J O H A N N A M I E T T I N E N
U N I V E R S I TA S WA S A E N S I S 2 0 0 8
ACTA WASAENSIA NO 197
B u S I N E S S A d m I N I S T r AT I O N 8 0
A C C O u N T I N g A N d F I N A N C E
The Effect of Audit Quality on
the Relationship between
Audit Committee Effectiveness and
Financial Reporting Quality
Reviewers Professor Hannu Schadéwitz
Department of Accounting and Finance
Turku School of Economics
Rehtorinpellonkatu 3
20500 Turku
Finland
Professor Stuart Turley
Manchester Business School
University of Manchester
Booth Street West
Manchester
M15 6PB
UK
III
Julkaisija Julkaisuajankohta
Vaasan yliopisto Joulukuu 2008
Tekijä(t) Julkaisun tyyppi
Monografia
Julkaisusarjan nimi, osan numero
Johanna Miettinen
Acta Wasaensia, 197
Yhteystiedot ISBN
978–952–476–244–1
ISSN
0355–2667, 1235–7871
Sivumäärä Kieli
Vaasan yliopisto
Kauppatieteellinen tiedekunta
Laskentatoimen ja rahoituksen laitos
PL 700
65101 Vaasa 169 englanti
Julkaisun nimike
Tilintarkastuksen laadun vaikutus tarkastusvaliokunnan tehokkuuden ja
taloudellisen tiedon laadun väliseen suhteeseen
Tiivistelmä
Tässä tutkimuksessa tarkastellaan tilintarkastuksen laadun vaikutusta tarkastusvaliokunnan ja taloudellisen
tiedon laadun väliseen suhteeseen. Monet tutkimukset ovat tarkastelleet tarkastusvaliokunnan tehokkuuden,
tilintarkastuksen laadun ja taloudellisen tiedon laadun välisiä yhteyksiä. Tämän tutkimuksen tarkoituksena on
yhdistää nämä erilliset tutkimussuuntaukset ja muodostaa malli, joka tarkastelee tarkastusvaliokunnan
tehokkuuden ja tilintarkastuksen laadun vaikutusta taloudellisen tiedon laatuun yhdessä. Näin ollen
tutkimuksessa testataan seuraavia aikaisemmasta kirjallisuudesta johdettuja hypoteeseja: 1) Tarkastus-
valiokunnan tehokkuus parantaa taloudellisen tiedon laatua, 2) Tarkastusvaliokunnan tehokkuus lisää tilin-
tarkastuksen laadun kysyntää, 3) Tilintarkastuksen laatu parantaa taloudellisen tiedon laatua ja 4) Tilintarkas-
tuksen laatu toimii välittävänä tekijänä tarkastusvaliokunnan tehokkuuden ja taloudellisen tiedon laadun
välisessä suhteessa.
Hypoteesien testaamiseksi käytettiin kahta toisiaan täydentävää menetelmää: Causal Steps –menetelmää ja
Sobelin testiä. Causal Steps –menetelmän periaatteiden mukaisesti S&P (Standard & Poor’s) 1500 indeksiin
kuuluvista yrityksistä koostuvaa aineistoa testattiin erilaisilla regressioanalyyseillä. Tämän lisäksi Sobelin
testiä käytettiin soveltuvin osin sen testaamiseksi, onko välittävä vaikutus tilastollisesti merkitsevä.
Tilastollisissa testeissä tarkastusvaliokunnan tehokkuutta mitataan kolmella tekijällä: tarkastusvaliokunnan
koolla (ACSIZE), tarkastusvaliokunnan asiantuntemusta kuvaavalla suhdeluvulla (ACEXP) ja
tarkastusvaliokunnan kokoustiheydellä (ACMEET). Tilintarkastuksen laatua mitataan tilintarkastajalle
maksetuilla palkkioilla (AUDITFEE). Lopuksi taloudellisen tiedon laatua mitataan harkinnanvaraisilla erillä
(ACC). Näin ollen seuraavia, mittareiden avulla esitettyjä malleja, testataan empiirisesti: ACSIZE
AUDITFEEACC, ACEXPAUDITFEEACC ja ACMEETAUDITFEEACC.
Tutkimuksen tulokset tukevat viimeistä mallia osoittaen, että tarkastusvaliokunnan kokoustiheydellä on sekä
suora että välitetty vaikutus taloudellisen tiedon laatuun tilintarkastuksen laadun kautta. Nämä tulokset
osoittavat, että tarkastusvaliokunnan kokoukset eivät ole ainoastaan symbolisia, mutta ne myötävaikuttavat
taloudellisen tiedon laatuun ja tilintarkastuksen laatuun. Tämän lisäksi tulokset osoittavat, että
tarkastusvaliokunnan tehokkuuden ja tilintarkastuksen laadun välille muodostuu jatkumo, joka vaikuttaa
edelleen taloudellisen tiedon laatuun. Näin ollen tutkimuksen tulokset tukevat hypoteesia, jonka mukaan
tilintarkastuksen laatu välittää tarkastusvaliokunnan tehokkuuden vaikutusta taloudellisen tiedon laatuun.
Tutkimuksen tulokset koskien malleja ACSIZEAUDITFEEACC ja ACEXPAUDITFEEACC ovat
kuitenkin epäjohdonmukaisia, eivätkä ne yleisesti ottaen tue tutkimuksen hypoteeseja.
Asiasanat
tarkastusvaliokunnan tehokkuus, tilintarkastuksen laatu, taloudellisen tiedon laatu
V
Publisher Date of publication
University of Vaasa December 2008
Author(s) Type of publication
Monograph
Name and number of series
Johanna Miettinen
Acta Wasaensia,
Contact information ISBN
978–952–476–244–1
ISSN
0355–2667, 1235–7871
Number of
pages
Language
University of Vaasa
Faculty of Business Studies
Department of Accounting and
Finance
P.O. Box 700
FI-65101 Vaasa, Finland 169 english
Title of publication
The effect of audit quality on the relationship between audit committee
effectiveness and financial reporting quality
Abstract
This study examines the role of audit quality on the relationship between audit committee effectiveness and
financial reporting quality. A steady stream of literature has examined relationships between audit committee
effectiveness, audit quality and financial reporting quality. The objective of this study is to connect these
various streams of research to produce an integrated model depicting the effect of audit committee
effectiveness and external audit quality on financial reporting quality. Thus, the following hypotheses,
derived from prior research, are tested: 1) Audit committee effectiveness improves financial reporting
quality, 2) Audit committee effectiveness increases the demand for audit quality, 3) Audit quality improves
financial reporting quality, and 4) Audit quality mediates the relationship between audit committee
effectiveness and financial reporting quality.
To provide insight on the above hypotheses two complementary methods were employed, namely the Causal
Steps Method and the Sobel Test. Thus, following the principles of the Causal Steps Method a series of
multiple regression analyses are employed for a sample of S&P (Standard & Poor’s) 1500 firms which had
their fiscal years ending during the calendar year 2006. In addition, the Sobel Test statistics are calculated to
examine the significance of the mediated effect when applicable. In the analyses audit committee
effectiveness is measured by three variables, namely audit committee size (ACSIZE), audit committee
expertise ratio (ACEXP) and audit committee meeting frequency (ACMEET). Audit quality is measured by
audit fees (AUDITFEE) paid to the incumbent auditor. Finally, financial reporting quality is measured by
discretionary accruals (ACC). Thus, in terms of operational measures following models were tested: ACSIZE
AUDITFEEACC, ACEXPAUDITFEEACC and ACMEETAUDITFEEACC.
The results support the last model, showing that audit committee meeting frequency has both a direct effect
as well as mediated effect through audit fees on discretionary accruals. These results imply that audit
committee meetings are not merely symbolic but that they contribute to financial reporting quality as well as
external audit quality. In addition, there seems to be a sequence from audit committee effectiveness to audit
quality which further contributes to financial reporting quality. Thus, the results regarding the model
ACMEETAUDITFEEACC support the hypothesized mediated effect of audit committee effectiveness
on financial reporting quality through audit quality. However, the results regarding models
ACSIZEAUDITFEEACC, and ACEXPAUDITFEEACC are inconclusive and they do not support
the hypothesized relationships.
Keywords
audit committee effectiveness, audit quality, financial reporting quality
VII
ACKNOWLEDGEMENTS
Doing academic research is like mountaineering. At times progress may feel
difficult but the goal to reach the summit motivates one to press on. I look back
on my climb towards a doctorate feeling great satisfaction. This climb would not
have been possible without a safety harness and sound guy ropes. Thus, I am full
of gratitude to those individuals who safeguarded my journey during my doctoral
studies.
Firstly, I would like to express my gratitude to my supervisor Professor Teija
Laitinen, who encouraged me to begin my doctoral studies. I am deeply indebted
to Professor Laitinen for believing in my abilities to survive in the academic
world. I also owe a debt of gratitude to the official pre-examiners, Professor
Stuart Turley of the University of Manchester and Professor Hannu Schadéwitz of
the Turku School of Economics for their constructive comments on the
manuscript.
During my doctoral studies I was fortunate to work at the Department of
Accounting and Finance at the University of Vaasa. I wish to thank the professors
and colleagues for creating such an excellent environment to do research. I would
especially like to thank Professor Timo Salmi for organizing research seminars
held jointly with the Department of Accounting and Finance and the Department
of Mathematics and Statistics. I owe gratitude to all seminar participants for their
valuable comments, suggestions and advice on the study. I am especially grateful
to Professor Erkki K. Laitinen, Professor Timo Salmi and Professor Sami
Vähämaa for sharing their expertise on miscellaneous occasions during the
research project. I am deeply indebted to Jarkko Miettinen, Lic.Soc.Sc.
(Statistics) for his invaluable advice on methodological issues in the study. I
would like to thank the audit research group, Dr. Annukka Jokipii, Mr Kim
Ittonen and Mr Tuukka Järvinen, for their continuous support and most
importantly for their friendship. I am also grateful to Tellervo Niemi for helping
with the dissertation process and to Virginia Mattila for language consulting.
Different versions of this study have been presented at several seminars and
conferences. I would like to thank the participants at the Doctoral Tutorial in
Accounting (Jyväskylä, August 2004), the Conference in Accounting and
Performance Management Perspectives in Business and Public Sector
Organizations (Tartu, September 2005), the KPMG European Doctoral Colloquim
in Accounting (Dublin, March 2006), the 1st International Conference in
VIII
Accounting and Finance (Thessaloniki, September 2006), the USBE Workshop
on Auditing and Financial Accounting Research (Umeå, January 2007), Doctoral
Tutorial in Accounting (Turku, June 2007), and the EARNet PhD workshop
(Aarhus, October 2007) for their comments. I am especially grateful to Professor
Jere Francis, Professor David Hay, Professor Hannu Schadéwitz, Professor
Michael Shields and Professor Stefan Sundgren, for their excellent comments and
suggestions which contributed significantly to my study.
The completion of this research has required financial support from several
institutions and organizations. I am grateful to the University of Vaasa, the
Finnish Cultural Foundation and the European Commission for their financial
support, which enabled me to work as a full-time doctoral student. The generous
financial support provided by the following organizations and institutions is also
gratefully acknowledged: the Finnish Institute of Authorized Public Accountants,
the Foundation for Economic Education, the Marcus Wallenberg Foundation, the
Oscar Öflund Foundation, the Ostrobothnia Chamber of Commerce, the
Foundation for Promoting Equity Markets in Finland, the Finnish Foundation for
Economic and Technology Sciences, the Evald and Hilda Nissi Foundation, the
Finnish Savings Banks Research Foundation, the Emil Aaltonen Foundation, and
the Gustaf Svanljung Foundation
I would like to thank the Manchester Business School at the University of
Manchester for giving me the opportunity to serve as a Marie Curie Fellow in
their Postgraduate Research Programme. I am especially indebted to the faculty
members of the Manchester Accounting and Finance Group. The exceptional
studying atmosphere of the Manchester Business School inspired me to write up
my thesis. I am also grateful to my fellow students with whom I had the
opportunity to share thoughts and interesting experiences during my time in
Manchester.
Finally, I owe my deepest gratitude to my family for providing me with an
excellent base camp where I always felt safe and loved. I would like to thank my
parents Hilma and Kari for their unwavering support and encouragement
throughout my life. I also wish to thank my brother Jarkko for setting and
excellent example to follow. My deepest gratitude goes to my fiancé Jari for his
love, dedication and understanding throughout these years. With you by my side I
feel ready to reach for new heights in the future.
IX
CONTENT
ACKNOWLEDGEMENTS ................................................................................. VII
LIST OF FIGURES............................................................................................. XII
LIST OF TABLES.............................................................................................. XIII
LIST OF ABBREVIATIONS ..............................................................................XIV
1 INTRODUCTION.......................................................................................... 1
1.1 Research problem .................................................................................. 2
1.2 Contribution .......................................................................................... 6
1.3 Structure of the study............................................................................. 7
2 THEORETICAL FRAMEWORK OF THE STUDY....................................... 10
2.1 Agency theory.......................................................................................... 10
2.1.1 Agency theory and audit committees......................................... 13
2.1.2 Agency theory and external auditing......................................... 14
2.2 Frameworks.............................................................................................. 15
2.2.1 Framework for corporate governance........................................ 15
2.2.2 Framework for audit committee effectiveness........................... 17
2.2.3 Frameworks for audit quality..................................................... 18
2.2.4 Regulatory framework................................................................20
2.2.4.1 Requirements related to audit committees................ 21
2.2.4.2 Requirements related to external auditing................. 23
2.3 Positioning of the study........................................................................... 24
3 DEFINITIONS AND OPERATIONAL MEASURES OF
KEY CONCEPTS............................................................................................. 27
3.1 Audit committee effectiveness................................................................. 27
3.1.1 Audit committee size................................................................. 28
3.1.2 Audit committee independence................................................. 29
3.1.3 Audit committee expertise......................................................... 30
3.1.4 Audit committee meeting frequency.......................................... 31
3.2 Audit quality.............................................................................................31
3.2.1 Audit firm size............................................................................32
3.2.2 Auditor industry specialization.................................................. 34
3.2.3 Auditor tenure............................................................................ 35
3.2.4 Audit fees................................................................................... 36
X
3.3 Financial reporting quality....................................................................... 40
3.3.1 Definition of earnings management........................................... 41
3.3.2 Types of earnings management..................................................42
3.4 Application of the definitions and operational measures......................... 43
4 DEVELOPMENT OF THE RESEARCH MODEL......................................... 46
4.1 Audit committee effectiveness and financial reporting quality.............. 46
4.2 Audit committee effectiveness and audit quality.................................... 48
4.3 Audit quality and financial reporting quality.......................................... 52
4.4 Effect types.............................................................................................. 54
4.4.1 Moderation effect...................................................................... 55
4.4.2 Mediation effect........................................................................ 56
4.5 Selection of effect type............................................................................ 58
5 METHODOLOGY AND SAMPLE…………………………………..……... 60
5.1 Statistical mediation…………………………………………………… 60
5.1.1 Causal Steps Method…………………………………………. 61
5.1.2 Sobel Test…………………………………………………….. 64
5.2 Operational measures………………………………………………….. 65
5.2.1 Measurement of audit committee effectiveness........................ 68
5.2.2 Measurement of audit quality.................................................... 69
5.2.3 Measurement of financial reporting quality............................... 69
5.2.4 Control variables........................................................................ 71
5.2.4.1 Control variables related to discretionary accruals....71
5.2.4.2 Control variables related to audit fees....................... 72
5.3 Description of the analytic techniques..................................................... 74
5.3.1 Adaptation of the Causal Steps Method……………………….74
5.3.2 Adaptation of the Sobel Test......................................................78
5.4 Sample selection and descriptive statistics.............................................. 80
6 RESULTS…………………………………………………………………….. 83
6.1 Results related to model ACSIZEAUDITFEEACC........................ 83
6.2 Results related to model ACEXPAUDITFEEACC………………. 85
6.3 Results related to model ACMEETAUDITFEEACC……...……... 86
6.4 Discussion of the main results…………………………………………..89
6.5 Robustness of the main results…………………………………………. 93
XI
6.5.1 Results for model ACMEETAUDITFEEACC using
winsorized data……………………………………………….. 93
6.5.2 Results for additional model specifications…………………... 94
6.5.3 Results for industry adjusted audit fees………………………. 97
6.5.4 Results for path analysis……………………………………….98
6.5.5. Results for unexpected fees…………………………………....99
6.5.6 Results for the moderation effect…………………………..... 101
7 CONCLUSIONS………………………………………………….………… 103
7.1 Discussion………………………………………………………..….... 103
7.2 Limitations……………………………………………………………. 107
7.3 Future research………………………………………………………... 108
REFERENCES......................................................................................................... 110
APPENDICES...........................................................................................................133
APPENDIX 1. Summary of corporate governance standards……………….. 133
APPENDIX 2. Summary of studies related to the development of the
research model………………………………………………. 135
APPENDIX 3. Definitions for variables used in the analyses………………..138
APPENDIX 4. Sample selection criteria……………………………………...139
APPENDIX 5. Descriptive statistics of variables used in statistical analyses..140
APPENDIX 6. Correlation matrix……………….…………………………... 141
APPENDIX 7. Main relationships as scatterplots………………………….... 143
APPENDIX 8. Companies grouped by industries………………………….... 145
APPENDIX 9. Results for model ACSIZEAUDITFEEACC…..….…....146
APPENDIX 10. Results for model ACEXPAUDITFEEACC……...…….147
APPENDIX 11. Results for model ACMEETAUDITFEEACC………….148
APPENDIX 12. Results for model ACMEETAUDITFEEACC
(winsorized data)…………………………………………….. 149
APPENDIX 13. All measures for audit committee effectiveness included…... 150
APPENDIX 14. All measures for audit committee effectiveness included
and control variables excluded……………………………….151
APPENDIX 15. Results for model ACMEETINDFEEACC…..…………152
APPENDIX 16. Results for path analysis……………………………………...153
APPENDIX 17. Results for unexpected fees…………………………………. 154
APPENDIX 18. Results for moderation effect………………………………... 155
XII
LIST OF FIGURES
Figure 1. The mediating role of audit quality on the relationship between
audit committee effectiveness and financial reporting quality……. 3
Figure 2. Phases of the study........................................................................... 8
Figure 3. Corporate governance mosaic and financial reporting
quality (Cohen et al. 2004)..............................................................16
Figure 4. Determinants of audit committee effectiveness
(DeZoort et al. 2002)……………………………………………...17
Figure 5. Perceived audit quality (DeAngelo 1981a; DeAngelo 1981b)....... 19
Figure 6. Determinants of audit quality (Watkins et al. 2004)……………...20
Figure 7. Moderation effect (Baron et al. 1986; Holmbeck 1997)…….........56
Figure 8. Mediation effect (Baron et al. 1986; Holmbeck 1997)…………... 57
Figure 9. Indirect effect (Streiner 2005)……………………………….........58
Figure 10. Mediation model diagrammatically and regression models to
test the mediated effect (MacKinnon et al. 2002)………………...63
Figure 11. Hypothesized relationships between measures of audit committee
effectiveness, audit quality and financial reporting quality............ 67
Figure 12. Summarized results for model ACSIZEAUDITFEEACC…. 84
Figure 13. Summarized results for model ACEXPAUDITFEEACC….. 86
Figure 14. Summarized results for model ACMEETAUDITFEE
ACC ……………………………............................................... 88
Figure 15. Summarized results for model ACMEETAUDITFEE
ACC using winsorized data……..……………………………….. 94
XIII
LIST OF TABLES
Table 1. Regression models required by the Causal Steps Method
(Baron et al. 1986)……………………………………………….. 62
Table 2. Regression models estimated to test model
ACSIZEAUDITFEEACC………...……….………………... 76
Table 3. Regression models estimated to test model
ACEXPAUDITFEEACC……..……………….………….....77
Table 4. Regression models estimated to test model
ACMEETAUDITFEEACC…………..………….…………..77
Table 5. Summary of the main results…………………………………….. 92
Table 6. Regression models estimated to test measures of audit
committee effectiveness simultaneously……………………….....95
Table 7. Regression models estimated to test measures of audit
committee effectiveness excluding control variables……………. 96
Table 8. Regression models estimated to test model
ACMEETINDFEEACC………...………………………….. 98
XIV
LIST OF ABBREVIATIONS
AMEX American Stock Exchange
CEO Chief Executive Officer
CPA Certified Public Accountant
ERC Earnings Response Coefficient
GAO Government Accountability Office
GAAP Generally Accepted Accounting Principles
GAAS Generally Accepted Auditing Standards
IPO Initial Public Offering
NASDAQ National Association of Securities Dealers Automated Quotation
System
NYSE New York Stock Exchange
SEC Securities and Exchange Commission
SOX Sarbanes-Oxley Act
1 INTRODUCTION
This study focuses on two principal actors of corporate governance, namely audit
committees and external auditors (Cohen, Krishnamoorthy & Wright 2004).
These corporate governance players have a common objective in ensuring
financial reporting quality. In addition, audit committees are responsible for
hiring and overseeing external auditors’ work (e.g. SOX 2002), which gives them
great authority over audit quality. When these responsibilities are taken into
consideration as a whole, audit quality can be considered to have an effect on the
relationship between audit committee effectiveness and financial reporting
quality.
As pointed out by Ball (2008) financial reporting is an important economic
activity. The demand for financial reporting arises from information asymmetry
between the managers and owners of the company (Jensen & Meckling 1976;
Healy & Palepu 2001). High quality of financial reporting is a prerequisite for an
efficient allocation of capital (Healy et al. 2001). Thus financial reporting quality
is of interest to those who use financial reports for decision-making. External
financial statement users, including current and potential investors, creditors, and
others need reliable financial information on which to base their resource
allocation decisions. Auditees, including management, audit committees, and
boards of directors have an interest in producing high quality financial reports, for
example, to help to reduce the cost of capital and to attract potential investors. In
addition, regulators and standard setters can increase the effectiveness of capital
markets by promulgating rules and regulations that help ensure financial reporting
quality (ISB 2000; Schipper & Vincent 2003).
One of the objectives of a company’s corporate governance system is to ensure
the quality of that company’s financial reporting (Abbott & Parker 2000; Abbott,
Parker & Peters 2004; Klein 2003; McMullen & Raghunandan 1996; Stewart &
Munro 2007). However, there have been concerns about corporate governance
quality in the present environment, where severe corporate failures have come to
light. It has been found that the perceived reliability of audited financial
information has declined. By contrast, the perceived relevance of audited
financial information has increased (Hodge 2003).
Due to these concerns, the impact of corporate governance on a company’s
financial reporting quality has attracted increasing emphasis among accounting
researchers in recent years (Pomeroy & Thornton 2008). Prior research has
2 Acta Wasaensia
indicated that both audit committees and external auditors are able to decrease
management discretion over accounting issues and therefore are able to enhance
financial reporting quality (e.g. Beasley, Carcello, Hermanson & Lapides 2000;
Frankel, Johnson & Nelson 2002; Geiger & Rama 2003; Abbott et al. 2004;
Bédard, Chtourou & Courteau 2004; Larcker & Richardson 2004; Bradbury, Mak
& Tan 2006). In addition, studies have shown that audit committees are associated
with the demand for high quality audit (e.g. Abbott et al. 2000; Abbott, Parker,
Peters & Raghunandan 2003a; Chen, Moroney & Houghton 2005).
The objective of this study is to investigate the interplay between audit
committees and external auditors in ensuring financial reporting quality. More
specifically, as indicated by prior research, it is hypothesized that both audit
committee effectiveness and audit quality contribute to financial reporting quality.
In addition, audit committee effectiveness is expected to increase audit quality.
Finally, these relationships are connected into a more comprehensive model
which suggests that audit quality may mediate the relationship between audit
committee effectiveness and financial reporting quality. The main contribution of
the study arises from the development of the mediation model as well as from its
empirical investigation.
1.1 Research problem
The role of external auditing in a company’s corporate governance function is a
complex one since the auditor interacts with several other actors of the corporate
governance function, such as the audit committee, the board of directors, the
internal auditors and the management (Cohen et al. 2004). From amongst this
complex net of interactions this study focuses on the relationship between audit
committees and external auditors in ensuring financial reporting quality. Although
earlier studies have recognized that audit committees and external auditors serve
as important determinants of financial reporting quality, the relationship between
these corporate governance actors has not been thoroughly explored. This is
because much of this research has adopted a direct or main effect approach and
less attention has been paid to the possibility of more complex effect types which
would enable a more thorough analysis of the underlying mechanisms of the
relationships. This approach enables research providing a more comprehensive
description of companies’ corporate governance function and is thus of greater
practical significance to interest groups in financial reporting.
Accordingly, this research develops and tests a model that establishes
relationships between: 1) audit committee effectiveness and financial reporting
Acta Wasaensia 3
quality, 2) audit committee effectiveness and audit quality, and 3) audit quality
and financial reporting quality. In the model developed, audit quality is expected
to have a mediating role in the relationship between audit committee effectiveness
and financial reporting quality. The mediating role maintains that the effect of
audit committee effectiveness on financial reporting quality goes through audit
quality, at least partly.
The model is summarized in Figure 1. The theoretical concepts of the model are
illustrated at the top of the figure. These are audit committee effectiveness, audit
quality and financial reporting quality. Audit committee effectiveness is modelled
as the independent variable, audit quality as the mediator and financial reporting
quality as the dependent variable in the model. Operational measures for the
variables are illustrated at the bottom of Figure 1. Audit committee effectiveness
is measured by three variables, namely audit committee size, audit committee
meeting frequency, and audit committee expertise ratio. Audit quality is measured
by audit fees paid to the incumbent auditor and financial reporting quality is
measured by discretionary accruals.
Audit committee effectiveness is the independent variable in the model and in the
empirical analyses it is measured by three variables. More specifically, audit
committee effectiveness is expected to increase along with audit committee size
(ACSIZE), expertise ratio (ACEXP), and meeting frequency (ACMEET) (e.g.
Bédard et al. 2004; Goodwin-Stewart & Kent 2006; Vafeas & Waegelein 2007).
The model suggests that audit committee effectiveness improves financial
reporting quality and increases the demand for audit quality. In addition, audit
committee effectiveness is expected to have a mediated effect on financial
reporting quality through audit quality.
Independent variable Mediator Dependent variable
Theoretical
concepts
Operational
measures
Audit committee
effectiveness
-Audit committee size
-Audit committee
meeting frequency
-Audit committee
expertise ratio
-Audit fees -Discretionary accruals
Audit quality Financial reporting
quality
Figure 1. The mediating role of audit quality on the relationship between audit
committee effectiveness and financial reporting quality.
4 Acta Wasaensia
Audit quality is the mediating variable in the model. Audit quality is measured by
audit fees (AUDITFEE) paid to the incumbent auditor. High levels of audit fees
are expected to indicate higher audit engagement effort and thus better audit
quality (e.g. Carcello, Hermanson, Neal & Riley 2002; Abbott et al. 2003a;
Srinidhi & Gul 2007) after controlling for other variables related to pricing of
audit services. Thus, audit quality as determined by audit fees, is expected to
improve financial reporting quality. In addition, audit quality is expected to
mediate the relationship between audit committee effectiveness and financial
reporting quality.
Financial reporting quality is the dependent variable in the model. Following
Watkins, Hillison and Morecroft (2004) financial reporting quality refers to how
well financial statement information reflects the true economic circumstances of
the company. Financial reporting quality is measured by discretionary accruals
(ACC)1 estimated using a modified Dechow and Dichev (2002) model. It is
proposed that a higher value of discretionary accruals indicates a greater level of
earnings management and thus, lower financial reporting quality.
Collectively, the model is used to test following hypotheses:
H1: Audit committee effectiveness improves financial reporting quality.
H2: Audit committee effectiveness increases the demand for audit quality.
H3: Audit quality improves financial reporting quality.
H4: Audit quality mediates the relationship between audit committee
effectiveness and financial reporting quality.
The model developed is tested with two complementary methods: the Causal
Steps Method and the Sobel Test. The Causal Steps Method (see Baron and
Kenny 1986) involves probing of four conditions which are analogous with the
hypotheses of the study. Thus, the Causal Steps Method involves a multistage
regression analysis which assesses following conditions for mediation: 1) the
independent variable must have a significant effect on the dependent variable, 2)
the independent variable must have a significant effect on the mediator, 3) the
mediator must have a significant effect on the dependent variable, and 4) the
independent variable should have no effect on the dependent variable when the
mediator is held constant (full mediation) or the effect of independent variable
1
Refers to accruals in which management has discretion over.
Acta Wasaensia 5
should become smaller when the mediator is held constant (partial mediation)
(Baron et al. 1986). If all conditions of the Causal Steps Method are met the Sobel
Test is also employed. In these situations the Sobel Test provides information
regarding the significance of the mediated effect.
The data for this study consists of a sample of S&P 1500 companies2. Data is
obtained from several sources. Data related to audit committee effectiveness are
obtained from Institutional Shareholder Services (ISS). Audit fee data are
obtained from the Audit Analytics Database. Finally, financial data is gathered
from Thomson Financial Database. The procedures of the Causal Steps Method
are carried through separately for the three measures of audit committee
effectiveness. Thus, the following models are tested:
1) ACSIZEAUDITFEEACC,
2) ACEXPAUDITFEEACC, and
3) ACMEETAUDITFEEACC.
The results of the Causal Steps Method as well as the Sobel Test provided support
for the last model, whereas the results concerning the first two models are
inconclusive. In general, the results show that variables related to audit committee
composition are not sufficient measures for audit committee effectiveness in the
US regulatory environment, likely because the US regulations allow little
variation in audit committee composition which results in companies setting up
homogeneous audit committees in terms of their size and expertise ratio3. Thus,
the relationships between audit committee composition measures and financial
reporting quality measure as well as audit committee composition measures and
audit quality measure cannot be observed with the data employed in the present
study.
However, the results show that audit committee meeting frequency can be used to
differentiate audit committee effectiveness between companies. More specifically
with regard to model ACMEETAUDITFEEACC the following results are
found. Firstly, the results reveal that audit committee meeting frequency has a
negative effect on discretionary accruals. This indicates that more active audit
committees are better able to restrict management influence over discretionary
2
Refers to the S&P (Standard & Poor’s) 1500 Composite Index which encompasses all stocks in
the S&P 500, S&P 400, and S&P 600 indices.
3
See e.g. Hay, Knechel & Ling (2008) for more discussion on this issue.
6 Acta Wasaensia
accruals and thus ensure financial reporting quality more effectively. Secondly,
audit committee meeting frequency is found to have a positive effect on audit
fees. This result has several plausible explanations. Audit committee meetings
may require more work by external auditors, which leads to higher audit fees.
Alternatively more active audit committees may require greater audit quality and
audit coverage, which leads to an increase in audit fees. Thirdly, audit fees are
found to have a modest negative effect on discretionary accruals. This result
implies that higher fees reflect greater audit effort, which leads to greater
monitoring provided by auditors and thus, to better financial reporting quality.
Finally, it was found that audit fees partially mediate the relationship between
audit committee meeting frequency and discretionary accruals. The fact that only
partial mediation was found indicates that there may be other control
mechanisms, currently beyond the scope of the model developed, which can
function as mediators in the relationship between audit committee effectiveness
and financial reporting quality. These control mechanisms include, for example,
internal auditing and the internal control mechanism of the company.
1.2 Contribution
This study adds to the existing knowledge regarding the interplay between audit
committees and external auditors in ensuring financial reporting quality. More
specifically this study develops a model in which audit quality mediates the
relationship between audit committee effectiveness and financial reporting
quality. This study contributes to the existing literature both theoretically as well
as empirically.
Firstly, the model developed can be placed in theoretical frameworks concerned
with corporate governance (Cohen et al. 2004), audit committee effectiveness
(DeZoort, Hermanson, Archambeault & Reed 2002) and audit quality (Watkins
et al. 2004). This study contributes to these frameworks by providing empirical
evidence for some of the specific aspects they address. Cohen et al. (2004)
discuss the interrelationships between various corporate governance actors
functioning inside and outside the company. This study focuses on the
interrelationship of two corporate governance actors, namely audit committees
and external auditors. The framework by DeZoort et al. (2002) addresses the
determinants of audit committee effectiveness. According to the framework this is
dependent upon composition, authority, resources and diligence of the audit
committee. Consistently the operational measures of audit committee
effectiveness employed in this study are related to composition and diligence
components of audit committee effectiveness. Finally, Watkins’ et al. (2004)
Acta Wasaensia 7
framework models the drivers, components and products of audit quality. The
framework maintains that there is a sequence from the drivers of audit quality to
components of audit quality and further to products of audit quality. The model
developed in the present study is analogous with this view: audit committee
effectiveness is expected to lead to audit quality, which is further expected to
result in financial reporting quality.
Secondly, the effect of audit committee effectiveness and external audit quality on
financial reporting quality is an area which has commanded considerable research
interest in empirical studies. In summary, prior research has determined
relationships between: 1) audit committee effectiveness and financial reporting
quality (e.g. Beasley et al. 2000; Abbott et al. 2004; Bédard et al. 2004), 2) audit
committee effectiveness and audit quality (e.g. Abbott & Parker 2001; Abbott et
al. 2003a; Vafeas et al. 2007), and 3) audit quality and financial reporting quality
(e.g. Nelson, Elliott & Tarpley 2002; Krishnan 2005; Srinidhi et al. 2007). This
study contributes to prior research theoretically by placing these relationships into
a more comprehensive model. The core of the model developed is the assumption
that audit quality mediates the relationship between audit committee effectiveness
and financial reporting quality.
Thirdly, the results of the earlier studies regarding the relationships between audit
committee effectiveness, audit quality and financial reporting quality have
naturally been obtained in several countries and at different times using different
sets of data. This study also contributes to earlier studies empirically by
examining whether these relationships can be found using a single set of data of
US companies from year 2006. If relationships can be found this study provides
further support for prior studies and shows that their results have not been driven,
for example, by special features in the data. Another empirical contribution of this
study arises from the analysis of the mediated effect. Prior audit research has not
addressed mediation models and thus, has not employed methods suitable to test
mediated effects. This study adopts methods used in other fields of social sciences
to test the mediation hypothesis.
