institutional dimen-
sion is, for example, strong in studies about value added accounting
(Burchell, Clubb, & Hopwood, 1985), inflation accounting (Robson,
1994; Thompson, 1987), and brand accounting (Power, 1992).
Such studies emphasise the roles of complex types of politics, claims
to expertise, and battles for jurisdiction (Burchell, Clubb, Hopwood,
Huges, & Nahapiet, 1980). However, the financial accountant seems
to be absent. In situations where financial accountants' choices are
discussed, the financial accountant is a Weberian ideal type who is a
one-sided exaggeration rather than an empirical person (Bayou,
Reinstein, & Williams, 2011; Macintosh, 2006, 2009). Institutional
voices are also loud in accounts of IFRS' relationships with processes
of financialization and globalisation that connect capitalism,
shareholder value, and political economy (Arnold, 2012; Froud,
Sukhdev, Leaver, & Williams, 2006; McGoun, 1997). Their pres-
ence is also felt, albeit in a more general form, in Power's (2010)
analysis of the conditions facilitating the rise of fair value ac-
counting (see also Bougen & Young, 2012; Macintosh et al., 2000;
McGoun, 1997) and in Hopwood's (1992, 2006, 2009) discussions
of the relationships between accounting, economics and finance. To
this institutional perspective also belongs research on the role of
standard setters in the development of IFRS (Erb & Pelger, 2015).
This body of research focuses on the action of institutions and
not much on financial accountants' practices. It is not possible to
know how the standard is translated into practice (Robson &
Young, 2009, p. 360). The study of financial accounting practice
concerns the work undertaken by financial accountants (Ahrens &
Chapman, 2007) when they attend to calculation (Miller, 2001) and
quantification (Espeland & Stevens, 2008; Porter, 1992, 1994a).
Such work is related to their engagement with the financial ac-
counting database, which is their primary instrument to make
financial calculations that communicate and construct the world.
Financial accounting databases consist of records that are traces
of economic events: a sales slip, a purchasing note, a production
record, a time sheet, a financial payment. The traces create links to
events but they are not the events; they are signs of activity that is
absent in time and/or space. Instead of storing the sales event and
the production event, the traces take their places. Through accu-
mulation, traces gradually turn into financial statements (Robson,
1992) or into subsequent financial analysis (Kalthoff, 2005). Since
the trace is not the event itself, its only quality is that it is produced
according to some procedure e it is an object that is guaranteed by
the rules that make it up. This is why it is constitutive more than
reflective (Hines, 1988). It is produced by an instrument that gua-
rantees these rules to be followed. A sales event is translated into a
trace by an instrument (e.g. by the cash register) that records the
revenues of this event and a date (but not the dynamics of the sales
activity between a sales person and the customer). Financial ac-
counting is concerned with finding, creating and organising traces
that can be accumulated in the financial database (the general
ledger and the sheet of accounts). In this way it traces economic
events such as sales activities. A trace is a sign that makes some-
thing traceable.
This principle can also be found in other areas of social and
economic activity. For example, in the case of genetically modified
organisms (GMO), traces can be developed which can identify
where and when such substances exist by tracing them from lab-
oratory, through production into markets (Lezaun, 2006). Traces
are here political objects that entice struggles about proper gov-
ernment because firms, regulators and activists disagree about the
appropriate identification of such substances. Therefore the
methods for tracing are political rules for identifying substances
rather than mere representations of these substances. The so-called
“transformation events” are the moments and places where GMOs
move from one existence to another, and the struggle is about how
the trace is to be made, including whether it is a pure scientific
reference or also one which imprints the uses of the GMO. The
properties of the trace are thus defined by the techno-legal-
economic properties of the GMO as settled by the administrative
instruments that produce the trace. Therefore, the trace takes the
place of the GMO rather than describes it.
In another area, such as science, traces are the textual outputs
from instruments and which go through accumulation and accel-
eration to convince sceptical audiences about the claim of a
spokesperson (cf. Latour, 1986, 1987). Traces are produced by in-
struments either in the form of small measurement devices such as
thermometer for temperature, ruler for length or watch for time, or
by constellations of devices such as whole laboratories. When
accumulated, traces turn into inscriptions that move into final ac-
ademic papers for example in the form of small diagrams. These are
the last accumulation and acceleration that the trace goes through.
Behind the inscription are traces. Behind traces there are
instruments.
For GMO and science, the trace is part of an administrative
process, or an organised procedure, where events are made trace-
able. The traces are made to show something by a process of
accumulation, similar to the process by which nation states attempt
to see the country and its population by a process of accumulating
statistical traces (Lam, 2011; Miller & Rose, 1990; Rose & Miller,
2008; Scott, 1998). If traces are disturbed, the things to be seen
remain ambiguous (Dambrin & Robson, 2011; Rottenburg, 2009).
There is a parallel in the case of financial accounting where traces
are the records found in financial databases.
This process of tracing is intuitively understandable in the case
of conventional, historical accounting with its emphasis on the
verifiability of completed events. With regard to IFRS, the event
poses a challenge, however, because, as Power (2010, p. 198) sug-
gests, in principle historical cost accounting is substituted by a
focus on the future, which implies a “contest between fundamen-
tally different notions of accounting reliability.” Reliability in the
transaction-based model is concerned with trails of evidence, while
for IFRS reliability is related more to investors' reliance on ac-
counting to make investment decisions. Market values would, in
principle, be helpful to discover “hidden economic substances” and
make them more transparent (Barth, 2007; Bromwich, 2007;
Whittington, 2008). This would be a set of financial accounts that
the investor could rely upon. The challenge for reliability is that
intuitively, the future has not yet made a trace; it is therefore likely
that traces are not located in the financial database.
Lezaun (2006) proposes a distinction between delimitation and
 J. Huikku et al. / Accounting, Organizations and Society 56 (2017) 68e83
demarcation to explain how it is possible to develop a procedure for
a not-yet-defined event. Delimitation concerns the possibilities of
articulating and proposing an object of interest. Demarcation, on
the other hand, concerns the practices and instruments used to
make the object administratively distinct. For example, IAS 36
proposes that goodwill accounting can be a value-in-use calcula-
tion and requires financial accountants to find cash flow forecasts
and develop the weighted cost of capital. Demarcation concerns the
mechanism by which traces about the empirical world can be
organised and made strong enough to pass as appropriate for the
net present calculation. In effect, delimitation is the accounting
standard and its supporting accounting principles; demarcation is
the search for traces that can stand for the future with adequate
strength.
The problem is that the future has not yet arrived and therefore
it is not obvious which traces can be allowed to stand for it. Hence
the research questions: How do financial accounting practices
produce goodwill impairment value, and how is the financial ac-
counting calculation reliable?
3. The study
3.1. The standard
Goodwill has no market value in itself as it is the residual from
acquisition value that cannot be allocated to existing material and
immaterial asset. IAS 36 requires firms to perform a goodwill
impairment test at least annually to ensure that its goodwill is
carried at no more than its recoverable value. An impairment loss is
the amount by which the carrying value exceeds the recoverable
value and is an expense in the income statement. The recoverable
value is tested separately in all cash generating units (CGU) and
their recoverable amount is the higher of its ‘value in use’ and ‘fair
value less costs to sell.’ Goodwill can only be impaired; it cannot be
re-appraised upwardly.
‘Value in use’ is widely used not only generally (McDonnell,
2005), but also in Finland. This was the only method mentioned
by interviewees in the empirical study. It is the present value of the
future cash flows from a CGU using a pre-tax discount rate. Esti-
mates of future cash flows have two origins: a period representing
1e5 years based on budgets and strategies, and the growth rate
prediction beyond this. According to IAS 36.55, a firm is required to
use a pre-tax discount rate that reflects current market assessments
of time value of money and risks.4
3.2. The research
The empirical material consists of 53 semi-structured in-
terviews with 55 financial accountants (CFOs, financial controllers,
financial standard advisers), external valuation experts, the finan-
cial supervisory authority, auditors, financial analysts, investors,
creditors, media, and academics (see Appendix 1). They were
conducted between March 2010 and April 2013 and developed 57 h
of recorded material. Appendix 2 lists the themes of the interviews.
Other empirical evidence includes annual reports and interim
reports of the listed companies (especially their notes on goodwill
impairment testing), publications and internal surveys of auditing
firms, reports produced by FIVA (Financial Supervisory Authority),
financial analysts, and books and blogs written by analysts. In
addition, confidential goodwill impairment testing material was
4
IAS 36.57 requires that the discount rate would be the entity's WACC, incre-
mental borrowing rate or market borrowing rate. Estimating WACC requires insight
into cost of equity, cost of debt, and target capital structure.
71
revealed to us by some of the companies. Goodwill was important
in the interviewees' firms, averaging 57% of equity.5 Generally, for a
third of Finnish publicly listed firms, goodwill accounted for more
than 50% of their equity.
In terms of method, interviews were organised around the
complexities of the production of goodwill impairment values. In-
terviewees brought up the idea that calculation is dispersed, and
that traces had to be found and made part of the calculation.
Questions were related to that translation and focused on chal-
lenges posed by goodwill impairment value as both being about the
future and about a non-separable asset. These questions concerned
the components of the calculation; the process of producing traces
relating to cash flows and WACC; the assurance of goodwill
impairment values; and participation in the goodwill impairment
calculation.
Secondly, the interview data was then organised in themes
about the cash flows, the WACC (which were a priori expected to be
important), and the process of assurance in and around the audit
committee. This latter theme came up as a more important
consideration than was first expected.
Thirdly, the thematic organisation around the elements of the
calculation of goodwill impairment value was reviewed and
lacunae were identified that were then addressed by additional
interviews or by returning to interviewees. This, for example,
concerned the roles of external experts whose position turned out
to be inadequately reflected by the first interviews. Fourthly, the
empirical material was then revised in relation to the research
questions, and the findings were distilled.
4. Empirical observations of the production of goodwill
impairment tests
The following empirical account describes how goodwill
impairment value is calculated, and in what way it is reliable. The
section is organised around the process of calculating goodwill
impairment value. IAS 36 delimits the annual impairment test
which is to be calculated as the present value of the future cash
flows from a CGU using a pre-tax discount rate. The calculation
involves estimations of future cash flows and a discount rate, as
financial economics would suggest.
An illustrative example from one of the firms' impairment value
calculations is provided in Fig. 1. This is the product of the financial
accounting office, the centre of calculation, which draws together
traces from near and far.
Fig. 1 is a delimitation of the object impairment value and it
shows the WACC and pre-tax discount rate for each CGU, and then
the enterprise value of the CGU is presented next to value on the
books. If enterprise value is the higher of the two, no impairment is
made. Thereafter, the calculation presents sensitivity analysis for
EBIT changes and sales growth changes. Last, there are notes that
explain some of its model parameters. The delimitation helps firms
to identify relevant traces that make the calculation possible.
This is a delimitation of impairment value, and like Lezaun's
(2006) observation about genetically modified organisms it is not
obvious how it is possible to demarcate and trace the events that
make the delimitation realistic. There has to be some extra work
involved in finding the traces of the future that is so strongly
5
The firms operated in the metal industry, diversified industry, telecommuni-
cation/ICT service, the media, chemical and plastics, and the retail trade, and they
were listed only in Finland. Financial institutions such as banks and insurance
companies were not included in this study. Financial markets are relatively small in
Finland: only 124 firms were publicly listed on the Helsinki Stock Exchange during
the study.
72 J. Huikku et al. / Accounting, Organizations and Society 56 (2017) 68e83