1.3 Structure of the study
The overall structure of the study is presented in Figure 2. In general, the chapters
of the study form four main phases which are as follows: 1) Introduction, 2)
Theory and prior literature, 3) Methodology and results, and 4) Conclusions. The
purpose of the first phase is to present the research problem area, the research
problem as well as the contributions of the study. The second phase explains the
8 Acta Wasaensia
theoretical foundations of the study and therefore agency theory and relevant
frameworks are discussed. This phase also provides definitions of key theoretical
concepts as well as operational measures for these concepts employed by prior
studies. Finally, the research model is derived from prior empirical research. The
third phase explains the statistical methods employed as well as the adaptation of
these methods. This phase also presents the results of this study. The fourth phase
provides concluding remarks including a discussion linking the results of the
present study with the existing literature. In addition, the implications of the
results and future research opportunities are discussed.
Figure 2. Phases of the study.
More specifically, this study consists of seven chapters organised in the following
way. The first chapter introduces the research problem and discusses the
contributions of the study. The second chapter introduces the theoretical
perspectives underlying the research problem area. Thus, this chapter introduces
Research phases
Phase 1: Introduction 1.Introduction -Introduce the research
problem area
- Introduce the research
problem
-Explain contributions of
the study
-Introduce agency theory
and relevant frameworks
-Provide definitions of the
key concepts
-Develop the research
model
2. Theoretical framework
of the study
3. Definitions and
operational measures of
key concepts
4. Development of the
research model
Phase 2: Theory and prior
literature
Main purposes Main chapters
Phase 4: Conclusion
Phase 3: Methodology and
results
7. Conclusions
5. Methodology and
sample
6. Results
-Discuss the connections
between present results and
prior literature
-Discuss implications of the
results of the study
-Discuss future research
-Introduce the methodology
and its adaptation
-Introduce the sample
-Present results of the main
analyses as well as
additional analyses
Acta Wasaensia 9
the basic premises of agency theory, which explains the demand for financial
reporting as well as corporate governance. This chapter also introduces
frameworks related to corporate governance, audit committee effectiveness and
audit quality. In addition the second chapter introduces the regulatory
environment of the study. This discussion is focused on regulations related to
audit committee effectiveness and audit quality. The third chapter introduces
definitions as well as measures for audit committee effectiveness, audit quality
and financial reporting quality used in prior studies. This discussion is based on
both theoretical and empirical research. The fourth chapter formulates the
hypotheses by reviewing studies focusing on relationships between audit
committee effectiveness, audit quality and financial reporting quality. This
chapter also discusses the alternative effect types which can be chosen to describe
the relationships. Finally, the research model is introduced. The fifth chapter
introduces the methodology for testing the mediation effect. In addition, this
chapter introduces the operational measures for variables and explains how
methods related to mediation effect are adopted in the present study. Chapter Six
presents the results of the analyses. Finally Chapter Seven provides concluding
remarks for the study.
10 Acta Wasaensia
2 THEORETICAL FRAMEWORK OF
THE STUDY
A company’s corporate governance function includes five main actors:
management, the board of directors, the audit committee, the external auditors
and the internal auditors (Cohen et al. 2004). One of the main objectives of
corporate governance is to ensure a company’s financial reporting quality. The
interaction among corporate governance actors is crucial to achieve this objective
(SOX 2002; Cohen et al. 2004). This study will focus on two of these corporate
governance actors, namely audit committees and external auditors. In particular
this study attempts to determine and analyse the type of relationship between
audit committees and external auditors in ensuring financial reporting quality.
The aim of this section is to introduce the underlying theoretical foundations for
this study which form the basis for the rest of the thesis. Firstly, agency theory
will be introduced. Agency theory is a general theory of accounting which
explains the demand for monitoring provided by audit committees and external
auditors. Secondly, theoretical frameworks regarding corporate governance, audit
committee effectiveness and audit quality are introduced. In addition, the
regulatory framework related to the research problem area is discussed. Finally,
the positioning of the present study into agency theory and the theoretical and
regulatory frameworks is explained.
2.1 Agency theory
The theoretical background of this study is based on agency theory, which
postulates that so-called agency problems emerge due to the separation of
ownership and control. Agency problems are further expected to have an impact
on financial reporting quality. This creates a need for monitoring of management
and thus produces the need for corporate governance including effective audit
committees and high quality external auditors (Jensen et al. 1976; Healy et al.
2001). An underlying notion behind agency theory is that the monitoring provided
by audit committees and external auditors will actually contribute to corporate
control, thereby increasing a company’s financial reporting quality. By contrast,
institutional theory, for example, states that many organizational structures such
as audit committees are merely symbolic and may be formed to conform to social
expectations without having any actual impact on financial reporting quality
(Kalbers & Fogarty 1998).
Acta Wasaensia 11
In agency theory emphasis is on rights established by contracts (Coase 1937;
Alchian & Demsetz 1972; Jensen et al. 1976; Fama & Jensen 1983). Jensen et al.
(1976) model the contract between the shareholder and the owner-manager,
which is called an agency relationship. An agency relationship is defined as a
contract under which one or more persons (the principals) engage another person
(the agent) to perform some service on their behalf which involves delegating
some decision making authority to the agent (Jensen et al. 1976). In the manager-
shareholder contract, the owner-manager is viewed as the agent and the
shareholder as the principal (Watts & Zimmerman 1986). Both the principals and
the agents are considered utility maximizers (Jensen et al. 1976).
Agency relationship contains two inherent aspects which, in combination, create
agency problems: 1) the potential conflicts of interests between owners and
managers which may cause managers to act against shareholders’ interests, and 2)
the imperfect observability of managerial actions by shareholders (DeFond 1992).
Agency problems can increase management’s propensity to produce substandard
financial information in order to conceal actions that have not been in the best
interest of the shareholders or debt-holders (Jensen et al. 1976). The agency
literature suggests that certain company specific characteristics increase
management incentives to act against shareholders’ or debt-holders’ interests,
thus increasing agency problems. The primary operational measures for agency
problems are leverage, management ownership and free cash flow (see e.g.
DeFond 1992).
Firstly, the agency problem of leverage postulates that managers (acting on behalf
of shareholders) have incentives to transfer wealth from debt-holders by taking
various actions such as paying dividends to shareholders at the expense of
profitable projects or restructuring of debt (Jensen et al. 1976; Chow 1982;
DeFond 1992; Parkash & Venable 1993). Some of these actions can result in a
decline in firm value because they involve suboptimal investment policies (Chow
1982). Moreover, the literature suggests that firms with high leverage are more
likely to face bankruptcy and such firms are more likely to engage in earnings
management since they are closer to debt covenant violations (Gul & Tsui 2001).
Secondly, agency literature recognizes that the level of management ownership
gives rise to an asymmetric information problem. This maintains that at low
levels of management ownership the manager may be better informed about the
activities and payoffs of the firm than the owner (Ng 1978; Ng & Stoeckenius
1979). Separation of ownership from management creates monitoring difficulties
giving the potential for management to take non-value-maximizing actions. Thus,
low management ownership creates an increased demand for accounting-based
12 Acta Wasaensia
contractual constraints which are used to discourage managers from non-value-
maximizing actions. Management may be motivated to mitigate these constraints
by strategically choosing accounting policies and determining accounting accruals
(Jensen et al. 1976). Accordingly it has been found that management ownership is
positively associated with earnings explanatory power for returns and negatively
related to the magnitude of discretionary accruals (Warfield, Wild & Wild 1995).
Thirdly, the agency problem of free cash flow postulates that in the presence of
high free cash flow, management has opportunities to make expenditures that
have negative Net Present Values (NPVs) rather than paying dividends to
shareholders or purchase stock. The free cash flow agency problem can be
implicated by a firm’s poor financial performance and consequently poor stock
market valuations. The free cash flow agency problem is also implicated by a
relation between company’s free cash flow and accrual activities. Managers in
firms with high free cash flow may have incentives to smooth earnings in order to
shirk the full impact of wasteful expenditures on earnings. Prior research has
documented a negative relation between free cash flow and the magnitude of
discretionary accruals. These results can be explained by the following rationale:
income-decreasing accruals occur if managers wish to shift profits to future years
when the full impact of expenditures hits earnings (Chung, Firth & Kim 2005;
Richardson 2006).
Agency theory maintains that there are two main ways in which shareholders can
mitigate agency problems. First, the shareholders can establish appropriate
incentives for the managers in such a way that their interests coincide with those
of the shareholders. Second, the shareholders can monitor the managements’
actions. Jensen et al. (1976) describe agency costs as the sum of these safeguards,
along with the effects of those abuses which could be prevented.
According to agency theory, the demand for financial reporting arises from the
manager’s needs to provide some description of the firm’s payoff for legal and
contractual reasons. However, financial reporting is of little use if its provision is
not monitored and enforced (Watts et al. 1986). Corporate governance actors,
such as audit committees and external auditors, provide monitoring whose main
value is dependent on its ability to decrease the likelihood that company’s
financial reports contain breaches. Agency theory predicts that as agency
problems become more severe, management will demand higher quality
monitoring in an effort to ensure financial reporting quality to shareholders, debt-
holders or other investors (Chow 1982; Francis & Wilson 1988; DeFond 1992;
Kalbers et al. 1998; Lennox 2005). Prior empirical studies have addressed this
notion and examined for example whether variables related to company’s agency
Acta Wasaensia 13
problems produce the need for effective audit committees (e.g. Menon &
Williams 1994; Collier & Gregory 1999) or high quality external audit (e.g.
Lennox 2005; Nikkinen & Sahlström 2004).
2.1.1 Agency theory and audit committees
Early studies focusing on the association between agency problems and audit
committees were conducted prior to the requirement for mandatory formation of
audit committees. Thus, studies such as Pincus, Rusbarsky and Wong (1989) and
Bradbury (1990) examined whether a company’s agency problems affect the
voluntary formation of audit committees. It was hypothesized that companies
with great agency problems are more likely to employ audit committees in order
to enhance the quality of financial reporting by management. The results of these
studies were somewhat mixed. Pincus et al. (1989) in their study of US
companies reported a number of significant relationships between variables
related to agency problems (i.e. leverage, company size, ownership structure) and
the formation of audit committee. Bradbury (1990), however, was unable to find
significant relationships between agency problem variables (i.e. number of
outside shareholders, leverage and assets-in-place) and formation of audit
committees for a sample of New Zealand companies.
Studies have also attempted to link agency problems with measures of audit
committee effectiveness, such as audit committee independence and activity
level. The results of these studies have also been mixed. For example, Menon et
al. (1994) found significant relations between selected agency variables (i.e.
outside directors on the board, auditor type and company size) and the existence
of audit committees, the percentage of outside directors on audit committees, or
the frequency of audit committee meetings using a sample of US companies.
Collier et al. (1999) attempted to replicate and extend the study by Menon et al.
(1994). More specifically they examined audit committee activity level in large
companies and by employing another measure of audit committee effectiveness,
namely the duration of audit committee meetings. Their results, however, failed
to support the findings of Menon et al. (1994) related to the impact of agency
variables on the number of audit committee meetings. They did find that the
(then) Big Six audit firms and leverage were positively related to audit committee
activity. In addition the results revealed that audit committee activity was reduced
in firms where the role of chairman and CEO were combined and where insiders
were included in the audit committee. Turpin and DeZoort (1998) found a
significant positive association between voluntary audit committee report
disclosure in annual reports and agency variables - company size, proportion of
14 Acta Wasaensia
outside directors, leverage, and trade on a major stock exchange. In contrast
Kalbers et al. (1998) investigated whether audit committee effectiveness is more
closely aligned with agency or institutional theory. Their results did not show a
strong link between audit committee effectiveness and agency variables, thus
providing indirect support for the institutional theory which states that the audit
committee is a symbolic structure formed to confirm to social pressures.
Studies have also examined the relationship between boards of directors and audit
committees. Since the audit committee is a subcommittee of the board it is
expected to have a significant effect on audit committee composition and
activities. Beasley and Salterio (2001) examined the effect of boards of directors
on voluntary improvements in audit committee composition. They found that
audit committee independence level and audit committee knowledge and
experience were positively associated with board size, proportion of outsiders on
the board, and the separation of board chair and CEO/president. Similarly Klein
(2002b) found that audit committee independence was positively associated with
board size and board independence and negatively associated with growth
opportunities and firms with losses. Klein (2002b) found no effect of leverage,
CEO on compensation committee and outside director holdings4 on audit
committee independence.
2.1.2 Agency theory and external auditing
Agency theory has also been applied to external auditing. These studies have
examined whether agency problems increase the demand for audit quality. Early
studies such as Chow (1982) and Watts and Zimmerman (1983) provide evidence
that firms voluntarily engage external auditing in situations of great agency
problems. Later studies used auditor reputation (audit firm size or brand name) as
a measure of audit quality and documented that companies facing agency
problems hire auditors with better reputation (Francis et al. 1988; DeFond 1992;
Lennox 2005; Fan & Wong 2005). More recent research has used audit fees as a
proxy for audit quality. Gul and Tsui (1998) examined the association between
free cash flow and audit fees. They presented evidence of a positive association
between free cash flow and audit fees for low growth firms. In addition it was
found that debt moderated this relationship. In a subsequent study Gul et al.
(2001) examined the association between free cash flow and audit fees for
different levels of management ownership. They found a positive association
4
Outside director holdings was the percentage of shares held by outside directors. Outside
directors were defined as directors having no affiliation with the firm other than serving as
directors in the board or audit committee whereas inside directors were current employees.
Acta Wasaensia 15
between free cash flow and audit fees. This association was stronger for
companies with low management ownership. In addition Nikkinen et al. (2004)
examined the relationship between agency problems and total fees paid to
incumbent auditors. They found a positive relation between free cash flow and
total fees and a negative relation between management ownership and total fees.
These results are consistent with the notion that management demands a higher
quality audit as firm’s agency problems increase. In addition, some other control
mechanisms such as debt holders may have an effect on the strength of this
relationship.
Prior research has also shown that audit clients distinguish between audit and
non-audit services when considering their effect on audit quality and especially
auditor independence. The notion behind these studies is that if auditees want to
signal audit quality and auditor independence to outsiders they restrict the
purchase of non-audit services from their incumbent auditor. This notion is
supported by Beck, Frecka and Solomon (1988a), Beck, Frecka and Solomon
(1988b), Parkash et al. (1993) and Firth (1997), who found that companies with
agency problems reduce the purchase of non-audit services from the incumbent
auditor. This can be explained by auditee’s wish to safeguard shareholders’
perceptions of auditor independence in situations where agency problems are
present (Parkash et al. 1993).
2.2 Frameworks
Prior literature includes frameworks which have been developed to improve the
understanding of the actors potentially influencing the effectiveness of corporate
governance including audit committees and external auditors. In this study the
frameworks are divided into three classes: 1) frameworks related to corporate
governance, 2) frameworks related to audit committees and 3) frameworks related
to audit quality. These frameworks are somewhat overlapping, although they
represent alternative theoretical approaches to analyse the functioning of
corporate governance, audit committee effectiveness and audit quality. The
frameworks are discussed in more detail in the following sections.
2.2.1 Framework for corporate governance
Cohen’s et al. (2004) corporate governance mosaic aims to describe how a
company’s corporate governance affects financial reporting quality. This mosaic
is presented in Figure 3. Firstly, the mosaic identifies actors and mechanisms for
16 Acta Wasaensia
the most part external to the company, which are expected to have an effect on the
effectiveness of the organizations corporate governance function. These actors
include regulators, legislators, financial analysts, stock exchanges, courts and the
legal system as well as the stockholders. Secondly, the main actors of corporate
governance are identified. These include board of directors, audit committees,
internal auditors, external auditors and management. These five actors are also
expected to have a more direct impact on a company’s financial reporting quality.
The framework maintains that there are interrelationships between the various
mechanisms and actors in the framework. More specifically, the effectiveness of a
company’s corporate governance function is dependent on proper communication
and interaction between corporate governance actors. This is consistent with SOX
(2002), which states that the effectiveness of corporate governance is dependent
on the interaction between board of directors, audit committees, external auditors,
internal auditors and management.
Financial reporting quality
Financial Analysts
Stockholders Stock Exchanges Regulators
Legislators Courts & Legal
System
Management Internal Auditors
Board of Directors
External Auditors
Audit Committee
Figure 3. Corporate governance mosaic and financial reporting quality (Cohen
et al. 2004).
Acta Wasaensia 17
2.2.2 Framework for audit committee effectiveness
DeZoort et al. (2002) provided a framework for evaluating audit committee
effectiveness. The framework is presented in Figure 4. This framework DeZoort et
al. (2002) consists of three levels of audit committee effectiveness, namely input,
process and output levels. The input level of audit committee effectiveness
includes components such as composition, authority and resources of the audit
committee. These factors create the basic premises for audit committee
effectiveness. Audit committee composition refers to audit committee members’
mental attributes such as expertise, independence, integrity and objectivity. On
the other hand authority refers to the responsibilities and influence of the audit
committee. In addition, resources involve audit committee members’ access to
management as well as internal and external auditors. The process level of audit
committee effectiveness includes diligence of the audit committee. It is suggested
that input level factors contribute to audit committee effectiveness only if audit
committee members are active and devote adequate time and effort to the
discharge of their duties regarding the functioning of the audit committee. The
input and process components are expected to have a joint effect on the output of
audit committee effectiveness. In the framework audit committee is considered
effective if it successfully fulfils its responsibilities.
Output
Process
Diligence
Input
Composition
(e.g. Expertise,
Independence)
Authority
(e.g. Responsibilities,
Influence)
Resources
(e.g. Access to
management, External and
Internal Auditors)
Audit committee effectiveness
Figure 4. Determinants of audit committee effectiveness (DeZoort et al. 2002).
18 Acta Wasaensia
2.2.3 Frameworks for audit quality
Frameworks related to audit quality include DeAngelo’s (1981a; 1981b) seminal
model for audit services as well as more recent description of determinants of
audit quality by Watkins et al. (2004). DeAngelo’s (1981a; 1981b) framework
defines determinants of perceived audit quality with a particular focus on auditor
independence. More recently, Watkins et al. (2004) developed DeAngelo’s
(1981a; 1981b) definition further. In comparison to DeAngelo’s (1981a; 1981b)
definition, which is concerned with perceived audit quality, Watkins et al. (2004)
make a distinction between actual and perceived audit quality.
DeAngelo’s (1981a; 1981b) definition of perceived audit quality is depicted in
Figure 5. DeAngelo (1981a; 1981b) defines audit quality as the market-assessed
probability that, given that the financial statements contain material errors, they
are discovered and reported. According to the definition the probability of
discovery depends on the auditor’s competence, whereas the probability of
reporting refers to the auditor’s independence from the auditee. According to the
framework independence is compromised if the auditor allows the client to use a
reporting policy that he or she believes would be viewed as an audit failure.
DeAngelo (1981a; 1981b) argues that auditor’s decision to retain his or her
independence would be impaired if the auditor fears dismissal. Losing a client
would mean that the auditor would lose the economic revenue that otherwise
would accrue to him or her from repeatedly auditing the same client. The
revenues are a result of gaining client specific knowledge. The revenue serves to
bind the auditor to the client because client specific knowledge results in audit
costs falling while audit fees rise over time (DeAngelo 1981a; DeAngelo 1981b).
However, potential loss of reputation from perceived non-independence is seen as
counteracting the bonding between auditor and client. Thus, auditor’s loss of
reputation can reduce the size of the auditors’ client portfolio. Ultimately the
decision to remain independent results from a comparison of the gains resulting
from choosing to lose one’s independence with those obtainable from remaining
independent (DeAngelo 1981a; DeAngelo 1981b). In addition, DeAngelo (1981a;
1981b) argues that large audit firms are better able to remain independent of the
audit client because they have more audit clients than small audit firms. Therefore
economic revenues received from one client are typically not as significant to a
large audit firm as to a small audit firm.
Acta Wasaensia 19
Perceived audit quality
Auditor competence
-Probability that auditor
discovers material errors in the
financial statements
Auditor independence
-Probability that auditor will
report discovered errors in the
financial statements
Figure 5. Perceived audit quality (DeAngelo 1981a; DeAngelo 1981b).
The framework by Watkins et al. (2004) extends the definition of audit quality
provided by DeAngelo (1981a; 1981b). The framework discusses drivers,
dimensions as well as products of audit quality. This framework is presented in
Figure 6. Drivers for audit quality are divided into demand and supply drivers.
Demand drivers include client risk strategies and agency conflicts and supply
drivers include auditor risk management strategies and audit fees. Audit quality is
divided into auditor reputation and auditor monitoring strength. Auditor
reputation refers to perceptions of audit quality and auditor monitoring strength
refers to actual audit quality. Consistent with DeAngelo (1981a; 1981b), both
auditor monitoring strength and auditor reputation can be divided into dimensions
of competence and independence. In other words, auditors’ monitoring strength
(reputation) is dependent on auditors’ actual (perceived) competence and actual
(perceived) independence. Monitoring strength and reputation are expected to be
determinants of information credibility and information quality. Consistently
information credibility refers to perceptions of financial reporting quality and
information quality refers to actual financial reporting quality in the framework.
20 Acta Wasaensia
Figure 6. Determinants of audit quality (Watkins et al. 2004).
2.2.4 Regulatory framework
Corporate failures and accounting scandals have provided US regulators with
strong impetus to re-evaluate requirements concerning corporate governance. In
the aftermath of these scandals the US Congress passed the Sarbanes-Oxley Act
in 2002 (hereafter referred to as the SOX 2002). In addition, all major US stock
exchanges have renewed their listing standards with regard to corporate
governance based on SOX (2002). In general, the purpose of these requirements
is to strengthen companies’ corporate governance including the functioning of
audit committees and external audit. The proper functioning of these actors is
believed to ensure a company’s financial reporting quality. The following
sections will discuss these requirements in more detail. The aim of this section is
to provide a description of the regulatory framework of this study.
Drivers for audit quality
Demand drivers Supply drivers
Audit quality
Products of audit quality
Client risk
strategies
Agency
conflicts
Auditor risk
management
strategies
Audit fees
Auditor reputation
-Perceived competence
-Perceived independence
Auditor monitoring strength
-Auditor competence
-Auditor independence
Information credibility Information quality
Financial statement
Acta Wasaensia 21
2.2.4.1 Requirements related to audit committees
The role of the audit committee in corporate governance has been a subject of
increasing regulatory interest. Currently, all firms listed on major US stock
exchanges (i.e. NYSE, AMEX or NASDAQ) are required to maintain audit
committees. Regulations place audit committees in a key position to ensure a
company’s financial reporting quality. Consistently, for example, the SOX (2002)
states that the purpose of an audit committee is to oversee the accounting and
financial reporting process of the company as well as the audits of the financial
statements of the company. In order to ensure audit committees effectiveness in
discharging its responsibilities, regulators have adopted requirements on the
functioning of audit committees in a number of areas including audit committee
composition and responsibilities. The main requirements contained by the SOX
(2002) and stock exchanges’ listing standards are summarized in Appendix 1 and
will be discussed in more detail next.
Firstly, the regulations emphasize the importance of audit committee composition
in achieving audit committee effectiveness. The objective of these requirements is
to ensure that audit committees have adequate resources and knowledge base to
fulfill their responsibilities. In general the regulations state that an effective audit
committee should comprise a sufficient number of directors. NYSE (2003),
AMEX (2003) and NASDAQ (2003) listing standards require that audit
committees should comprise at least three members. It is further stated that all
audit committee members must be independent of the company as well as
financially literate. In addition, regulations state that at least one audit committee
member must be a financial expert. For example, NYSE (2003) defines audit
committee member independence as freedom from relationships to the company
that may interfere with the exercise of the director’s independence of the
management and the company. Financial literacy refers to audit committee
members’ ability to understand fundamental financial statements, including a
company's balance sheet, income statement, and cash flow statement (AMEX
2003). On the other hand financial expertise refers to a director’s employment
experience in finance or accounting or in other comparable experience which
results in the individual’s financial sophistication (e.g. NASDAQ 2003).
Secondly, the regulations recognize that proper audit committee composition does
not necessarily ensure audit committee effectiveness. Therefore current
regulations include requirements related to audit committee responsibilities. In
general the regulations maintain that audit committees are responsible for
assessing the quality of a company’s financial reporting by evaluating the
implementation of accounting principles as well as changes in them. To fulfil this
22 Acta Wasaensia
responsibility the audit committee is required to communicate regularly with the
board, management, external auditors, and internal auditors. With respect to
external auditing, the current regulations emphasise the audit committee’s
position as a main mechanism to ensure proper communication between auditor
and the company. The literature has traditionally assumed that management has
considerable influence on the audit mandate, including hiring and firing the
auditor, as well as negotiating the audit contract, and the audit fees (O’Keefe,
Simunic & Stein 1994; Mikol & Standish 1998; Beasley et al. 2001). The
literature has suggested that management control over the audit mandate poses a
potential threat to audit quality and particularly auditor independence because an
auditor’s financial dependence on the auditee has depended heavily on
management’s power to hire and fire the auditor (Ashbaugh 2004; Mayhew &
Pike 2004) 5. However, after the enactment of SOX (2002) management’s
influence on the external audit mandate has been considerably reduced. This is
because under SOX (2002) audit committees are responsible for the appointment,
compensation, retention and oversight of external auditors. In addition SOX
(2002) expects communication between auditor and audit committee in several
issues including a timely report of 1) all critical accounting policies, 2) alternative
accounting treatments and disclosures, and 3) a management letter. In addition,
the audit committee is required to establish procedures for handling complaints
received by the company regarding accounting, internal controls over financial
reporting, or auditing matters including confidential submission by company
5
Empirical research on opportunistic behaviour by management has concentrated on
management’s ability to change the auditor. There is evidence that some auditor switches are
motivated by auditor-management disagreement over auditors’ reporting decisions. Specifically,
auditees switch auditors more frequently after receiving a qualified audit report (Chow & Rice
1982; Smith 1986). In addition, studies have analysed the discretionary accruals of firms that
changed auditors. The results showed that discretionary accruals were significantly income
decreasing during the last year of the predecessor auditor, and insignificant during the first year of
the successor auditor. These results indicate that conservative accounting choices preferred by
auditors give auditees an incentive to change auditor (DeFond & Subramanyam 1998). These
findings are related to auditee management ability to “opinion shop”, that is to switch auditors in
order to avoid unfavourable audit reports. Successful opinion shopping is harmful to audit quality
because auditors may avoid issuing qualified audit reports in order to retain incumbency.
However, the research evidence on auditees’ ability to opinion shop is conflicting. Some studies
have concluded that companies can opinion shop because they would have received unfavourable
reports more often had they not switched auditors (Lennox 2000), while others suggest that
companies do not engage successfully in opinion shopping because post-switch opinions are not
more favourable than pre-switch opinions (Krishnan 1994; Krishnan & Stephens 1995). Research
on perceived independence has examined management control over audit mandate in several
respects. These studies have shown that perceptions of auditor independence are negatively
affected when management has: 1) control of auditors’ appointment or remuneration (Beattie et al.
1999), 2) the ability to seek a second opinion on contentious issues (Beattie, Brandt & Fearnley
1999) and 3) the ability to negotiate audit fees or determine the deadline for submitting the audit
report (Emby & Davidson 1998).
Acta Wasaensia 23
employees of concerns regarding questionable accounting or auditing matters.
Audit committees must also pre-approve all audit services and permitted non-
audit services provided by external auditors. In order to perform these duties
effectively the regulations state that audit committees should meet relatively
frequently. For example, both NYSE (2003) and AMEX (2003) require that audit
committees must have periodic meetings. More specifically, NYSE (2003)
requires that the audit committee must hold separate meetings with management,
with internal auditors and with external auditors. NYSE (2003) considers that
separate meetings between these parties are more productive than joint sessions
for considering issues warranting committee attention.
2.2.4.2 Requirements related to external auditing
One of the main objectives of SOX (2002) is to strengthen audit quality,
particularly auditor independence. Thus, the SOX (2002) includes several
requirements regarding external auditors. The requirements which are relevant to
this research are as follows: requirement to disclose audit and non-audit services
fees, mandatory rotation of audit partners, and restrictions regarding non-audit
services provided by incumbent auditors.
Disclosure of audit and non-audit services fees is considered a way to enhance
audit quality by safeguarding auditors from significant financial dependence. This
is because the disclosure gives transparency to the auditor-auditee relationship
and is therefore expected to be a sufficient way to inform shareholders, investors,
and other parties of the auditor’s incentives to compromise their independence
(Firth 1997; SEC 2000; Firth 2002). On the other hand, the disclosure may
enhance an audit firms’ independence because they may be reluctant to provide
the types of services or charge the level of fees that might be perceived as threats
to independence (Hillison & Kennelley 1988; Firth 1997). Currently the SEC
(2000; 2003) requires the disclosure of the amount of all audit and non-audit
services fees in proxy statements. SEC (2003) states that fees paid should be
divided into the following categories: 1) audit fees, 2) audit-related fees, 3) tax
fees and 4) all other fees.
Mandatory auditor rotation has been suggested as a solution to the independence
threat caused by long-term audit tenures. It is argued that mandatory rotation
gives auditors greater incentives to resist management pressure, thus increasing
independence since the tenure period becomes limited. Proponents of rotation
argue that it reduces audit failures, forces audit clients to adopt conservative
accounting practices and results in more complete financial statement disclosures.
24 Acta Wasaensia
In addition, rotation is expected to ensure that a company’s accounting choices,
particularly those in subjective and judgmental areas, are reviewed by different
auditors (Catanach & Walker 1999). Currently, the SOX (2002) and SEC’s
(2003) final rule requires a rotation of all partners on the audit engagement team
after five years of services.
The prohibition of non-audit services is expected to reduce issues related to audit
quality and auditor independence presented by the provision of non-audit services
to auditees (e.g. economic dependence, self-review threat to independence).
Currently, the SOX (2002) prohibits auditors from supplying the following types
of non-audit services: 1) bookkeeping, 2) financial information systems, 3)
appraisal, 4) actuarial, 5) internal audit outsourcing, 6) management or human
resources, 7) broker or dealer, 8) legal and expert and 9) any other services
specified by the SEC. Despite these prohibitions, registrants may still purchase
many types of non-audit services from the incumbent auditor such as tax
compliance and consulting, employee plan audits, consulting on accounting
matters, mergers and acquisition consulting, and consulting on new debt and
equity issues (Raghunandan, Read & Whisenant 2003)
2.3 Positioning of the study
This section introduced agency theory as well as some relevant frameworks
related to corporate governance, audit committee effectiveness and audit quality.
In addition the regulatory framework of this study was discussed. The present
study can be positioned into agency theory and frameworks as follows.
Firstly, this study adopts one of the basic premises of agency theory, which
maintains that corporate governance in general and audit committees and external
auditors in particular are important in ensuring financial reporting quality. In
addition, this study subscribes to an underlying notion that certain company
specific characteristics which create agency problems drive the demand for
monitoring provided by audit committees and external auditors. Accordingly
audit committees and external auditors are expected to provide assurance to
shareholders that a company’s financial statements are in accordance with GAAP.
This notion is contradictory, for example, to institutional theory, which states that
these monitoring mechanisms are developed as a response to social norms and
regulations but do not necessarily improve financial information quality.
Secondly, Cohen’s et al. (2004) corporate governance mosaic suggests that
interrelationships between corporate governance actors are important for its
Acta Wasaensia 25
effectiveness. The present study does not take into consideration all the
interrelationships suggested in Cohen’s et al. (2004) framework but focuses on
two important corporate governance actors: audit committees and external
auditors. In other words, this study examines what type of relationship audit
committees and external auditors have in determining financial reporting quality.
To achieve this objective the approach is to model the interrelationship between
audit committees and external auditors. More specifically, it is suggested that
audit quality mediates the relationship between audit committee effectiveness and
financial reporting quality.
Thirdly, the framework by DeZoort et al. (2002) focuses on audit committee
effectiveness. This framework suggests that audit committee effectiveness is
dependent on various aspects related to audit committee composition as well as
audit committee diligence. The operational measures for audit committee
effectiveness employed in this study are consistent with the framework of
DeZoort et al. (2002). This is because the operational measures are related to both
audit committee composition as well as audit committee activity level. The
operational measures will be discussed in more detail in the subsequent sections
of the study.
Fourthly, this study adopts DeAngelo’s (1981a; 1981b) and Watkins’ et al. (2004)
definition of audit quality, which states that audit quality consists of two
components: auditor independence and auditor competence. In addition, this study
adopts Watkins’ et al. (2004) view that audit quality can be divided into auditor
reputation and auditor monitoring strength. In particular this study focuses on
auditor monitoring strength by testing the effect of measure of audit quality on
financial reporting quality. This study also contributes to Watkins’ et al. (2004)
framework by considering audit committee effectiveness as an important driver
for audit quality.
Finally, in addition to the theoretical framework, the US regulatory environment
creates a specific research setting where severe scrutiny is applied to audit
committee effectiveness and audit quality. Regulations regarding audit committee
effectiveness include requirements related to audit committee composition as well
as the activities of the audit committee. Under the current regulations all audit
committee members are required to be independent of the company and to be
financially literate. In addition, at least one audit committee member is expected
to possess financial expertise. According to the regulations the audit committee’s
main objective is to ensure the company’s financial reporting quality. With
respect to external auditing, audit committees are responsible for hiring, firing
and compensating external auditors. The primary measures for audit committee
26 Acta Wasaensia
effectiveness can be derived from these regulations as well as from prior
empirical research. Thus, it is suggested that audit committee effectiveness can be
measured by audit committee size, audit committee expertise ratio and audit
committee meeting frequency. The underlying notion behind this study is that the
regulations provide the minimum requirements for audit committee composition
and activities and some companies may strive to maintain quality differentiated
audit committees. That is, companies with audit committees with stronger
attributes than the minimum requirements are expected to be more effective. The
next section will provide further justification for the audit committee
effectiveness measures employed in this study.
In addition to the requirements regarding audit committees the regulations related
to audit quality are relevant for this study. This is because the regulatory
environment is expected to affect the interaction between the companies and their
external auditors. The discussion here is focused on SOX (2002) because one of
its main objectives is to enhance audit quality and particularly auditor
independence. Therefore the SOX (2002) includes several requirements regarding
external auditors that are relevant for this study. These include requirements for
mandatory auditor rotation, the prohibition of non-audit services and disclosure of
audit and non-audit services fees. Mandatory auditor rotation ensures that audit
quality is not affected by long audit mandates. In addition, prohibition of specific
types of non-audit services ensures that auditor independence is not threatened by
these services. The disclosure of audit and non-audit services fees provides a
measure for audit quality employed in this study. It is suggested that audit fees
represent audit effort and therefore higher audit fees are expected to indicate
better audit quality. This notion is related, for example, to the GAO (2003) report,
which documented an increase in audit fees after the enactment of SOX (2002)6.
GAO (2003) asserted that the fee development is, at least partly, caused by
changes in audit environment and increases in the oversight of publicly disclosed
financial information. Thus, it is argued that enactment of SOX (2002) has
resulted in increase in audit effort and further audit fees. In addition, prior
empirical research has linked audit fees with increased financial reporting quality.
6
Prior research reveals that audit fees stayed flat or decreased slightly from the late 1980s to the
mid-1990s (Firth 1997; Menon & Williams 2001).
Acta Wasaensia 27
3 DEFINITIONS AND OPERATIONAL
MEASURES OF KEY CONCEPTS
The purpose of this section is to introduce the definitions for audit committee
effectiveness, audit quality and financial reporting quality provided by prior
literature. In addition, the operational measures for these variables employed in
prior studies will be discussed. The primary measures for audit committee
effectiveness used in prior literature are audit committee size, audit committee
independence, audit committee expertise and audit committee meeting frequency.
The most commonly used measures for audit quality are audit firm size, audit
firm industry specialization, audit tenure and audit fees. Measures for financial
reporting quality can be divided into two groups: those related to financial
reporting within GAAP and those related to financial reporting outside GAAP.
The section will be concluded with a discussion regarding the definitions and the
measures derived from the prior literature. The aim of the discussion is to explain
which definitions and measures are regarded as adequate to be employed in the
present study. These choices require, among other things, taking into
consideration the regulatory environment of this study.
3.1 Audit committee effectiveness
An audit committee is a subcommittee of the board of directors which is
particularly designated to oversee the company’s financial reporting process.
According to Wolnizer (1995) and DeZoort (1997) the responsibilities of audit
committees fall into areas of: 1) financial reporting (including internal controls),
2) auditing, and 3) other corporate governance (e.g. facilitate communications
between the board and the external auditors). Audit committee effectiveness is
often associated with an audit committee’s ability to fulfil these responsibilities
(Kalbers & Fogarty 1993; DeZoort et al. 2002). There is a common understanding
that the mere existence of an audit committee does not guarantee that it will be
effective. Thus, the literature determines several attributes which are needed to
achieve audit committee effectiveness. According to DeZoort et al. (2002) audit
committee effectiveness is dependent on its composition (the independence and
expertise of its members), its authority (responsibilities and influence) and its
resources (number of members and access to other governance parties).
Moreover, audit committees need to be diligent to effectively discharge their
responsibilities. In addition, the current US regulations recognize that audit
committee effectiveness is dependent on proper audit committee composition and
28 Acta Wasaensia
diligence. Consistently, the primary operational measures employed by prior
empirical research include audit committee size, audit committee independence,
audit committee expertise and audit committee activity level. The first three
measures are concerned with audit committee composition and the last is
concerned with audit committee diligence. The following discussion will provide
further rationalization for these measures.
3.1.1 Audit committee size
The literature suggests that audit committee size measured as the number of audit
committee members will have a positive effect on audit committee effectiveness.
This is because it is likely that audit committees with a sufficient number of
members have better resources than smaller audit committees (DeZoort et al.
2002). In addition, the decision-making literature has indicated that increasing the
number of people involved in an activity substantially increases group
performance and decreases the opportunity for wrongdoing because collusion
becomes more difficult (e.g. Cummings, Huber & Arendt 1974; Burton, Pathak &
Zigli 1977). Thus, it can be argued that decision-making in larger audit
committees is of better quality than in smaller audit committees7.
Audit research has also provided evidence that audit committee size is an
important determinant of audit committee effectiveness. Vafeas et al. (2007)
found a positive relationship between audit committee size and audit fees. This
result indicates that the demand for audit increases as the size of the audit
committee increases. In addition Archambeault & DeZoort (2001) examined
suspicious auditor switches8. They found that companies with suspicious auditor
switches had smaller audit committees than non-suspicious switching companies.
This implies that larger audit committees are better able to safeguard auditors
from being unfairly dismissed than smaller audit committees.
7
It can also be argued that audit committee size has a nonlinear effect on audit committee
effectiveness. Initially, adding members to the audit committee is likely to enhance audit
committee effectiveness because it ensures that audit committee has required knowledge to make
decisions regarding its responsibilities. However, it is likely that audit committee effectiveness
may suffer if it becomes too large. This is because large group may create process losses and
diffusion of responsibility.
8
Analysis was conducted by matching switching and non-switching companies with suspicious
circumstances (i.e. disclosure of reportable event, qualified audit opinion or other recent auditor
switch).
Acta Wasaensia 29
3.1.2 Audit committee independence
Audit committee members’ independence is considered to be an essential
determinant of audit committee effectiveness (DeZoort et al. 2002; Pomeroy et al.
2008). For example, the SOX (2002) considers that an audit committee member is
independent if he or she does not receive any compensation from the company or
its affiliates except in the capacity of audit committee member. Independent audit
committee members are expected to be better able to oppose management
pressure in conflict situations and are therefore expected to contribute to audit
committee effectiveness. Prior empirical research has provided evidence
supporting this notion.
Several studies have found that audit committee independence is associated with
higher earnings quality (Klein 2002a; Bédard et al. 2004; Vafeas 2005; Bradbury
et al. 2006). Klein (2002a) found a significant negative association between an
audit committee with a majority of independent directors and discretionary
accruals. In a similar vein, Bédard et al. (2004) reported a significant reduction in
aggressive earnings management when the audit committee was 100 %
independent. Bradbury et al. (2006) using data from Singapore and Malaysia
found that audit committee independence is associated with lower abnormal
working capital accruals. Vafeas (2005) found that audit committees with a higher
percentage of insiders on the audit committee are more likely to report small
earnings increases. These results indicate that independent audit committees have
a constraining effect on managerial behaviour in earnings management.
Several studies have also examined the association between audit committee
independence and audit quality. For example, Mitra, Hossain and Deis (2007) and
Vafeas et al. (2007) found a significant and positive association between the level
of audit committee independence and audit fees indicating that independent audit
committees are interested in ensuring high quality service provided by external
auditors. Carcello and Neal (2000) examined the relationship between distressed
firms’ audit committee independence and the likelihood of receiving going-
concern audit reports. They found that distressed firms with more independent
audit committees were more likely to receive a going-concern audit report. This
result may indicate that more independent audit committees provide greater
support for auditors in their reporting decisions than less independent audit
committees. Moreover, Archambeault et al. (2001) found that more independent
audit committees are better able to protect auditors from being suspiciously
switched than less independent audit committees.
30 Acta Wasaensia
3.1.3 Audit committee expertise
Audit committee effectiveness is expected to increase as the proportion of experts
in the audit committee increases. Expertise refers to an audit committee member’s
knowledge and experience in the areas of accounting and financial reporting,
internal controls and auditing (e.g. SOX 2002). In general, it is argued that audit
committee members’ expertise in these areas enables a better understanding of
financial statements and the audit process. This may lead to an enhancement of
audit committee effectiveness in several respects. Firstly, experts are expected to
provide greater oversight on the financial reporting quality than non-expert
members. Secondly, experts are expected to understand the risks and benefits
associated with audit quality better than their non-expert colleagues. Thirdly,
experts are expected to be better equipped to understand auditor judgments and
evaluate the substance of disagreements between management and the external
auditor.
Several studies have provided evidence supporting these arguments. Firstly,
studies have found that expertise affects audit committee member’s decision-
making. DeZoort (1998) found that audit committee members’ experience related
to audit and internal control evaluation resulted in internal control judgments
more in line with auditors than members’ lacking such experience. McDaniel,
Martin and Maines (2002) found that expert and financially literate audit
committee members’ evaluation of the quality of financial reporting items differ.
This result implies that the inclusion of financial experts in audit committees is
likely to have an effect on audit committee’s assessment of a company’s financial
reporting in general. Secondly, Bédard et al. (2004) focused on earnings
management and found that audit committees with more expert members are
better equipped to restrict earnings management. Thirdly, with regard to external
auditors DeZoort (1998) and DeZoort & Salterio (2001) found that audit
committee members possessing auditing knowledge provide more support for
auditors in auditor management disagreements over an ambiguous accounting
issue than members lacking such knowledge. Experts were also found to
safeguard auditors from being unfairly dismissed (Archambeault et al. 2001).
Finally, Vafeas et al. (2007) found that audit committee member expertise has a
positive relationship with audit fees indicating that experts demand higher quality
service by external auditors.
Acta Wasaensia 31
3.1.4 Audit committee meeting frequency
The discussion above reveals that variables related to audit committee
composition have a significant influence on audit committee effectiveness. In
addition to these variables, audit committee activity level has also been
recognized as an important process factor needed to achieve audit committee
effectiveness (Kalbers et al. 1993; DeZoort et al. 2002). A common measure for
audit committee activity level is the number of meetings held by the audit
committee. More specifically, audit committee effectiveness is expected to
improve with more frequent audit committee meetings (Menon et al. 1994;
Abbott et al. 2000; DeZoort et al. 2002; Lee & Mande 2005; Stewart et al. 2007).
Thus, in order for the audit committee to be effective audit committee members
must be willing to invest a substantial amount of time and energy in the
functioning of the audit committee (Kalbers et al. 1993; Lee, Mande & Ortman
2004).
With regard to auditing, regular meetings between the audit committee and the
external auditor can be expected to contribute to audit committee knowledge
about relevant accounting and auditing issues. Thus it can be argued that audit
committees that meet frequently are more diligent in the discharge of their duties
(Abbott et al. 2003a; Raghunandan, Read & Rama 2001). More diligent audit
committees can also be expected to be more concerned with audit quality. This
notion is supported by Goodwin-Stewart et al. (2006), who found that more
frequent audit committee meetings were associated with higher audit fees. In
addition, Archambeault et al. (2001) reported that audit committees with frequent
meetings are more likely to protect auditors from being switched under suspicious
circumstances than audit committees with less frequent meetings. Prior studies
have also documented that audit committee meeting frequency is associated with
higher earnings quality (Xie, Davidson & Dadalt 2003; Vafeas 2005).
3.2 Audit quality
Practitioners often define audit quality relative to auditors’ ability to meet legal
and professional requirements (Francis 2004; Watkins et al. 2004). In light of this
definition audit quality can be considered dichotomous. Audit is considered to be
of poor quality if an auditor fails to comply with legal and professional
requirements. In other words an audit failure occurs.9 On the other hand all audits
9
According to Francis (2004) audit failures can occur in two situations. Firstly, audit failure
occurs when an auditor fails to enforce GAAP. Secondly, audit failure occurs when an auditor fails
to issue an appropriate type of audit report: an auditor issues a qualified audit report when
32 Acta Wasaensia
which are legally satisfactory can be considered to be of sufficient quality
(Francis 2004). Many audit quality studies, however, have an underlying
assumption that most audits meet minimal legal and professional requirements.
Thus, these studies focus on audit quality that exceeds minimal legal and
regulatory requirements. These studies consider audit quality as a continuum
ranging from very poor quality to very good quality. In this approach audit
failures occur at the extreme low end of the continuum. In addition, it is
considered that legally satisfactory audits can be quality differentiated (Francis
2004).
Researchers provide definitions for audit quality which can be applied when audit
quality is considered to be a continuum. As discussed in Section 2.2.3 researchers
generally link audit quality to both auditor independence and auditor competence.
There is a widespread view in the audit literature that audit quality cannot be
directly observed by outside parties or measured by researchers. (DeAngelo
1981a; DeAngelo 1981b; Palmrose 1988; Willekens & Simunic 2007). Therefore,
prior studies have used several surrogates for audit quality. The primary
surrogates for audit quality are audit firm type (i.e. size or brand name), audit firm
industry specialization, auditor tenure and audit fees. These measures will be
discussed in more detail in the subsequent sections.
3.2.1 Audit firm size
Audit firm size is a conventional measure for audit quality.10 Audit quality is
expected to increase with audit firm size because large audit firms are expected to
have superior resources to conduct an audit and they are expected to be better able
to remain independent from the auditee than smaller audit firms (Goldman &
Barlev 1974; DeAngelo 1981a; Shockley 1981). A related stream of research
argues that large audit firms have greater incentives to provide high-quality audits
because they have more to lose from an audit failure in terms of their pre-
established reputations (i.e. brand name) than smaller audit firms (Francis et al.
1988). Based on these rationalizations several lines of empirical research have
focused on the dichotomy between large and small audit firms and reported
supporting results for the above notions.
inappropriate (false positive reporting) or auditor issues an unqualified audit report when
inappropriate (false negative reporting) (Francis 2004).
10
In this study, the majority of sample companies employed a Big Four audit firm (95 %), thus
leaving little opportunity to evaluate audit quality in terms of audit firm size.
Acta Wasaensia 33
Firstly, prior research has provided evidence that audit firm size is associated
with better financial reporting quality. A majority of these studies have focused
on discretionary accrual paradigm (Jones 1991) and examined whether large audit
firms are better able to detect and oppose to managers’ opportunistic earnings
management than small audit firms. Becker, DeFond, Jiambalvo and
Subramanyam (1998) and Francis, Maydew and Sparks (1999) found supporting
evidence and suggested that clients of the (then) Big Six auditors reported lower
discretionary accruals than clients of non-Big Six auditors. Consistently Nelson’s
et al. (2002) results show that large audit firms are able to detect earnings
management attempts and will object to opportunistic earnings management by
managers. Studies also provide evidence that the markets link audit firm size with
better financial reporting quality. For example, Beatty’s (1989) results showed
that the price paid by investors is higher for IPOs involving a large audit firm.
Similarly, Teoh and Wong (1993) reported a positive association between audit
firm size and the ERC. Contrary to these findings, Piot & Janin (2007) did not
find evidence of Big Five audit quality differentiation with respect to earnings
management in France.
Secondly, prior studies have also addressed the relationship between audit firm
size and auditor reporting decisions. Some studies have shown that the size of the
audit firm does not affect the likelihood of a qualified audit report (Wines 1994;
Sharma & Sidhu 2001; Craswell, Stokes & Laughton 2002). However Francis and
Krishnan (1999) report that large audit firms have lower thresholds for issuing
qualified audit reports, which indicates that large audit firms issue more
conservative reports than small audit firms. Lennox (1999) also provided
evidence of a positive relationship between auditor size and auditor accuracy.
Lennox (1999) found using UK data that large auditors issue reports that are more
accurate and include more informative signals of financial distress than audit
reports issued by small auditors. In a similar vein, Weber and Willenborg (2003)
examined audit reports issued before IPO. They found that audit reports issued by
Big Four audit firms have better predictive accuracy in terms of future stock
returns and subsequent delistings than audit reports issued by small audit firms.
Thirdly, studies focusing on legal actions as well as disciplinary actions by
professional bodies against auditors have also provided evidence that large audit
firms provide high quality audits. For example, Palmrose (1988) found that Big
Four audit firms face legal actions less frequently than small audit firms. In
addition, results by Feroz and Pastena (1991) showed that Big Four audit firms
are sanctioned less frequently by the SEC than other audit firms. A counter-
argument for these results is that large audit firms are not really better but they
34 Acta Wasaensia
have better resources to oppose legislators and professional bodies (Francis
2004).
3.2.2 Auditor industry specialization
In addition to audit firm size, the audit literature suggests that industry
specialization11 of an audit firm also contributes to audit quality. It is expected
that industry specialization increases the quality of auditor’s performance due to
the auditor’s superior knowledge regarding the specific industry. Empirical
studies have addressed this notion and provided supportive evidence. For example
Solomon, Shields and Whittington (1999) and Owhoso, Messier and Lynch
(2002) provided some evidence that industry-experienced auditors detect errors
more effectively within their industry specialisation than outside their
specialisation.
In general research regarding audit firm industry specialization is parallel to the
audit firm size research. Thus, prior research has investigated whether industry
specialization of audit firms is associated with higher financial reporting quality.
Balsam, Krishnan and Yang (2003) compared the discretionary accruals of clients
of specialist and non-specialist auditors and found that the discretionary accruals
of clients of industry specialist auditors were lower than those of clients of non-
specialist auditors, thereby supporting the notion that industry specialisation is
associated with better financial reporting quality. In addition, Krishnan (2005)
found that companies with an industry specialist auditor report more conservative
earnings according to Basu’s (1997) framework. Research has also indicated that
an industry specialist auditor signals higher audit quality as well as financial
reporting quality to the markets. An early study by Shockley and Holt (1983)
found that audit firms with the largest market share are perceived as higher
quality suppliers by the chief financial officers of banks. Balsam et al. (2003)
found evidence that clients of industry specialist auditors have higher ERCs
(Earnings Response Coefficient) than clients of non-specialist auditors. Similarly,
Krishnan (2003) found that market reactions to earnings surprises are more
positive for companies with industry specialized auditors.
11
Industry market share is commonly used as a proxy for audit firm industry specialization
(Francis 2004).
Acta Wasaensia 35
3.2.3 Auditor tenure
Auditor tenure has also been recognized as an important determinant of audit
quality. The literature, however, does not provide a clear consensus as to how
auditor tenure affects audit quality. On the other hand it can be suggested that
long-term audit tenures increase audit quality due to auditee-specific knowledge
gained by the auditors over time (Simon & Francis 1988) while the counter-
argument is that auditor’s long term association with the auditee poses a threat to
auditor independence12 thus leading to lower level of audit quality (Mautz &
Sharaf 1961; SOX 2002; IFAC 2005). The results of prior studies are also
inconclusive since they provide support for both arguments.
For example Myers, Myers and Omer (2003) provided evidence supporting long-
term audit tenures reporting that auditors are more likely to restrain management
from making extreme reporting decisions concerning accruals when auditor’s
tenure is longer. In a similar vein, Piot et al. (2007) using French data did not find
any evidence that auditor tenure would lead to decrease in earning quality.
Furthermore, Ghost and Moon (2005) found a positive association between audit
tenure and investor perceptions of earnings quality using ERC as a proxy. On the
other hand several studies indicate that long-term audit tenures lead to decline in
audit quality. These studies have found that long-term tenures: 1) decrease the
likelihood of the auditor to issuing a qualified audit report (Vanstraelen 2000),
decreases compliance with GAAS (Deis & Giroux 1992) and, increase the
likelihood of receiving a substandard audit (Copley & Doucet 1993). Moreover,
the length of audit tenure has been found to have an impact on perceived audit
quality. Knapp (1991) found that audit committee members perceived that audit
quality declines as auditor tenure lengthens. Long periods of tenure have also
been found to have a negative effect on perceived auditor independence (Beck et
al. 1988b; Teoh & Lim 1996).
Although prior research does not provide a clear understanding of the effect of
audit tenure on audit quality, US regulators adopted a view that auditor tenure
should be restricted with mandatory auditor rotation. The SOX (2002)
requirements regarding auditor rotation were discussed in Section 3. Empirical
research has provided some support for the benefits of rotation. Dopuch, King and
Schwartz (2001) found that mandatory rotation requirements decreased auditor’s
willingness to issue biased reports in favour of management. Additionally, the
research evidence indicates that the requirements for auditors to be reappointed
12
This is because long-term audit tenures may result in too close a relationship between the
auditor and the management, which results in the auditor losing his or her independence (Mautz et
al. 1961).
36 Acta Wasaensia
annually enhance perceptions of auditor independence (Beattie et al. 1999;
Hussey & Lan 2001).
3.2.4 Audit fees
Prior studies have also used audit fees as a measure of audit quality. The notion
behind these studies is that audit fees reflect the magnitude of audit effort: higher
audit fees are expected to indicate more auditing work, which is expected to
contribute to audit quality. However, the relationship between audit quality and
audit fees is a complex one because audit fees are jointly determined by both
demand and supply side drivers (Mitra et al. 2007).
From the supply side it can be stated that audit fees reflect the economic costs of
the efficient auditor. More specifically, auditors seek to balance their resources
between costs arising from additional audit work and losses arising from legal
liability. Additional audit effort decreases the probability that the auditor will face
liability losses and thus the required audit coverage varies considerably with the
characteristics of the company, more risky and complex clients requiring more
audit effort. In general, the auditor provides a quantity of audit work that
decreases to an acceptable level the probability that the auditor will suffer liability
losses (Simunic 1980; Simunic & Stein 1996).
Audit fees may also be affected by demand side drivers. The literature provides
several complementary explanations for the demand for audit. In summary the
following four explanations can be identified: monitoring, information, insurance,
and organizational control (Wallace 1980; Dye 1993; Hay et al. 2004). Firstly, the
monitoring explanation for the demand of audit is based on agency theory, which
states that audit services are demanded to reduce agency problems arising from
conflict of interests between owners and managers (Jensen et al. 1976; Watts et
al. 1983; Chow 1982; Blackwell, Noland & Winters 1998; Carey, Simnett &
Tanewski 2000). Secondly, the information explanation emphasizes that investors
demand audited financial statements because they provide information that is
useful in their investment decisions (Wallace 1980; DeAngelo 1981a; Beatty
1989; Dye 1993; Willenborg 1999). Thirdly, the insurance or “deep pocket”
explanation posits that the demand for audit arises from auditors’ extended legal
liability, which enables the full or partial recovery of investor losses resulting
from financial statement misrepresentations by suing auditors (Dye 1993; Hillison
& Pacini 2004; Menon et al. 1994; Baber, Kumar & Verghese 1995; Willenborg
1999; Lennox 1999). Finally, the organizational control explanation maintains
Acta Wasaensia 37
that owners regard audits as a compensatory control system for the organizational
loss of control in hierarchical organizations (Hay et al. 2004; Abdel-Khalik 1993).
Based on the discussion above it can be concluded that audit fees are affected by
both supply side and demand side drivers. From the supply side audit fees reflect
auditor’s assessment of required audit coverage based on client riskiness and
complexity. On the other hand, from the demand side audit fees reflect the
demand for audit coverage. Regardless of the viewpoint, since audit fees are
expected to reflect audit coverage, they can be linked with audit quality. That is,
higher fees reflect a more through audit and so also higher audit quality. Audit
fees have commanded considerable research interest and several research streams
provide support for this argument.
Firstly, early audit pricing studies examined the association between audit fees
and non-audit services fees in order to find evidence of “knowledge spillovers”
which are transfers of knowledge from non-audit to audit services and vice versa.
Knowledge spillovers are expected to increase the quality of both services and
therefore lead to higher fees. These arguments were supported by several studies
reporting a significant positive association between audit services and non-audit
services fees (Simunic 1984; Simon 1985; Palmrose 1986a; Turpen 1990; Davis,
Ricchiute & Trompeter 1993; Butterworth & Houghton 1995; Craswell, Francis
& Taylor 1995; Ezzamel, Gwilliam & Holland 1996; Firth 1997; Firth 2002;
Antle, Gordon, Narayanamoorthy & Zhou 2006). These studies conclude that the
benefits of knowledge spillovers are generally retained by the auditor as higher
fees. By contrast, some studies find no evidence of a positive association between
audit fees and non-audit services fees (Simon et al. 1988; Abdel-Khalik 1990;
Barefield, Gaver & O’Keefe 1993; O’Keefe et al. 1994; Whisenant,
Sankaraguruswamy & Raghunandan 2003).
Secondly, audit pricing studies have provided evidence suggesting that auditees
are willing to pay a price premium for expected high quality service. For
example, Francis and Stokes (1986), Palmrose (1986b), and Simon et al. (1988)
detected a price premium for large audit firms likely arising from a higher quality
of audit offered by these firms. In addition prior studies have identified a fee
premium for industry specialist audit firms. Craswell et al. (1995) found that
industry specialist (then) Big Eight auditors charged a 34 % premium over non-
specialist Big Eight auditors. DeFond, Francis and Wong (2000) found evidence
indicating that the three top industry leaders earn a price premium relative to
other large audit firms in Hong Kong. A more recent study by Ferguson and
Stokes (2002) also provided evidence that the top two industry leaders are able to
earn price premiums over other large audit firms in Australia. Basioudis and
38 Acta Wasaensia
Francis (2007) also found a price premium for city-specific industry leaders
relative to other Big Four auditors in the UK. Similarly Ferguson, Francis and
Stokes (2003) as well as Francis, Reichelt and Wang (2005) found a price
premium for audit firms that were joint national and city-level industry leaders.
Thirdly, studies have linked higher audit fees with a greater demand for audit
quality. These studies suggest that the demand for audit quality increases audit
effort, which is further reflected in higher audit fees. For example, Gul et al.
(1998), Gul et al. (2001) and Nikkinen et al. (2004) found a positive relation
between measures of agency problems and fees paid to the incumbent auditor.
Similarly, studies on corporate governance have documented a positive
association between measures of board or audit committee effectiveness and audit
fees (Abbott et al. 2003a; Lee et al. 2005; Goodwin-Stewart et al. 2006; Mitra et
al. 2007).
Finally, studies have documented a positive relation between audit fees and
financial reporting quality. Firstly, Frankel et al. (2002), Larcker et al. (2004) and
Srinidhi et al. (2007) linked higher audit fees with smaller discretionary accruals.
Frankel et al. (2002) reported that audit fees are negatively related with small
earnings surprises and discretionary accruals. In addition they found that non-
audit fees were positively associated with these measures of earnings
management. These results indicate that audit fees and non-audit fees have an
opposite effect on audit quality: higher audit fees are related to higher audit
quality whereas higher non-audit services fees are associated with lower audit
quality. In a similar vein Srinidhi et al. (2007) found that audit fees result in
higher earnings quality whereas non-audit fees result in economic bonding and
thus decrease earnings quality. Somewhat contradictorily Larcker et al. (2004)
found a negative relation between both audit and non-audit fees and discretionary
accruals after identifying clusters of firms with homogeneous regression structure.
The strongest relation was found for firms with weak corporate governance
structure. According to the authors these results indicate that reputation concerns
are an important determinant of auditor behaviour and they enhance auditors’
incentives to restrict unusual accounting practices by clients. Secondly, Geiger et
al. (2003) examined the association between audit and non-audit fees and auditor
reporting decisions for financially distressed firms. They found a positive
association between audit fees and qualified audit opinions, which implies that
additional audit effort results in more accurate audit opinions. In addition, it was
reported that non-audit fees did not have a statistically significant effect on audit
opinions.
Acta Wasaensia 39
In contrast to the above studies, a substantial amount of audit research has
examined the effect of financial dependence on auditor independence. This
literature suggests that the auditor’s incentives to violate independence increase as
the economic bond between the auditor and auditee increases (DeAngelo
1981a)13. Studies focusing on independence in fact have investigated whether
financial dependence generated by audit fees, non-audit services fees or total fees
enables earnings management (e.g. Frankel et al. 2002; Ashbaugh, LaFond &
Mayhew 2003; Chung & Kallapur 2003; Reynolds, Deis & Francis 2004; Mitra
2007), increases the number of restated financial statements (e.g. Kinney et al.
2004), decreases auditors’ propensity to issue qualified audit reports (e.g.
DeFond, Raghunandan & Subramanyam 2002; Geiger et al. 2003 ), increases the
length of audit tenure (e.g. Barkess & Simnett 1994) or exposes an auditor to
litigation (Antle, Griffen, Teece and Williamson 1997). Overall, this body of
research indicates that auditor independence in fact is not threatened by fees
generated by auditees thus providing indirect support for the possible positive link
between audit fees and audit quality.
A large body of research has also examined the effect of financial dependence on
perceived auditor independence. Firstly, studies have investigated whether the
joint provision of audit and non-audit services causes a negative stock market
reaction (Frankel et al. 2002; Ashbaugh et al. 2003), has an effect on the bond
rating process (Brandon, Crabtree & Maher 2004) or has an impact of shareholder
ratification of the auditor (Glezen & Millar 1985; Raghunandan 2003;
Raghunandan & Rama 2003; Mishra, Raghunandan & Rama 2005). Secondly,
research has used questionnaires to investigate a wide variety of subjects’
perceptions of auditor independence using several measures of financial
dependence (e.g. Knapp 1985; McKinley, Pany & Reckers 1985; Gul 1991; Gul
& Tsui 1992; Bartlett 1993; Teoh et al. 1996; Beattie et al. 1999). Financial
dependence has been hypothesized to develop from audit fees, non-audit services
fees or total fees, although some studies do not clearly articulate whether audit
fees or total fees are under examination. In general, the results of these studies
indicate that financial dependence causes more problems for perceived
independence than actual independence.
13
In general, regulators have not been concerned that audit fees might be too high or that the audit
fee itself might result in financial dependence of the auditor on the auditee (Kinney, William,
Palmrose & Scholz 2004). On the other hand, regulators in the US (SEC 2000; SEC 2003), Europe
(European Commission 2002) and globally (IFAC 2005) have voiced their concerns that the
provision of non-audit services to auditees can create economic dependence and thus poses a
threat to auditor independence. In addition, regulators recognise that total fees paid by the auditee,
regardless of their origin, may compromise independence.
40 Acta Wasaensia
To summarize, the results of the above studies imply that higher audit fees may
be associated with higher audit quality, either through more audit effort (more
hours) or through superior expertise of the auditor (higher billing rates). It can be
argued that a higher audit fee per se does not necessarily ensure a higher quality
audit, particularly if accounting firms have pricing power over clients (Francis
2004). However, studies on audit outcomes also provide evidence that higher
audit fees are related to better financial reporting quality (e.g. Frankel et al. 2002;
Geiger et al. 2003; Larcker et al. 2004; Srinidhi et al. 2007).
3.3 Financial reporting quality
Corporate disclosure may assume a variety of forms: companies may provide
disclosure through regulated financial reports and also engage in voluntary
communication. Among these different forms of disclosures financial reporting is
an important means for management to communicate, among other things, the
company’s performance to external stakeholders. Financial reporting quality is a
prerequisite for efficient capital markets because several individuals and groups
base their resource allocation decisions on financial information (Healy et al.
2001).
The objective of regulators and standard setters is to promulgate rules and
regulations that help ensure financial reporting quality. Financial reporting quality
has also commanded considerable research interest. However, neither researchers
nor current US regulations provide a clear definition of what constitutes financial
reporting quality (Pomeroy et al. 2008). For example the SOX (2002) requires
audit committees and auditors to discuss the quality of the financial reporting
methods of the company. However, SOX (2002) does not define what is meant by
financial reporting quality and therefore this requirement remains vague (Cohen
et al. 2004)14. In addition, Watkins et al. (2004) provides a broad definition
stating that financial reporting quality refers to how well a company’s financial
information reflects the true economic circumstances of the company. This
definition also highlights the unobservable nature of financial reporting quality
(Pomeroy et al. 2008).
Prior research has used several proxies for financial reporting quality (Pomeroy et
al. 2008). Many of the studies examining financial reporting quality have focused
on situations where there are concerns regarding financial information quality.
These situations can be divided into two classes: misstatements outside GAAP
14
See e.g. Pomeroy et al. (2008) for a discussion on this issue.
Acta Wasaensia 41
and misstatements within GAAP (Jiambalvo 1996). Evidence related to
misstatements outside GAAP includes financial restatements (Raghunandan et al.
2003; Kinney et al. 2004), litigation (Antle et al. 1997; Bonner, Palmrose &
Young 1998), SEC enforcements (Dechow, Sloan & Sweeney 1996) and business
failures (Palmrose 1987; Francis & Krishnan 2002). Misstatements within GAAP
are related to the choices of accounting methods or earnings quality measures.
Common indicators of earnings quality are accruals (e.g. Frankel et al. 2002;
Ashbaugh et al. 2003; Chung et al. 2003), earnings informativeness (Earnings
Response Coefficients) (Teoh et al. 1993; Balsam et al. 2003; Ghost et al. 2005;
Wang 2006), predictability of cash flows, measures of income smoothing and
measures of timely loss recognition (Jiambalvo 1996). Audit research has also
used audit report accuracy as an indicator of the quality of the financial reporting
process (DeFond et al. 2002; Geiger et al. 2003).
This study focuses on US listed companies and it is fairly unlikely that they issue
financial reports that do not meet the minimum requirements of GAAP. This
suggests that the present study focuses on the variation of financial reporting
quality within the boundaries of GAAP. Consistent with the literature, the present
study focuses on earnings management as an indicator of financial reporting
quality.
3.3.1 Definition of earnings management
Earnings management can be defined as a purposeful intervention in the external
financial reporting process the purpose of which is to obtain private gain for
shareholders or managers (Schipper 1989). Shareholders will gain from earnings
management if it is used to signal managers’ private information (Healy & Palepu
1995; Subramanyam 1996), to avoid costly debt re-contracting or to reduce
political costs (Watts et al. 1986). On the other hand earnings management may
be costly to shareholders because managers may use earnings management to
pursue their personal gains such as increased compensation (Healy 1985;
Holthausen, Larcker & Sloan 1995) or reduced likelihood of dismissal when
performance is poor (Weisbach 1988; Peasnell, Pope & Young 2005). This aspect
of earnings management may result in financial reports which mislead
stakeholders about the underlying economic performance of the firm (Healy &
Wahlen 1999).