Summary of Goodwill impairment tes ng results Tes ng date 30.9.2010

Forecast period 2010-2014 Break-even Sales info

CGU WACC Pre-tax EV Net assets Results EBIT change Compound
% discount rate % €M €M threshold % growth rate %
CGU A 8,87 15,67 860 355 Not impaired -10,0 5,0
CGU B 8,87 15,94 240 200 Not impaired -2,0 -2,3
CGU C 9,09 14,95 175 143 Not impaired 2,5 1,2
CGU D 9,22 15,20 354 110 Not impaired -25,0 2,6
CGU E 8,87 13,90 78 55 Not impaired 11,0 3,2
CGU F 8,76 14,26 331 170 Not impaired -6,5 4,1
CGU G 8,76 11,90 155 33 Not impaired 5,0 9,6
CGU H 8,76 14,10 30 10 Not impaired 3,0 3,8
Headoffice 8,87 N/A -132 N/A N/A N/A N/A
Total EV 2091

Cash flow based on the latest updated forecast for 2010

Each CGU (segment) has provided data for 2011 - 2013 (based on the strategy review by management in 09/2010)
2014 es mated based on 2013 performance
When WACC applied in discoun ng, cash flows are post-tax-Pre-tax discount rate also iterated for IFRS purposes
Net Assets of CGUs as of 30.9.2010
Terminal growth rate 1,8%
When EV (Enterprise value) > net assets, there is no need for impairment
CONCLUSION: No impairment detected
Break-even analysis calculates yearly EBIT change from 2010 t0 2014 required to reach break even posi on (EV = Net assets)
(When the EBIT level is nega ve in 2010, the required annual EBIT change cannot be calculated)
Compound growth rate of Net sales from 2010 to 2013 as forecasted by the CGU (Segment) and 2014 es mated

Fig. 1. Example of a template for reporting goodwill impartment tests used in a firm (actual numbers are disguised).

present in the delimitation. The following empirical sections are
oriented to understanding how demarcation of traces happens in
Finland. This involves three steps: first a section on the traces
needed to develop cash flow forecasts, thereafter a section on the
development of the WACC, and then a final section on assurance by
the audit committee.
4.1. Demarcating traces to forecast the cash flows
In the Finnish firms, goodwill impairment testing was part of
other pre-existing organisational processes such as budgeting and
strategy-making. This would be helpful since goodwill impairment
calculations then were not taken “out of the blue”, as one auditor
(Partner, Audit firm 3) argued:
“It is a completely clear process. And it is, as such, good. It is not an
exercise where you just take numbers out of the blue. It is closely
related to the budgeting process in the autumn. Goodwill impair-
ment testing is really based on the latest cash flow estimates.”
many businesses that the business growth is in line with economic
growth plus/ minus something.” (CFO, Firm 7)
The traces of “historical events” influenced expectations about
growth rates, which would not be reliable if they were abnormal.
Even “dummies” would be able to see and recognise this. For a trace
to attract reliability it had to have continuity with the past. Erratic
movements were understood as noise and traces had to be
immediately recognisable or plausible. Reliability would increase if
firms' forecasts would follow recognisable, general economic
growth rates.
Firms already prepared cash flow forecasts through budgeting
for the 1e2 years of the first five years, and the rest were justified
by strategy. A CFO explained:
“When we test, we use our established long-term plans, strategic
figures. They are there in the background, and we don't make them
particularly for this [goodwill impairment testing].” (CFO of Firm
8)

Being part of a wider procedure added reliability to traces such
as budget numbers. There were already traces in the form of past
cash flows, and cash flows of the testing year had been defined
during the budgeting process. These traces would increase the
reliability of the goodwill impairment calculation as they were
based on “historical events”:
“If we talk about growth of sales volume, we usually start to
examine historical events. It gives a base for analyses. If sales vol-
ume has increased from two to three percent, and then there is
suddenly an estimated growth of 10%; even dummies can see
whether there is any sense. In addition, we need to reflect the fig-
ures with economic expectations and forecasts. It is typical for
Goodwill impairment tests were organisationally embedded in
budgets and in strategic and long-term planning processes. Thus,
traces that had several uses beyond the goodwill impairment test
would be more reliable than single-purpose traces. Traces were not
manufactured only for the purpose of impairment testing but were
more generally involved in managing the business; this would
strengthen reliability. Single-purpose traces would have too few
strings attached and multi-purpose traces related to strategy and
budgets were organisationally embedded and tolerated by many
actors across the organisation. Budgets were compromises between
different concerns and this would increase reliability. The CFO of
Firm 10 explained:
 J. Huikku et al. / Accounting, Organizations and Society 56 (2017) 68e83
“The companies have a tendency to be optimistic in all their
planning activities. It inherently belongs to our role as business
managers that we think that our improvements and investments
lead to good outcomes. In practice, lots of negative issues can
happen, which are not integrated into our budgets: competition
can become fiercer, raw material costs will increase, and currency
exchange rates change surprisingly. Many times we do not take
these into account and accordingly the starting point for our
numbers is that they integrate positive thinking.”
Budget numbers were rarely simple predictions, but negotiated
between business divisions (CGUs) and top management.6 Business
managers often proposed more optimistic ambitions than did top
management. It was suggested that “market views are not taken out
of the blue, but commonly they include more positive growth than a
little bit of pessimism would require” (Financial standard adviser,
Firm 4). An alternative was also possible; namely that business
management “did not show much confidence in the favourable future
development of their business … the level of ambition has to be much
higher” (Auditor 1, Audit firm 1). As a rule, business managers
produce “this ordinary story: Namely, those units that are really
underperforming want to show too optimistic figures. They want to
play for time and avoid being sacked” (CFO, Firm 5). So, budgeting
was a process of moderation introduced to make forecasts some-
what more in line with past performance so that the budget would
be more recognisable and therefore understood to be more reliable.
Top management had to be persuaded, and they typically moder-
ated budgeted growth because the past would, it was presumed,
rarely be completely overtaken by new strategies, and the aspira-
tions of ambitious business managers would have to be curtailed by
traces of their past which added weight to the reliability of budgets.
There was therefore a strong urge to embed budgets
organisationally.
This organisational embeddedness was reinforced by the strict
boundary that was set between the roles of business managers who
produced cash flow forecasts and the financial accountants who
put the calculation together in a goodwill impairment value. The
CFO of Firm 1 explained as follows:
“Business units update business plans [strategy] and provide us
with cash flows. After that, testing with the model, determination of
the discount rate, everything takes place here [in the financial ac-
counting office]. We do the testing and tell the business units the
outcomes. We do not let them do the calculation because we do not
want them to play with the figures. So we do it here. Hence, they
provide us with only the cash flows, and they should not be able to
recognise whether it will lead to impairment or not.”
Separation of responsibilities for the different traces of the
calculation would also increase reliability because actors would not
be able to “play with the figures”. This separation and distribution of
elements of the calculative practice among calculating actors who
would not regularly be in contact with each other would make the
resulting financial accounting calculations more impersonal and
therefore more reliable (Pentland, 1993; Power, 1996, 2003). The
calculation was not the product of any particular individual; it was a
procedure that organised and mobilised people from many corners
6 “There is the problem that it is so easy just to say that you have to
There is a separate concern, too big to discuss systematically here, in devising
how many CGUs take part in the cash flow stream: “We discuss continuously what
kinds of assets and liabilities should be allocated to this and that CGU. Is there extra
cash or something that does not belong to this business? It is a huge job, but very
central” (Auditor 1). The design of CGUs can influence how cash flows are taken into
account to the point that it can make the difference between impairment or not.
73
of the firm and beyond:
“The companies I've worked for before hired consultants to devise
scenarios for the business concerning the growth rate, changes in
the sales prices, or purchase prices as far as the principal raw
materials or their main products were concerned. And this, of
course, is the best way, as it doesn't represent the viewpoint of the
managers, but comes from outside. These are anyway the most
important factors, and they are mostly connected with the devel-
opment of cost or of the sales prices. We have hired consultants in
my current company, too.” (CFO, Firm 7)
The reliability of the budget would increase if traces “came from
the outside” e a “view from nowhere” (Nagel, 1986; Porter, 1992,
1994b). This does not mean that external experts would be better
at judging future price and cost; it only means that it was more
opportune to allow them to do this. Internally generated forecasts
would be understood as subjective and political traces, while
external voices, which were at a distance, would be understood to
speak more impersonally. They would thus increase the reliability
of forecasts.
External traces also took the form of “expectations and forecasts
for the whole economy. It is, after all, typical in many business areas
that the business grows hand in hand with the economy, give or take a
little” (CFO, Firm 7). Traces of the “whole economy” would also make
the forecasts more reliable as there would be a limit to how
different a firm could be from the whole economy; the firm had to
have traces that could be recognised from the outside. The CFO of
Firm 11 proposed:
“The process of defining the growth rates for terminal value periods
is not that demanding. We expect that the growth cannot deviate
much from GDP growth. We follow carefully what the other com-
panies use in their disclosures. Additionally, sometimes we have
received benchmarking material through our auditors.”
The ambition not to stand out too much and to be in line with
general developments was proposed as a source of reliability.
Estimated growth rates were traced from many and diverse places,
many, even most, of which would have external origins that pro-
posed growth not for the individual firm but for the industry or the
economy (Vollmer, Mennicken, & Preda, 2009). Such impersonal
traces were understood to be more reliable than propositions made
by individual firms.
Firms' top managers often also preferred external traces because
internal budgets were typically of only little assistance with regard
to predicting the future, as explained by the CFO of Firm 9:
“You make a calculation in 2005 according to which net sales
would be this and that much. When we look backwards in 2008
and the outcome differs, the alarm bells should already ring. Wait a
moment! You don't have a clue about the future of your business. It
is a good way to make comparisons. We always make it four years
backwards. We want to show that our forecasts do not show
growth of 10% and profit improvement of 15% while the real out-
comes are only 2% and 3%.”
Internal budgets were often mistaken and strategy was also
much in doubt for the reliability of forecasts:
forecast your future. Business is required to be forecasted for the
next 12 months, which is extremely difficult. These components are
so subtle. I haven't seen any strategy that was realised as it was
forecasted. By using these figures based on assumptions you will
74 J. Huikku et al. / Accounting, Organizations and Society 56 (2017) 68e83