42 Acta Wasaensia
3.3.2 Types of earnings management
Earnings management may be either income increasing or income decreasing.
The literature links income increasing earnings management with situations when
earnings fall below certain thresholds. The literature has addressed three
thresholds: 1) avoiding reporting a loss, 2) reporting a growth in profits, and 3)
meeting analysts’ forecasts. Burgstahler and Dichev (1997) and Degeorge, Patel
and Zeckhauser (1999)15 found using US data higher-than-expected frequency of
firms which had slightly positive reported earnings and lower-than-expected
frequency of firms which had slightly negative reported earnings. These results
are consistent with managements’ attempts to beat the benchmarks in question
(Peasnell et al. 2005).
Managers may also engage in income decreasing earnings management. There are
several plausible explanations why managers may have incentives to engage in
income decreasing earnings management. Income decreasing earnings
management can be driven by managers’ desires to shift abnormal positive
earnings forward in time in order to make the benchmarks easier to cross in the
future (Peasnell et al. 2005). This notion is supported by Degeorge et al. (1999),
who found that managers manipulate reported earnings downwards when pre-
managed earnings substantially exceed benchmark earnings. Alternatively
managers may be reluctant to report large earnings because it might result in
increased earnings-based performance targets in the future (Peasnell et al. 2005).
Consistently, Healy (1985), Gaver, Gaver and Austin (1995), and Holthausen et
al. (1995) found evidence of income-decreasing earnings management when
managers’ accounting-based bonuses were at their maximum using a sample of
US firms.
Management may employ two methods for earnings manipulation: management
can manipulate accounting numbers or change the way the firm does business
(Peasnell et al. 2005). The former method is likely to involve discretionary
accruals because they are regarded as an area where management will use its
discretion to manipulate accounting numbers. Managers may prefer accruals to
manipulate earnings because generally they do not have direct cash flow
consequences and they are relatively difficult to detect. The latter method offers
several options: management may, for example, boost reported profit by cutting
back on advertising and research and development (Bushee 1998), sell assets
(Bartov 1993; Poitras, Wilkins & Kwan 2002), or cut back on staff development
15
Degeorge et al. (1999) also found that there is a hierarchy to the benchmarks reporting a profit
being the most important one followed by reporting growth in earnings. Meeting analysts’
forecasts was of importance only if the other two thresholds had been met.
Acta Wasaensia 43
and essential equipment maintenance. These methods, however, are costly to the
company because they are likely to have negative effects on the firm’s future cash
flows as well as firm value. Therefore it is expected that managers would rather
manipulate accruals than make chances in investment and operating activities
(Peasnell et al. 2005).
3.4 Application of the definitions and operational
measures
This section introduced definitions as well as empirical measures for audit
committee effectiveness, audit quality and financial reporting quality. The aim of
this section was to provide a broad overview of the definitions and measures
suggested by prior literature. The objective of the present study is not to apply all
the definitions and measures discussed but to adopt those that best represent the
variables of interest in the study’s environment. Thus, the present study adapts the
definitions and measures as follows.
Firstly, audit committee effectiveness is generally linked to audit committee’s
ability to fulfil its responsibilities. This study adopts this view and focuses on
audit committees’ ability to enhance financial reporting quality and audit quality.
Prior empirical research has used several surrogates for audit committee
effectiveness, the most essential being audit committee size, independence,
expertise and activity level. Following prior research as well as current US
regulations, this study focuses on audit committee effectiveness generated by the
size of the audit committee, the expertise of audit committee members and the
activity level of the audit committee. In addition it is recognized that audit
committee independence is a crucial component of audit committee effectiveness
(e.g. Beasley et al. 2000; Abbott et al. 2003a). However, the regulations
concerning the sample companies require that all audit committee members need
to be independent of management (e.g. SOX 2002). Therefore, this variable is
constant for all sample companies and is not included in the scope of this study.
The sample companies are expected to comply with the exchanges’ and SOX’s
(2002) requirements concerning the maintenance, composition and
responsibilities of the audit committee. However, this study makes a distinction
between the form (meeting regulatory requirements) and the substance (the
effectiveness of audit committees) of the audit committee. In other words, it is
suggested that regulations set the primary requirements for the maintenance and
responsibilities of an audit committee. Audit committees which exceed these
primary requirements are expected to be more effective to a certain level.
44 Acta Wasaensia
Secondly, following prior audit quality research this study adopts the basic
premise that most audits conform to the applicable legal and professional
requirements. In other words it is expected that audits are generally legally
satisfactory. In addition, it is suggested that audit clients may demand quality
differentiated audits and audit committee effectiveness explains, at least partly,
this demand. The primary measures for audit quality used in existing studies are
audit firm size, audit firm industry specialization, audit tenure and audit fees. This
study uses audit fees as a measure of audit quality. It is argued that audit firm
size, industry specialization and auditor tenure do not provide sufficient measures
for audit quality in the current US audit market. This is because the current US
audit market is heavily concentrated and the Big Four audit firms dominate the
market for auditing services with a market share of 78 % of all US public
companies. In other words, the Big Four firms constitute a tight oligopoly (GAO
2003). Due to the audit market situation, it is suggested that audit firm
characteristics (i.e. size or industry expertise) may not be a sufficient measure of
audit quality. In addition, the SOX (2002) issues restrictions on auditor tenure.
The SOX (2002) requires a rotation of all audit partners after five years’ service.
Given that audit partner tenure is restricted to five years its ability to have an
impact on audit quality is also limited. However, it is reasonable to expect that
some companies may still demand a differentially higher audit quality. It is
argued that audit fees provide a sufficient measure of audit quality. This is
because audit fees are expected to reflect additional audit effort, which leads to a
higher level of audit quality by increasing the probability that auditors detect
potential problems in the company’s financial reporting (DeAngelo 1981a;
DeAngelo 1981b; Caramanis & Lennox 2008).16 Thus, higher audit fees are
expected to be related to better financial reporting quality (Carcello et al. 2002;
Frankel et al. 2002; Abbott et al. 2003a; Larcker et al. 2004; Srinidhi et al. 2007).
Finally, following the literature this study suggests that discretionary accruals
indicate earnings management and can be used as a measure for financial
reporting quality. This study focuses on the negative aspect of earnings
management and takes the perspective that earnings management is undesirable
because it can be costly to shareholders. Consistent with Peasnell et al. (2005) this
study adopts a view that both income increasing and income decreasing earnings
manipulation may impose costs on shareholders and other external parties of the
company. This is because earnings management may result in financial reports
which are misleading to shareholders and other external parties. Thus it is
16
Caramanis et al. (2008) provide direct evidence linking audit effort with audit quality. More
specifically, they examine the effect of audit effort, measured by audit hours, on earnings
management. Their results show that high audit effort decreases the extent to which managers
report aggressively high earnings.
Acta Wasaensia 45
suggested that audit committees and external auditors should be concerned with
both income decreasing and income increasing manipulations. In addition, it is
suggested that managers prefer using discretionary accruals to manipulate
earnings because they do not necessarily have direct cash flow consequences and
are difficult for outsiders to discover.
46 Acta Wasaensia
4 DEVELOPMENT OF THE RESEARCH MODEL
The aim of this section is to develop the research model which will be examined
empirically in this study. This section is organized as follows. Firstly, literature
examining relationships between audit committee effectiveness, audit quality and
financial reporting quality will be introduced. This literature is divided into three
lines as follows. The first line of research examines the relationship between audit
committee effectiveness and financial reporting quality. The second line of
research examines the relationship between audit committee effectiveness and
audit quality. Finally, the third line of research investigates the relationship
between audit quality and financial reporting quality. These separate lines of
research form the basis for the model developed. The studies that are most
significant for the development of the research model are summarized in
Appendix 2. Secondly, the various effect types which can be used to connect the
separate lines of research are introduced. The effect types are derived from other
fields of social sciences and include moderated and mediated effects. Finally, a
model combining the separate lines of research will be introduced. The
construction of the model involves considering the compatibility between
corporate governance and audit literature and the underlying theoretical
perspectives surrounding the alternative effect types.
4.1 Audit committee effectiveness and financial
reporting quality
As discussed in previous sections, the main purpose of the audit committee is to
contribute to corporate control. More specifically, audit committees are expected
to improve a company’s financial reporting quality. The effectiveness of audit
committees in monitoring control has also commanded considerable research
interest. Prior studies have primarily used an archival approach to assess the links
between measures of audit committee effectiveness and financial reporting
quality (Kalbers et al. 1998). These studies are discussed in more detail in the
subsequent sections.
Early studies examined the association between presence of an audit committee
and fraudulent financial reporting. Some of these early studies indicate that audit
committees do not entirely prevent fraudulent reporting by auditees (Verschoor
1989; Verschoor 1990; Beasley 1996). For example, Beasley (1996) found that
the presence of an audit committee does not have an effect on the likelihood of
Acta Wasaensia 47
financial statement fraud. However, the role of audit committees has changed
considerably since this study and more recent studies are generally consistent in
their findings that audit committees appear to be effective in preventing fraud.
Accordingly Dechow et al. (1996) found that firms subject to enforcement actions
by the SEC were less likely to have an audit committee. Similarly McMullen et
al. (1996) found that the presence of an audit committee is associated with fewer
SEC enforcement actions and illegal acts. In addition, Abbott et al. (2000) found
that companies with independent and active audit committees were less likely to
be sanctioned by the SEC for fraudulent or misleading financial reporting.
Beasley et al. (2000) investigated fraudulent companies and their no-fraud
benchmarks in three industries: technology, health care and financial services.
They found that in general fraudulent companies have weaker governance
mechanisms relative to their no-fraud benchmarks. More specifically, fraud
companies in technology and financial services industries had fewer audit
committees. In addition, fraudulent companies in all industries had less
independent audit committees and boards. Audit committee effectiveness has also
been examined in relation to financial restatements. For example Abbott et al.
(2004) found that audit committee independence and activity level decrease the
occurrence of restatements. In addition, audit committee expertise was found to
have a negative association with restatements. Also consistent with the notion that
effective corporate governance is associated with better financial reporting quality
Karamanou and Vafeas (2005) found that companies with more effective boards
and audit committees were more likely to make or update an earnings forecast.
Several studies have investigated the relation between audit committee
effectiveness and earnings management. Bédard et al. (2004) found that audit
committee effectiveness measured by expertise, independence and responsibilities
of the audit committee restrict aggressive earnings management. Similarly Klein
(2002a), and Bradbury et al. (2006), reported a negative relation between audit
committee independence and company’s income increasing discretionary
accruals. Xie et al. (2003) found that audit committee activity level and its
members’ financial sophistication constrain the propensity of managers to engage
in earnings management. In addition, some studies have examined whether the
presence or absence of an audit committee has an effect on financial reporting
quality. Results of Peasnell et al. (2005) showed that the presence or absence of
an audit committee does not have an effect on earnings management among UK
firms. However, Jaggi and Leung (2007) using Hong Kong firms as a sample
found that voluntarily established audit committees play a significant role in
constraining earnings management. Piot et al. (2007) in a French setting also
found that audit committees control for income increasing earnings management.
Wild (1996) investigated whether audit committees have an effect on perceived
48 Acta Wasaensia
earnings quality. He found that market’s reaction to earnings reports increased
after the formation of an audit committee indicating that audit committees
improve perceptions of financial reporting quality.
The above findings lead to the first hypothesis:
H1: Audit committee effectiveness improves financial reporting
quality.
4.2 Audit committee effectiveness and audit quality
In general the literature posits that a company’s corporate governance and
external auditing are complements, whereas a company’s internal control and
external auditing are considered substitutes (Hay, Kneckel & Ling 2008). The
relationship between audit committee effectiveness and audit quality is also
considered to be a complex one. Consistent with the more general discussion, the
literature provides two possible rationales which may explain how audit
committee effectiveness and external audit quality are related: audit committees
and external auditors can be considered as either substitutes or complements for
each other17.
According to the substitute rationale audit committee effectiveness should be
negatively associated with audit quality and vice versa. This is because the
oversight provided by either audit committees or external auditors is expected to
be sufficient to ensure financial reporting quality in its own right. According to
this rationale both the demand for and supply of assurance provided by the
auditor, for example, should be reduced when a company has an effective audit
committee. Although the substitution rationale is initially appealing, it does not
take into consideration the incentives of audit committee members. It can be
argued that audit committee members may wish to invest in external auditing in
order to protect their reputational capital (Abbott et al. 2000; Knechel &
Willekens 2006) or mitigate the risk of financial liability arising from financial
17
Previously the roles of audit committees and external auditors were viewed somewhat
differently. Prior to the requirements for mandatory audit committee formation it was common
that external auditors aided their clients in forming audit committees. This was because audit
committees were viewed as a means to enhance the perceptions of auditor independence (Menon
et al. 1994). Accordingly early studies such as Eichenseher and Shields (1985) and Menon et al.
(1994) found that audit firm type affects the formation of audit committee. More specifically they
reported that companies employing a big audit firm were more likely to voluntarily form an audit
committee. In addition Collier et al. (1999) found that companies with Big Six audit firm had
more active audit committees than companies with non-Big Six audit firm.
Acta Wasaensia 49
statements (Abbott et al. 2000). Thus, the complement rationale suggests a
positive association between audit committee effectiveness and audit quality. This
is because effective audit committees are expected to consist of directors with
higher incentives to demand the assurance provided by external auditors (Hay et
al. 2008).
The research so far is inconclusive as to which rationale is more appropriate in
describing the relationship between audit committee effectiveness and audit
quality. Both rationales have been supported by prior research: studies focusing
on the auditor’s decision-making have provided support for the substitute
rationale, while studies regarding auditor choice and remuneration have provided
support for the complement rationale. These studies will be discussed in more
detail next.
Research supporting the substitute rationale includes studies by Cohen and Hanno
(2000), Bédard and Johnstone (2004), Lee et al. (2004), and Stewart et al. (2007).
Cohen et al. (2000) examined whether audit planning judgments are affected by
the strength of company’s corporate governance structure. They found that audit
planning judgments were less favourable for companies with audit committees
lacking resources and technical experience. In a similar vein, Bédard et al. (2004)
reported that increased earnings manipulation risk is associated with increased
planned audit hours and billing rates. In addition, an experimental study by
Stewart et al. (2007) showed that auditors assess the level of audit risk lower for
companies with audit committees, more frequent committee meetings and higher
auditor’s attendance at meetings. However, the impact of these variables on audit
testing and audit efficiency was perceived by auditors to be minimal. Lee et al.
(2004) provided further collaborative evidence by examining auditor resignations.
They found that auditors are less likely to resign when the company has an
effective independent audit committee, which indicates that audit committee
effectiveness affects auditors’ assessment of audit risk and willingness to continue
the audit assignment.
Research related to the complement rationale provides evidence indicating that
audit committees have a critical role to play in enhancing audit quality at several
stages in external auditing (Lee et al. 2004). Research examining the relationship
between audit committee effectiveness and audit quality can be divided into the
following lines of research (Abbott et al 2003a): 1) auditor selection, 2) audit
coverage and audit fees, and 3) audit process and 4) audit opinion. In addition,
studies have examined the association between audit committees and perceived
audit quality.
50 Acta Wasaensia
The first line of research provides consistent evidence supporting the notion that
effective audit committees demand higher level of audit quality. Abbott et al.
(2000) focused on an association between audit committee effectiveness and audit
firm size and found evidence that companies with independent audit committees
that meet frequently were more likely to select a (then) Big Six audit firm when
switching auditors. Audit committee effectiveness is also found to have an effect
on the selection of an industry specialist auditor. Abbott et al. (2001) reported that
audit committees which are independent and active are more likely to use an
industry specialist audit firm. In a similar vein, Chen et al. (2005) found that the
proportion of independent audit committee members increases the likelihood of
employing a specialist audit firm.
The second line of research examines whether audit committee effectiveness
affects the demand for greater quantity of audit effort from the incumbent auditor
in order to improve audit quality. These studies hypothesize that the amount of
the audit fees reflects audit quantity and further audit quality. Thus, a positive
association between audit committee effectiveness and audit fees is expected.
This positive association can also be expected because the audit committee has
the important duty to ensure that audit hours are not reduced to a level that
compromises the quality of an audit (Stewart et al. 2007). Research results have
generally been consistent with these views. An early study by Collier and
Gregory (1996) found that UK companies with audit committees have higher
audit fees. However, a later study by Goddard and Masters (2000) did not find
this relation. Abbott et al. (2003a) examined the association between audit fees
and audit committee independence, expertise and meeting frequency. Abbott et al.
(2003a) found a positive association between audit committee independence as
well as expertise and audit fees. Number of meetings was not significantly related
to audit fees. Similarly, Lee et al. (2005) found a significant association between
measures of audit committee effectiveness (independence, activity and expertise)
and audit fees. Goodwin-Stewart et al. (2006) also found that the existence of an
audit committee, more frequent committee meetings and increased use of internal
audit were related to higher audit fees. In addition it was found that the expertise
of audit committee members was associated with higher audit fees when meeting
frequency and independence was low. Knechel et al. (2006) reported that audit
fees are higher when a company has an audit committee. Mitra et al. (2007)
demonstrated that more independent audit committees pay higher fees for
auditors. However, other measures of audit committee effectiveness including
audit committee meeting frequency and audit committee expertise did not have an
effect on audit fees. In addition Vafeas et al. (2007) found that audit committee
size, independence level and expertise were positively associated with audit fees.
Finally Hay et al. (2008) examined the effect of internal auditing, corporate
Acta Wasaensia 51
governance and concentration of ownership on audit fees in a New Zealand
setting. They reported that all variables examined, including the existence of an
audit committee, were positively related to audit fees in a less regulated
environment. However, it was found that these relationships did not hold in a
highly regulated environment in which companies have relatively homogeneous
control arrangements.
A related vein of research focuses on the relationship between audit committee
effectiveness and non-audit services fees and provides further evidence that audit
committee effectiveness is related to audit quality. The notion behind these
studies is that effective audit committees restrict the purchase of non-audit
services from their incumbent auditor in order to ensure auditor independence.
Abbott, Parker, Peters and Raghunandan (2003b) supported this notion and
showed that non-audit services fees were lower in companies with independent
and active audit committees. Lee et al. (2005) also suggested that effective audit
committees seek to enhance auditor independence by reducing the non-audit
services provided by the incumbent auditor.
The third research stream focuses on the audit committee’s role in auditor-
management disputes. The underlying notion behind these studies is that an audit
committee member’s characteristics affect his or her willingness to support the
auditor in conflict situations with management. The results of these studies are
somewhat inconclusive. Some studies suggest that audit committee members who
are current managers of companies (Knapp 1987), possess financial expertise
(DeZoort et al. 2001) or are CPAs (Certified Public Accountant) (DeZoort,
Hermanson & Houston 2003a) are more likely to support auditors in audit-
management disagreements. However, DeZoort et al. (2001) found that audit
committee members’ concurrent experience as a board member and as a member
of senior management was associated with greater support for management, while
DeZoort et al. (2003a) did not. DeZoort, Hermanson and Houston (2003b) found
that audit committee members who were CPAs were less likely to support
adjustments proposed by the auditor.
The fourth research stream focuses on audit committees’ ability to mitigate the
threat of management replacing the auditor. It is suggested that by mitigating the
threat of management replacing the auditor the audit committee can affect the
level of audit coverage and promote more truthful reporting by auditors. Carcello
et al. (2000) as well as Carcello and Neal (2003b) examined the relation between
audit committee effectiveness and auditor’s decision to issue a going-concern
report. They found that audit firms are more likely to issue going-concern
opinions to financially distressed firms with independent audit committees. In
52 Acta Wasaensia
addition, Carcello et al. (2003) found that auditors who issue initial going-concern
modifications are less likely to be terminated when audit committee members are
independent, possess governance expertise and have low stockholding.
Archambeault et al. (2001) examined suspicious auditor switches. They found
that companies that made suspicious auditor switches had 1) less independent
audit committees, 2) less experienced audit committee members, 3) smaller audit
committees, and 4) less active audit committees than matched counterparts.
In addition to the above studies, a related research stream links audit committee
effectiveness with perceived audit quality and independence. For example
Raghunandan et al. (2003) investigated the association between audit committee
effectiveness and shareholder ratification of auditor. The results indicated that in
companies with high non-audit fee ratios, shareholders were less likely to vote
against auditor ratification if the audit committee had solely independent
directors. Raghunandan et al. (2003) concluded that audit committee composition
is associated with shareholders’ perceptions of audit quality and independence.
Teoh et al. (1996) investigated the perceptions of Malaysian accountants and
found that the formation of the audit committee had a strong positive impact on
enhancing perceived auditor independence. Beattie et al. (1999) also reported
similar results on UK data. Goodwin and Seow (2002) investigated the
effectiveness of audit committees in enhancing the quality of financial reporting
and auditing perceived by auditors and directors in Singapore. A strong audit
committee was found to have a significant impact on perceived audit
effectiveness and the quality of financial reporting
The above discussion leads to the second hypothesis:
H2: Audit committee effectiveness increases the demand for audit
quality.
4.3 Audit quality and financial reporting quality
The role of the external audit is to ensure financial reporting quality (Cohen et al.
2004). Several empirical studies have examined whether external audit fulfils this
role and contributes to financial reporting quality. These studies have used several
measures for both audit quality and financial reporting quality. Based on
measures of financial reporting quality these studies can be divided into two
groups: 1) studies focusing on actual financial reporting quality and 2) studies
focusing on perceived financial reporting quality. Overall, this line of research
provides evidence that external audit quality contributes to both actual and
Acta Wasaensia 53
perceived financial reporting quality. These studies will be discussed next in more
detail.
There is an extensive body of research focusing on the relationship between audit
quality and actual financial reporting quality. Many of these studies examine the
association between audit quality and earnings quality. These studies suggest that
audits of higher quality are more effective in restricting management discretion
over accounting issues than audits of lower quality, and thus result in better
earnings quality. These studies consider accruals as instruments which
management prefers to use in order to manage earnings. Research has provided
extensive evidence that audit quality measured by audit firm size (Becker et al
1998; Francis et al. 1999), audit firm industry specialization (Balsam et al. 2003;
Krishnan 2005), auditor tenure (Myers et al. 2003; Piot et al. 2007) or audit fees
(Frankel et al. 2002; Larcker et al. 2004; Srinidhi et al. 2007) increases earnings
quality (i.e. decreases discretionary accruals). In addition to accruals, audit quality
has been linked to several other measures of financial reporting quality, such as
restated financial statements (e.g. Kinney et al. 2004), audit reports (e.g. Geiger et
al. 2003), and litigation (Antle et al. 1997). These studies mainly examine the
relationship between auditor remuneration and financial reporting quality on the
basic premise that financial dependence threatens auditor independence and thus
decreases financial reporting quality. Overall, this line of research fails to support
this premise and thus provides indirect support for the notion that audit fees can
be considered an indicator of auditor’s effort to conduct a high quality audit
Many studies have also examined how audit quality affects perceived financial
reporting quality. These studies are also quite consistent in their findings,
reporting a positive link between audit quality and perceived financial reporting
quality. Studies focusing on ERC have found that audit firm size (Teoh et al.
1993), industry specialization (Balsam et al. 2003) and tenure (Ghost et al. 2005)
have an effect on perceived earnings quality. In other words, these studies have
documented a positive association between audit quality and ERCs. Similarly
studies focusing on IPOs (Initial Public Offering) have found that higher audit
quality indicated by audit firm size result in more favorable firm valuations
(Beatty 1989). In addition, a study by Krishnan (2003) focused on earnings
surprises and reported collaborating evidence showing more positive market
reactions to earnings surprises for companies with industry specialized auditors.
The above discussion leads to the third hypothesis:
H3: Audit quality improves financial reporting quality.
54 Acta Wasaensia
4.4 Effect types
As evidenced by the foregoing discussion the relationships between audit
committee effectiveness, audit quality and financial reporting quality have
aroused considerable research interest. The objective of the present study is to
place the separate relationships emerging from the literature in a more
comprehensive model. Since corporate governance and audit research have
traditionally focused on direct effects between variables the model is bound to
include more complex effect types. Thus the objective of the study involves
selection of an effect type which adequately describes the relationships between
the variables of interest and is in accordance with the theory and results of
empirical research.
The alternative effect types are ultimately derived from other fields of social
sciences, in which researchers have found it necessary to invoke conceptual
models that include various effect types to describe for example human behavior
or decision making18. The main effect types addressed in these studies are
mediated effects and moderated effects. These effect types are also addressed by
management accounting research, which has examined, for example, the effects
of moderator and mediator variables on the relationships between company’s
contingency characteristic and company performance (see e.g. Gerdin & Greve
2004; Jokipii 2006).
Before the selection of an appropriate effect type it is important to make a
distinction between these effect types, because they have different theoretical
starting points and they answer different research questions. In general, a variable
can serve either function depending on the theoretical model under investigation.
However, a variable cannot serve both functions in the same analysis. In other
words, if a variable is tested as both a moderator and a mediator in competing
models and both models yield significant results this should be considered to be a
contradictory result (see e.g. Baron et al. 1986; Holmbeck 1997; Muller, Judd &
Yzerbyt 2005; Jokipii 2006). The aim of the following sections is to define the
concepts of moderator and mediator and provide the basis for the choice
regarding the effect type in the model developed.
18
See e.g. MacKinnon, Lockwood, Hoffman, West and Sheets (2002) and MacKinnon, Fairchild
and Fritz (2007) for a discussion regarding these research questions.
Acta Wasaensia 55
4.4.1 Moderation effect
Moderated effect focuses on factors having an effect on the strength and/or
direction of the relationship between the independent variable and the dependent
variable. A moderation variable is often introduced in order to examine the
conditions under which the strength of the effect of an independent variable on a
dependent variable varies. Alternatively, a moderation variable can be introduced
in order to examine the condition under which the direction of effect varies
(Muller et al. 2005). Baron et al. (1986) define a moderator variable more
specifically as follows:
“a qualitative (e.g., sex, race, class) or quantitative…variable that
affects the direction and/or strength of a relation between an
independent or predictor variable and a dependent or criterion
variable…a basic moderator effect can be represented as an
interaction between a focal independent variable and a factor (the
moderator) that specifies the appropriate conditions for its
operation…Moderator variables are typically introduced when
there is an unexpectedly weak or inconsistent relation between a
predictor and a criterion variable (p. 1174, 1178).”
The moderation effect is presented diagrammatically in Figure 7. In the figure X
refers to an independent variable, Mo refers to a moderation variable and Y refers
to a dependent variable. In the figure moderator variable Mo moderates the
relationship between the independent variable X and the dependent variable Y
(Baron et al. 1986; Holmbeck 1997). As is evident from the figure, the
independent variable X and the moderator variable Mo are independent of each
other. That is, the independent variable should not have an effect on the
moderator variable in the moderation model (Muller et al. 2005).
56 Acta Wasaensia
Figure 7. Moderation effect (Baron et al. 1986; Holmbeck 1997).
4.4.2 Mediation effect
The mediation effect consists of a chain of relations where an independent
variable affects a mediator variable which then affects the dependent variable.
The mediation effect is built on the assumption that there is a significant
association between the independent variable and the dependent variable before
testing for a mediated effect. The mediator variable is introduced in order to
examine whether the effect of an independent variable on a dependent variable
goes through a mediating variable. In other words a mediator variable specifies
how (or the mechanism by which) a given effect occurs. Baron et al. (1986)
describe the mediator variable as follows:
“the generative mechanism through which the focal independent
variable is able to influence the dependent variable of
interest…(and) Mediation …is best done in the case of a strong
relation between the predictor and the criterion variable (p. 1173,
1178).
The mediation effect is depicted diagrammatically in Figure 8. In the figure X
refers to the independent variable, Me refers to the mediator variable and Y refers
to the dependent variable. In the figure the mediator variable Me mediates the
relationship between the independent variable X and the dependent variable Y
(Baron et al. 1986; Holmbeck 1997).
Mo
X Y
Notes:
The variables are defined as follows:
Y= Dependent variable
X= Independent variable
Mo= Moderating variable
Acta Wasaensia 57
Figure 8. Mediation effect (Baron et al. 1986; Holmbeck 1997).
The mediated effect may be either full or partial. A full mediation occurs when the
mediator eliminates the effect of the independent variable on the dependent
variable, whereas a partial mediation occurs when the mediator significantly
decreases the effect of independent variable on dependent variable. It should be
noted that a partial mediation may indicate the operation of multiple mediating
factors (Baron et al. 1986; Holmbeck 1997).
It should be noted that a distinction can be made between indirect effect and
mediated effect, although these concepts are frequently used as synonyms in the
literature. Figure 9 depicts an indirect effect. In the figure X denotes the
independent variable, I the intervening variable and Y the dependent variable.
The difference between indirect and mediated effects arises from the initial
relationship between the independent variable and the dependent variable.
Following Holmbeck (1997) it is asserted that the mediated effect is possible only
if the independent variable has an initial effect on the dependent variable. The
mediator variable is expected to account for the relationship between the
independent and the dependent variable. However, an indirect effect may occur if
significant relationships are found between the independent variable and
intervening variable as well as the intervening variable and dependent variable. In
the case of indirect effects it cannot be claimed that the independent variable and
the dependent variable are significantly related and thus interpretations
concerning such effects should be conservative (Streiner 2005; Holmbeck 1997).
Me
X Y
Notes:
The variables are defined as follows:
Y= Dependent variable
X= Independent variable
Me= Mediating variable
58 Acta Wasaensia
Figure 9. Indirect effect (Streiner 2005).
4.5 Selection of effect type
The literature suggests that the relationships between corporate governance actors
are complex (e.g. Cohen et al. 2004). Thus, a substantial amount of research has
focused on examining relationships between audit committee effectiveness, audit
quality and financial reporting quality. Collectively the existing research provides
evidence for relationships between: 1) audit committee effectiveness and financial
reporting quality, 2) audit committee effectiveness and audit quality, and 3) audit
quality and financial reporting quality. The evidence shows that audit committee
effectiveness is associated with better financial reporting quality as well as with
greater demand for external audit quality. The results also indicate that audit
quality enhances financial reporting quality. No previous research, however, has
attempted to place these relationships in a more comprehensive model.
Thus, the objective of this research is to incorporate the above relationships and
develop a model which provides a more comprehensive understanding of the
relationships. This objective involves selection of an effect type which
sufficiently describes the underlying dynamics of the relationships. This study
considers two alternative effect types, namely the moderation effect and the
mediation effect. These effect types are derived from other fields of social
sciences, where they are frequently examined (see e.g. Baron et al. 1986;
Holmbeck 1997; Holmbeck 2002; Muller et al. 2005). The main focus of the
model developed is on the role audit quality plays in the relationship between
audit committee effectiveness and financial reporting quality.
Y X I
Notes:
The variables are defined as follows:
Y= Dependent variable
X= Independent variable
I= Intervening variable
Acta Wasaensia 59
Firstly, audit quality could be modelled as a moderator which alters the strength
of the relationship between audit committee effectiveness and financial reporting
quality. This effect type can be based either on the complement or the substitute
rationale. Based on a complement rationale it can be suggested that audit quality
strengthens the relationship between audit committee effectiveness and financial
reporting quality. This notion maintains that external auditors contribute to the
monitoring of financial reporting quality provided by audit committees. On the
other hand from the substitute perspective the effect could be the opposite. This is
because an effective audit committee would decrease the need for high quality
audit service and vice versa. However, in the moderation model the independent
variable and the moderator cannot be related. The literature provides evidence of
a positive effect between audit committee effectiveness and audit quality.
Therefore, it is suggested that the moderation effect, on a theoretical basis, cannot
be employed to describe the role of audit quality on the relationship between audit
committee effectiveness and financial reporting quality (Muller et al. 2005).
Secondly, audit quality may function as a mediator in the relationship between
audit committee effectiveness and financial reporting quality. This effect type is
in accordance with the complement rationale: audit committee effectiveness can
be expected to improve audit quality which further improves financial reporting
quality. The mediation effect maintains that there are significant relations
between 1) independent variable and dependent variable, 2) independent variable
and mediator variable, and 3) mediator and dependent variable. If audit
committee effectiveness is modelled as an independent variable, audit quality as a
mediator variable and financial reporting quality as a dependent variable, the
empirical studies provide evidence for all the required relationships. In other
words, combining the relationships reported in the existing research forms a
sequence from audit committee effectiveness to audit quality and further to
financial reporting quality. Thus, empirical research and the theory concerning the
mediation effect are analogous. Therefore it is suggested that audit quality
functions as a mediator in the relationship between audit committee effectiveness
and financial reporting quality.
Based on the above propositions a fourth hypothesis is proposed:
H4: Audit quality mediates the relationship between audit
committee effectiveness and financial reporting quality.
60 Acta Wasaensia
5 METHODOLOGY AND SAMPLE
This research focuses on the interplay between audit committees and external
auditors in ensuring financial reporting quality. Following the literature this study
examines the relationships between 1) audit committee effectiveness and financial
reporting quality, 2) audit committee effectiveness and audit quality, 3) audit
quality and financial reporting quality. In addition, as a novel approach, these
relationships are combined into a more comprehensive model which suggests that
audit quality mediates the relationship between audit committee effectiveness and
financial reporting quality. The rationalization for the model developed was
presented in the previous section.
This section will introduce the methodology for examining the research model
proposed. The section is organized as follows. Firstly, the analytic techniques for
testing the mediation effect are explained. Secondly, operational measures for
audit committee effectiveness, audit quality and financial reporting quality as well
as control variables are introduced. This part also involves a restatement of the
study’s hypotheses in terms of operational measures. Thirdly, adaption of analytic
techniques is explained. Finally, sample selection criteria and descriptive
statistics are presented.
5.1 Statistical mediation
Prior audit research has mainly tested direct effects between measures of audit
committee effectiveness, audit quality and financial reporting quality using
multiple regression analysis. Hence, no prior research has addressed the
possibility of more sophisticated effect types between the variables. To bridge this
apparent gap in the literature the present study investigates an alternative effect
type, the mediation effect. More specifically, it is suggested that audit quality
functions as a mediator in the relationship between audit committee effectiveness
and financial reporting quality. The statistical methods used to test mediation
effect are derived from other fields of social sciences, where this type of effect is
frequently tested (see e.g. MacKinnon, Lockwood, Hoffman, West & Sheets
2002).
Several methods for testing mediation can be derived from the literature.
According to MacKinnon, Fairchild and Fritz (2007) 19 the three major statistical
19
See e.g. MacKinnon et al. (2007) and Mackinnon et al. (2002) for a review of these methods.
Acta Wasaensia 61
approaches for testing mediation hypothesis are: 1) the Causal Steps Method, 2)
the Difference in Coefficients Method and 3) the Product of Coefficients Method.
The first method, the Causal Steps Method, originates from the seminal work by
Baron et al. (1986). This method involves testing series of relationships (or paths)
in the proposed mediation model. The second approach, the Difference in
Coefficients Method, focuses on the difference between the regression coefficient
before and after adjustment for the mediating variable into a regression model
regarding a relationship between the independent variable and the dependent
variable. The third approach, the Product of Coefficients Method, attempts to test
the significance of the mediated effect (MacKinnon et al. 2002).
The above methods are somewhat overlapping and it has been suggested that
combining the Causal Steps Method and the Product of Coefficients Method
provides a more thorough analysis of the mediation effect than any single method
(see Holmbeck 1997). Thus, following Holmbeck (1997) both the Causal Steps
Method and a significance test for the mediated effect (i.e. the Sobel Test)
determined by the Product of Coefficients method are employed in this study. The
next sections will discuss these methods in more detail.
5.1.1 Causal Steps Method
As discussed in Section 4.4.2 the mediation hypothesis suggests a series of
relations where the independent variable affects a mediating variable, which
affects the dependent variable (MacKinnon et al. 2007) with an initial assumption
that the independent variable affects the dependent variable (Holmbeck 1997).
This study employs the Causal Steps Method for examining the presence of the
mediated effect. This method was introduced by Baron et al. (1986) and is the
most commonly used method for testing mediation, for example, in psychological
literature (MacKinnon et al. 2002; MacKinnon et al. 2007). The Causal Steps
Method involves the estimation of three regression models which are presented in
Table 1 (e.g. Baron et al. 1986; Muller et al. 2005).
62 Acta Wasaensia
Table 1. Regression models required by the Causal Steps Method (Baron et al.
1986).
Condition Equation no
Condition 1:
Y=β10+β11X+ε1
(1)
Condition 2:
Me=β20+β21X+ε2
(2)
Conditions 3 and 4:
Y=β30+β31X+ β32Me+ε3
(3)
The variables are defined as follows:
Y= Dependent variable
X= Independent variable
Me= Mediating variable
The above regression models can also be placed in the mediation model diagram
as presented in Figure 10. The first regression model is related to the direct effect
between independent variable and dependent variable. The second regression
model tests the relation between the independent variable and the mediator. The
third regression model tests the effect of the independent variable on dependent
variable after the mediator has been included in the regression model. In specific
terms the Causal Steps Method requires that four conditions must be met for a
variable to be considered a mediator. These conditions are the following (Baron et
al. 1986):
1. In Equation 1, the independent variable X must have an effect on the
dependent variable Y (β11 is significant).
2. In Equation 2, the independent variable X must have an effect on the
mediator Me (β21 is significant).
3. In Equation 3, the mediator Me must have an effect on the dependent
variable Y controlling for the independent variable X (β32 is significant).
4. In Equation 3, the effect of the independent variable X on the dependent
variable Y (β31) should be smaller, in absolute value, than the effect in
Equation 1 (β11).
Acta Wasaensia 63
Me
X Y
Y X
β11
β21 β32
β31
Y=β10+β11X+ε1
Me=β20+β21X+ε2
Y=β30+β31X+ β32Me+ε3
Figure 10. Mediation model diagrammatically and regression models to test the
mediated effect (MacKinnon et al. 2002).
The literature has debated whether the first condition is necessary for a mediated
effect to occur (Muller et al. 2005). The present study adopts the view that the
first condition must be met in order to test the mediation effect. This view is
consistent, for example, with Shrout and Bolger (2002) who state that:
“experimentalists who wish to elaborate the mechanism of an experimental effect
need to first establish that the effect exists” (p. 430). In addition, the literature
provides several alternatives as to how the fourth condition is established. In
general, the results are interpreted as indicating full mediation (i.e. the effect of
the independent variable is completely transmitted through the mediating
variable) if the independent variable coefficient β31 does not differ significantly
from zero when the mediating variable is included in the regression model. The
literature also recognizes the possibility of partial mediation where the effect of
the independent variable is transmitted through the mediating variable only
partially (Baron et al. 1986). Thus, a result indicating that │ β31│<│ β11│is
consistent with the existence of partial mediation. Following, for example, Baron
et al. (1986) this study adopts the view that both full and partial mediation effects
are acceptable.
Notes:
The variables are defined as follows:
Y= Dependent variable
X= Independent variable
Me= Mediating variable
64 Acta Wasaensia
5.1.2 Sobel Test
Holmbeck (2002) argues that a single implementation of the Causal Steps Method
may lead to false-negative or false-positive conclusions regarding the occurrence
of the mediation effect. False-negative conclusions refer to the rejection of the
mediation hypothesis when it should be accepted and false-positive conclusions
refer to acceptance of the mediation hypothesis when it should be rejected.
According to Holmbeck (2002) false conclusions can be reduced by testing the
significance of the mediated effect.
The literature proposes several tests for the significance of the mediated effect.
Related to these tests the mediated effect can be determined by the following
equality relationship which exists among the parameters of the regression models
1-3 (see e.g. MacKinnon, Warsi & Dwyer 1995; Muller et al. 2005):
β11- β32 = β21* β32
Thus, tests for the significance of the mediated effect involve testing whether the
parameter difference on the left side of the above equality departs from zero or
whether the product on the right side does so. Following Holmbeck (2002) this
study employs the Sobel Test to examine the significance of the mediated effect
and thus, the focus is on the latter.
More specifically the Sobel Test involves calculating the coefficient for the
mediated effect (i.e. indirect effect)20. This is achieved by multiplying coefficients
β21β32. The coefficient of the mediated effect is further divided by its standard
error seβ21β32. This test value is compared to the standard normal distribution.
Thus, in terms of equations the Sobel Test can be presented as follows (Sobel
1982; Baron et al. 1986):
3221
3221
3221
ββ
ββ
ββ
se
z
⋅
= (4)
( ) ( )2212322322213221 ββββββ sesese ⋅+⋅= (5)
20
Refer to regression models 2 and 3 for the notations.
Acta Wasaensia 65
The variables are defined as follows:
β32= Unstandardized regression coefficient for the association between the
mediator and the dependent variable.
β21= Unstandardized regression coefficient for the association between the
independent variable and the mediator.
seβ32 = Standard error of the mediator β32.
seβ21= Standard error of the independent variable β21.
As a summary, the mediated effect is tested as follows. Firstly, the Causal Steps
Method is employed to test the necessary relations for the mediated effect. If the
variables in a tested mediation model fulfill all four conditions of the Causal
Steps Method the results are considered to support the mediation effect. In this
study both full and partial mediation effects are considered to be an acceptable
result. Secondly, if the Causal Steps Method supports the mediation model Sobel
Test is applied to test the significance of the mediated effect. If Sobel Test
statistics are greater than 1.96 the mediated effect is considered to be statistically
significant at the 5% level.
5.2 Operational measures
This study develops a model in which audit quality mediates the relationship
between audit committee effectiveness and financial reporting quality. Analysis
of the model involves testing four hypotheses which are analogous to the
conditions of the Causal Steps Method and must be reached in order for the
mediation effect to occur. The hypotheses establish the predicted signs of the
relationships. Thus, the following hypotheses are tested:
H1: Audit committee effectiveness improves financial reporting quality.
H2: Audit committee effectiveness increases the demand for audit quality.
H3: Audit quality improves financial reporting quality.
H4: Audit quality mediated the relationship between audit committee
effectiveness and financial reporting quality.
66 Acta Wasaensia
Initially the model developed focuses on five variables. Firstly, there are three
independent variables which serve as measures of audit committee effectiveness,
namely audit committee size, audit committee expertise ratio and audit committee
meeting frequency. Secondly, there is the dependent variable, which is a measure
of financial reporting quality. In this study discretionary accruals are suggested to
represent financial reporting quality in general and earnings quality in particular.
Finally, there is a potential mediating variable which is a measure of audit
quality. In this study audit quality is measured by audit fees paid to the external
auditor. Further rationalization for these variables is presented in Sections 5.2.1-
5.2.3. The hypotheses discussed above can be restated as follows in terms of
operational measures:
H1: There is a negative relationship between audit committee size (audit
committee expertise ratio and audit committee meeting frequency
respectively) and discretionary accruals.
H2: There is a positive relationship between audit committee size (audit
committee expertise ratio and audit committee meeting frequency
respectively) and audit fees.
H3: There is a negative relationship between audit fees and discretionary
accruals.
H4: Audit fees mediate the relationship between audit committee size
(audit committee expertise ratio and audit committee meeting
frequency respectively) and discretionary accruals.
Figure 11 summarizes the operational measures of the developed model as well as
hypothesized relationships between variables. As can be seen from the top of the
figure, variables measuring audit committee effectiveness (audit committee size,
audit committee expertise ratio and audit committee meeting frequency) are
expected to have a negative effect on discretionary accruals. This is because more
effective audit committees are expected to decrease management discretion over
accounting issues and thus lead to a lower level of discretionary accruals than less
effective audit committees. These relationships are addressed by Hypothesis 1. In
addition, as can be seen from the bottom of the figure, measures of audit
committee effectiveness are expected to have a positive effect on audit fees. The
notion behind this relationship is that audit committee effectiveness is expected to
increase the demand for external audit effort, which leads to an increase in audit
fees. In turn, increase in external audit effort is expected to be positively related to
audit quality. This relationship is consistent with Hypothesis 2. The bottom of the
Acta Wasaensia 67
figure also indicates that audit fees have a negative effect on discretionary
accruals. This is because greater audit effort and better audit quality are expected
to constrict earnings management, and thus to be related to better financial
reporting quality. This is the notion underlying Hypothesis 3. Finally, measures of
audit committee effectiveness, audit quality and financial reporting quality are
connected to form a model in which the effect of audit committee effectiveness
on financial reporting quality goes through audit quality. The bottom of Figure 11
illustrates this effect. The mediation effect is addressed by Hypothesis 4.
Figure 11. Hypothesized relationships between measures of audit committee
effectiveness, audit quality and financial reporting quality.
As discussed previously, the hypotheses of the study are analogous with the four
conditions of the Causal Steps Method. Thus, in the subsequent sections the term
condition refers to both conditions of the Causal Steps Method and these
hypotheses. If the condition is fulfilled the related hypothesis is also supported.
The following sections will introduce the operational measures of audit
committee effectiveness, audit quality and financial reporting quality in more
(- )
(+)
(- )
( - )
(+)
(- )
(+
A udit committee
size
Audit committee
meeting frequency
A udit committee
size
A udit committee
expertise ratio
A ud it fees
Discretionary
accruals
Discretionary
accruals
A udit committee
expertise ratio
Audit committee
meeting frequency
Notes:
Expected signs of the relationships are presented in parentheses
68 Acta Wasaensia
detail. In addition to the variables of interest, this study employs control variables
for both discretionary accruals and audit fees. These variables are also introduced
in the coming sections. Appendix 3 provides a summary of the measures used in
the empirical analyses.
5.2.1 Measurement of audit committee effectiveness
The operational measures for audit committee effectiveness are drawn from
empirical research and from current US regulations concerning corporate
governance and audit committees. In general the literature suggests that audit
committee effectiveness is dependent on its composition as well as its activity
level (e.g. DeZoort et al. 2002; SOX 2002; Beasley et al. 2000). Accordingly a set
of three audit committee characteristics is suggested to have an effect on audit
committee effectiveness namely, audit committee size, audit committee expertise
ratio, and audit committee meeting frequency.21 The first two measures are
related to audit committee composition and the last one is related to audit
committee activity level.
A substantial amount of research has provided evidence that these audit
committee effectiveness measures are associated with financial reporting quality
(e.g. Xie et al. 2003; Bédard et al. 2004). In addition prior studies provide evidence
consistent with the notion that these audit committee effectiveness measures
enhance the demand for external audit quality (Abbott et al. 2003a, Lee et al.
2005, Goodwin-Stewart et al. 2006; Vafeas et al. 2007). More specifically, audit
committee size (ACSIZE) is measured as the number of directors serving on the
audit committee. Audit committee expertise ratio (ACEXP) is the ratio of
financial experts on the audit committee. Audit committee meeting frequency
(ACMEET) is measured as the number of audit committee meetings held during
the fiscal year. Following the literature these measures are expected to have a
negative effect on discretionary accruals and positive effect on audit fees.
21
Although the literature regards audit committee independence as an important determinant of
audit committee effectiveness it is not included in the scope of this study. This is because current
US regulations (e.g. SOX 2002) require that all audit committee members must be independent of
the company. Therefore all sample companies are expected to have independent audit committees
and this variable would not have variance to conduct statistical analysis.
Acta Wasaensia 69
5.2.2 Measurement of audit quality
The audit research has suggested several measures for audit quality including
audit firm size, audit firm industry specialization, audit tenure and audit fees.
However, this study is situated in a highly regulated as well as concentrated audit
environment. Thus, most US listed companies are expected to employ Big Four
audit firms as well as industry specialist audit firms. In addition the SOX (2002)
restricts audit partner tenure to five years. Therefore, audit firm size, industry
specialization and tenure are expected to have a limited ability to reflect audit
quality. However, it is suggested that companies may still require quality
differentiated audits and audit fees are expected to provide a sufficient indicator
of audit quality. Following the literature it is suggested that audit fees reflect audit
effort, which further benefits auditor’s decision-making and thus improves the
quality of services provided by the external auditor (e.g. Carcello et al. 2002;
Frankel et al. 2002; Abbott et al. 2003a; Larcker et al. 2004; Srinidhi et al. 2007;
Caramanis et al. 2008). In addition it is expected that certain drivers of audit
quality, such as audit committee effectiveness, are associated with variations in
the level of audit fees. This is because these drivers may affect audit effort during
the planning of an audit as well as during the course of the audit (Simunic 1980;
Hay et al. 2004; Watkins et al. 2004).
The above arguments are supported by several lines of empirical research linking
audit fees with knowledge spillovers (e.g. Firth 1997; Firth 2002), free premiums
(e.g. Simon et al. 1988; Ferguson et al. 2002), demand for audit quality (e.g.
Abbott et al. 2003a) and financial reporting quality (e.g. Srinidhi et al. 2007). In
general, these studies suggest that audit fees can be associated with audit quality
through either greater audit competence or more audit work (Francis 2004). Thus,
according to the literature audit quality is measured by the natural logarithm of
audit fees (AUDITFEE) paid to the incumbent auditor. It is suggested that audit
fees are related to an improvement in financial reporting quality, and thus have a
negative effect on discretionary accruals.
5.2.3 Measurement of financial reporting quality
As discussed in Section 3.3 financial reporting quality has not been
unambiguously defined by either regulators or prior research. Due to the
vagueness of the definition, prior research has employed a number of measures
for financial reporting quality. In general studies have concentrated on situations
where financial reporting quality may be impaired. These situations can be
divided into misstatements outside GAAP and misstatements within GAAP.
70 Acta Wasaensia
This study focuses on US listed companies and it is expected that these
companies rarely issue financial reports which do not meet minimum GAAP
requirements. Therefore this study focuses on financial reporting within GAAP.
More specifically, discretionary accruals are used as a measure of financial
reporting quality in general and earnings quality in particular. The basic premise
is that discretionary accruals capture earnings management and therefore provide
an inverse measure of earnings quality. The low level of discretionary accruals is
expected to indicate good earnings quality (e.g. Healy 1985; DeAngelo 1986;
Jones 1991; Dechow, Sloan & Sweeney 1995).
Several methods have been proposed for estimating discretionary accruals. Early
studies use the change in total accruals as a measure of discretionary accruals
(Healy 1985; DeAngelo 1986) whereas more recent research uses linear
discretionary accruals models to decompose accruals into discretionary and
nondiscretionary components (Jones 1991; Dechow et al. 1995, Dechow et al.
2002). However, it has been argued that the conventional accruals models fail to
recognize the nonlinear nature of the accounting accruals process (Ball &
Shivakumar 2005). To address this issue, a modified Dechow et al. (2002) model
employed by Srinidhi et al. (2007) and Francis, LaFond, Olsson and Schipper
(2005) is adopted in this study. In the original Dechow et al. (2002) model current
accruals are the dependent variable and cash flows in previous, current and
subsequent years are independent variables. Following McNichols (2002), Francis
et al. (2005) and Srinidhi et al. (2007) the change in sales revenues (∆Revt) and
gross property, plant and equipment (PPEt) are included in the model as
additional control variables. Thus, the model used to estimate discretionary
accruals (ACC) is as follows:
TCAt= β1+ β2OCFt-1+ β3OCFt+ β4OCFt+1+ β5∆Revt+ β6 PPEt+ εt (6)
The variables are defined as follows:
TCA= ∆CA-∆Cash-(∆CL-∆STDeb)
∆CA= Change in current assets
∆Cash= Change in cash balance
∆CL= Change in current liabilities
Acta Wasaensia 71
∆STDebt= Change in short term debt included in current liabilities
OCF=Operating cash flow from the cash flow statement
∆Rev= Change in revenues
PPE= Gross property, plant and equipment
In line with Srinidhi et al. (2007) and Francis et al. (2005) all variables depicted
above are scaled by average total assets in t and t-1. In addition, all changes
presented above are between period t and t-1. The model is estimated separately
for each industry (two-digit SIC code)22 with a minimum of 20 observations.
Discretionary accruals (ACC) are then calculated as the residuals from the above
regression model. The interpretation of the residual is as follows: a higher value
of the residual is expected to indicate a greater level of earnings management and
lower earnings quality.
5.2.4 Control variables
As discussed previously, the Causal Steps Method involves a phased analysis
technique in which mediator variable (AUDITFEE) is a dependent variable in one
regression and becomes an independent variable in another. In addition the
mediation model involves independent variables (ACSIZE, ACEXP and
ACMEET) and a dependent variable (ACC). Due to the analytic technique,
statistical analyses employ two sets of control variables. The first set involves
control variables related to discretionary accruals (ACC). The second set of
variables involves control variables related to audit fees (AUDITFEE). These
control variables are discussed separately in the following sections.
5.2.4.1 Control variables related to discretionary accruals
Following prior research, the analysis involves selected firm characteristics which
are expected to have an impact on earnings management. These firm
characteristics proxy for company size, uncertainty of operations, systematic risk
and growth opportunities (see e.g. Becker et al. 1998; Reynolds & Francis 2000;
Dechow et al. 2002; Cheng & Warfield 2005; Wang 2006; Srinidhi et al. 2007).
The logarithm of total assets (TA) is a proxy for company size. Following prior
22
Industries with less than twenty observations were excluded from the sample (see e.g. Francis et
al. 2005).
72 Acta Wasaensia
research, large companies are expected to have systematically lower discretionary
accruals due to operating characteristics such as greater stability and
diversification of portfolio of activities (Becker et al. 1998; Dechow et al. 2002;
Srinidhi et al. 2007). Thus, a negative relationship between TA and ACC is
expected. Uncertainty of operations is measured by operating cycle (OPCYCLE)
which is calculated as follows (see e.g. Srinidhi et al. 2007):
(7)
For companies with no inventories OPCYCLE is calculated as follows:
(8)
Dechow et al. (2002) and Srinidhi et al. (2007) argue that longer operating cycle
is associated with more uncertainty and more estimation, thus leading to lower
earnings quality. Therefore OPCYCLE is expected to have a positive effect on
ACC. Growth rate in net sales (SALESG) is used to capture the effect of growth
opportunities on discretionary accruals. High growth firms have high equity
incentives and thus have greater incentives to manage earnings than low growth
companies (Antle et al. 2006). Therefore, SALESG is expected to be positively
associated with ACC. Occurrence of loss (LOSS) serves as a proxy for systematic
risk. LOSS is an indicator variable which takes a value of 1 if the net income of
the fiscal year is negative and otherwise 0. The literature suggests that riskier
firms which are financially distressed may be more prone to use accruals to
manage earnings upwards (Dechow et al. 2002; Antle et al. 2006 Srinidhi et al.
2007). Thus, LOSS is expected to have a positive effect on ACC.
5.2.4.2 Control variables related to audit fees
According to Hay, Knechel and Wong (2006) variables related to audit fees
involve both “supply” variables and “demand” variables. Supply variables refer to
company characteristics which may have an effect on auditor’s planning decisions
regarding the level of audit effort. On the other hand, demand variables refer to
Acta Wasaensia 73
company characteristics which may influence the demand for greater audit effort
and audit quality.
Supply variables related to audit fees serve as proxies for company size,
complexity, inherent risk, profitability and leverage. Company size is measured as
the natural logarithm of total assets (TA). The research has shown that company
size has a very high power to explain audit fees (e.g. Carcello et al. 2002; Abbott
et al. 2003a; Abbott et al. 2003b; Lee et al. 2005; Goodwin-Stewart et al. 2006;
Knechel et al. 2006; Mitra et al. 2007). Company size controls for several
attributes such as risk, earnings persistence, profitability, regulatory costs,
accounting practices and information environment (see e.g. Hay et al. 2006).
Given that larger companies have more complex systems and a wider range of
activities, auditors are prone to devote more audit hours to large companies than
to small companies (Palmrose 1986a; Palmrose 1986b; Barkess et al. 1994). Thus,
TA is expected to have a positive effect on AUDITFEE. Inventory and
receivables (INVREC) are used as proxies for inherent risk. INVREC is
calculated as a sum of total inventories and total receivables scaled by total assets.
Given that inventories and receivables require greater audit effort, INVREC is
expected to be positively associated with AUDITFEE (Carcello et al. 2002; Lee et
al. 2005; Mitra et al. 2007). Complexity is measured as a ratio of foreign sales to
total sales (FOROPR). Companies with foreign operations are expected to require
greater audit effort due to more heterogeneous information and business
complexity and therefore a positive association between FOROPR and
AUDITFEE is expected (Lee et al. 2005; Mitra et al. 2007). Quick ratio (QR)
serves as a measure of leverage. QR is calculated as a ratio of current assets less
inventory to current liabilities. Firms with high leverage are expected to be more
risky and require more audit effort (Antle et al. 2006). Thus, quick ratio (QR) is
expected to have a negative effect on AUDITFEE. Finally, profitability is
measured by the loss variable (LOSS), which is an indicator variable taking the
value of 1 if the net income of the fiscal year is negative and otherwise 0. In
addition to profitability, LOSS is also a measure of audit risk because it reflects
the possibility of the auditor being exposed to loss in the event that a client is not
financially viable (Simunic 1980; Hay et al. 2006). Thus, LOSS is expected to be
positively associated with AUDITFEE (Carcello et al. 2002; Goodwin-Stewart et
al. 2006; Mitra et al. 2007; Vafeas et al. 2007).
In addition to the supply variables, the analysis involves three demand variables
related to board effectiveness, namely board size, board independence and board
diligence. Board size (BSIZE) is measured as the number of directors serving on
74 Acta Wasaensia
the board. Board independence23 (BIND) is measured as the percentage of
independent outsiders on the board. Board diligence (BMEET) is measured as the
number of meetings held during fiscal year. These variables are expected to
increase board effectiveness and therefore increase the demand for audit quality.
Measures of board effectiveness are included in the analysis as control variables
because it is expected that the audit committee does not absolve the whole board
from its responsibilities concerning financial reporting. The board may have
interests in the quality of external audits because audited accounting figures are
used in various decision making-situations. These include management
compensation, reviews of operating issues and investment decisions (Peasnell et
al. 2005). In addition, board members may strive to protect their reputations and
demand greater assurance of financial reporting quality from external auditors.
Therefore the measures of board effectiveness, BSIZE, BIND and BMEET, are
expected to have a positive effect on AUDITFEE.
5.3 Description of the analytic techniques
The relationships hypothesized in the model developed are tested statistically in
two stages. Firstly, the Causal Steps Method is used to test for the occurrence of
the mediated effect. The three regression models required to test the conditions of
the Causal Steps Method will be estimated separately for the three measures of
audit committee effectiveness (ACSIZE, ACEXP and ACMEET). Thus, this part
of the analysis involves the estimation of a total of nine regression models.
Secondly, the significance of the potential mediated effect(s) is tested using the
Sobel Test. This part of the analysis is conditional upon the results of the Causal
Steps Method. That is the Sobel Test statistic is calculated only for models which
fulfill all conditions of the Causal Steps Method. The Sobel Test is used in order
to provide further support regarding the significance of the potential mediated
effect.
The adaptation of the above methods will be explained in more detail in the
following subsections. In addition to the main analysis, several additional
analyses are carried out in order to provide further evidence of the accuracy of the
model and consistency of the results. The methods used in these analyses will be
explained together with the results.
23
The data provided by Institutional Shareholder Services classifies each director on the board as
an 1) insider, 2) affiliated, or 3) independent outsider.
Acta Wasaensia 75
5.3.1 Adaptation of the Causal Steps Method
According to Holmbeck (1997) there are two main statistical approaches to
testing the mediation effect, namely the regression approach and the SEM
approach. Although the SEM approach has some advantages compared to the
regression approach24, its use in the present study would have resulted in a highly
complex structural model whose results could have been difficult to interpret.
This is because the present study’s statistical analyses involve five variables of
main interest as well as two sets of control variables. Because the regression
approach provides a well-established procedure for the treatment of control
variables it is regarded as more appropriate than the SEM approach for the
purposes of this study. Moreover, the regression approach provides the same
information regarding the individual relationships as the SEM approach.
The analysis regarding the occurrence of the mediated effect is conducted
separately for each independent variable serving as a measure for audit committee
effectiveness. More specifically following models are examined:
1) ACSIZEAUDITFEEACC,
2) ACEXPAUDITFEEACC, and
3) ACMEETAUDITFEEACC.
The analysis of the above models involves examination of the four conditions of
the Causal Steps Method discussed in Chapter 5.1.1. If the variables in a
particular model are found to satisfy all four conditions of the Causal Steps
Method, the mediation effect is supported. This analysis is also analogous with
the study’s hypotheses and thus, in such a situation Hypotheses 1-4 are supported.
In addition to the variables of interest, the regression models include control
variables for the dependent variable ACC and the mediator AUDITFEE. The
regression models related to conditions 1 and 3 include four control variables for
ACC: Company size (TA), operating cycle (OPCYCLE), growth opportunities
(SALESG) and profitability (LOSS). TA is expected to decrease earnings
management and therefore have a negative effect on discretionary accruals,
whereas OPCYCLE, SALESG and LOSS are expected to increase earnings
24
For example the SEM approach allows a range of relationships to be included in a single
analysis. In addition it provides indices which provide information as to how well the structural
model fits the data (Kline 1998; Baines et al. 2003; Jokipii 2006).
76 Acta Wasaensia
management, thus resulting in a positive relation. In addition, the regression
model related to condition 2 includes control variables for AUDITFEE.
Following prior research audit fees are expected to be positively associated with
the log of total assets (TA), total inventories and total receivables to total assets
(INVREC), foreign sales to total sales (FOROPR), and whether the company has
incurred a loss during the fiscal year (LOSS). Quick ratio (QR) is expected to
have a negative effect on audit fees. In addition, BSIZE, BIND and BMEET are
expected to be positively related to audit fees. The variables used in the analyses
are summarized in Appendix 3. The regression models employed to test the
conditions of the Causal Steps Method are presented in Tables 2-425 below.
Table 2. Regression models estimated to test model ACSIZEAUDITFEE
ACC.
Condition Equation no
Condition 1:
ACC=β0+β1TA+β2OPCYCLE+β3SALEG+β4LOSS+ β5ACSIZE+ε
(9)
Condition 2:
AUDITFEE=β0+β1TA+β2INVREC+β3FOROPR+β4QR+β5LOSS+β6BSIZE+β7BIND+
β8BMEET+β9ACSIZE+ε
(10)
Conditions 3 and 4:
ACC= β0+β1TA+β2OPCYCLE+β3SALEG+β4LOSS+β5ACSIZE+ β6AUDITFEE +ε
(11)
The variables are defined as follows:
ACC= Discretionary accruals (Residual from the regression model 6 scaled by 100)
ACSIZE= Number of audit committee members
AUDITFEE= Natural logarithm of audit fees
TA= Natural logarithm of total assets
OPCYCLE= Operating cycle (Calculated as presented in Equations 7-8)
SALEG= Growth rate in net sales
LOSS= Indicator variable, 1 if the net income of the fiscal year is negative, otherwise 0
INVREC= Total inventories and total receivables to total assets
FOROPR= Foreign sales to total sales
QR= Quick ratio
BSIZE= Number of board members
BIND= Board independence %
BMEET= Number of board meetings during fiscal year
25
The variables of main interest are marked in bold in the equations.
Acta Wasaensia 77
Table 3. Regression models estimated to test model ACEXPAUDITFEE
ACC.
Condition Equation no
Condition 1:
ACC= β0+β1TA+β2OPCYCLE+β3SALEG+β4LOSS+ β5ACEXP+ε
(12)
Condition 2:
AUDITFEE=β0+β1TA+β2INVREC+β3FOROPR+β4QR+β5LOSS+β6BSIZE+β7BIND+
β8BMEET+β9ACEXP+ε
(13)
Conditions 3 and 4:
ACC= β0+β1TA+β2OPCYCLE+β3SALEG+β4LOSS+β5ACEXP+ β6AUDITFEE+ε
(14)
The variables are defined as follows:
ACC= Discretionary accruals (Residual from the regression model 6 scaled by 100)
ACEXP= Ratio of financial experts on the audit committee
AUDITFEE= Natural logarithm of audit fees
TA= Natural logarithm of total assets
OPCYCLE= Operating cycle (Calculated as presented in Equations 7-8)
SALEG= Growth rate in net sales
LOSS= Indicator variable, 1 if the net income of the fiscal year is negative, otherwise 0
INVREC= Total inventories and total receivables to total assets
FOROPR= Foreign sales to total sales
QR= Quick ratio
BSIZE= Number of board members
BIND= Board independence %
BMEET= Number of board meetings during fiscal year
Table 4. Regression models estimated to test model ACMEETAUDITFEE
ACC.
Condition Equation no
Condition 1:
ACC= β0+β1TA+β2OPCYCLE+β3SALEG+β4LOSS+β5ACMEET+ε
(15)
Condition 2:
AUDITFEE=β0+β1TA+β2INVREC+β3FOROPR+β4QR+β5LOSS+β6BSIZE+β7BIND+
β8BMEET+β9ACMEET+ε
(16)
Conditions 3 and 4:
ACC= β0+β1TA+β2OPCYCLE+β3SALEG+β4LOSS+β5ACMEET+ β6AUDITFEE +ε
(17)
The variables are defined as follows:
ACC= Discretionary accruals (Residual from the regression model 6 scaled by 100)
ACMEET= Number of audit committee meetings during fiscal year
AUDITFEE= Natural logarithm of audit fees
TA= Natural logarithm of total assets
OPCYCLE= Operating cycle (Calculated as presented in Equations 7-8)
SALEG= Growth rate in net sales
LOSS= Indicator variable, 1 if the net income of the fiscal year is negative, otherwise 0
INVREC= Total inventories and total receivables to total assets
FOROPR= Foreign sales to total sales
QR= Quick ratio
BSIZE= Number of board members
BIND= Board independence %
BMEET= Number of board meetings during fiscal year
78 Acta Wasaensia
5.3.2 Adaptation of the Sobel Test
The Sobel Test is employed to complement and further verify the results of the
Causal Steps Method. The Sobel Test statistics are calculated in order to rule out
false-negative and false-positive conclusions, which are possible if only the
Causal Steps Method is used to test the occurrence of the mediated effect (see
Holmbeck et al. 2002). This part of the analysis is conditional upon the results of
the Causal Steps Method. That is, the Sobel Test statistics are calculated only for
those models (i.e. ACSIZEAUDITFEEACC, ACEXPAUDITFEEACC,
or ACMEETAUDITFEEACC) which fulfil the four conditions of the Causal
Steps Method. In other words, if the Causal Steps Method indicates that the
mediated effect occurs, the significance of the mediated effect is examined further
with the Sobel Test. Regression models 9-17 provide the necessary information to
calculate the Sobel Test statistics. The following equations present the calculation
of the Sobel Test statistics for the potential mediation models:
Calculation of Sobel Test statistics for model ACSIZEAUDITFEEACC26
AUDITFEEACSIZE
AUDITFEEACSIZE se
z AUDITFEEACSIZE
69
69
69
ββ
ββ
ββ ⋅
= (18)
( ) ( )2926262969 ACSIZEAUDITFEEAUDITFEEACSIZE sesese AUDITFEEACSIZE ββββββ ⋅+⋅= (19)
The variables are defined as follows:
β6AUDITFEE= Unstandardized regression coefficient for the association between
AUDITFEE and ACC.
β9ACSIZE= Unstandardized regression coefficients for the association between
ACSIZE and AUDITFEE.
seβ6AUDITFEE = Standard error of the mediator β6AUDITFEE.
seβ9ACSIZE= Standard error of the independent variable β9ACSIZE.
26
Refer to regression models 9-11 for the notations.
Acta Wasaensia 79
Calculation of Sobel Test statistics for model ACEXPAUDITFEEACC27
AUDITFEEACEXP
AUDITFEEACEXP se
z AUDITFEEACEXP
69
69
69
ββ
ββ
ββ ⋅
= (20)
( ) ( )2926262969 ACEXPAUDITFEEAUDITFEEACEXP sesese AUDITFEEACEXP ββββββ ⋅+⋅= (21)
The variables are defined as follows:
β6AUDITFEE= Unstandardized regression coefficient for the association between
AUDITFEE and ACC.