just get some outcome but they are the assumptions you make at
that moment. However, there are so many components that they
will never be recognised the way they were forecast.” (Financial
standard adviser, Firm 4)
IFRS' requirement to calculate goodwill impairment value
assumed more certainty about budgets and strategy than firms
would normally attribute to them. This is why managers would call
the process a “calculative” one:
“It has been very challenging to forecast next year in this sort of a
global financial situation [recession]. Well, we forecast our strategy
as financial numbers for the next three years but just as she
[Financial standard adviser] already said, the terminal value has
the greatest impact and how it continues in the future, what is the
[terminal] value then … ? This illustrates how calculative this
testing process is.” (Senior Vice President, Financial Control, Firm
4)
had to be negotiated and settled between business and top man-
agers before it could be accepted as more than pure subjectivity;
two people's compromise would be more reliable. The budget also
became more reliable if it were aligned with historical performance
since it would not seem plausible if it were possible to outperform
history. Likewise, growth rates used for predicting terminal value
could be the whims of business managers, but when juxtaposed
with external reports about industry or GDP growth they were
understood to be more reliable traces for the calculation of good-
will impairments. The further the trace was from the place of the
business manager whose entity would produce the work (products
and services) that would result in cash flows, the more reliable it
was understood to be. The surprising point is that the trace would
gain reliability by moving away from the place of business activity.
4.2. Demarcating traces to estimate the WACC

As with the forecast of cash flows, the discount rate e WACC e

This was “calculative”, i.e. a simulation, because “terminal value
was also determined as part of the impairment testing process.
makes about 70e80 per cent of an outcome. Whether you have two or
Companies would use the same WACC for all accounting calcula-
five per cent as your growth rate will have a huge influence on the
tions, e.g. for capital budgeting and for impairment testing, as a CFO
outcome” (Senior Vice President, Finance, Firm 1). Forecasting was
(Firm 3) said:
important, but managers' experience was that traces such as bud-
gets, strategic plans, and cash flow forecasts would rarely be pre- “We test goodwill for impairment once a year if there are no sig-
cise. This is why the process of arriving at cash flow forecasts and nificant events during the year. We use the same WACC figures also
growth rates for impairment testing required managers to make in capital budgeting and other [management] calculations.”
the process less personal and therefore less the result of their own
guesses. Managers preferred external traces such as industry av-
This procedure dampened controversies around the WACC value
erages, which would relieve firms from the ambiguous task of
and here also the multiple uses of WACC components increased
budgeting precisely. This was often understood as increasing
reliability. Many actors, such as auditors, saw WACC determination
reliability:
as both easy and difficult, as explained by an auditor (Partner, Audit
“Once we now have a lot of impairment data, it would be inter- firm 1):
esting to see if ‘industry average’ figures will be used some day. It
“It is a very difficult issue, or in fact it is on the one hand difficult,
would make the production of these calculations remarkably easier
but on the other hand easy. You know the components of WACC.
if these industry average values would be generally known. Let's
They are there, and the major issues are related to the choice of debt
say, for example, that everybody would know how high growth
maturity e 10 or 30 years e and the capital structure to be used.
rates in the airlines industry are. That would be a way to get closer
Will we use the current capital structure or a targeted capital
to how reliable these calculations are. It would also diminish
structure set by the board of directors? Additionally, the cost of debt
pressure for accountants to manage these figures because all would
and equity are different in different companies and thus it is
know the industry averages. If we differ remarkably from that, we
difficult to make comparisons between them. Nevertheless, we can
should have a good reason” (Senior Vice President, Finance, Firm
compare the development within the company historically. Then
1).
we have beta. The problem may be the lack of a reference group for
beta values if your company is very unique.”
Mobilising traces across and beyond the firm added reliability.
The historical past, the multiple uses of budgets, the dampening of
WACC was typically calculated by the financial accounting office,
optimism, the second opinions of external experts, and the role of
not by business units. Often this would be based on knowledge
industry averages and of country growth rates all contributed to
about other companies via a reference group. Yet, it was also clear
making singular business managers' forecasts more reliable
that it would be problematic to use such a trace because there
because they were generally recognisable.
would be a “lack of a reference group” because even if there were “a
The traces used to propose cash flows could not be found in
big group of companies we use as a benchmark for determining our
financial accounting databases but had to be found elsewhere, such
WACCs, the problem is that these are companies with different capital
as in management accounting budgets (Quagli, 2011;
structures and financial status” (CFO of Firm 11). The call for a
Weißenberger & Angelkort, 2011), or in traces found across orga-
reference group for beta values (riskiness of the firm) was impor-
nisations, markets and the economy (Vollmer et al., 2009). The
tant because they would be the external traces of an item that
calculation required by IFRS asked financial accountants to organise
would be difficult to calculate. There was a preference for external
others inside and outside the firm so that their traces could be
traces: “We sometimes contact financial analysts that analyse our firm
assembled and summarised in the financial accounting office. The
or the corporate finance unit of our auditors in order to get bench-
cash flow forecast could not rely on readily available traces in the
marking data for our peer WACC determination” (SVP, Finance of firm
financial accounting infrastructure and these had to be discovered
1). The benchmark data was important even if the reference group
elsewhere. These traces would be more reliable if they were views
would be a problem, as mentioned above.
from nowhere. There was a series of procedures to increase reli-
Alternatively, firms could ask for help to check the quality of
ability. For example, the budget was relevant for cash flows, but it
their own calculations, as the CFO of Firm 5 explained:
 J. Huikku et al. / Accounting, Organizations and Society 56 (2017) 68e83
“First we discuss in our firm what we think would be a risk-free
rate, our capital structure, betas et cetera et cetera. Often we
check the components of WACC with an expert. This is not our own
auditor, but another auditing firm. We ask them how they see risk
premiums in certain circumstances, betas, and risk-free rates. We
do not ask directly what our WACC should be. Then we show these
[WACCs] to our own auditors. Then they comment on whether it is
ok or not. So far it has been ok for them.”
Firms could often calculate the WACC themselves and the
external check was frequently a ritual. It confirmed what firms
already knew and typically external checks would not change a lot.
Yet, it was a useful procedure because this was a way to “play it safe”
when there could be controversies about the quality of the traces:
“People widely acknowledge that there are many small firms listed
in Finland with less liquid shares. However, people have very
different views on how much the beta and other figures of these
companies need to be rectified when estimating WACC. Many listed
firms buy their WACC estimation from an external consulting firm,
because they e the board and the financial management e want to
play it safe” (Auditor 2).
The concern to “play it safe” was a result of the traces being
ambiguous because estimates of the elements of WACC were based
on a more average and industry-wide capital structure and eco-
nomic outlook in a basket of benchmark companies than those
facing the particular firm. This reliance on benchmarks could be
even stronger than that which CFOs considered reasonable,
because the reference group might be skewed, as a CFO (Firm 11)
explained:
“I think that at the moment we have a little bit of a stupid relation
between equity and debt in our WACC calculation, but I cannot
change it because we have to follow our peer companies. There are
financially distressed companies and companies with zero gearing
in our peer group. If I try to discuss that I would like to leave this
company out, they [auditors and FIVA] say, hey, are you now trying
to manipulate the calculation. I say to them, forget it, let's then use
the WACC we have now.”
While managers were concerned with the firm, auditors were
interested in the link to the industry average, which, however,
would be an average across very different types of capital struc-
tures. The variance between firms was significant. Managers gave
up their objections not because they were convinced, but because it
7
Sometimes managers did object, such as when potential recognition of
impairment losses would jeopardise the solidity of the company: “Then it [goodwill
impairment loss] strikes of course the equity. You make losses and lose your equity, and
this may affect your capacity to pay dividends”. (Senior Vice President, Finance, Firm
1); when it affected a company's capacity to pay dividends: “And with regard to this
impairment loss, if a company can manage it without a winding up situation, it has an
effect on profit distribution. A company may suffer because it cannot pay dividends”.
(Executive Vice President, Finance, Firm 2); when it triggered covenants: “We have
gearing ratio as a covenant: debt versus equity. The debt ratio increases when equity
decreases [due to the impairment losses]”. (Senior Vice President, Finance, Firm 1).
Such effects appear not to have been typical, though, because “When impairment
loss is recorded, the market has already earlier discounted it to the value of the com-
pany” (Senior Vice President, Financial Control, Firm 4). Managers refrained from
strong objections because impairments may not be too surprising. Studies by ac-
ademics, consulting firms, and the business press suggest that business acquisitions
and mergers are more likely to destroy than create shareholder value. It has been
reported that about 50e83% of M&As fail, which indicates the dilemma that it is
difficult to forecast future outcomes of M&As (de Camara & Renjen, 2004, p. 10;
Nguyen & Kleiner, 2003, p. 447).
75
was not important enough for them to insist on their views, which
auditors understood as manipulation. Here, reliability would in-
crease because managers refrained from objecting as much as they
could and allowed auditors' voices to be heard clearly. Criticism
towards the use of industry average WACC in the goodwill
impairment calculation was presented as follows by a financial
standard advisor (Firm 4):
“In the IFRS approach, WACC should be better linked to the firm in
question. If we think about these factors, such as WACC, they should
be considered from the business perspective, and not some absolute
and overall fair value for markets. It is not possible in reality.”
“In reality” it was not possible to produce traces of the firm's
capital structure. The WACC calculation was therefore understood
to be fragile by managers, and instead of increasing objection to
what was considered a dubious procedure,7 managers were com-
forted by others taking the risk. Expert propositions, benchmarks,
and quality checks were such risk management mechanisms. This
mechanism called on external experts to have a role even if firms
believed themselves to be quite capable of producing the goodwill
impairment test in-house, as a Senior Vice President, Finance (Firm
1) suggested:
“In my previous firm we used external financial advisors for
goodwill impairment testing at the beginning of 2000 to get the
expertise. Gradually we strengthened our in-house expertise by
recruiting employees from investment bankers to our corporate
finance unit and from auditing companies to our accounting
department. Accordingly, we no longer have a need to use these
external consultants.”
At the time of the research, external expertise was not so much
used as technical knowledge,8 as an Auditor (Audit firm1) said: “At
the beginning [firms] were seeking expertise about which models to
use and now … [they] conduct goodwill impairment testing in-house.”
It was, in many situations, possible for firms to calculate WACCs, but
their own calculations would be understood as lacking reliability.
When external sounding boards and experts tolerated it, it became
more reliable.
As with cash flow forecasts, WACC also required distribution. To
make traces more reliable, firms had to “play it safe.” Playing it safe
was a strategy for sharing the risk of judgments involved in the
calculation, e.g. of beta values. External experts were not only there
for their separate expertise (Gendron et al., 2007; Power, 1995;
Smith-Lacroix et al., 2012). They were also there to engage an un-
solvable conundrum e “in reality,” as it was articulated, it was
difficult to imagine a calculation that would take the firm's situa-
tion into account and not primarily be a calculation of the average
market. External valuation experts were called upon to handle the
conflict between auditors and managers e the conflict between
recognisability and business sense. Their solution was built on their
tables of betas and their judgment, but these only resolved the
problem of reliability and not the problem of the business sense.
8
External experts were in demand at the beginning of IAS 36 implementation,
around 2005, when no-one had sufficient knowledge about impairment calcula-
tions: For example, a Valuator commented: “At the beginning when these new
goodwill testing procedures were introduced, we made much more testing than
nowadays … [now] our expertise is more needed for purchase price allocation [PPA]
and acquisition price valuation purposes. If you compare PPA to goodwill impairment
testing, very often PPA is made by externals. It is a bit more specific; you need more
valuation expertise on intangible assets. It is not something that the in-house team does
every day” (Valuator 1, Managing Director of a major international financial advi-
sory firm).
76 J. Huikku et al. / Accounting, Organizations and Society 56 (2017) 68e83
Their solution fell on the side of recognisability where the firm
would be understood to be an example of the wider industry, while
managers retracted and decided that their objection was not
serious enough to suffer the consequences of attacks by auditors
and regulatory authorities.
4.3. Assuring the calculation
The audit committee had to be convinced of the calculation. It
tested goodwill impairment values, or their relative reliability, by
interrogating auditors and the CFO (Gendron & Be
dard, 2006;
Gendron, Be
dard, & Gosselin, 2004). The test of the goodwill
impairment test was comprehensive, as an auditor (Partner, Audit
firm 1) described:
“We have some fundamental issues to analyse. Are they really
committed to their business plan, are the goodwill impairment
figures related to it? What happens in the business environment?
Comparison to the previous years will be made. We do mechanical
testing using our own calculation models and compare, for
example, whether we have the same WACC. We do sensitivity
analysis and compare with their analysis. We check all the com-
ponents. How realistic and likely they are. Additionally, we discuss
with corporate management about the basic assumptions in the
testing. And last but not least, we add up all the units [CGUs] and
check whether the firm value generated by this exercise correlates
with the market value. I am also present at all the meetings of the
audit committees.”
Auditors checked traces and calculations and carried out
sensitivity analysis. In a sense they repeated financial accountants'
procedures to arrive at the impairment value, not just taking the
effort to calculate but also to review strategies and budgets. It was,
as suggested, analysis of “fundamental issues” and auditors intro-
duced these to the audit committee. This was the place where
everything came together for the second time (after the financial
accounting office) and was tested for reliability. This was where, as
Auditor 1 said:
“We have very detailed discussions [in the audit committee], what
we have done, what we have noticed in the calculations, which
kinds of issues are sensitive and have to be followed carefully, and
also we go through the quality and completeness of the
disclosures.”
The “detailed discussions” produced a concern with the “quality
and completeness” of financial accounting and its reputation was at
stake. The financial accounting office had to be able to stand trial for
its undertakings.
Auditors were prepared for audit committee meetings. Before
the meetings, they checked traces and made reasonableness tests.
They went through the strategy documents and related this to
sensitivity analysis and the firm's market value. The process of
convincing the audit committee would be “very, very serious” as the
CFO of Firm 5 said:
“To somehow illustrate the significance of goodwill, you can have a
look at these 25 pages of analysis about goodwill impairment
testing that we have gone through with our audit committee.
Discussing this material will take about 60% of the meeting time.
This is very, very, serious … As I already told you, our auditor has
said that if there is something he must audit, it is definitely the
goodwill value. And when he has done this, he sleeps well. Of course
they check other issues, too, but this is very important.”
These 25 pages of analysis were indications that care had been
taken to review the impairment value. The efforts developed to
review the calculation were often about the past, historical traces,
as explained by Auditor 4:
“Especially important is to go through the historical development;
how precise are the forecasts they [firms] have done in the past and
what have been the real outcomes of these forecasts; whether there
are variances or not. It is very fundamental to look at this even at
an annual level and see how the estimations of the previous year
and the result of impairment testing fit with the actual outcome.
Historical success and failure in this process are indications of the
capability of the firm's management to show whether they are able
to keep their promises or not. This is the first step in auditing
goodwill impairment tests.”
Auditors reviewed managers' and financial accountants' ability
to keep their promises about the future. This post hoc evaluation
was a test of forecast accuracy. The step was to test documents
against documents e traces against traces e and hardly the busi-
ness per se. Therefore, they checked less whether forecasts were
reasonable, whether the terminal value was reasonable, or whether
the WACC was reasonable, because these were business concerns.
They checked more whether, when put together, the impairment
value followed rules (Pentland, 1993; Power, 1997). In this sense the
calculation was distributed; it summed up calculations made
elsewhere, as explained by Auditor 4 (Audit firm 4):
“The auditing of goodwill impairment testing is very detailed. We
follow ISA 540.9 Internally we use our corporate finance guys for
valuation purposes. We ensure that the calculations are mathe-
matically correct by using our own calculation models. We
benchmark and analyse the WACC parameters. We consult the
industry experts to understand market views. We check whether it
is true when the firm says that the market will start to improve:
How it is related to the available data about the markets? I can say
that this testing is taken very carefully and seriously.”
The auditors' role in goodwill accounting was to review the
firm's strategy by testing whether “managers are committed to their
budgets and [cash flow] forecasts, and that these figures are linked
with the goodwill impairment calculations … we check whether we
have the same WACC … from external sources … and then check with
market values” (Auditor 1, Audit firm 1). Auditors checked the
process, e.g. via additional non-human devices such as the ISA
540.10 They focussed on checking the mechanics of the calculation
and asked external experts, consultants and advisors to provide the
business inputs for the calculation. Auditors assessed the consis-
tency of the narrative and the mathematical correctness of the
calculation. Auditors knew that they faced the risk “that somebody
will come and say, hey, this impairment loss recognition should have
been booked two years ago” (Director of corporate finance, Audit
9
International standard on auditing: Auditing accounting estimates, including
fair value accounting estimates, and related disclosures.
10
This intensive checking of calculation of value in use was super-checked by the
Finnish Financial Supervisory Authority (FIVA). Annually, FIVA checked listed firms
with high goodwill compared to their equity, and selected other firms for random
checks. If calculations were not satisfactory, FIVA would give a list of requirements
to be implemented for the subsequent year. In this way FIVA influenced the
emergence of a set of recommended practices for a goodwill impairment testing
process generally, and the determination of WACC specifically. In other words, even
though FIVA emerged after the publication of numbers, their scrutiny would in-
fluence future procedures. This was understood to improve reliability.
 J. Huikku et al. / Accounting, Organizations and Society 56 (2017) 68e83
firm 1), but “in practice, auditors cannot challenge future cash flows.
The figures are according to the plans that have been approved by our
board of directors” (Senior Vice President of Financial Control, Firm
4). Therefore, the role of auditors could not be to go beyond the
traces of cash flows because these had been negotiated elsewhere.
Auditors recounted financial accountants' work and evaluated
its “completeness and quality.” Yet, it was not a place where auditors
felt quite at home because they had reservations about numbers
with regard to the future; such numbers created discomfort
(Justesen & Skærbæk, 2005):
“For us auditors this is really difficult. We are not allowed to say
that they [cash flow forecasts and WACC] are guesses; they are
estimates. We are talking about future estimates. It is not an exact
science at all. Everything is based on judgments and estimates and
for us it is extremely difficult to say what the right figures would be”
(Partner, auditor; Audit Firm 3).
Such discomfort with numbers was present even if all traces
were checked. There was still doubt:
“In the audit committee the CFO presents the [goodwill impairment
testing] calculations. The members ask questions and try to figure
out the quality of the testing. I am also present there. I say what we
have done, and sometimes I have to explain to them more what this
testing is about. Sometimes the members want to hear the com-
ments about the testing from the auditor in order to ensure that the
CFO and the auditor tell the same story” (Auditor 2, Audit firm 2).
Auditors' doubts gave them two tasks in the audit committee.
One was to show the process e to recount financial accountants'
work e including having “to say what the testing is all about” which
was not obvious to audit committee members. Their work was
subtle and evasive. Another task was to develop a narrative that the
CFO, the auditor, and the audit committee could agree upon. Since
there was a limit as to how precise the auditor could be, the task
was also to assure agreement about a story. Not even auditors could
see through all the numbers because these were about the future.
The assurance of the impairment value repeated the financial
accounting process. In many ways, it was a mirror activity to the
financial accounting activity. Both processes were oriented towards
finding, qualifying, stabilizing and calculating traces. Both pro-
cesses were oriented towards increasing reliability by increasing
traces, by assembling many people and by trusting traces with
multiple purposes. Managers, auditors, external experts and
financial accountants refrained, however, from excessively strong
statements about the future.11 Instead, they placed traces in the
form of benchmarks, industry reports, and economic forecasts to
propose growth rates and WACC elements; they also returned to
the history of the firms, the business units, and the CGUs to allow
traces about the past to frame the future.
Yet, this reliability was relative and not absolute. Actors were
still worried because the future was still a problem irrespective of
all the checking and all the composition of the traces. Auditors
worried that they would not know enough about the business.
Managers worried that not enough of the business would be pre-
sent in the calculation. External experts worried about their role to
11
Research has noted that net present value calculations are only parts, and
sometimes even small parts, of the justification of strategic investments in firms;
other strategic qualities and organisational politics play important roles (Arnold &
Hatzopoulos, 2000; Bower, 1970; Brunzell, Liljeblom, & Vaihekoski, 2013; Miller &
O'Leary, 1997, 2005a, 2005b, 2007). This poses a problem for goodwill impairment
accounting where net present value calculations are assumed to be strong.
77
assure numbers that managers did not quite believe in, and
financial accountants worried that they had not been able to reach
far enough in their discovery of traces. The audit committee
worried whether it could develop a narrative that would explain to
inquisitive others that goodwill was not a feeble asset such as a
‘goodwill bomb’ as journalists often claimed. All this worrying in-
dicates that reliability was relative. Reliability was not absolute
because if such were the case this would efface worry. Reliability
was relative and the goal of the audit committee was to agree on a
narrative that was (hoped to be) reliable enough to be tolerable.
5. Relatively reliable and recognisable IFRS numbers:
goodwill impairment testing and calculation as process
This study investigates the practice of calculating goodwill
impairment value that is reliable enough. The detailed analysis of a
calculative practice e the goodwill impairment value calculation e
illustrates that a calculation develops by finding, qualifying, stabi-
lizing and calculating traces. This calculation carries with it a large
array of human actors and non-human traces that take part in its
realisation and make it reliable enough. Reliability is never cate-
gorically present/non-present; there are always hiccups, disbeliefs,
shortcuts and disagreements that are not comforting, yet are
tolerable. Therefore, reliability is relative.
5.1. Seeing value requires looking elsewhere
Goodwill impairment value is the result of a calculative practice
that is organised by financial accountants who find, develop and
mobilise traces that are made by many hands, minds and tools
within and beyond organisations. This is a process that turns traces
into financial statements through calculation. Here, goodwill is a
particularly difficult asset because it is a non-separable asset that
requires net present value calculation. While Barker and Schulte
(2016) suggest that it is, in principle, impossible to carry out
impairment tests because there are no institutions in support of it,
the present study suggests a possible way out of this problem by
focusing less on institutions and more on the practices of calcu-
lating. This practice makes the task of calculating a financial
number less oriented towards the mathematical operations on
traces (Vollmer, 2003, 2007) since there is a requirement of a net
present value calculation as delimited by IAS 36. The practice is
more concerned with identifying and organising traces that have to
be discovered in and selected from various places across time and
space within and beyond the firm.
The mathematics of calculation e how to make the world visible
e is a minor concern because it has already been defined by
regulation. IAS 36 develops a delimitation e a template of a
financial model e which is not easily given up as it is mandated. The
drift noted for example by Quattrone and Hopper (2001, 2005),
requires that accounting is understood as malleable, seems not to
be as urgent in the case of goodwill accounting where the delimi-
tation is clearer.
The delimitation defined by IAS 36 gives financial accountants a
task which is to find and support a single authoritative statement
about financial value. There seems to be more controversy about
demarcation concerning the selection of traces that make it
possible to end up in a calculation that is reliable enough. As the
study shows, this concern turns financial accountants into discov-
erers of traces more than calculators on traces. Their task is to
search for traces that add reliability which leads them to search
further away. Distance matters; at the distance are traces which are
not tainted by the hands of financial accountants and managers. As
effect, when readers use financial statements to see the firm they
look at traces about other things inside or outside the firm than its
78 J. Huikku et al. / Accounting, Organizations and Society 56 (2017) 68e83
entrepreneurial activities themselves. Accounts of entrepreneurial
activities are too singular and idiosyncratic to be recognised as
reliable because such traces are understood as personal. To ascer-
tain reliability, the study shows, calculative practices draw together
traces from within and beyond a firm's financial accounting data-
base (Vollmer et al., 2009), it is a mechanical production process
(Pentland, 1993; Power, 2003), it arranges a plethora of different
traces and cascading devices (Qu & Cooper, 2011), and by that
process it operates a fluid infrastructure rather than a stable one
(Lezaun, 2006). Each of these traces increases reliability because
they are understood as recognisable by readers. They are under-
stood as not manufactured for a particular calculative episode.
This is a strangely cumbersome process because financial ac-
countants are quite capable of performing the calculation them-
selves. They know all the aspects of the calculation because they
manage the delimitation that is used to find, qualify, stabilise and
calculate traces. The problem is that even if they do master the
process, their own calculation is subjective. Financial accountants
would be understood as speaking on their own (or their managers')
behalf rather than providing an impersonal (Porter, 1995) view
from nowhere (Latour, 1987; Nagel, 1986). To increase their ability
to speak for the many, financial accountants design strategies to
increase reliability. Generally, such strategies decentre individuals
and substitute them for traces that extend well beyond the firm's
financial database to industry outlook, benchmark numbers, na-
tional GDP projections; and they extend into human expertise such
as institutionalised expert positions and auditors' reviews. No one
carries the whole calculation e it is distributed much wider than
the financial accounting office e and many attempt to influence it
within and beyond the firm.
Reliability is related to the recognisability of the traces and to
the impersonality of the calculation. The three mechanisms already
mentioned (increasing number of external traces; increasing the
number of human allies; increasing multi-uses of traces) make
reliability stronger. This explains the paradox that external experts
can be necessary and superfluous at the same time, and that
managers may not be willing to insist on what they believe to be
business sense but auditors would believe to be manipulation. This
happens, for example, when financial accountants accept a WACC
that might have been more appropriate if the firm's actual capital
structure had been taken into account, rather than the average
capital structure reported in benchmark reports. It also happens
when ambitious business managers accept that they do not follow
their idiosyncratic growth aspirations related to their entrepre-
neurial activities but rather follow industry-wide historical growth-
rates.
This is where firms, by giving up their preferred WACCs and
growth-rates, substitute their views with those of external experts
and auditors. They may not believe experts' views but they can
maintain that they are at least not personal and therefore not
subjective. This also explains why calculative practices such as
financial accounting are less liberal than other accounts that
emphasise people's personal power to transform them largely at
will (Kalthoff, 2005; Quattrone & Hopper, 2005; Vollmer, 2007). It
also explains why the goodwill calculation may not be stabilised by
an infrastructure of referentiality which would provide a “singular
and unambiguous referent” (Lezaun, 2006, p. 499). Financial ac-
counting practices, as described above, negotiate frictions due to
concerns about recognisability and impersonality. This is a social
activity rather than a personal one. Therefore, as a centre of
calculation, the financial accounting office is a site that invites and
assembles actors; it is not mainly a cognitive mind or primarily an
interpretive process (Hines, 1988; McKernan, 2007). Nor does it
operate an unambiguous bureaucratic referent, because financial
accountants and others worry all the time about where a new
proposed trace may come from. The referent is an obligation or a
task for financial accountants who consider it ambiguous and frail.
Actors face the problem that they would like to be able to
recognise e i.e. create the ability to see e economic events, but this
creates the following paradox: the things people see when they
observe financial accounting are associated less with the firm's
entrepreneurial activities and more with things outside the firm.
They imagine that seeing the firm requires them to look at its past,
at its competitors, at industrial outlook, and at the statistical bu-
reaus that compile information on the economic development of
industries and countries; they may also have to listen to valuation
experts and auditors. Seeing the firm requires actors to look else-
where. This paradox makes the firm's calculation much more
average than it would be expected given the promises of IFRS.
These would most likely imply a much more differentiated and
singular notion of the individual firm’ entrepreneurial activities
(Barth, 2007; Bromwich, 2007; Whittington, 2008). For market-
based research, which cannot find singularity, this is a conun-
drum (Ramanna & Watts, 2012); for a network perspective on
calculative practices it is understandable.
This average is not necessarily conservative in an accounting
sense, however, because it relies on general economic outlooks and
prospects. Such traces can be more aggressive than historical traces
found in a particular firm because the economy may be proposed to
turn from slow growth to high growth. Average growth-rates
therefore may not be conservative. This is also clear when top
managers insist that divisional managers are too reluctant to
commit to growth to anticipate accountability. So, reliability is not
similar to conservatism even if it produces the tendency towards
the average growth rates for a firm, an industry, or a nation.
5.2. The proliferation of calculating agencies
When they draw things together, financial accountants compose
a large network of human actors and non-human traces. While
financial accountants occupy a position as a centre of calculation,
the list of allies that help to prepare the financial statement is much
larger. Preparers are not even restricted to within the firm, but have
positions across time and space. They are all those human actors
and non-human traces developed by calculating agencies beyond
the firm that help financial accountants make the goodwill calcu-
lation more reliable.
The study shows that there are many other calculating agencies
than the financial accounting office (Czarniawska, 2004). These
other agencies are scattered within and outside the firm such as
managers' negotiation of budgets, statistical bureaus and consul-
ting firms who all produce calculations. The financial accounting
office is a centre of calculation for the impairment test, which the
other calculating agencies are not; it draws together other calcu-
lating agencies' calculations. This is also partly what the Road Map
(Miller & O'Leary, 2007) does when it arranges firms in relation to
investments in technology and then develops acts of coordination.
The financial template does not intervene on the other calculating
agencies as the Road Map does, though, because the financial
template only requires other calculating agencies' calculation and
not their subsequent acceptance. The financial accounting office
moves things (such as calculations) into the firm and summarises
this in one number; the Road Map moves things (such as
commitment to invest) out of the firm into several instances of
autonomous decision-making.
Other calculating agencies add degrees of reliability because
their calculations are not seen as personal accounts. Particularly in
relation to IFRS, traces become a concern because of the require-
ment to calculate a future-oriented net present value for a non-
separable asset. This asset may be identified by the financial
 J. Huikku et al. / Accounting, Organizations and Society 56 (2017) 68e83
infrastructure but its value is not defined by it. This requires
attention to the details of finding traces, of qualifying traces, of
stabilizing traces and of calculating them, which may be more
cumbersome than most extant research reveals. There is a signifi-
cant burden in the production of traces, which indicates that even if
it is possible to make calculation a malleable practice, there is a
continuing worry about what the trace can help to explain. For
example, Dambrin and Robson (2011) point out that in some in-
stances it may be impossible to follow through and find the origin
of traces. Likewise, Frandsen (2009) illuminates the difficulties
encountered in tracing the translations between medical treat-
ments and financial numbers, and Preston (2006) shows that when
calculations are put to work they disregard many aspects of the
world that may return and haunt the effects of the calculation.
Therefore, the traces are possibly weaker than imagined by actors.
There is a limit with regard to how much can be known about what
they retain because the principles of recording and calculating can
both be multiple and difficult to monitor (see also Nobes, 2013;
Nobes & Stadler, 2013). Many traces lack their own history and
are guaranteed primarily by the reputation of their calculating
agencies (e.g. institutions such as valuation experts and auditors,
statistical bureaus and industrial association, and government
agencies) rather than by specific knowledge about their charac-
teristics and qualities, and therefore the only promise is that they
have been produced by a procedure even if this procedure may
itself be dimly lit or foggy. Sometimes it may be a source of ob-
jection such as when managers and auditors disagree on what to
take into account: recognisability or business sense. Therefore, the
calculations performed on top of such traces may seem to be much
clearer and more powerful than the traces would justify by
themselves.
In effect, the calculation (of goodwill) brings the firm into ex-
istence and makes it visible in a peculiar way (Hines, 1988; Kalthoff,
2005). The orientation towards the future which IFRS obligate
financial accountants to take into account brings to them a chal-
lenge to justify the future reasonably reliably. As the future cannot
be known the obligation to account for it is not trivial. The obli-
gation to account for the future harbours the demise of the calcu-
lation because the future will probably be different from the
calculation's predictions. Therefore, the productive power of the
calculation is ambiguous. It brings the firm into existence in a form
that most people do not believe strongly.
Financial accountants produce an ambiguous calculation but
they share the burden of reliability with many other calculating
agencies. Other agencies develop calculations which are taken as
input to the financial accounting office's work. Agencies such as
statistical bureaus, industry organisations, consulting firms with
valuation expertise, government agencies and international orga-
nisations such as EU and OECD all produce calculations of growth
rates, beta-values and interest rates. These agencies all produce
calculations and since these calculations are not produced for the
purpose of the individual firm's impairment test, they are under-
stood as impersonal. When the financial office invites remote
calculating agencies into its own calculative work, it makes many
agencies carry the burden of reliability. Financial accountants
organise a process but this is not only or primarily a process of
adding, subtracting, multiplying and dividing numbers (Vollmer,
2003, 2007) which is largely defined by the delimitation of IAS
36. It is more importantly a process of demarcation by identifying
the traces and their associated calculating agencies which are
hoped to be recognised by readers as having enough reliability.
The study thus shows that reliability is relative. It is not relative
to the world as such; it is relative to degrees of recognisability and
impersonality via the work to find, qualify, stabilise and calculate
traces. This work takes time and effort, and is never final.
79
6. Conclusions
This paper reports a qualitative study of financial accounting in
the area of goodwill impairment testing. Specifically, the study
investigates how a calculative practice produces goodwill impair-
ment value, and how the calculation is reliable. The study discusses
the calculative practice involved in producing goodwill impairment
values that are recognisable and impersonal enough to pass the test
of reliability.
It contributes and adds to the literature by firstly providing a
qualitative study of financial accounting practices, which is rare in
the literature.12 It shows that the financial accounting calculation
(under the influence of IFRS) is a distributed affair. The list of traces
that help in the preparation of financial accounts is much larger
than the financial accounting office, and even the firm. This list
includes human actors and non-human traces from beyond the
firm. Secondly, it contributes by showing how reliability is relative.
It is relative not to the world but to the degree of recognisability and
impersonality of traces.
Such a relative view on reliability is important vis-a-vis studies
of quantification that lead to institutionalised infrastructures of
referentiality (Espeland & Stevens, 2008; Lam, 2011; Porter, 1995).
Even highly institutionalised infrastructures, such as sheets of ac-
counts and general ledgers, have their ambiguities. These in-
frastructures are leaky (Callon & Latour, 1981) because in spite of
commensuration and institutionalisation (Espeland & Stevens,
1998) they are not comforting for practicing financial accoun-
tants. Through IFRS, fair market value is connected with the
financial accounting database in the sense that the general ledger
makes room for the item goodwill, but its value is found elsewhere.
It is not possible to demarcate clearly the possible entities that can
speak for a (future) market. The financial accountant cannot know
where to look for the future because it can be claimed by many
different traces. The structure of the calculation (delimitation) may
be stable but the set of traces (demarcation) that can or does make
it up is potentially endless. With IFRS the uncertainty about the
boundaries of the calculation has increased.
IFRS create two challenges for financial accounting. The first
challenge is that IFRS have created a leak in the financial accounting
database. It is less authoritative than before, when it was a stron-
g(er) infrastructure of referentiality (Lezaun, 2006) and, now, its
institutionalised power is reduced because it only develops a space
for an asset but not its value. The second challenge is that uncer-
tainty has been produced as to who creates the calculation. The
calculation is made from a heterogeneous network of people and
traces. This network cannot be restricted by the delimitation of an
infrastructure. There are human actors and non-human traces, each
of which can challenge the infrastructure and the calculation (of
goodwill), such as new reports from various calculating agencies
from governments or industrial associations through international
organisations and statistical bureaus that paint pictures of future
growth rates, and differences in ‘subjectivity’ between managers,
top managers, auditors and experts. These concerns may be part of
the heterogeneous calculating network, but they are hardly
delimited members of the infrastructure. Even when the calcula-
tion is stable (such as the goodwill impairment calculation) there is
potential for endless numbers of traces each of which can be
12
This is one qualitative study of financial accounting and more studies of this
kind are useful. The study has shown the general complexity of calculation, but
since the empirical evidence is developed from many different companies and sites,
it is relevant to study, as case studies, parallel processes within particular firms. This
would add insight to the differences between situations and not least more insight
to the relations between financial accounting and broader organisational processes
of planning, strategizing, and management accounting.
80 J. Huikku et al. / Accounting, Organizations and Society 56 (2017) 68e83