β9ACEXP= Unstandardized regression coefficients for the association between
ACEXP and AUDITFEE.
seβ6AUDITFEE = Standard error of the mediator β6AUDITFEE.
seβ9ACEXP= Standard error of the independent variable β9ACEXP.
Calculation of Sobel Test statistics for model ACMEETAUDITFEE
ACC28
AUDITFEEACMEET
AUDITFEEACMEET se
z AUDITFEEACMEET
69
69
69
ββ
ββ
ββ ⋅
= (22)
( ) ( )2926262969 ACMEETAUDITFEEAUDITFEEACMEET sesese AUDITFEEACMEET ββββββ ⋅+⋅=
(23)
The variables are defined as follows:
β6AUDITFEE= Unstandardized regression coefficient for the association between
AUDITFEE and ACC.
27
Refer to regression models 12-14 for the notations.
28
Refer to regression models 15-17 for the notations.
80 Acta Wasaensia
β9ACMEET= Unstandardized regression coefficients for the association between
ACMEET and AUDITFEE.
seβ6AUDITFEE = Standard error of the mediator β6AUDITFEE.
seβ9ACMEET= Standard error of the independent variable β9ACMEET.
5.4 Sample selection and descriptive statistics
The data for this study are obtained from three sources. Firstly, data concerning
audit fees are obtained from the Audit Analytics Database. Secondly, data related
to audit committees are obtained from Institutional Shareholder Services (ISS).
Finally, financial data is obtained from the Thomson Financial Database.
Appendix 4 provides the selection criteria for the sample. The initial sample
consists of S&P 1500 firms which had fiscal years ending any time during the
calendar year 2006. Following prior research financial institutions, (SIC codes
6000-6999) are excluded from the sample due to their special regulatory
environment. Firms in industries with insufficient data to estimate discretionary
accruals (i.e. industries with less than 20 observations) are also excluded. These
sample selection criteria yield a final sample of 1000 firms. All remaining
missing observations in the data are replaced by variable mean in the analyses.
Table 1 of Appendix 5 presents descriptive statistics for variables employed in the
main analysis. Statistics show that ACSIZE varies from a minimum of 2 members
to maximum of 8 members. ACEXP varies from minimum of 0 to a maximum of
1. ACMEET has a wide range in the data distribution with a minimum of 0
meetings and maximum of 31 meetings. The descriptive statistics regarding
measures of audit committee effectiveness indicate that not all sample companies
comply with the SOX’s (2002) or the stock exchanges’ rules related to audit
committee composition and activities29. Firstly, NYSE (2003), AMEX (2003) or
NASDAQ (2003) require that audit committees should have at least three
members. Some sample companies have smaller audit committees than required:
the smallest audit committee in the sample has 2 members. However, the mean
audit committee size (3.74) complies with the requirements. Secondly, the SOX
(2002) and all stock exchanges require that audit committees should have at least
one member who can be considered to have expertise in accounting and related
matters. However, some companies in the sample have not specified whether they
29
An alternative explanation for these outliers is that some companies have not reported matters
regarding audit committee composition and activities sufficiently in their SEC filings.
Acta Wasaensia 81
have an expert in the audit committee with an expertise ratio of 0. However, the
mean (0.49) expertise ratio indicates that many companies have set up audit
committees which exceed the minimum requirements. Thirdly, for example
NYSE (2003) and AMEX (2003) require that audit committees should meet on a
quarterly basis. Some companies in the sample do not meet this requirement,
having no audit committee meetings. Again the mean audit committee meeting
frequency (9.02) exceeds the minimum requirements. In addition, the AUDITFEE
variable has a wide range in the data. The variable ranges from 12.00 to 18.27
which in monetary terms means that audit fees range from a minimum of
$162,750 to maximum of $85,800,000. The dependent variable ACC ranges from
-49.04 to 48.68.
Descriptive statistics regarding the initial variables imply that the data may
contain outliers which could have an effect on the results of the analyses. The
potential effect of outliers is addressed by winsorizing the data at 2.5 % level
from both tails. Table 2 of Appendix 5 presents descriptive statistics for
winsorized variables. After winsorizing the data ACSIZE ranges from a minimum
of 3 to a maximum of 6. ACEXP ranges from 0.17 to 1.00. ACMEET ranges
from a minimum of 4 to a maximum of 17. AUDITFEE ranges from 12.93 to
16.74 which, in monetary terms, equals approximately $376,000 to $18,600,000.
ACC varies from -21.31 to 26.39. After winsorization of the data all audit
committees effectiveness measures in the sample fulfil SOX (2002) and stock
exchanges requirements regarding audit committee composition and activities. In
addition, the range of AUDITFEE and ACC in the data set is reduced
considerably after winsorization. To ensure that the initial results are not driven
by outliers in the data the main analysis is repeated using winsorized data. The
results of this analysis are reported in the additional analysis section of the study.
Appendix 6 presents a correlation matrix for the variables used in the empirical
tests. The correlations show that ACMEET and AUDITFEE are negatively and
significantly correlated with ACC. ACSIZE is also negatively correlated with
ACC but this correlation is not statistically significant. In addition, all measures
of audit committee effectiveness, ACSIZE, ACMEET and ACEXP, are positively
and significantly correlated with AUDITFEE. These correlations provide some
initial support for the hypothesized relationships between measures of audit
committee effectiveness, audit quality and financial reporting quality. The
correlations also reveal that all variables of main interest (ACSIZE, ACEXP,
ACMEET, AUDITFEE and ACC) are significantly related to company size (TA)
ACSIZE, ACEXP, ACMEET, AUDITFEE having a positive relation and ACC
having a negative relation. This indicates that larger companies have more
effective audit committees as well as higher quality external auditors. Larger
82 Acta Wasaensia
companies also seem to have better financial reporting quality. These correlations
highlight the fact that company size is an important control variable which has an
effect on the company’s control environment. In addition to the hypothesized
relationships the review of the correlations shows significant relations between
measures of audit committee effectiveness. More specifically ACEXP is strongly
correlated with both ACSIZE and ACMEET, which indicates that
multicollinearity may have an impact on the results if the three measures of audit
committee effectiveness are introduced simultaneously into a regression model.
This indicates that the separate analysis of the audit committee effectiveness
measures is warranted.
In addition to the correlation matrix, the main relationships of interest are
presented as scatterplots in Appendix 7. The scatterplots are consistent with
findings reported in the correlation matrix. The graphs related to relationships
between measures of audit committee effectiveness (ACSIZE, ACEXP and
ACMEET) and financial reporting quality (ACC) show that ACSIZE and ACEXP
do not have a strong linear relationship with ACC. On the other hand, the graphs
show a negative relation between ACMEET and ACC. The weak relationship
between ACSIZE (ACEXP respectively) and ACC may be due to the fact that
sample companies have fairly homogeneous audit committees in terms of their
composition. This is likely to be a consequence of the strict US regulatory
environment related to corporate governance within the company. Due to the
regulatory environment of the study, audit committee meetings might be the only
measure of audit committee effectiveness with sufficient variance in the data. In
addition according to the graphs AUDITFEE has a negative association with
ACC as hypothesized. Finally, the graphs related to relationships between
ACSIZE (ACEXP and ACMEET respectively) and AUDITFEE show a positive
association between variables. That is, the demand for audit coverage and audit
quality seems to increase as audit committees become more effective.
Appendix 8 presents sample companies grouped by industry. Business services
(SIC 73) is the leading industry with 140 companies in the sample. Electronic and
other electrical equipment and components, except computer equipment (SIC 36)
is the second largest industry represented in the sample with 114 companies.
Printing, publishing and allied industries (SIC 27) and Wholesale trade non-
durable goods (SIC 51) industries are least represented in the sample with 21
companies.
Acta Wasaensia 83
6 RESULTS
This section will introduce the results regarding the model developed using the
methodology introduced in the previous section. This section is organized as
follows. Firstly, the main results related to models ACSIZEAUDITFEE
ACC, ACEXPAUDITFEEACC and ACMEETAUDITFEEACC are
presented. Analysis of a single model involves estimation of three regression
models required by the Causal Steps Method to examine the occurrence of the
mediated effect.30 The results of the regression models are summarized in Figures
12-14 and presented in more detail in Appendices 8-10. In addition, the results
reported in the Appendices provide the Sobel Test statistics for models which
fulfill the conditions of the Causal Steps Method. Secondly, the implications of
the main results are discussed. Finally, additional analyses are conducted in order
to test the robustness of the main results.
6.1 Results related to model ACSIZEAUDITFEE
ACC
Model ACSIZE-AUDITFEE-ACC is based on the underlying notion that audit
committee effectiveness increases as the size of the audit committee increases. In
addition, audit quality is expected to increase as audit fees increase. Thus the
hypothesized relationships for the model are as follows: 1) ACSIZE has a
negative effect on ACC, 2) ACSIZE has a positive effect on AUDITFEE, 3)
AUDITFEE has a negative effect on ACC and 4) AUDITFEE mediates the
relationship between ACSIZE and ACC. The regression models estimated to test
the above relationships are presented in Table 2.
Figure 12 summarizes and Appendix 9 provides detailed results for the model
ACSIZEAUDITFEEACC. With regard to the first condition ACSIZE is
regressed on ACC along with control variables. Contrary to the hypothesis,
ACSIZE has a significant positive effect on ACC. In this regression control
variables TA, OPCYCLE, SALESG and LOSS are all statistically significant and
TA and OPCYCLE have expected signs. To test for the second condition
ACSIZE is regressed on AUDITFEE with a set of control variables. The results
show that the effect of ACSIZE on AUDITFEE is not statistically significant. All
30
As mentioned previously, the conditions of the Causal Steps Method are analogous with the
hypotheses of the study. Thus, in the subsequent sections the term condition refers to both these
hypotheses and the conditions of the Causal Steps Method.
84 Acta Wasaensia
control variables TA, INVREC, FOROPR, QR, LOSS, BSIZE, BIND and
BMEET are statistically significant with expected signs. To test for the third and
fourth conditions ACSIZE, AUDITFEE and a set of control variables are
regressed on ACC. The results reveal that AUDITFEE has a significant negative
effect on ACC as hypothesized. ACSIZE has a significant positive effect on ACC
which is inconsistent with the hypothesis.
(-) 1,039***
(+) -,031
(-)-1,329**
(-) 1,040***
AUDITFEE
ACSIZE
ACC ACSIZE
ACC
Figure 12. Summarized results for model ACSIZEAUDITFEEACC.
To conclude the model ACSIZEAUDITFEEACC does not fulfill conditions
of the Causal Steps Method. This is because audit committee size does not have
the hypothesized effect on either discretionary accruals or audit fees. On the
contrary, the results show a positive association between audit committee size and
discretionary accruals. This indicates that audit committees become less effective
in ensuring financial reporting quality as their size increases. This result may be
due to multicollinearity among the explanatory variables in the regression model.
This is evidenced when audit committee size is regressed on discretionary
accruals excluding control variables. In this case the effect is negative as
hypothesized but statistically insignificant. Alternatively audit committee size
may have a nonlinear effect on audit committee effectiveness. That is, audit
committee effectiveness increases along with its size to a certain extent but starts
Notes:
***, **, * denote significance at 1%, 5% and 10% level
respectively
Expected signs of the relationships are presented in
parentheses
Reported numbers are coefficients
The variables are defined as follows:
ACSIZE= Number of audit committee members
AUDITFEE= Natural logarithm of audit fees
ACC= Discretionary accruals (Residual from the regression
model 6 scaled by 100)
Acta Wasaensia 85
to decrease if the audit committee becomes too large. In addition audit committee
size is found to have a positive but not statistically significant effect on audit fees.
With regard to the third condition, audit fees are found to have a negative effect
on discretionary accruals. To conclude, since only the third condition of the
Causal Steps Method was verified by model ACSIZEAUDITFEEACC the
mediated effect from audit committee size to discretionary accruals through audit
fees is not supported.
6.2 Results related to model ACEXPAUDITFEE
ACC
The underlying notion for model ACEXPAUDITFEEACC is that audit
committee effectiveness increases as the proportion of expert members in the
audit committee increases. Audit quality is also expected to increase as audit fees
increase. Therefore the following relationships are proposed: 1) ACEXP has a
negative effect on ACC, 2) ACEXP has a positive effect on AUDITFEE, 3)
AUDITFEE has a negative effect on ACC and 4) AUDITFEE mediates the
relationship between ACEXP and ACC. Table 3 presents the regression models
estimated to test the above relationships.
The results for model ACEXPAUDITFEEACC are summarized in Figure 13
and presented comprehensively in Appendix 10. To test for the first condition
ACEXP as well as a set of control variables are regressed on ACC. It is found that
ACEXP has a positive effect on ACC. However, this effect is not statistically
significant. Thus, the first condition is not supported. Control variables TA,
OPCYCLE, SALESG and LOSS are significant and TA and OPCYCLE have
expected signs. In order to test for the second condition ACEXP is regressed on
AUDITFEE along with a set of control variables. In this regression specification
ACEXP is found to have a positive but insignificant effect on AUDITFEE. Thus,
the results fail to support the second condition. The control variables in this
regression are all significant and they have predicted signs. Finally, to test the
third and fourth conditions AUDITFEE and ACEXP are regressed on ACC with a
set of control variables. The results show that AUDITFEE has a negative effect
on ACC as expected. Thus condition three is fulfilled. In this regression
specification ACEXP has a negative and insignificant effect on ACC. Control
variables TA, OPCYCLE, SALESG and LOSS are all significant and TA as well
as OPCYCLE have expected signs. However, the fourth condition cannot be
assessed because ACEXP did not initially have a statistically significant effect on
ACC.
86 Acta Wasaensia
(-) ,793
(+) ,041
(-) -1,337**
(-) ,860
AUDITFEE
ACEXP
ACC ACEXP
ACC
Figure 13. Summarized results for model ACEXPAUDITFEEACC.
In summary, model ACEXPAUDITFEEACC does not fulfill all the
conditions of the Causal Steps Method. This is because the audit committee
expertise ratio does not have a statistically significant effect on either
discretionary accruals or audit fees, albeit the directions of these effects are as
hypothesized. Thus, conditions 1 and 2 are not fulfilled. With regard to the third
condition audit fees are found to have a negative effect on discretionary accruals
and therefore condition 3 is fulfilled. To conclude, since the model
AXECPAUDITFEEACC fulfills only condition 3, the mediated effect of
audit committee expertise ratio through audit fees on discretionary accruals is not
supported by the results.
6.3 Results related to model ACMEET
AUDITFEEACC
Model ACMEETAUDITFEEACC is based on the assumption that audit
committee effectiveness increases along with audit committee meeting frequency.
Moreover, audit quality is expected to increase along with audit fees. Thus, the
Notes:
***, **, * denote significance at 1%, 5% and 10% level
respectively
Expected signs of the relationships are presented in
parentheses
Reported numbers are coefficients
The variables are defined as follows:
ACEXP= Ratio of financial experts on the audit committee
AUDITFEE= Natural logarithm of audit fees
ACC= Discretionary accruals (Residual from the regression
model 6 scaled by 100)
Acta Wasaensia 87
following relationships are proposed: 1) ACMEET has a negative effect on ACC,
2) ACMEET has a positive effect on AUDITFEE, 3) AUDITFEE has a negative
effect on ACC and 4) AUDITFEE mediates the relationship between ACMEET
and ACC. The regression models estimated to test the above relationships are
shown in Table 4.
Figure 14 summarizes and Appendix 11 presents the detailed results related to
model ACMEETAUDITFEEACC. In order to test the first condition
ACMEET is regressed on ACC with a set of control variables. It is found that
ACMEET has a significant negative effect on ACC, thus the first condition is
fulfilled. In this regression all control variables are statistically significant and TA
and OPCYCLE have the expected signs. To test for the second condition
ACMEET and a set of control variables are regressed on AUDITFEE. The results
show that ACMEET has a significant positive effect on AUDITFEE. Therefore
the second condition is fulfilled. In this regression all control variables are
statistically significant and they have expected signs. In order to test the third and
fourth conditions both ACMEET and AUDITFEE as well as control variables are
regressed on ACC. It is found that AUDITFEE has a significant negative effect
on ACC and thus the third condition is fulfilled. The fourth condition is addressed
by assessing whether AUDITFEE decreases the effect of ACMEET on ACC. It is
found that the effect of ACMEET on ACC is smaller in the third regression than
the first one (-.246**<-.279***). Thus, condition 4 is also met. The control
variables are found to be statistically significant and TA as well as OPCYCLE
have the expected signs. In addition to the results of Causal Steps Method the
Sobel Test statistics (-1.784*) indicates that the mediated effect is statistically
significant at the 10% level.
88 Acta Wasaensia
(-) -,279***
(+) ,026***
(-) -1,073*
(-) -,246**
AUDITFEE
ACMEET
ACC ACMEET
ACC
Figure 14. Summarized results for model ACMEETAUDITFEEACC.
As a summary, the results indicate that model ACMEETAUDITFEEACC
fulfils all four conditions of the Causal Steps Method. That is, audit committee
meeting frequency is found to have a negative effect on discretionary accruals and
a positive effect on audit fees. In addition, audit fees are found to have a negative
effect on discretionary accruals. Finally, the results show that audit fees decrease
the effect of audit committee meeting frequency on discretionary accruals. Thus,
the results appear to support the hypothesis that audit fees mediate the
relationship between audit committee meeting frequency and discretionary
accruals. The mediated effect is only partial since audit fees do not reduce the
effect of audit committee meeting frequency on discretionary accruals to non-
significance. This result may indicate that the effect of audit committee meeting
frequency on discretionary accruals may also be mediated by other control
mechanisms, such as internal auditing, which are currently out of the scope of the
model.
Notes:
***, **, * denote significance at 1%, 5% and 10% level
respectively
Expected signs of the relationships are presented in
parentheses
Reported numbers are coefficients
The variables are defined as follows:
ACMEET= Number of audit committee meetings during
fiscal year
AUDITFEE= Natural logarithm of audit fees
ACC=Discretionary accruals (Residual from the regression
model 6 scaled by 100)
Acta Wasaensia 89
6.4 Discussion of the main results
The main analysis involves examination of three separate models, namely
ACSIZEAUDITFEEACC, ACEXPAUDITFEEACC and ACMEET
AUDITFEEACC. The underlying notion in these models is that more effective
audit committees will increase financial reporting quality and also demand better
audit quality. In addition, audit quality is expected to contribute to financial
reporting quality. When these effects are combined, audit quality is modeled as a
mediator in the relationship between audit committee effectiveness and financial
reporting quality. The main analytic technique employed to test the models is the
Causal Steps Method, which is accompanied by the Sobel Test when applicable.
Table 5 provides a summary of the main results.
The results regarding models ACSIZEAUDITFEEACC, and
ACEXPAUDITFEEACC do not meet the conditions of the Causal Steps
Method and therefore the mediation hypothesis is not supported. This is because
neither audit committee size nor audit committee expertise ratio has the
hypothesized effect on discretionary accruals or audit fees. These results imply
that the audit committee composition depicted by audit committee size or
expertise ratio may not be sufficient indicators as to how effective audit
committees will be in discharging their responsibilities. This may be due to the
fact that audit committee composition is strictly regulated in the USA and there is
not enough variation in these variables. Due to these results further discussion
will be focused on model ACMEETAUDITFEEACC.
The results for the model ACMEETAUDITFEEACC indicate that audit fees
mediate the relationship between audit committee meeting frequency and
discretionary accruals as hypothesized. Both the findings of the Causal Steps
Method and Sobel Test support this result. In relation to the Causal Steps Method
the following significant relationships are found: 1) a negative relationship
between audit committee meeting frequency and discretionary accruals, 2) a
positive relationship between audit committee meeting frequency and audit fees,
and 3) a negative relationship between audit fees and discretionary accruals. In
addition, the results show that audit fees reduce the effect of audit committee
meeting frequency on discretionary accruals, which is consistent with the
mediation hypothesis. However, since the effect of audit committee meeting
frequency on discretionary accruals is not reduced to non-significance, only
partial mediation is supported. Furthermore, the Sobel Test statistic (-1.784*)
indicates that the mediated effect of audit committee meetings on discretionary
accruals through audit fees is statistically significant at the 10% level. Thus these
90 Acta Wasaensia
results in total are indicative of a partial mediated effect.
The results regarding model ACMEETAUDITFEEACC have several
implications. In general the results for the model are consistent with the
predictions of agency theory. This is because both audit committee meeting
frequency and audit fees are found to have a negative effect on discretionary
accruals, which suggests that both audit committee effectiveness and external
audit quality contribute to a company’s financial reporting quality. The results are
also consistent with the notion that audit committees and external auditors
complement each other. That is, audit committee meetings are found to increase
audit fees, which are further found to decrease discretionary accruals. In addition,
the individual relationships constituting the model have the following
implications.
Firstly, the negative relationship between audit committee meeting frequency and
discretionary accruals suggests that more active audit committees are better able
to restrict management discretion over accounting issues. This may be because
more active audit committees are better informed about the state of affairs in the
company and thus are better able to monitor its accounting practices. For
example, regular meetings between audit committees and external auditors or
audit committees and internal auditors may enhance information flow between
these parties and therefore result in better decision making by audit committees.
Active audit committees may also be staffed by individuals who are motivated to
devote their time to the functioning of the audit committee and such individuals
may exercise closer monitoring over management thereby restricting earnings
management more effectively.
Secondly, the positive relationship between audit committee meeting frequency
and audit fees has several plausible explanations. The positive relationship
between audit committee meetings and audit fees may exist because preparation
and attendance at meetings requires additional work by auditors, causing higher
audit fees (see Stewart et al. 2007). On the other hand more active audit
committees may be more concerned about audit quality and therefore demand
greater quantity of audit effort, which is again reflected in an increase in audit
fees. Since these explanations are not mutually exclusive both of them might
apply to the present findings.
Thirdly, a negative effect between audit fees and discretionary accruals is found.
This result implies that greater audit effort reflected in higher audit fees leads to
closer monitoring of a company’s accounting issues, and thus decreases
management discretion over accounting choices. These results may also be
Acta Wasaensia 91
related to auditor independence discussion. These results imply that higher audit
fees do not compromise auditor independence, which would result in less rigorous
monitoring by auditors. On the contrary, the results suggest that higher audit fees
may indicate that sufficient audit effort and audit hours have been allocated to the
client. The results moreover suggest that protection of reputation may lead to
auditor reporting conservatism: auditors who are paid more allow their clients less
discretion over discretionary accruals than auditors who are paid less.
Finally, audit fees are found to function as a mediator in the relationship between
audit committee meeting frequency and discretionary accruals. The mediation
model consists of the relationships discussed above. The mediation model
maintains that although audit committee meetings contribute to financial
reporting in their own right, part of this effect goes through external auditing. The
model is consistent with prior literature stating that audit committees are
responsible for both ensuring a company’s financial reporting quality and
ensuring external audit quality. Thus, some of the monitoring activities are bound
to be transmitted from audit committees to external auditors. Since the results do
not imply full mediation, it is likely that other monitoring mechanisms of the
company, such as internal auditing and internal controls, may also function as
mediators between audit committee effectiveness and financial reporting quality.
These control mechanisms, however, are beyond the scope of the present study.
92 Acta Wasaensia
Table 5. Summary of the main results.
So
be
l T
es
t
n
o
t
ca
lc
u
la
te
d
n
o
t c
al
cu
la
te
d
-
1.
78
4*
C
o
n
di
tio
n
4
n
o
t a
ss
es
se
d
n
o
t a
ss
es
se
d
│
.
24
6│
<
│
.
27
9│
Co
ef
fic
ie
n
t
-
1.
32
9*
*
-
1.
33
7*
*
-
1.
07
3*
C
o
n
di
tio
n
3
Pr
ed
ic
te
d
sig
n
-
-
-
Co
ef
fic
ie
n
t
-
.
03
1
.
04
1
.
02
6*
*
*
C
o
n
di
tio
n
2
Pr
ed
ic
te
d
sig
n
+
+
+
Co
ef
fic
ie
n
t
1.
03
9*
*
*
.
79
3
-
.
27
9*
*
*
C
o
n
di
tio
n
1
Pr
ed
ic
te
d
sig
n
-
-
-
M
o
de
l
A
CS
IZ
E
A
U
D
IT
FE
E
A
CC
A
CE
X
P
A
U
D
IT
FE
E
A
CC
A
CM
EE
T
A
U
D
IT
FE
E
A
CC
N
o
te
s:
*
*
*
,
*
*
,
*
de
n
o
te
sig
n
ifi
ca
n
ce
at
1%
,
5%
an
d
10
%
le
v
el
re
sp
ec
tiv
el
y
Th
e
va
ri
a
bl
es
a
re
de
fin
ed
a
s
fol
lo
w
s:
A
CS
IZ
E=
N
u
m
be
r
o
f a
u
di
t c
o
m
m
itt
ee
m
em
be
rs
A
CE
X
P=
R
at
io
o
f f
in
an
ci
al
ex
pe
rt
s
o
n
th
e
au
di
t c
o
m
m
itt
ee
A
CM
EE
T=
N
u
m
be
r
o
f a
u
di
t c
o
m
m
itt
ee
m
ee
tin
gs
du
rin
g
fis
ca
l y
ea
r
A
U
D
IT
FE
E=
N
at
u
ra
l l
o
ga
rit
hm
o
f a
u
di
t f
ee
s
A
CC
=
D
isc
re
tio
n
ar
y
ac
cr
u
al
s
(R
es
id
u
al
fro
m
th
e
re
gr
es
sio
n
m
o
de
l 6
sc
al
ed
by
10
0)
Acta Wasaensia 93
6.5 Robustness of the main results
This section aims at testing the robustness of the results provided by the main
analysis. Since the main results only supported model ACMEETAUDITFEE
ACC the additional analyses are mainly focused on this model. The additional
analyses are organized as follows. Firstly, the main analysis regarding model
ACMEETAUDITFEEACC is repeated using winsorized data. Secondly, the
robustness of the main results is tested by employing two additional model
specifications. Thirdly, an additional model specification in which the mediating
variable AUDITFEE is replaced by industry adjusted audit fees (INDFEE) is
tested. Thus the following model ACMEETINDFEEACC is examined.
Fourthly, a model specification including a composite measure for audit
committee effectiveness is tested using path analysis employing AMOS statistical
package. Fifthly, the effect of unexpected fees on discretionary accruals is tested.
Finally, a competing effect type, moderation effect, is examined.
6.5.1 Results for model ACMEETAUDITFEEACC using winsorized data
This part of the analysis examines whether the main results related to model
ACMEETAUDITFEEACC are affected by outliers in the initial data. This is
achieved by repeating the analysis related to the Causal Steps Method using
winsorized data31. The descriptive statistics for the winsorized variables were
presented in Section 5.4. The regression models employed are consistent with
models 15-17. In addition, Sobel Test statistics are calculated as presented in
Equations 22-23.
The results for the winsorized data are summarized in Figure 15 and presented in
full in Appendix 12. The findings of the analysis are consistent with the results
reported in the main analysis. Thus the results reveal that ACMEET has a
significant negative effect on ACC and thus the first condition is fulfilled. The
results also show that ACMEET has a significant positive effect on AUDITFEE
thereby fulfilling the second condition. AUDITFEE is also found to have a
significant negative effect on ACC, which meets the requirements of the third
condition. In addition, it is found that AUDITFEE reduces the effect of ACMEET
on ACC. This result fulfils the fourth condition. These results therefore imply that
AUDITFEE mediates the relationship between ACMEET and ACC. This result is
further supported by the Sobel Test statistic (-2.01**) which shows that the
mediated effect is significant at the 5% level. Based on these findings it can be
31
Data is winsorized by 2.5% from both tails.
94 Acta Wasaensia
concluded that the results of the main analysis are not driven by outliers in the
initial data.
(-) -,282***
(+) ,027***
(-) -1,135**
(-) -,244**
AUDITFEE
ACMEET
ACC ACMEET
ACC
Figure 15. Summarized results for model ACMEETAUDITFEEACC using
winsorized data.
6.5.2 Results for additional model specifications
The objective of this part of the analysis is to test the robustness of the main
results by employing two additional model specifications. In the first model
specification the conditions of the Causal Steps Method are tested simultaneously
for all three measures of audit committee effectiveness including control
variables. This regression specification reveals whether the main results can be
achieved by including all measures of audit committee effectiveness
simultaneously into the analysis or whether multicollinearity between the
measures has an impact on the results. The regression models required for the
analysis are presented in Table 6.
Notes:
***, **, * denote significance at 1%, 5% and 10% level
respectively
Expected signs of the relationships are presented in
parentheses
Reported numbers are coefficients
The variables are defined as follows:
ACMEET= Number of audit committee meetings during
fiscal year
AUDITFEE= Natural logarithm of audit fees
ACC= Discretionary accruals (Residual from the regression
model 6 scaled by 100)
Acta Wasaensia 95
Table 6. Regression models estimated to test measures of audit committee
effectiveness simultaneously.
Condition Equation no
Condition 1:
ACC= β0+β1TA+β2OPCYCLE+β3SALEG+β4LOSS+β5ACSIZE+β6ACEXP +
β7ACMEET+ε
(24)
Condition 2:
AUDITFEE=β0+β1TA+β2INVREC+β3FOROPR+β4QR+β5LOSS+β6BSIZE+β7BIND+
β8BMEET+β9ACSIZE+ β10ACEXP + β11ACMEET +ε
(25)
Conditions 3 and 4:
ACC= β0+β1TA+β2OPCYCLE+β3SALEG+β4LOSS+β5ACSIZE+ β6ACEXP + β7ACMEET+
β8AUDITFEE +ε
(26)
The variables are defined as follows:
ACC= Discretionary accruals (Residual from the regression model 6 scaled by 100)
ACSIZE= Number of audit committee members
ACEXP= Ratio of financial experts on the audit committee
ACMEET= Number of audit committee meetings during fiscal year
AUDITFEE= Natural logarithm of audit fees
TA= Natural logarithm of total assets
OPCYCLE= Operating cycle (Calculated as presented in Equations 7-8)
SALEG= Growth rate in net sales
LOSS= Indicator variable, 1 if the net income of the fiscal year is negative, otherwise 0
INVREC= Total inventories and total receivables to total assets
FOROPR= Foreign sales to total sales
QR= Quick ratio
BSIZE= Number of board members
BIND= Board independence %
BMEET= Number of board meetings during fiscal year
The findings of the analysis are presented in Appendix 13. The results of these
regression model specifications are consistent with the results obtained from the
main analysis. It is found that the first condition is fulfilled by ACMEET, which
has a significant negative effect on ACC. Consistent with the results of the main
analysis, ACSIZE and ACEXP do not have the hypothesized effect on ACC and
therefore the first condition is not met by these variables. In this regression all
control variables are significant and TA as well as OPCYCLE have expected
signs. The results also reveal that the second condition is met by ACMEET but
not by ACSIZE and ACEXP. That is ACMEET has a positive effect on
AUDITFEE whereas ACSIZE and ACEXP are not statistically significantly
related to AUDITFEE. In this regression all control variables except BMEET are
significant and the direction of their effect is as hypothesized. The results further
show that the third condition is fulfilled since AUDITFEE has a significant
negative effect on ACC. Due to the above results the fourth condition can be
addressed only with regard to ACMEET. It is found that adding AUDITFEE to
the model decreases the coefficient of ACMEET slightly, thus supporting the
mediation hypothesis. The Sobel Test statistic (-1.818*) provides further evidence
96 Acta Wasaensia
that the mediated effect from ACMEET to ACC through AUDITFEE is also
statistically significant in these regression specifications. Based on the results it
can be concluded that multicollinearity between measures of audit committee
effectiveness does not have an effect on the results.
In the second model specification the conditions of the Causal Steps Method are
tested simultaneously for all three measures of audit committee effectiveness
excluding control variables. This regression specification is tested in order
examine whether the results are affected by multicollinearity between measures of
audit committee effectiveness and control variables. The estimated regression
models are presented in Table 7.
Table 7. Regression models estimated to test measures of audit committee
effectiveness excluding control variables.
Condition Equation no
Condition 1:
ACC=β0 +β1ACSIZE+β2ACEXP+β3ACMEET+ε
(27)
Condition 2:
AUDITFEE=β0+β1ACSIZE+β2ACEXP+β3ACMEET+ε
(28)
Conditions 3 and 4:
ACC=β0+β1ACSIZE+β2ACEXP+β3ACMEET+β4AUDITFEE+ε
(29)
The variables are defined as follows:
ACC= Discretionary accruals (Residual from the regression model 6 scaled by 100)
ACSIZE= Number of audit committee members
ACEXP= Ratio of financial experts on the audit committee
ACMEET= Number of audit committee meetings during fiscal year
AUDITFEE= Natural logarithm of audit fees
The findings of the above models are presented in Appendix 14. The results of
these model specifications are largely consistent with the results provided by the
previous analyses. Testing of the first condition reveals that ACMEET has a
statistically significant negative effect on ACC. ACSIZE and ACEXP are also
negatively related to ACC, albeit these relations are not statistically significant.
Thus, only ACMEET fulfils the first condition. Testing for the second condition
shows that all ACSIZE, ACEXP and ACMEET variables have a statistically
significant and positive effect on AUDITFEE. Therefore all measures of audit
committee effectiveness fulfil the second condition in this model specification.
The third condition is also met since AUDITFEE is found to have a significant
negative effect on ACC. Due to the results for conditions 1-3 only ACMEET can
be taken into consideration when the fourth condition is addressed. The results
reveal that including AUDITFEE in the model reduces the effect of ACMEET on
Acta Wasaensia 97
ACC. Thus, the mediated effect of ACMEET on ACC through AUDITFEE is
supported. The Sobel Test statistic (-4.694***) also indicates that AUDITFEE
significantly mediates the relationship between ACMEET and ACC. These results
imply that multicollinearity between measures of audit committee effectiveness,
particularly ACSIZE and ACEXP, and control variables may have had some
influence on the main results. This is because the second condition is met by these
variables when control variables are excluded from the regression models.
6.5.3 Results for industry adjusted audit fees
This part of the analyses involves an alternative mediating variable, namely
industry adjusted audit fees (INDFEE). Thus the model ACMEETINDFEE
ACC is examined. The mediating variable INDFEE is calculated as a ratio of a
company’s audit fees relative to the mean audit fees of the industry. Mean audit
fees for the industry are calculated based on two digit SIC codes. This analysis is
conducted in order to test whether companies with more frequent audit committee
meetings demand better audit quality than other companies in the same industry
on average and thus pay higher industry adjusted audit fees. In addition, this
analysis shows whether higher industry adjusted audit fees result in better
financial reporting quality. The regression models estimated to test model
ACMEETINDFEEACC are presented in Table 8.
The results for the model are reported in Appendix 15. The results are consistent
with the main analysis. Firstly, the results show that ACMEET has a negative
effect on ACC, which fulfils the first condition. Secondly, the results reveal that
ACMEET has a positive effect on INDFEE, which is consistent with the second
condition. Thirdly, INDFEE is found to have a negative effect on ACC, which
satisfies the third condition. Finally, INDFEE is found to mediate the relationship
between ACMEET and ACC. That is, the effect of ACMEET on ACC is reduced
after INDFEE is included in the regression. The Sobel Test statistic (-1.877*) also
indicates that the mediated effect is significant at the 10% level. These results
imply that companies with active audit committees demand better audit quality
than other companies in the same industry and therefore pay relatively higher
audit fees. The results also show that industry adjusted audit fees may reflect
audit quality: auditors who are paid more than other auditors in the same industry
on average seem to decrease management discretion over accruals more
effectively than auditors who are paid less.
98 Acta Wasaensia
Table 8. Regression models estimated to test model ACMEETINDFEE
ACC.
Condition Equation no
Condition 1:
ACC= β0+β1TA+β2OPCYCLE+β3SALEG+β4LOSS+β5ACMEET+ε
(30)
Condition 2:
INDFEE=β0+β1TA+β2INVREC+β3FOROPR+β4QR+β5LOSS+β6BSIZE+β7BIND+
β8BMEET+β9ACMEET +ε
(31)
Conditions 3 and 4:
ACC= β0+β1TA+β2OPCYCLE+β3SALEG+β4LOSS+ β5ACMEET + β6INDFEE +ε
(32)
The variables are defined as follows:
ACC= Discretionary accruals (Residual from the regression model 6 scaled by 100)
ACMEET= Number of audit committee meetings during fiscal year
INDFEE = Ratio of audit fees relative to the industry mean
TA= Natural logarithm of total assets
OPCYCLE= Operating cycle (Calculated as presented in Equations 7-8)
SALEG= Growth rate in net sales
LOSS= Indicator variable, 1 if the net income of the fiscal year is negative, otherwise 0
INVREC= Total inventories and total receivables to total assets
FOROPR= Foreign sales to total sales
QR= Quick ratio
BSIZE= Number of board members
BIND= Board independence %
BMEET= Number of board meetings during fiscal year
6.5.4 Results for path analysis
This analysis tests a variation of the model developed using path analysis. This
analysis is conducted using AMOS statistical package (see Holmbeck 2002).
There are several advantages in this approach. Firstly, path analysis provides an
opportunity to test multiple relationships simultaneously. Secondly, this approach
provides measures for the overall goodness-of-fit statistics of the proposed model.
This study uses the Chi-square test (χ2) to assess the goodness-of-fit of the model.
Chi-square test statistics are interpreted as follows: a non-significant (P>0.05)
value indicates that the model fits the data satisfactorily. Finally, path analysis
provides information regarding the significance of the relationship between the
variables (Kline 1998; Baines & Langfield-Smith 2003).
The path model is formed as follows. In the path model, a composite measure of
audit committee effectiveness (ACSUM) is modeled to have a path leading to
AUDITFEE, which is further modeled to have a path leading to ACC. ACSUM is
formed by calculating standardized values of ACSIZE, ACEXP and ACMEET.
Calculated standardized values are then summed to form a variable ACSUM. It
should be noted that this part of the analysis relaxes the first condition of the
Acta Wasaensia 99
Causal Steps Method. That is the direct effect of audit committee effectiveness on
financial reporting quality is not tested. Therefore, the path analysis tests whether
ACSUM has an indirect effect on ACC through AUDITFEE rather than testing
for the mediated effect. This interpretation is consistent with Holmbeck’s (1997)
definition of mediating and intervening effects.
The model is presented in Appendix 16. The goodness-of-fit statistics [χ2(1)=
.204 (P=.652)] indicate that the model fits the data satisfactorily. In addition it is
found that ACSUM has a positive and significant effect on AUDITFEE, which is
further found to have a significant negative effect on ACC. The results of this
model also indicate that audit quality may function as an intervening variable in
the relationship between audit committee effectiveness and financial reporting
quality. This issue should be examined more thoroughly in future studies.
6.5.5 Results for unexpected fees
This analysis employs an optional approach to test the developed model by
examining unexpected audit fees. Prior audit studies have suggested that audit
quality may be influenced by the amount of audit fees relative to their expected
amount rather than their realized amounts (Craswell et al. 1995; Gul et al. 1998;
Tsui, Jaggi & Gul 2001; Srinidhi et al. 2007). The underlying notion behind this
argument is that audit quality is associated with unexpectedly high or low audit
fees: unexpectedly high audit fees indicating a more thorough audit performance
and thus superior audit quality.
Examining unexpected audit fees involves the following two-step procedure.
Firstly, two alternative regression specifications are used to estimate unexpected
audit fees (UNEXPFEE_1 and UNEXPFEE_2). These regression specifications
include measures of audit committee effectiveness as well as other variables
found in earlier studies to have an effect on audit fees. The difference between the
models is as follows. The first model includes all three measures of audit
committee effectiveness as well as a set of control variables found by prior
studies to have an effect on audit fees whereas the second model includes only a
measure of audit committee meeting frequency and a set of control variables. The
latter model is estimated because the main analysis indicated that audit committee
meeting frequency is the most important audit committee related explanatory
variable for audit fees. More specifically, unexpected audit fees are estimated as
residuals from the following alternative regression models:
100 Acta Wasaensia
AUDITFEE=β0+β1TA+β2INVREC+β3FOROPR+β4QR+β5LOSS+β6BSIZE+
β7BIND+β8BMEET+β9ACSIZE+β10ACEXP+β11ACMEET+ε (33)
AUDITFEE=β0+β1TA+β2INVREC+β3FOROPR+β4QR+β5LOSS+β6BSIZE+
β7BIND+β8BMEET+β9ACMEET+ε (34)
The variables are defined as follows:
AUDITFEE= Natural logarithm of audit fees
TA= Natural logarithm of total assets
INVREC= Total inventories and total receivables to total assets
FOROPR= Foreign sales to total sales
QR= Quick ratio
LOSS= Indicator variable, 1 if the net income of the fiscal year is negative,
otherwise 0
BSIZE= Number of board members
BIND= Board independence %
BMEET= Number of board meetings during fiscal year
ACSIZE= Number of audit committee members
ACEXP= Ratio of financial experts on the audit committee
ACMEET= Number of audit committee meetings during fiscal year
Secondly, in order to test whether UNEXPFEE_1 or UNEXPFEE_2 are related to
accrual quality they are regressed on ACC. If higher unexpected audit fees
indicate higher engagement effort and thus higher audit quality, a negative
association between UNEXPFEE_1 or UNEXPFEE_2 and ACC should be
observed. The results of the analysis are presented in Appendix 17. The results
show that both UNEXPFEE_1 and UNEXPFEE_2 have a significant negative
Acta Wasaensia 101
effect on ACC. The results provide further support for the model proposed. This
is because the results indicate that measures of audit committee effectiveness can
be used to predict unexpected audit fees. In addition, unexpected audit fees are
found to function as determinants of accrual quality. That is higher level of
unexpected audit fees seems to result in better accrual quality and more generally
better financial reporting quality.
6.5.6 Results for the moderation effect
As discussed in Chapter 4, one of the core issues of this study is the selection of
an effect type which sufficiently describes the relationships between variables of
interest. After considering the theory related to the variables as well as effect
types it was concluded that audit quality is likely to function as a mediator in the
relationship between audit committee effectiveness and financial reporting
quality. The main results provide support for this effect type for model
ACMEETAUDITFEEACC. Although the mediation hypothesis was derived
from the literature the possibility of the alternative effect type, moderation, cannot
be entirely ruled out without testing it empirically.
This part of the analysis therefore tests whether AUDITFEE moderates the
relationship between ACMEET and ACC. Following Holmbeck (1997) the
moderation effect is tested using multiple regression analysis by entering the main
effects of AUDIFEE and ACMEET into the ACC model along with control
variables first, following the interaction term AUDITFEE*ACMEET of
ACMEET and AUDITFEE. A significant effect of the interaction term means that
the moderated effect is supported. Since mediation and moderation effects cannot
appear simultaneously (see Section 4.4.) it is expected that the interaction term
will not be statistically significant. Thus the following regression is employed to
test the potential moderation effect of AUDITFEE on the relationship between
ACMEET and ACC:
ACC= β0+β1TA+β2OPCYCLE+β3SALEG+β4LOSS+β5ACMEET+β6AUDITFEE+
β7AUDITFEE*ACMEET+ε (35)
The variables are defined as follows:
ACC= Discretionary accruals (Residual from the regression model 6 scaled by
100)
102 Acta Wasaensia
TA= Natural logarithm of total assets
OPCYCLE= Operating cycle (Calculated as presented in Equations 7-8)
SALEG= Growth rate in net sales
LOSS= Indicator variable, 1 if the net income of the fiscal year is negative,
otherwise 0
ACMEET= Number of audit committee meetings during fiscal year
AUDITFEE= Natural logarithm of audit fees
AUDITFEE*ACMEET= Interaction term of ACMEET and AUDITFEE
The results of the above analysis are presented in Appendix 18. The results show
that the interaction term AUDITFEE*ACMEET does not have a statistically
significant effect on ACC, which is consistent with the expectations. In this
regression specification neither ACMEET nor AUDITFEE has a statistically
significant effect on ACC. This result indicates that moderated effect type does
not describe the relationship of ACMEET and AUDITFEE on ACC sufficiently.
Specifically, the results imply that AUDITFEE does not alter the strength or
direction of the effect of ACMEET on ACC. This may be due to the fact that
ACMEET has a significant effect on ACC whereas moderation effect is
traditionally suggested for relationships which are unexpectedly weak (see Baron
et al. 1986). Thus, it can be concluded that the results related to moderation effect
provide indirect support for the mediation model analyzed in this research.
Acta Wasaensia 103
7 CONCLUSIONS
This study focuses on two important corporate governance actors whose purpose
is to ensure a company’s financial reporting quality, namely audit committees and
external auditors. This study developed a model suggesting that audit quality
functions as a mediator in the relationship between audit committee effectiveness
and financial reporting quality. The purpose of this section is to summarize the
study and to discuss its limitations and possibilities for future research. The
section is organized as follows. Firstly, the underlying premises, contributions as
well as main results of the developed model are discussed. Secondly, limitations
of the study are reviewed. Finally, suggestions for future research are proposed.
7.1 Discussion
This study develops a model which suggests that audit quality mediates the
relationship between audit committee effectiveness and financial reporting
quality. The theoretical foundation of the model developed is based on agency
theory, which states that monitoring mechanisms such as audit committees and
external auditors contribute to corporate control and thus have an effect on a
company’s financial reporting quality (e.g. Jensen et al. 1976). In addition, the
developed model can be related to prior frameworks focusing on corporate
governance, audit committee effectiveness and audit quality. Firstly, Cohen’s et
al. (2004) corporate governance mosaic highlights interrelationships between
various corporate governance actors. This study contributes both theoretically as
well as empirically to the framework by Cohen et al. (2004) by focusing on the
interrelationship between two crucial corporate governance actors, namely audit
committees and external auditors. Secondly, DeZoort et al. (2002) discusses the
determinants of audit committee effectiveness and states that it is conditional
upon audit committee composition, authority, resources and diligence. The
operational measures for audit committee effectiveness employed in this study are
consistent with DeZoort’s et al. (2002) framework: audit committee size and
expertise ratio relating to audit committee composition and audit committee
meeting frequency relating to audit committee diligence. Finally, the framework
by Watkins et al. (2004) presents the determinants of audit quality including
drivers, components as well as products of audit quality. This study adds to this
framework by proposing that audit committee effectiveness is an important
demand side driver of audit quality. This demand is expected to be reflected in
104 Acta Wasaensia
real audit quality (i.e. auditor monitoring strength) and further in financial
reporting quality, which is a product of audit quality.
In addition to the theoretical literature, current US regulations (e.g. SOX 2002)
regarding audit committees and external audits provide a regulatory framework
which determines the minimum requirements for audit committee effectiveness
and audit quality. This study focuses on companies which exceed the minimal
regulatory requirements regarding audit committee effectiveness or audit quality.
The regulatory framework also guides the choices of operational measures of the
variables of interest. More particularly, the current US regulations include
requirements regarding audit committee composition as well as its
responsibilities. In addition, particularly the SOX (2002) objective is to ensure
external audit quality and requirement to disclose audit fees provides a measure
for audit quality.
The elements of the model developed are derived from prior empirical studies
focusing on the relationships between audit committee effectiveness, audit quality
and financial reporting quality. The research model has an underlying notion that
more effective audit committees provide better monitoring over accounting
choices of the company (e.g. Beasley et al. 2000; Abbott et al. 2004; Bédard et al.
2004) and also have an interest in investing more on external auditing (e.g.
Abbott et al. 2001; Abbott et al. 2003a; Vafeas et al. 2007). In addition, external
audit is expected to contribute to a company’s financial reporting quality (e.g.
Carcello et al. 2002; Frankel et al. 2002; Nelson et al. 2002; Abbott et al. 2003a;
Krishnan 2005; Larcker et al. 2004; Srinidhi et al. 2007). This study extends the
work of the above studies both theoretically and empirically. The theoretical
contribution arises when the separate relationships are combined into a
comprehensive model: audit quality is modeled as a mediator in the relationship
between audit committee effectiveness and financial reporting quality. This effect
type has not been addressed by prior research focusing on audit committees or
external audits. The empirical contribution arises from two sources. Firstly, the
examination of the proposed effect type involves methodology which has not
been addressed by prior studies. Secondly, the present study examines the
relationships using a single set of US data. This examination reveals whether the
results reported in prior studies hold in the present US regulatory environment.
The mediation model is tested using two complementary methods. Firstly, the
Causal Steps Method was used to test the individual relationships in the model
and examine the occurrence of the mediated effect. The regression analyses
required by the method were employed separately for the three measures of audit
Acta Wasaensia 105
committee effectiveness and thus the following models were examined:
ACSIZEAUDITFEEACC, ACEXPAUDITFEEACC, and ACMEET
AUDITFEEACC. Secondly, the Sobel Test statistics were calculated for
models supported by the Causal Steps Method in order to verify the significance
of a potential mediation effect.
The main results of this study support model ACMEETAUDITFEEACC.
However, the results regarding other two models, ACSIZEAUDITFEEACC
and ACEXPAUDITFEEACC, are inconclusive. The latter result is due to the
fact that that neither audit committee size nor audit committee expertise ratio have
the hypothesized effect on discretionary accruals or audit fees. These results are
inconsistent with prior studies, which have linked audit committee size and audit
committee expertise with outcomes of audit committee effectiveness (e.g.
Archambeault et al. 2001; Bédard et al. 2004; Vafeas et al. 2007). The
inconsistency of the results can be explained by the difference in the regulatory
environment. The present study was conducted in a highly regulated environment
which requires companies to maintain homogeneous audit committees in terms of
their composition. It is likely that prior studies have been conducted in less
regulated environments in which companies’ audit committees have been less
homogeneous32.
With regard to model ACMEETAUDITFEEACC the results are consistent
with those of prior studies. More specifically, audit committee meeting frequency
is found to have a negative effect on discretionary accruals (e.g. Xie et al. 2003;
Vafeas 2005) and a positive effect on audit fees (e.g. Goodwin-Stewart et al.
2006). Consistent with prior studies audit fees are also found to have a negative
effect on discretionary accruals (e.g. Frankel et al. 2002; Larcker et al. 2004;
Srinidhi et al. 2007). In addition, audit fees are found to decrease the effect of
audit committee meeting frequency on discretionary accruals, and the mediation
hypothesis is therefore supported. The Sobel Test statistics verify this relationship
by showing that the mediated effect is significant at the 10 % level. In addition to
the main analysis several additional analyses provide further support for model
ACMEETAUDITFEEACC.
The results for model ACMEETAUDITFEEACC have several implications.
Firstly, the results indicate that more active audit committees are better able to
restrict earnings management, and thus contribute to a company’s financial
reporting quality. This may be due to the fact that active audit committees are
better informed about the state of affairs of the company or that active audit
32
See Hay et al. (2008) for a discussion on this issue.
106 Acta Wasaensia
committees are staffed with individuals who are dedicated in investing their time
and efforts into the functioning of the audit committee which results in closer
monitoring of company’s reporting decisions by the audit committee.
Secondly, the results show that more frequent audit committee meetings lead to an
increase in audit fees. This result has several plausible explanations. The positive
relation may be due to the fact that external auditor’s preparation and attendance
at audit committee meetings leads to increase in audit fees. Alternatively, more
active audit committees may require greater assurance by external auditors. In
turn greater assurance requires more audit effort, which leads to increase in audit
fees.
Thirdly, a negative relationship between audit fees and discretionary accruals is
reported. This result implies that audit effort reflected in audit fees leads to closer
monitoring by auditors, thereby decreasing management’s opportunities to use
discretionary accruals to manage earnings. The results also show that audit fees
do not compromise auditor independence, which would decrease auditor’s
willingness to oppose management attempts to manage earnings. On the contrary,
the results imply that auditors’ reputation protection leads to auditor reporting
conservatism and thus audit clients are left with less discretion with respect to
discretionary accrual when audit fees are high.
Finally, the results show that there is a sequence from audit committee meeting
frequency to audit fees to discretionary accruals. Thus, the results support the
research model developed which states that audit quality mediates the relationship
between audit committee effectiveness and financial reporting quality. This result
is satisfying for both audit committees as well as external auditors, since they
indicate that both of these control mechanisms fulfil their role as assurers of a
company’s financial reporting quality. The results also imply that the cooperation
between audit committee members and external auditors is beneficial and leads to
better financial reporting quality. In other words the results provide support for
the notion that audit committees and external auditors are complementary
contributors to financial reporting quality.
In addition to the academic implications, the results of this study have important
practical significance for preparers, users as well as auditors of the financial
statements. From the preparer point of view the results of this study imply that
companies can signal their financial reporting quality to outsiders by setting up
audit committees which are in compliance with the regulations and by
encouraging their audit committees to be active. Companies can moreover invest
in external auditing by demanding a greater scope of external audit work. On the
Acta Wasaensia 107
other hand, the results of this study imply that users of financial statements can, to
certain extent, use audit committee meeting frequency and audit fees as proxies
for audit committee effectiveness and external audit quality as they attempt to
assess financial reporting quality. The results of this study also have implications
for auditors. The results of the study imply that higher audit fees are associated
with better financial reporting quality: higher audit fees ensure that proper
resources have been invested in the audit and audit quality is therefore enhanced.
Thus, the results encourage auditors to resist possible budget pressures from
clients.
7.2 Limitations
This study is concerned with aggregate and average. That is, it seeks to examine
what type of generalized model can be established to describe the relationships
between audit committee effectiveness, audit quality and financial reporting
quality33. This approach involves using measures for audit committee
effectiveness, audit quality and financial reporting quality which can be derived
from publicly available sources using cross-sectional data.
The research method applied has inherent limitations, which should be taken into
consideration when making interpretations of the results. It can be argued that
publicly available sources provide only crude proxies for audit committee
effectiveness, audit quality and financial reporting quality: the measures
employed may not capture all aspects of the variables of interest. More
importantly, their ability to explain the interrelationships and communication
between audit committee members and external auditors in determining financial
reporting quality is limited. It should also be noted that the methodology and data
used does not allow demonstrating cause-effect relationships between variables.
Therefore it is difficult to distinguish between alternative explanations of the
results, thus the inferences of the results must be cautious. Furthermore, as
discussed by Francis (2004) it is possible that “good companies” with high
financial reporting quality may also have strong incentives to invest in audit
committee effectiveness and audit quality. Although the present study attempts to
control for company differences and test the endogeneity of the variables, this
rationale cannot be ruled out entirely as an alternative explanation for the results.
33
See Turley et al. (2007) for a critique of this approach.
108 Acta Wasaensia
7.3 Future research
This study opens up several alternative avenues for future research. Firstly, this
study employs a cross-sectional data from a single country and, as discussed
above, this creates limitations for the analysis of the model. Future research could
examine the model developed using different types or sets of data. More
specifically, future studies could seek to employ longitudinal data or data from
different countries to test the model. The advantage of longitudinal data is that it
would provide more power for the statistical analyses. In addition, studies using
longitudinal data could examine how companies’ audit committees and external
audits adapt to changes in regulatory and other environmental requirements.
Future research could also examine the model developed using data from
different countries. As discussed previously, the results did not support all the
hypothesized relationships between variables of interest. This may be because the
present study was conducted in a highly regulated environment which restricts
companies’ choices regarding audit committees, external auditors as well as
financial reporting quality. Therefore, future studies could test the model with
data derived from a less-regulated environment in order to examine whether the
hypothesized relationships are present in such an environment. It is also possible
that the relationship between audit committees and external auditors is affected by
cultural differences and therefore it would be important to test the model using
data from different countries and different cultural settings.
Secondly, in the present study the operational measures for audit committee
effectiveness, audit quality and financial reporting quality were obtained from
publicly available sources. It is reasonable to acknowledge that the measures
employed may not capture all aspects of audit committee effectiveness, audit
quality and financial reporting quality sufficiently. Thus future research could aim
to develop and establish alternative measures for audit committee effectiveness,
audit quality as well as financial reporting quality. This could be achieved by
using, for example, a questionnaire. This approach would enable measures to be
derived which are non-regulated and not publicly available. Such measures would
enable a more thorough understanding of the relationships between the variables
of interest.
Thirdly, audit committees and external auditors are only some of the many
potential monitoring mechanisms operating to ensure a company’s financial
reporting quality. Some of these mechanisms, such as a company’s full board of
directors and internal control mechanisms are within the company and
specifically designed for this purpose. Some control mechanisms are external to
Acta Wasaensia 109
the company and exercise monitoring in order to safeguard their own interests.
These parties include, for example, shareholders34 and debt holders. Thus it is
evident that examining the relationship between audit committee effectiveness
and audit quality in isolation provides an incomplete description of the
determinants of financial reporting quality. It is therefore suggested that future
research could include the various parties in the model examined in this study and
so develop a broader understanding of the interaction between these parties35.
Finally, due to the fact that audit committees and external auditors operate in a
highly complex environment it can be argued that quantitative research methods
may not be able to capture all aspects of the relationships between audit
committees and external auditors. Thus, as suggested by Turley and Zaman
(2004), Turley and Zaman (2007) and Stewart et al. (2007) there is a need for
more qualitative research regarding the functioning of audit committees as well as
external audits. Therefore, it is suggested that a case study methodology and
interviews could provide a complementary means to examine the model
developed. The main advantage of this approach is that it would enable a more
thorough analysis of the model proposed and its underlying dynamics.
34
In general the literature suggests that institutional shareholders have sufficient power to monitor
the company and its financial reporting process whereas individual shareholders’ ability to execute
such monitoring is regarded as limited.
35
See e.g. Hay et al. (2008) for a discussion regarding the relationship between internal controls
and external auditing.
110 Acta Wasaensia
REFERENCES
Abbott, L. & S. Parker (2000). Auditor selection and audit committee
characteristics. Auditing: A Journal of Practice & Theory 19:2, 47–66.
Abbott, L. & S. Parker (2001). Audit committees and auditor selection. Journal of
Accountancy 191:6, 95–95.
Abbott, L., S. Parker, G. Peters & K. Raghunandan (2003a). The association
between audit committee characteristics and audit fees. Auditing: A
Journal of Practice & Theory 22:2, 17–32.
Abbott, L., S. Parker, G. Peters & K. Raghunandan (2003b). An empirical
investigation of audit fees, nonaudit fees, and audit committees.
Contemporary Accounting Research 20:2, 215–234.
Abbott, L., S. Parker & G. Peters (2004). Audit committee characteristics and
restatements. Auditing: A Journal of Practice & Theory 23:1, 69–87.
Abdel-Khalik, A. R. (1990). The jointness of audit fees and demand for MAS: A
self-selection analysis. Contemporary Accounting Research 6:2-1, 295–
322.
Abdel-Khalik, A. R. (1993). Why do private companies demand auditing? A case
for organizational loss of control. Journal of Accounting, Auditing &
Finance 8:1, 31–52.
Alchian, A. & H. Demsetz (1972). Production, information costs, and economic
organization. American Economic Review 62:5, 777–795.
AMEX (2003). Corporate Governance Requirements. Part 8, AMEX listing
requirements.
Antle, R., E. Gordon, G. Narayanamoorthy & L. Zhou (2006). The joint
determination of audit fees, non-audit fees, and abnormal accruals. Review
of Quantitative Finance and Accounting 27:3, 235–266.
Acta Wasaensia 111
Antle, R., P. Griffin, D. Teece & O. Williamson (1997). An economic analysis of
auditor independence for a multi-client, multi-service public accounting
firm. Appendix B: Serving the Public Interest: A New Conceptual
Framework for Auditor Independence, a White Paper submitted to the
Independence Standards Board on behalf of the AICPA, October 20, 1997.
Archambeault, D. & F. Dezoort (2001). Auditor opinion shopping and the audit
committee: an analysis of suspicious auditor switches. International
Journal of Auditing 5:1, 33–52.
Ashbaugh, H. (2004). Ethical issues related to the provision of audit and non-
audit services: Evidence from academic research. Journal of Business
Ethics 52:2, 143–148.
Ashbaugh, H., R. LaFond & B. Mayhew (2003). Do nonaudit services
compromise auditor independence? Further evidence. The Accounting
Review 78:3, 611–639.
Baber, W., K. Kumar & T. Verghese (1995). Client security price reactions to the
Leventhol and Horwath bankruptcy. Journal of Accounting Research
33:2, 385–395.
Baines, A. & K. Langfield-Smith (2003). Antecedents to management accounting
change: A structural equation approach. Accounting, Organizations and
Society 28:7–8, 675–698.
Ball, R. (2008). What is the Actual Economic Role of Financial Reporting? Essay
based on a presentation at the Panel on “Big” unanswered questions in
accounting. Annual Meeting of the American Accounting Association,
Chicago, August 2007.
Ball, R. & L. Shivakumar (2005). Earnings quality in UK private firms:
Comparative loss recognition timeliness. Journal of Accounting &
Economics 39:1, 83–128.
Balsam, S., J. Krishnan & J. Yang (2003). Auditor industry specialization and
earnings quality. Auditing: A Journal of Practice & Theory 22:2, 71–97.
112 Acta Wasaensia
Barefield, R., J. Gaver & T. O’Keefe (1993). Additional evidence on the
economics of attest: Extending results from the audit market to the
market for compilations and reviews. Auditing: A Journal of Practice &
Theory 12:1, 76–88.
Barkess, L. & R. Simnett (1994). The provision of other services by auditors:
Independence and pricing issues. Accounting and Business Research,
24:94, 99–108.
Baron, R. & D. Kenny (1986). The moderator-mediator variable distinction in
social psychological research: Conceptual, strategic, and statistical
considerations. Journal of Personality and Social Psychology 51:6,
1173–1182.
Bartlett, R. (1993). A scale of perceived independence: New evidence on an old
concept. Accounting, Auditing & Accountability Journal 6:2, 52–67.
Bartov, E. (1993). The timing of asset sales and earnings manipulation. The
Accounting Review 68:4, 840–855.
Basioudis, l. & J. Francis (2007). Big 4 audit fee premiums for national and
office-level industry leadership in the United Kingdom. Auditing: A
Journal of Practice & Theory 26:2, 143–166.
Basu, S. (1997). The conservatism principle and the asymmetric timeliness of
earnings. Journal of Accounting & Economics 24:1, 3–37.
Beasley, M. (1996). An empirical analysis of the relation between the board of
director composition and financial statement fraud. The Accounting
Review 71:4, 443–465.
Beasley, M., J. Carcello, D. Hermanson & P. Lapides (2000). Fraudulent financial
reporting: Consideration of industry traits and corporate governance
mechanisms. Accounting Horizons 14:4, 441–454.
Beasley, M. & S. Salterio (2001). The relationship between board characteristics
and voluntary improvement in audit committee composition and
experience. Contemporary Accounting Research 18:4, 539–570.
Acta Wasaensia 113
Beattie, V., R. Brandt & S. Fearnley (1999). Perceptions of auditor independence:
U.K. evidence. Journal of International Accounting, Auditing and
Taxation 8:1, 67–107.
Beatty, R. (1989). Auditor reputation and the pricing of initial public offerings.
The Accounting Review 64:4, 693–709.
Beck, P., T. Frecka & I. Solomon (1988a). A model of the market for MAS and
audit services: Knowledge spillovers and auditor-auditee bonding.
Journal of Accounting Literature 7, 50–64.
Beck, P., T. Frecka, & I. Solomon (1988b). An empirical analysis of the
relationship between MAS involvement and auditor tenure: Implications
for auditor tenure. Journal of Accounting Literature 7, 65–84.
Becker, C., M. DeFond, J. Jiambalvo & K. R. Subramanyam (1998). The effect of
audit quality on earnings management. Contemporary Accounting
Research 15:1, 1–24.
Bédard, J., S. Chtourou & L. Courteau (2004). The effect of audit committee
expertise, independence, and activity on aggressive earnings
management. Auditing: A Journal of Practice & Theory 23:2, 13–35.
Bédard, J. & K. Johnstone (2004). Earnings manipulation risk, corporate
governance risk, and auditors’ planning and pricing decisions. The
Accounting Review, 79:2, 277–304, 28.
Blackwell, D., T. Noland & D. Winters (1998). The value of auditor assurance:
Evidence from loan pricing. Journal of Accounting Research 36:1, 57–
70.
Bonner, S., Z. Palmrose & S. Young (1998). Fraud type and auditor litigation: An
analysis of SEC accounting and auditing enforcement releases. The
Accounting Review 73:4, 503–532.
Bradbury, M. (1990). The incentives for voluntary audit committee formation.
Journal of Accounting and Public Policy 9:1, 19–36.
Bradbury, M., Y. Mak & S. Tan (2006). Board characteristics, audit committee
characteristics and abnormal accruals. Pacific Accounting Review 18:2,
47–68.
114 Acta Wasaensia
Brandon, D., A. Crabtree & J. Maher (2004). Nonaudit fees, auditor
independence, and bond ratings. Auditing: A Journal of Practice &
Theory 23:2, 89–103.
Burgstahler, D. & I. Dichev (1997). Earnings management to avoid earnings
decreases and losses. Journal of Accounting & Economics 24:1, 99–126.
Burton, G., D. Pathak & R. Zigli (1977). Using group size to improve the
decision-making ability of nominal groups. Academy of Management
Proceedings 37, 52–56
Bushee, B. (1998). The influence of institutional investors on myopic R&D
investment behaviour. The Accounting Review 73:3, 305–333.
Butterworth, S. & K. A. Houghton (1995). Auditor switching: The pricing of
audit services. Journal of Business Finance & Accounting 22:3, 323–
344.
Caramanis C. & C. Lennox (2008). Audit effort and earnings management.
Journal of Accounting and Economics 45:1, 116–138.
Carcello, J., D. Hermanson, T. Neal & R. Riley (2002). Board characteristics and
audit fees. Contemporary Accounting Research 19:3, 365–384.
Carcello, J. & T. Neal (2000). Audit committee composition and auditor
reporting. The Accounting Review 75:4, 453–467.
Carcello, J. & T. Neal (2003a). Audit committee characteristics and auditor
dismissals following “new” going-concern reports. The Accounting
Review 78:1, 95–117.
Carcello, J. & T. Neal (2003b). Audit committee independence and disclosure:
Choice for financially distressed firms. Corporate Governance 11:4,
289–299.
Carey, P., R. Simnett & G. Tanewski (2000). Voluntary demand for internal and
external auditing by family businesses. Auditing: A Journal of Practise
and Theory 19: International Symposium On Audit Reseach, 37–51.
Acta Wasaensia 115
Catanach, A. & P. Walker (1999). The international debate over mandatory
auditor rotation: A conceptual research framework. Journal of
International Accounting, Auditing and Taxation 8:1, 43–66.
Chen, Y., R. Moroney & K. Houghton (2005). Audit committee composition and
the use of an industry specialist audit firm. Accounting & Finance 45:2,
217–239.
Cheng, Q. & T. Warfield (2005). Equity incentives and earnings management.
The Accounting Review 80:2: 441–476.
Chow, C. (1982). The demand for external auditing: Size, debt and ownership
influences. The Accounting Review 57:2, 272–291.
Chow, C. & S. Rice (1982). Qualified audit opinions and auditor switching. The
Accounting Review 57:2, 326–336.
Chung, R., M. Firth & J. Kim (2005). FCF agency costs, earnings management,
and investor monitoring. Corporate Ownership & Control 2:4, 51–61.
Chung, H. & S. Kallapur (2003). Client importance, nonaudit services, and
abnormal accruals. The Accounting Review 78:4, 931–955.
Coase, R. H. (1937). The nature of the firm. Economic 4:16, 386–405.