opened and new traces will be found. Nobody knows what a
calculation of goodwill impairment contains before it is challenged.
If unchallenged it is a black box that functions cleanly; when
challenged, for example in relation to recognisability and imper-
sonality, new traces can become important. The effect of this is that
reliability of a calculation and its traces is relative.
Considering Robson and Young (2009, p. 360) suggestion that
“studies of the performativity of new standards and accounting
calculations may help develop further insights into the construc-
tion and reconstruction of accounting and economic agency,” this
study of the goodwill impairment calculation shows the task facing
practicing financial accountants. They have to produce calculations
that bring things into existence. A goodwill impairment calculation
may be an extreme case because goodwill is a residual, non-
separable asset. This illustrates that the work produced by finan-
cial accounting is not to represent the world; it is to substitute the
world with a calculation that is recognisable and impersonal.
Preparers of financial statements may not be able to see the
future but they can have a look at budgets, predictions, experts'
opinions etc. These traces exist. These traces suggest that some-
thing does happen and that this something seems to have sub-
stance. Precisely what this something is, however, is ambiguous. At
least it is ironic that in order to see the firm e or its goodwill e
people have to look elsewhere around and beyond the firm. It is
also ironic that the traces that thus attract attention are those that
have become normalised so that the firm becomes an instance of
wider processes taking place in an industry or in the economy. The
result is an average firm e an average goodwill impairment value.
This may not have been the intention behind IFRS as it would
expect singular values for singular firms (Barth, 2007; Bromwich,
2007; Whittington, 2008).
The present study of the goodwill impairment calculation has
begun to illustrate some of the concerns that face financial ac-
countants. Financial accountants are collectors, producers and
managers of traces. At every corner they can expect, but not
necessarily experience, traces that are ascribed even more recog-
nisability and impersonality. Relative reliability makes the possi-
bility of new traces a continuing worry.
Acknowledgements
We appreciate the helpful comments of Christopher Chapman,
Seppo Ika€heimo, Juha Kinnunen, Martin Kornberger, Anna-Maija
Lantto, Richard Macve, Andrea Mennicken, Anette Mikes, Yuval
Millo, Michael Power, Keith Robson, Joni Young, two particularly
engaged anonymous reviewers and the editors. We also acknowl-
edge comments from participants of the AOS workshop at the LSE
(2011), EAA's Annual Congress in Ljubljana (2012), IPA Conference
in Cardiff (2012), and accounting seminars at University of Turku
(Turku School of Economics), Aalto University School of Business,
Auckland University of Technology, University of Auckland, Uni-
versity of New South Wales, and Macquarie University. Addition-
ally, we want to thank the people interviewed. The financial
support granted to Huikku and Silvola by Jenny and Antti Wihuri
Foundation (Grant numbers 2010/100091 and 2010/100499),
Marcus Wallenberg Foundation, Foundation for Economic Educa-
tion (Grant number 32138), and HSE Foundation is acknowledged
with gratitude.
Appendix 1. Interviews

Actor group First round Second round

Duration in minutes Duration in minutes

Academics:
1. Professor A, University 1 60
2. Professor B, University 1, A Chairman of the Board 65
3. Professor (Adjunct) C, University 1 60
4. Professor, University 2 30
5. Professor, University 3 40
6. Professor, University 4 40
Auditors (big-4):
1. Partner, auditor; Audit firm 1 90 93 þ 70
2. Partner, auditor; Audit firm 2 77
3. Partner, auditor; Audit firm 3 51
4. Partner, auditor; Audit firm 4 80
Financial supervisory authority:
1. Head of division, Financial reporting 120
2. IFRS expert A, Financial reporting 110 92
3. IFRS expert B, Financial reporting 87
Companies:
1. Senior Vice President, Finance, Firm 1 65 14*
2. Executive Vice President, CFO, Firm 2 46
3. Executive Vice President, Firm 3 56
4. Senior Vice President, Financial Control, Firm 4 80
5. Financial standard adviser, Firm 4 17*
6. CFO, Firm 5 53
7. CFO, Firm 6 100
8. CFO, Firm 7 50
9. CFO, Firm 8 55
10. CFO, Firm 9 60
11. CFO, Firm 10 90
12. CFO, Firm 11 36
13. Director, Group Accounting and Taxation, Firm 12 74
14. Group Financial Controller, Firm 12
Financial analysts and investors:
1. Senior Equity Analyst, Analyst 1 57 10*
2. Head of Strategies, Analyst 2 52 13*
 J. Huikku et al. / Accounting, Organizations and Society 56 (2017) 68e83 81

(continued )

Actor group First round Second round

Duration in minutes Duration in minutes

3. Portfolio Manager, Investor 1 114 11*

4. Portfolio Manager, Equities, Investor 2 40 10*
5. Managing Director, Investor 3 34
6. Senior Analyst, Analyst 3 61 16*
7. Head of Equities, Direct Equities, Investor 4 52 11*
8. Chief Executive Officer, Investor 5 35
9. Deputy Chief Investment Officer, Investor 6 44 10*
10. Analyst (Equity Research), Analyst 4 69 12*
11. Equity Analyst, Analyst 5 37
12. Analyst, Analyst 6 68
13. Managing Director, Investor 7 34
14. Analyst, Analyst 7 47
15. Head of Trading and Capital Markets, Analyst 8 34
16. Equity Analyst, Analyst 9 63
Media:
1. Journalist, Magazine 1 60
2. Journalist, Magazine 2 54
3. Analyst, Magazine 2
Financial advisors:
1. Partner, Director, Corporate Finance, Audit firm 1 127
2. Manager, Corporate Finance, Audit firm 1
3. Managing Director, Financial advisor firm 1 20
Creditors:
1. Chief Risk Officer, Bank 1 37
2. Head of Credit & Industry Analysis, Bank 2 75
3. Vice President, Credit Risks, Bank 3 65
4. Senior Credit Manager, Bank 2 36
5. Vice President, Head of Credit analysis, Bank 3 80
6. Senior Credit Analyst, Bank 3

The interview data comprises 53 semi-structured interviews with 55 interviewees and 10 follow-up (telephone) conversations (marked by asterisks) (57 h in total).

Appendix 2. Summarised interview structure (case-specific Challenges in goodwill impairment testing

for the actor groups)
Interviewee and goodwill impairment testing: the role and opinions
 What is your general opinion about the (goodwill) impairment
testing standard (IAS36)?
 What is your involvement with goodwill impairment testing?
 When does goodwill matter?
 How does goodwill matter?
 What are the benefits of goodwill impairment testing for
different actor groups?
 What does recognition of an impairment loss signal?
 Have attitudes towards goodwill and goodwill amortization
changed: now vs. the pre-IFRS era?
 How do your colleagues perceive goodwill impairment testing?
 Are you interested in goodwill and goodwill impairment
testing?
 Is the information obtained in goodwill impairment testing
useful?
 Is the information reported about goodwill impairment testing
understandable?
 Are analysts/investors interested in your goodwill (for firms)?
 Can the measurement of intangible assets be reliable?
 Do goodwill and recognition of an impairment loss have an
impact on company valuation?
 Does goodwill and a recognition of an impairment loss have an
impact on managerial behaviour?
 How could the standards or procedures related to goodwill be
improved?
 What are the challenges in goodwill impairment testing?
 How could a company hide a need to recognise an impairment
loss?
 What are the factors diminishing the potential for hiding a need
for a recognition of an impairment loss
 Why would hiding/delaying recognition of an impairment loss
be advantageous to companies/managers?
 Should there be more recognition of impairment losses during a
recession/depression?
Descriptions: processes and procedures related to goodwill
impairment testing
 Goodwill impairment testing process in companies
 Auditing of goodwill impairment testing
 Role of financial supervisory authorities
 Company valuation process
References
Ahrens, T., & Chapman, C. S. (2007). Management accounting as practice. Ac-
counting, Organizations and Society, 32(1e2), 1e27.
Anderson-Gough, F., Grey, C., & Robson, K. (2001). Tests of time: organizational
time-reckoning and the making of accountants in two multi-national ac-
counting firms. Accounting, Organizations and Society, 26(2), 99e122.
Arnold, P. J. (2012). The political economy of financial harmonization: the East Asian
financial crisis and the rise of international accounting standards. Accounting,
Organizations and Society, 37(6), 361e381.
Arnold, G. C., & Hatzopoulos, P. D. (2000). The theory-practice gap in capital
budgeting: evidence from the United Kingdom. Journal of Business Finance and
82 J. Huikku et al. / Accounting, Organizations and Society 56 (2017) 68e83
Accounting, 25(5/6), 603e626.
Barker, R., & Schulte, S. (2016). Representing the market perspective: fair value
measurement for non-financial assets. Accounting, Organizations and Society (in
press).
Barth, M. E. (2007). Standard-setters, measurement issues, and the relevance of
research. Accounting and Business Research, 37(3), 7e15.
Bayou, M. E., Reinstein, A., & Williams, P. F. (2011). To tell the truth: a discussion of
issues concerning truth and ethics in accounting. Accounting, Organizations and
Society, 36(2), 109e124.
Beatty, A., & Weber, J. (2006). Accounting discretion in fair value estimate: an ex-
amination of SFAS 142 goodwill impairments. Journal of Accounting Research,
44(2), 257e288.
Bougen, P. D., & Young, J. J. (2012). Fair value accounting: simulacra and simulation.
Critical Perspectives on Accounting, 23(4e5), 390e402.
Bower, J. (1970). Managing the resource allocation process. Boston, MA: Harvard
Business School.
Bromwich, M. (2007). Fair values, imaginary prices and mystical markets. In
P. Walton (Ed.), The Routledge companion to fair value and financial reporting (pp.
46e68). London: Routledge.
Brunzell, T., Liljeblom, E., & Vaihekoski, M. (2013). Determinants of capital budg-
eting methods and hurdle rates in Nordic firms. Accounting and Finance, 53(1),
85e110.
Burchell, S., Clubb, C., & Hopwood, A. G. (1985). Accounting in its social context:
towards a history of value added in the United Kingdom. Accounting, Organi-
zations and Society, 10(4), 381e413.
Burchell, S., Clubb, C., Hopwood, A. G., Huges, J., & Nahapiet, J. (1980). The roles of
accounting in organizations and society. Accounting, Organizations and Society,
5(1), 1e25.
Callon, M., & Latour, B. (1981). Unscrewing the big Leviathan: how actors macro-
structure reality and sociologists help them to do so. In K. Knorr-Cetina, &
A. V. Cicourel (Eds.), Advances in social theory and methodology. Towards an
integration of micro- and macro-sociologies (pp. 277e303). London: Routledge
and Kegan Paul.
de Camara, D., & Renjen, P. (2004). The secrets of successful mergers: dispatches
from the front lines. Journal of Business Strategy, 25(3), 10e14.
Cooper, D. J., Greenwood, R., Hinings, B., & Brown, J. L. (1998). Globalization and
nationalism in a multinational accounting firm: the case of opening new
markets in Eastern Europe. Accounting, Organizations and Society, 23(5e6),
531e548.
Czarniawska, B. (2004). On time, space, and action nets. Organization, 11(6),
773e791.
Dambrin, C., & Robson, K. (2011). Tracing performance in the pharmaceutical in-
dustry: ambivalence, opacity and the performativity of flawed measures. Ac-
counting, Organizations and Society, 36(7), 428e455.
Durocher, S., & Gendron, Y. (2011). IFRS: on the docility of sophisticated users in
preserving the ideal of comparability. European Accounting Review, 20(2),
233e262.
Erb, C., & Pelger, C. (2015). ‘Twisting words’? A study of the construction and
reconstruction of reliability in financial reporting standard-setting. Accounting,
Organizations and Society, 40, 13e40.
Espeland, W. N., & Stevens, M. L. (1998). Commensuration as process. Annual Review
of Sociology, 24, 313e343.
Espeland, W. N., & Stevens, M. L. (2008). A sociology of quantification. European
Journal of Sociology/Archives Europ
eennes De Sociologie, 49(03), 401e436.
Frandsen, A. (2009). From psoriasis to a number and back. Information and Orga-
nization, 19, 103e128.
Froud, J., Sukhdev, J., Leaver, A., & Williams, K. (2006). Financialization and strategy,
narrative and numbers. London: Routledge.
Gendron, Y., & Be
dard, J. (2006). On the constitution of audit committee effec-
tiveness. Accounting, Organizations and Society, 31(3), 211e239.
Gendron, Y., Be
dard, J., & Gosselin, M. (2004). Getting inside the black box: a field
study of practices in “effective” audit committees. Auditing: A Journal of Practice
& Theory, 23(1), 153e171.
Gendron, Y., Cooper, D. J., & Townley, B. (2007). The construction of auditing
expertise in measuring government performance. Accounting, Organizations and
Society, 32(1e2), 101e129.
Hines, R. D. (1988). Financial accounting: in communicating reality, we construct
reality. Accounting, Organizations and Society, 13(3), 251e261.
Hines, R. D. (1991). The FASB's conceptual framework, financial accounting and the
maintenance of the social world. Accounting, Organizations and Society, 16(4),
313e331.
Hopwood, A. G. (1992). Accounting calculation and the shifting sphere of the
economic. European Accounting Review, 1(1), 125e143.
Hopwood, A. G. (2000). Understanding financial accounting practice. Accounting,
Organizations and Society, 25(8), 763e766.
Hopwood, A. G. (2006). Accounting calculation and the shifting sphere of the
economic. European Accounting Review, 125e143.
Hopwood, A. G. (2009). Exploring the interface between accounting and finance.
Accounting, Organizations and Society, 34(5), 549e550.
Jordan, C. E., Clark, S. J., & Vann, C. E. (2007). Using goodwill impairment to effect
earnings management during SFAS no. 142's year of adoption and later. Journal
of Business and Economic Research, 5(1), 23e30.
Justesen, L., & Skærbæk, P. (2005). Performance auditing and the production of
discomfort. In S. Jo € nsson, & J. Mouritsen (Eds.), Northern lights e Accounting
research in Scandinavia (pp. 321e343). Malmø: Liber.
Kalthoff, H. (2005). Practices of calculation. Theory, Culture & Society, 22(2), 69e97.
Kettunen, J. (2014). Interlingual translation of international financial reporting stan-
dards: Understanding the problems and procedures of translation. Working paper:
University of Jyvaskyla.
Kornberger, M., Justesen, L., & Mouritsen, J. (2011). When you make manager,
we put a big mountain in front of you: an ethnography of managers in a
big 4 accounting firm. Accounting, Organizations and Society, 36(8),
514e533.
Lam, T. (2011). A passion for facts. Social surveys and the construction of the Chinese
nation-state, 1900e1949. Berkeley: University of California Press.
Latour, B. (1986). Visualisation and cognition: thinking with eyes and hand.
Knowledge and Society: Studies in the Sociology of Culture and Present, 6,
1e40.
Latour, B. (1987). Science in action; how to follow scientists and engineers through
society. Cambridge, Mass: Harvard University Press.
Latour, B. (1989). Clothing the naked truth. In H. Lawson, & L. Appignanesi (Eds.),
Dismantling truth. Reality in the post-modern world (pp. 101e126). London:
Weidenfeld and Nicolson.
Latour, B. (2005). Reassembling the social. Oxford: Oxford University Press.
Laux, C., & Leuz, C. (2009). The crisis of fair-value accounting: making sense of the
recent debate. Accounting, Organizations and Society, 34(6e7), 826e834.
Lezaun, J. (2006). Creating a new object of government: making genetically
modified organisms traceable. Social Studies of Science, 36(4), 499e531.
Macintosh, N. B. (2006). Commentary: the FASB and accounting for economic re-
ality; accounting - truth, lies, or “bullshit”? A philosophical investigation. Ac-
counting and the Public Interest, 6, 22e36.
Macintosh, N. B. (2009). Accounting and the truth of earnings reports: philosophical
considerations. European Accounting Review, 18(1), 141e175.
Macintosh, N. B., Shearer, T., Thornton, D. B., & Welker, M. (2000). Accounting as
simulacrum and hyperreality: perspectives on income and capital. Accounting,
Organizations and Society, 25(1), 13e50.
Massoud, M. F., & Raiborn, C. A. (2003). Accounting for goodwill: are we better off?
Review of Business, 24(2), 26e32.
Masters-Stout, B., Costigan, M. L., & Lovata, L. M. (2008). Goodwill impairments and
chief executive officer tenure. Critical Perspectives on Accounting, 19(8),
1370e1383.
McDonnell, J. (2005). Impairment of assets. Accountancy Ireland, 37(6), 17e19.
McGoun, E. G. (1997). Hyperreal finance. Critical Perspectives on Accounting, 8(1e2),
97e122.
McKernan, J. F. (2007). Objectivity in accounting. Accounting, Organizations and
Society, 32(1e2), 155e180.
Mennicken, A., & Millo, Y. (2012). Testing values: financialization and the emer-
gence of impairment rules. In Interdisciplinary perspectives on accounting con-
ference, 11e13 July, Cardiff, UK.
Miller, P. (2001). Governing by numbers: why calculative practices matter. Social
Research, 68(2), 379e396.
Miller, P., & O'Leary, T. (2007). Mediating instruments and making markets: capital
budgeting, science and the economy. Accounting, Organizations and Society,
32(7e8), 701e734.
Miller, P., & O'Leary, T. (1997). Capital budgeting practices and complementarity
relations in the transition to modern manufacture. Journal of Accounting
Research, 35(2), 257e271.
Miller, P., & O'Leary, T. (2005a). Capital budgeting, coordination and strategy: a field
study of inter-firm and intrafirm mechanisms. In C. S. Chapman (Ed.), Control-
ling strategy: Management accounting and performance measurement (pp.
151e182). Oxford: Oxford University Press.
Miller, P., & O'Leary, T. (2005b). Managing operational flexibility in investment
decisions. Journal of Applied Corporate Finance, 17(2), 18e24.
Miller, P., & Rose, N. (1990). Governing economic life. Economy and Society, 19(1),
1e31.
Nagel, T. (1986). The view from nowhere. Oxford: Oxford University Press.
Nguyen, H., & Kleiner, B. H. (2003). The effective management of mergers. Leader-
ship and Organization Development Journal, 24(8), 447e454.
Nobes, C. (2013). The continued survival of international differences under IFRS.
Accounting and Business Research, 43(2), 83e111.
Nobes, C., & Stadler, C. (2013). How arbitrary are international accounting classi-
fications? Lessons from centuries of classifying in many disciplines, and ex-
periments with IFRS data. Accounting, Organizations and Society, 38(8),
573e595.
Pentland, B. T. (1993). Getting comfortable with the numbers: auditing and the
micro-production of macro-order. Accounting, Organizations and Society,
18(7e8), 605e620.
Porter, T. M. (1992). Quantification and the accounting ideal in science. Social Studies
of Science, 22, 633e652.
Porter, T. M. (1994a). Making things quantitative. In M. Power (Ed.), Accounting and
science. Cambridge: Cambridge University Press.
Porter, T. M. (1994b). Objectivity as standardization: the rhetoric of impersonality in
measurement, statistics, and cost-benefit analysis. In A. Megill (Ed.), Rethinking
objectivity. Duke and London: Duke University Press.
Porter, T. M. (1995). Trust in numbers: The pursuit of objectivity in science and public
life. Princeton, New Jersey: Princeton University Press.
Power, M. (1992). The politics of brand accounting in the United Kingdom. European
Accounting Review, 1(1), 39e68.
Power, M. (1995). Auditing, expertise and the sociology of technique. Critical Per-
spectives on Accounting, 6(4), 317e339.
 J. Huikku et al. / Accounting, Organizations and Society 56 (2017) 68e83
Power, M. (1996). Making things auditable. Accounting, Organizations and Society,
21(2e3), 289e315.
Power, M. (1997). The audit society: Rituals of verification. Oxford: Oxford University
Press.
Power, M. (2003). Auditing and the production of legitimacy. Accounting, Organi-
zations and Society, 28(4), 379e394.
Power, M. (2010). Fair value accounting, financial economics and the transformation
of reliability. Accounting and Business Research, 40(3), 197e210.
Preston, A. M. (2006). Enabling, enacting and maintaining action at a distance: an
historical case study of the role of accounts in the reduction of the Navajo herds.
Accounting, Organizations and Society, 31(6), 559e578.
Quagli, A. (2011). Goodwill accounting as missing link between financial accounting
and management accounting: literature review and research agenda. Financial
Reporting, 17e39.
Quattrone, P., & Hopper, T. (2001). What does organisational change mean? Spec-
ulations on a taken for granted category. Management Accounting Research,
12(4), 403e435.
Quattrone, P., & Hopper, T. (2005). A ‘timeespace odyssey’: management control
systems in two multinational organisations. Accounting, Organizations and So-
ciety, 30(7e8), 735e764.
Qu, S. Q., & Cooper, D. J. (2011). The role of inscriptions in producing a balanced
scorecard. Accounting, Organizations and Society, 36(6), 344e362.
Ramanna, K. (2008). The implications of unverifiable fair-value accounting: evi-
dence from the political economy of goodwill accounting. Journal of Accounting
and Economics, 45(2e3), 253e281.
Ramanna, K., & Watts, R. L. (2012). Evidence on the use of unverifiable estimates in
required goodwill impairment. Review of Accounting Studies, 17(4), 749e780.
Robson, K. (1991). On the arenas of accounting change: the process of translation.
Accounting, Organizations and Society, 16(5e6), 547e570.
Robson, K. (1992). Accounting numbers as ‘inscription’: action at a distance and the
development of accounting. Accounting, Organizations and Society, 17(7),
685e708.
Robson, K. (1994). Inflation accounting and action at a distance: the sandilands
episode. Accounting, Organizations and Society, 19(1), 45e82.
Robson, K., & Young, J. J. (2009). Socio-political studies of financial reporting and
standard setting. In C. S. Chapman, D. J. Cooper, & P. B. Miller (Eds.), Accounting,
organizations and institutions. Essays in honour of Anthony Hopwood (pp.
83
341e366). Oxford: Oxford University Press.
Rose, N., & Miller, P. B. (2008). Governing the present. Administering economic, social
and personal life. Cambridge: Cambridge University Press.
Rottenburg, R. (2009). Far-fetched facts: A parable of development aid. Cambridge,
MA, USA: MIT Press.
Scott, J. C. (1998). Seeing like a state, how certain schemes to improve the human
condition have failed. New Haven, USA: Ebrary, Inc.
Smith-Lacroix, J., Durocher, S., & Gendron, Y. (2012). The erosion of jurisdiction:
auditing in a market value accounting regime. Critical Perspectives on Ac-
counting, 23(1), 36e53.
Suddaby, R., Cooper, D. J., & Greenwood, R. (2007). Transnational regulation of
professional services: governance dynamics of field level organizational change.
Accounting, Organizations and Society, 32(4e5), 333e362.
Suddaby, R., Gendron, Y., & Lam, H. (2009). The organizational context of profes-
sionalism in accounting. Accounting, Organizations and Society, 34(3e4),
409e427.
Thompson, G. (1987). Inflation accounting in a theory of calculation. Accounting,
Organizations and Society, 12(5), 523e543.
Vollmer, H. (2003). Bookkeeping, accounting, calculative practice: the sociological
suspense of calculation. Critical Perspectives on Accounting, 14(3), 353e381.
Vollmer, H. (2007). How to do more with numbers: elementary stakes, framing,
keying, and the three-dimensional character of numerical signs. Accounting,
Organizations and Society, 32(6), 577e600.
Vollmer, H., Mennicken, A., & Preda, A. (2009). Tracking the numbers: across ac-
counting and finance, organizations and markets. Accounting, Organizations and
Society, 34(5), 619e637.
Weißenberger, B. E., & Angelkort, H. (2011). Integration of financial and manage-
ment accounting systems: the mediating influence of a consistent financial
language on controllership effectiveness. Management Accounting Research,
22(3), 160e180.
Whittington, G. (2008). Fair value and the IASB/FASB conceptual framework project:
an alternative view. Abacus, 44(2), 139e168.
Wines, G., Dagwell, R., & Windsor, C. (2007). Implications of the IFRS goodwill ac-
counting treatment. Managerial Auditing Journal, 22(9), 862e880.
Young, J. J. (2006). Making up users. Accounting, Organizations and Society, 31(6),
579e600.