Cohen, J. & D. Hanno (2000). Auditors’ consideration of corporate governance
and management control philosophy in preplanning and planning
judgments. Auditing: A Journal of Practice & Theory 19:2, 133–146.
Cohen, J., G. Krishnamoorthy & A. Wright (2004). The corporate governance
mosaic and financial reporting quality. Journal of Accounting Literature
23, 87–152.
Collier, P. & A. Gregory (1996). Audit committee effectiveness and the audit fee.
The European Accounting Review 5:2, 177–198.
Collier, P. & A. Gregory (1999). Audit committee activity and agency costs.
Journal of Accounting and Public Policy 18:4/5, 311–332.
116 Acta Wasaensia
Copley, P. & M. Doucet (1993). Auditor tenure, fixed fee contracts, and the
supply of substandard single audits. Public Budgeting and Finance 13:3,
23–35.
Craswell, A., J. Francis & S. Taylor (1995). Auditor brand name reputations and
industry specialization. Journal of Accounting & Economics 20:3, 297–
322.
Craswell, A., D. Stokes & J. Laughton (2002). Auditor independence and fee
dependence. Journal of Accounting & Economics 33:2, 253–275.
Cummings, L., G. Huber & E. Arendt (1974). Effects of size and spatial
arrangements on group decision making. The Academy of Management
Journal 17:3, 460–475.
Davis, L., D. Ricchiute & G. Trompeter (1993). Audit effort, audit fees and the
provision of nonaudit services to audit clients. The Accounting Review
68:1, 135–150.
DeAngelo, L. (1981a). Auditor size and audit quality. Journal of Accounting &
Economics 3:3, 183–199.
DeAngelo, L. (1981b). Auditor independence, ‘low balling’, and disclosure
regulation. Journal of Accounting & Economics 3:2, 113–127.
DeAngelo, L. (1986). Accounting numbers as market valuation substitutes: A
study of management buyouts of public stockholders. The Accounting
Review 61:3, 400–420.
Dechow, P. & I. Dichev (2002). The quality of accruals and earnings: The role of
accrual estimation errors. The Accounting Review 77:Supplement, 35–59.
Dechow, P., R. Sloan & A. Sweeney (1995). Detecting earnings management.
The Accounting Review 70:2, 193–225.
Dechow, P., R. Sloan & A. Sweeney (1996). Causes and consequences of
earnings manipulation: An analysis of firms subject to enforcement
actions by the SEC. Contemporary Accounting Research 13:1, 1–36.
Acta Wasaensia 117
DeFond, M. (1992). The association between changes in client firm agency costs
and auditor switching. Auditing: A Journal of Practice & Theory 11:1,
16–31.
DeFond, M., J. Francis & T. Wong (2000). Auditor industry specialization and
market segmentation: Evidence from Hong Kong. Auditing: A Journal of
Practice & Theory 19:1, 49–66.
DeFond, M., K. Raghunandan & K. R. Subramanyam (2002). Do non-audit
services fees impair auditor independence? Evidence from going concern
audit opinions. Journal of Accounting Research 40:4, 1247–1274.
DeFond, M. & K. R. Subramanyam (1998). Auditor changes and discretionary
accruals. Journal of Accounting & Economics 25:1, 35–67.
Degeorge, F., J. Patel & R. Zeckhauser (1999). Earnings management to exceed
thresholds. Journal of Business 72:1, 1–33.
Deis, D. & G. Giroux (1992). Determinants of audit quality in the public sector.
The Accounting Review 67:3, 462–479.
DeZoort, F. T. (1997). An investigation of audit committees’ oversight
responsibilities. Abacus 33:2, 208–227.
DeZoort, F. T. (1998). An analysis of experience effects on audit committee
members’ oversight judgments. Accounting, Organizations and Society
23:1, 1–22.
DeZoort, F. T., D. Hermanson, D. Archambeault & S. Reed (2002). Audit
committee effectiveness: a synthesis of the empirical audit committee
literature. Journal of Accounting Literature 21, 38–75.
DeZoort, F. T., D. Hermanson & R. Houston (2003a). Audit committee support
for auditors: The effects of materiality justification and accounting
precision. Journal of Accounting and Public Policy 22:2, 175–199.
DeZoort, F. T., D. Hermanson & R. Houston (2003b). Audit committee member
support for proposed audit adjustments: A source credibility perspective.
Auditing: A Journal of Practice & Theory 22:2, 189–205.
118 Acta Wasaensia
DeZoort, F. T. & S. Salterio (2001). The effects of corporate governance
experience and financial-reporting and audit knowledge on audit
committee members’ judgments. Auditing: A Journal of Practice &
Theory 20:2, 31-47.
Dopuch, N., R. King & R. Schwartz (2001). An experimental investigation of
retention and rotation requirements. Journal of Accounting Research
39:1, 93–117.
Dye, R. (1993). Auditing standards, legal liability, and auditor wealth. Journal of
Political Economy 101:5, 887–914.
Eichenseher, J. & D. Shields (1985). Corporate director liability and monitoring
preferences. Journal of Accounting and Public Policy 4:1, 13–31.
Emby, C. & R. Davidson (1998). The effect of engagement factors on auditor
independence: Canadian evidence. Journal of International Accounting,
Auditing and Taxation 7:2, 163–179.
European Commission (2002). Statutory Auditors’ Independence in the EU: A Set
of Fundamental Principles. Brussels: The European Communities.
Ezzamel, M., D. Gwilliam & K. Holland (1996). Some empirical evidence from
publicly quoted UK companies on the relationship between the pricing of
audit and non-audit services. Accounting and Business Research 27:1, 3–
16.
Fama, E. & M. Jensen (1983). Separation of ownership and control. Journal of
Law & Economics 26:2, 301–325.
Fan, J. & T. J. Wong (2005). Do external auditors perform a corporate
governance role in emerging markets? Evidence from East Asia. Journal
of Accounting Research 43:1, 35–72.
Ferguson, A., J. Francis & D. Stokes (2003). The effects of firm-wide and office-
level industry expertise on audit pricing on audit pricing. The Accounting
Review 78:2, 429–448.
Ferguson, A. & D. Stokes (2002). Brand name audit pricing, industry
specialization, and leadership premiums post-Big 8 and Big 6 mergers.
Contemporary Accounting Research, 19:1, 77–110.
Acta Wasaensia 119
Feroz, E. & P. Pastena (1991). The financial and market effects of the SEC’s
accounting and auditing enforcement releases. Journal of Accounting
Research 29:Supplement, 107–142.
Firth, M. (1997). The provision of nonaudit services by accounting firms to their
audit clients. Contemporary Accounting Research 14:2, 1–21.
Firth, M. (2002). Auditor-provided consultancy services and their association
with audit fees and audit opinions. Journal of Business Finance &
Accounting 29:5/6, 661–693.
Francis, J. (2004). What do we know about audit quality? The British Accounting
Review 36:4, 345–368.
Francis, J. & J. Krishnan (1999). Accounting accruals and auditor reporting
conservatism. Contemporary Accounting Research 16:1, 135–165.
Francis, J. & J. Krishnan (2002). Evidence on auditor risk-management strategies
before and after the private securities litigation reform act of 1995. Asia-
Pacific Journal of Accounting and Economics 9:2, 135–165.
Francis, J., R. LaFond, P. Olsson & K. Schipper (2005). The market pricing of
accruals quality. Journal of Accounting & Economics 39:2, 295–327.
Francis, J., E. Maydew & H. C. Sparks (1999). The role of big 6 auditors in the
credible reporting of accruals. Auditing: A Journal of Practice & Theory
18:2, 17–34.
Francis, J., K. Reichelt & D. Wang (2005). The pricing of national and city-
specific reputations for industry expertise in the U.S. audit market. The
Accounting Review 80:1, 113–136.
Francis, J. & D. Stokes (1986). Audit prices, product differentiation, and scale
economies: Further evidence from the Australian market. Journal of
Accounting Research 24:2, 383–393.
Francis, J. & E. Wilson (1988). Auditor changes: A joint test of theories relating
to agency costs and auditor differentiation. The Accounting Review 63:4,
663–682.
120 Acta Wasaensia
Frankel, R., M. Johnson & K. Nelson (2002). The relation between auditors’ fees
for nonaudit services and earnings management. The Accounting Review
77:Supplement, 71–105.
GAO (2003). Public Accounting Firms: Mandated Study on Consolidation and
Competition. United States General Accounting Office.
Gaver, J., K. Gaver & J. Austin (1995). Additional evidence on bonus plans and
income management. Journal of Accounting & Economics 19:1, 3–28.
Geiger, M. & D. Rama (2003). Audit fees, nonaudit fees, and auditor reporting on
stressed companies. Auditing: A Journal of Practice & Theory 22:2, 53–
69.
Gerdin, J. & J. Greve (2004). Forms of contingency fit in management accounting
research – a critical review. Accounting, Organizations and Society
29:3–4, 303–326.
Ghost, A. & D. Moon (2005). Auditor tenure and perceptions of audit quality. The
Accounting Review 80:2: 585–612.
Glezen, G. W. & J. Millar (1985). An empirical investigation of stockholders
reaction to disclosures required by ASR No. 250. Journal of Accounting
Research 23:2, 859–870.
Goddard, A. & C. Masters (2000). Audit committees, Cadbury Code and audit
fees: An empirical analysis of the UK companies. Managerial Auditing
Journal 15:7, 358–371.
Goldman, A. & B. Barlev (1974). The auditor-firm conflict of interests: Its
implications for independence. The Accounting Review 49:4, 707–718.
Goodwin, J. & J. Seow (2002). The influence of corporate governance
mechanisms on the quality of financial reporting and auditing:
Perceptions of auditors and directors in Singapore. Accounting &
Finance 42:3, 195–223.
Goodwin-Stewart J. & P. Kent (2006). Relation between external audit fees, audit
committee characteristics and internal audit. Accounting & Finance 46:3,
387–404.
Acta Wasaensia 121
Gul, F. (1991). Size of audit fees and perceptions of auditors’ ability to resist
management pressure in audit conflict situations. Abacus 27:2, 162–172.
Gul, F. & J. Tsui (1992). An empirical analysis of Hong Kong bankers’
perceptions of auditor ability to resist management pressure in an audit
conflict situation. Journal of International Accounting Auditing &
Taxation 1:2, 177–190.
Gul, F. & J. Tsui (1998). A test of free cash flow and debt monitoring hypotheses:
Evidence from audit pricing. Journal of Accounting & Economics 24:2,
219–237.
Gul, F. & J. Tsui (2001). Free cash flow, debt monitoring, and audit pricing:
Further evidence on the role of director equity ownership. Auditing: A
Journal of Practice & Theory 20:2, 71–84.
Hay, D. & D. Davis (2004). The voluntary choice of an auditor of any level of
quality. Auditing: A Journal of Practice & Theory 23:2, 37–53.
Hay, D., W. R. Knechel & H. Ling (2008). Evidence on the impact of internal
control and corporate governance on audit fees. International Journal of
Auditing 12:1, 9–24.
Hay, D., W. R. Knechel & N. Wong (2006). Audit fees: A meta-analysis of the
effect of supply and demand attributes. Contemporary Accounting
Research 23:1, 141–191.
Healy, P. (1985). The effect of bonus schemes on accounting decisions. Journal
of Accounting & Economics 7:1–3, 85–107.
Healy, P. & K. Palepu (1995). The challenges of investor communication: The
case of CUC International, Inc. Journal of Financial Economics 38: 2,
111–140.
Healy, P. & K. Palepu (2001). Information asymmetry, corporate disclosure, and
the capital markets: A review of the empirical disclosure literature.
Journal of Accounting and Economics 31:1–3, 405–440.
Healy, P. & J. Wahlen (1999). A review of the earnings management literature
and its implications for standard setting. Accounting Horizons 13:4, 365–
383.
122 Acta Wasaensia
Hillison W. & M. Kennelley (1988). The economics of nonaudit services.
Accounting Horizons 2:3, 32–40.
Hillison, W. & C. Pacini (2004). Auditor reputation and the insurance hypothesis:
The information content of disclosures of financial distress of a major
accounting firm. Journal of Managerial Issues 16:1, 65–86.
Hodge, F. (2003). Investors’ perceptions of earnings quality, auditor
independence, and the usefulness of audited financial information.
Accounting Horizons 17:Supplement, 37–48.
Holmbeck, G. (1997). Toward terminological, conceptual, and statistical clarity in
the study of mediators and moderators: examples from the child-clinical
and pediatric psychology literatures. Journal of Consulting and Clinical
Psychology 65:4, 599–610.
Holmbeck, G. (2002). Post-hoc probing of significant moderational and
meditational effects in studies of pediatric populations. Journal of
Pediatric Psychology 27:1, 87–96.
Holthausen, R., D. Larcker & R. Sloan (1995). Annual bonus schemes and the
manipulation of earnings. Journal of Accounting & Economics 19:1, 29–
74.
Hussey, R. & G. Lan (2001). An examination of auditor independence issues
from the perspectives of U.K. finance directors. Journal of Business
Ethics 32:2, 169–179.
IFAC (2005). IFAC Code of Ethics for Professional Accountants [online].
International Federation of Accountants [cited 18.10. 2002]. Available
from World Wide Web: .
ISB (2000). A Conceptual Framework for Auditor Independence. Independence
Standards Board (DM 00-1).
Jaggi, B. & S. Leung (2007). Impact of family dominance on monitoring of
earnings management by audit committees: Evidence from Hong Kong.
Journal of International Accounting, Auditing and Taxation 16:1, 27–50.
Acta Wasaensia 123
Jensen, M. & W. Meckling (1976). Theory of the firm: Managerial behavior,
agency costs and ownership structure. Journal of Financial Economics
3:4, 305–360.
Jiambalvo, J. (1996). Discussion of the “Causes and consequences of earnings
manipulation: an analysis of firms subject to enforcement actions by the
SEC”. Contemporary Accounting Research 13:1, 37–47.
Jokipii, A. (2006). The Structure and Effectiveness of Internal Control: A
Contingency Approach. Acta Wasaensia No. 166, Business
Administration 69, Accounting and Finance.
Jones, J. (1991). Earnings management during important relief investigations.
Journal of Accounting Research 29:2, 193–228.
Kalbers, L. & T. Fogarty (1993). Audit committee effectiveness: An empirical
investigation of the contribution of power. Auditing: A Journal of
Practice & Theory 12:1, 24–49.
Kalbers, L. & T. Fogarty (1998). Organizational and economic explanations of
audit committee oversight. Journal of Managerial Issues 10:2, 129–150.
Karamanou, I. & N. Vafeas (2005). The association between corporate boards,
audit committees, and management earnings forecasts: An empirical
analysis. Journal of Accounting Research 43:3, 453–486.
Kinney, W., Z. Palmrose & S. Scholz (2004). Auditor independence, non-audit
services, and restatements: Was the U.S. government right? Journal of
Accounting Research 42:3, 561–588.
Klein, A. (2002a). Audit committee, board of director characteristics, and
earnings management. Journal of Accounting & Economics 33:3, 375–
400.
Klein, A. (2002b). Economic determinants of audit committee independence. The
Accounting Review 77:2, 435–452.
Klein, A. (2003). Likely effects of stock exchange governance proposals and
Sarbanes-Oxley on corporate boards and financial reporting. Accounting
Horizons 17:4, 343–355.
124 Acta Wasaensia
Kline, R. (1998). Principles and Practice of Structural Equation Modeling. New
York: The Guilford Press.
Knapp, M. (1985). Audit conflict: An empirical study of the perceived ability of
auditors to resist management pressure. The Accounting Review 60:2,
202–211.
Knapp, M. (1987). An empirical study of audit committee support for auditors
involved in technical disputes with client management. The Accounting
Review 62:3, 578–588.
Knapp, M. (1991). Factors that audit committee members use as surrogates for
audit quality. Auditing: A Journal of Practice & Theory 10:1, 35–52.
Knechel, W. R. & M. Willekens (2006). The role of risk management and
governance in determining audit demand. Journal of Business Finance &
Accounting 33:9/10, 1344–1367.
Krishnan, J. (1994). Auditor switching and conservatism. The Accounting Review
69:1, 200–214.
Krishnan, G. (2003). Does big 6 industry expertise constrain earnings
management? Accounting Horizons 17:Supplement, 1–16.
Krishnan, G. (2005). The association between Big 6 auditor industry expertise
and the asymmetric timeliness of earnings. Journal of Accounting,
Auditing and Finance 20:3, 209–228.
Krishnan, J. & R. Stephens (1995). Evidence on opinion shopping from audit
opinion conservatism. Journal of Accounting and Public Policy 14:3,
179–201.
Larcker, D. & S. Richardson (2004). Fees paid to audit firms, accrual choices, and
corporate governance. Journal of Accounting Research 42:3, 625–658.
Lee, H. & V. Mande (2005). The relationship of audit committee characteristics
with endogenously determined audit and non-audit fees. Quarterly
Journal of Business & Economics 44:3/4, 93–112.
Acta Wasaensia 125
Lee, H., V. Mande & R. Ortman (2004). The effect of audit committee and board
of director independence on auditor resignation. Auditing: A Journal of
Practise and Theory 23:2, 131–146.
Lennox, C. (1999). Are large auditors more accurate than small auditors?
Accounting and Business Research 29:3, 217–227.
Lennox, C. (2000). Do companies successfully engage in opinion-shopping?
Evidence from the UK. Journal of Accounting & Economics 29:3, 321–
337.
Lennox, C. (2005). Management ownership and audit firm size. Contemporary
Accounting Research 22:1, 205–227.
MacKinnon, D., A. Fairchild & M. Fritz (2007). Mediation analysis. The Annual
Review of Psychology 58, 593–614.
MacKinnon, D., C. Lockwood, J. Hoffman, S. West & V. Sheets (2002). A
comparison of methods to test mediation and other intervening variable
effect. Psychological Methods 7:1, 83–104.
MacKinnon, D., G. Warsi & J. Dwyer (1995). A simulation study of mediated
effect measures. Multivariate Behavioral Research 30:1, 41–62.
Mautz, R. K. & H. Sharaf (1961). The Philosophy of Auditing. American
Accounting Association: Monograph No. 6.
Mayhew, B. & J. Pike (2004). Does investor selection of auditors enhance auditor
independence? The Accounting Review 79:3, 797–822.
McDaniel, L., R. Martin & L. Maines (2002). Evaluating financial reporting
quality: The effects of financial expertise vs. financial literacy. The
Accounting Review 77: Supplement, 139–167.
McKinley, S., K. Pany & P. Reckers (1985). An examination of the influence of
CPA firm type, size, and MAS provision on loan officer decisions and
perceptions. Journal of Accounting Research 23:2, 887–896.
McMullen, D. & K. Raghunandan (1996). Enhancing audit committee
effectiveness. Journal of Accountancy 182:2, 79–81.
126 Acta Wasaensia
McNichols, M. (2002). Discussion of the quality of accruals and earnings: The
role of accrual estimation errors. The Accounting Review,
77:Supplement, 61–69.
Menon, K. & J. Williams (1994). The use of audit committees for monitoring.
Journal of Accounting and Public Policy 13:2, 121–139.
Menon, K. & D. Williams (2001). Long-term trends in audit fees. Auditing: A
Journal of Practice & Theory 20:1, 115–136.
Mikol, A. & P. Standish (1998). Audit independence and nonaudit services: A
comparative study in differing British and French perspectives.
European Accounting Review 7:3, 541–569.
Mishra, S., K. Raghunandan & D. Rama (2005). Do investors’ perceptions vary
with types of nonaudit fees? Evidence from auditor ratification voting.
Auditing: A Journal of Practice & Theory 24:2, 9–25.
Mitra, S. (2007). Nonaudit service fees and auditor independence: Empirical
evidence from the oil and gas industry. Journal of Accounting, Auditing
& Finance 22:1, 85–107.
Mitra, S., M. Hossain & D. Deis (2007). The empirical relationship between
ownership characteristics and audit fees. Review of Quantitative Finance
& Accounting 28:3, 257–285.
Muller, D., C. Judd & V. Yzerbyt (2005). When moderation is mediated and
mediation is moderated. Journal of Personality and Social Psychology
89:6, 852–863.
Myers, J., L. Myers & T. Omer (2003). Exploring the term of the auditor-client
relationship and the quality of earnings: a case for mandatory auditor
rotation. The Accounting Review 78:3, 779–799.
NASDAQ (2003). Corporate Governance. Rule 4350A, Listing requirements for
Nasdaq national market and issuers except for limited partnerships.
Nelson, M., J. Elliott & R. Tarpley (2002). Evidence from auditors about
managers’ and auditors’ earnings management decisions. The
Accounting Review 77:Supplement, 175–202.
Acta Wasaensia 127
Ng, D. (1978). An information economics analysis of financial reporting and
external auditing. The Accounting Review 53:4, 910–920.
Ng D. & J. Stoeckenius (1979). Auditing: Incentives and truthful reporting.
Journal of Accounting Research 17:Supplement, 1–24.
Nikkinen J. & P. Sahlström (2004). Does agency theory provide a general
framework for audit pricing? International Journal of Auditing 8:3, 253–
262.
NYSE (2003). Corporate Governance Rules. Section 303A, NYSE listed
company manual.
O’Keefe, T., D. Simunic & M. Stein (1994). The production of audit services:
Evidence from a major public accounting firm. Journal of Accounting
Research 32:2, 241–261.
Owhoso, V., W. Messier & J. Lynch (2002). Error detection by industry-
specialized teams during sequential audit review. Journal of Accounting
Research 40:3, 883–900.
Palmrose, Z. (1986a). The effect of nonaudit services on the pricing of audit
services: Further evidence. Journal of Accounting Research 24:2, 405–
411.
Palmrose, Z. (1986b). Audit fees and auditor size: Further evidence. Journal of
Accounting Research 24:1, 97–110.
Palmrose, Z. (1987). Litigation and independent auditors: The role of business
failures and management fraud. Auditing: A Journal of Practise &
Theory 6, 90–103.
Palmrose, Z. (1988). An analysis of auditor litigation and audit service quality.
The Accounting Review 63:1, 55–73.
Parkash, M. & C. Venable (1993). Auditee incentives for auditor independence:
The case of non-audit services. The Accounting Review 68:1, 113–133.
Peasnell, K. V., P. F. Pope & S. Young (2005). Board monitoring and earnings
management: Do outside directors influence abnormal accruals? Journal
of Business Finance & Accounting 32:7/8, 1311–1346.
128 Acta Wasaensia
Pincus, K., M. Rusbarsky & J. Wong (1989). Voluntary formation of corporate
audit committees among NASDAQ firms. Journal of Accounting and
Public Policy 8:4, 239–265.
Piot, C. & R. Janin (2007). External auditors, audit committees and earnings
management in France. European Accounting Review 16:2, 429–454.
Poitras, G., T. Wilkins & Y. Kwan (2002). The timing of asset sales: Evidence of
earnings management. Journal of Business Finance & Accounting
29:7/8, 903–934.
Pomeroy, B. & D. Thornton (2008). Meta-analysis and the accounting literature:
The case of audit committee independence and financial reporting
quality. European Accounting Review 17:2, 305–330.
Raghunandan, K. (2003). Nonaudit services and shareholder ratification of
auditors. Auditing: A Journal of Practice & Theory 22:1, 155–163.
Raghunandan, K. & D. Rama (2003). Audit committee composition and
shareholder actions: Evidence from voting on auditor ratification.
Auditing: A Journal of Practice & Theory 22:2, 253–263.
Raghunandan, K., W. Read & D. Rama (2001). Audit committee composition,
“gray directors”, and interaction with internal auditing. Accounting
Horizons 15:2, 105–118.
Raghunandan, K., W. Read & J. S. Whisenant (2003). Initial evidence on the
association between nonaudit fees and restated financial statements.
Accounting Horizons 17:3, 223–234.
Reynolds, J. K., D. Deis & J. Francis (2004). Professional service fees and auditor
objectivity. Auditing: A Journal of Practice & Theory 23:1, 29–52.
Reynolds, J. K. & J. Francis (2000). Does size matter? The influence of large
clients on office level auditor reporting decisions. Journal of Accounting
& Economics 30:3, 375–400.
Richardson, S. (2006). Over-investment of free cash flow. Review of Accounting
Studies 11:2/3, 159–189.
Schipper, K. (1989). Earnings management. Accounting Horizons 3:4, 91–102.
Acta Wasaensia 129
Schipper, K. & L. Vincent (2003). Earnings quality. Accounting Horizons,
17:Supplement, 97–110.
SEC (2000). Final rule: Revision of commission’s auditor independence
requirements. Washington, D.C.: SEC. Available from World Wide
Web: .
SEC (2003). Final Rule: Strengthening the commission’s requirements regarding
auditor independence. Washington, D.C.: SEC. Available from World
Wide Web: .
Sharma, D. & J. Sidhu (2001). Professionalism vs commercialism: The
association between non-audit services (NAS) and audit independence.
Journal of Business Finance & Accounting 28:5/6, 595–629.
Shockley, R. (1981). Perceptions of auditors’ independence: An empirical
analysis. The Accounting Review 56:4, 785–800.
Shockley, R. & R. Holt (1983). A behavioral investigation of supplier
differentiation in the market for audit services. Journal of Accounting
Research 21:2, 545–564.
Shrout, P. & N. Bolger (2002). Mediation in experimental and non-experimental
studies: New procedures and recommendations. Psychological Methods
7:4, 422–445.
Simon, D. (1985). Audit services market: Additional empirical evidence.
Auditing: A Journal of Practice & Theory 5:1, 71–77.
Simon, D. & J. Francis (1988). The effect of auditor change on audit fees: Tests
of price cutting and price recovery. The Accounting Review 63:2, 255–
269.
Simunic, D. (1980). The pricing of audit services: Theory and evidence. Journal
of Accounting Research 18:1, 161–190.
Simunic, D. (1984). Auditing, consulting and auditor independence. Journal of
Accounting Research 22:2, 679–702.
130 Acta Wasaensia
Simunic, D. & M. Stein (1996). The impact of litigation risk on audit pricing: A
review of the economics and the evidence. Auditing: A Journal of
Practise and Theory 15:Supplement, 119–134.
Smith, D. (1986). Auditor “subject to” opinions, disclaimers, and auditor changes.
Auditing: A Journal of Practice & Theory 6:1, 95–108.
Sobel, M. (1982). Asymptotic confidence intervals for indirect effects in
structural equation models. Sociological methodology 13, 290–312.
Solomon, I., M. Shields & O. R. Whittington (1999). What do industry-specialist
auditors know? Journal of Accounting Research 37:1, 191–208.
SOX (2002). Sarbanes-Oxley Act.
Srinidhi, B. & F. Gul (2007). The differential effects of auditors’ nonaudit and
audit fees on accrual quality. Contemporary Accounting Research 24:2,
595–629.
Stewart, J. & L. Munro (2007). The impact of audit committee existence and
audit committee meeting frequency on the external audit: Perceptions of
Australian auditors. International Journal of Auditing 11:1, 51–69.
Streiner, D. (2005). Finding our way: An introduction to path analysis. The
Canadian Journal of Psychiatry 50:2, 115–122.
Subramanyam, K. R. (1996). The pricing of discretionary accruals. Journal of
Accounting & Economics 22:1-3, 250–281.
Teoh, H. & C. Lim (1996). An empirical study of the effects of audit committees,
disclosure of nonaudit fees, and other issues on audit independence:
Malaysian evidence. Journal of International Accounting, Auditing and
Taxation 5:2, 231–248.
Teoh, S. & T. J. Wong (1993). Perceived auditor quality and the earnings
response coefficient. The Accounting Review 68:2, 346–366.
Tsui, J., B. Jaggi & F. Gul (2001). CEO domination, growth opportunities, and
their impact on audit fees. Journal of Accounting, Auditing & Finance
16:3, 189–208.
Acta Wasaensia 131
Turley, S. & M. Zaman (2004). The corporate governance effects of audit
committees. Journal of Management and Governance 8:3, 305–332.
Turley, S. & M. Zaman (2007). Audit committee effectiveness: Informal
processes and behavioural effects. Accounting, Auditing &
Accountability Journal 20:5, 765–788.
Turpen, R. (1990). Differential pricing on auditors’ initial engagements: Further
evidence. Auditing: A Journal of Practice & Theory 9:2, 60–76.
Turpin, R. & F. T. DeZoort (1998). The characteristics of firms that voluntarily
include an audit committee report in their annual report. International
Journal of Auditing 2:1, 35–48.
Vafeas, N. (2005). Audit committee, boards, and the quality of reported earnings.
Contemporary Accounting Research 22:4, 1093–1122.
Vafeas, N. & J. Waegelein (2007). The association between audit committees,
compensation incentives, and corporate audit fees. Review of
Quantitative Finance and Accounting 28:3, 241–255.
Vanstraelen, A. (2000). Impact of renewable long-term audit mandates on audit
quality. The European Accounting Review 9:3, 419–442.
Verschoor, C. (1989). A case study of audit committee effectiveness at Sunstrand.
Internal Auditing 4:4, 11–19.
Verschoor, C. (1990). MiniScribe: A new example of audit committee
ineffectiveness. Internal Auditing 5:4, 13–19.
Wallace, W. (1980). The economic role of the audit in free and regulated markets.
University of Rochester. Available on the internet at <
http://raw.rutgers.edu/raw/wallace/homepage.html >.
Wang, D. (2006). Founding family ownership and earnings quality. Journal of
Accounting Research 44:3, 619–656.
Warfield, T., J. Wild & K. Wild (1995). Managerial ownership, accounting
choices, and informativeness of earnings. Journal of Accounting &
Economics 20:1, 61–91.
132 Acta Wasaensia
Watkins, A., W. Hillison & S. Morecroft (2004). Audit quality: A synthesis of
theory and empirical evidence. Journal of Accounting Literature 23,
153–193.
Watts, R. & J. Zimmerman (1983). Agency problems, auditing, and the theory of
the firm: Some evidence. Journal of Law & Economics 26:3, 613–633.
Watts, R. & J. Zimmerman (1986). Positive Accounting Theory. Prentice-Hall.
Weber, J. & M. Willenborg (2003). Do expert international intermediaries add
value? Evidence from auditors in microcap IPOs. Journal of Accounting
Research 41:4, 681–720.
Weisbach, M. (1988). Outside directors and CEO turnover. Journal of Financial
Economics 20:1/2, 413–460.
Whisenant, S., S. Sankaraguruswamy & K. Raghunandan (2003). Evidence on the
joint determination of audit and non-audit fees. Journal of Accounting
Research 41:4, 721–744.
Wild, J. (1996). The audit committee and earnings quality. Journal of Accounting,
Auditing & Finance 11:2, 247–276.
Willekens, M. & D. Simunic (2007). Precision in auditing standards: Effects on
auditor and director liability and the supply and demand for audit
services. Accounting and Business Research 37:3, 217–232.
Willenborg, M. (1999). Empirical analysis of the economic demand for auditing
in the initial public offerings market. Journal of Accounting Research
37:1, 225–238.
Wines, G. (1994). Auditor independence, audit qualifications and the provision of
non-audit services: A note. Accounting & Finance 34:1, 75–86.
Wolnizer, P. W. (1995). Are audit committees red herrings? Abacus 31:1, 45–66.
Xie, B., W. Davidson & P. DaDalt (2003). Earnings management and corporate
governance: The roles of the board and the audit committee. Journal of
Corporate Finance 9:3, 295–316.
Acta Wasaensia 133
APPENDIX 1. Summary of corporate governance standards36.
Rule
Regulator
SOX (2002) NYSE (2003) AMEX (2003) NASDAQ (2003)
Audit committee
size
n/a Minimum of three
members
At least three
members
At least three
members
Audit committee
independence
Audit committee
member may not
accept, other than
in his or her
capacity as a
member of the
audit committee,
the board of
directors or any
other board
committee, any
consulting,
advisory or other
compensatory fee
from the issuer or
be an affiliated
person of the
issuer or any
subsidiary
thereof.
To be
independent, an
audit committee
member must
have no material
relationship with
the listed
company either
directly or as a
partner,
shareholder, or
officer of an
organization that
has a relationship
with the company.
Audit committee
member may not
other than in his or
her capacity as a
member of the
audit committee,
the board of
directors, or any
other board
committee:1)
accept directly or
indirectly any
consulting,
advisory, or other
compensatory fee
from the issuer or
any subsidiary
thereof, or 2) be an
affiliated person of
the issuer or any
subsidiary thereof.
Consistent with
SOX (2002) audit
committee
members are
prohibited from
receiving any
payment other than
payment for board
or committee
service. Affiliated
persons of the
company or its
subsidiaries may
not be audit
committee
members. In
addition, audit
committee
members may not
own or control 20
percent or more of
the issuer's voting
securities, or such
lower number as
established by the
SEC.
Audit commitee
members’
experience
Issuer must
disclose whether
its audit
committee
includes at least
one “financial
expert,” and if
not, why not. To
qualify as a
financial expert,
audit committee
member must
have: 1) an
understanding of
GAAP and
financial
statements, 2)
experience
applying GAAP
in connection
with and in
preparing or
auditing financial
statements and
All audit
committee
members must be
financially
literate, and at
least one must
have accounting
or financial
management
expertise.
All audit
committee
members must be
able to read and
understand
fundamental
financial
statements,
including a
company's balance
sheet, income
statement, and cash
flow statement. At
least one member
of the audit
committee must be
financially
sophisticated, in
that he or she has
past employment
experience in
finance or
accounting,
requisite
All audit
committee
members must be
able to read and
understand
fundamental
financial
statements,
including a
company’s balance
sheet, income
statement, and cash
flow statement.
Additionally, each
issuer must certify
that it has, and will
continue to have,
at least one
member of the
audit committee
who has past
employment
experience in
finance or
36
The above table summarizes the SOX (2002) requirements as well as the stock exchanges’
standards regarding corporate governance. Therefore it does not inclusively offset all the corporate
governance requirements in these regulations.
134 Acta Wasaensia
Continued
Rule
Regulator
SOX (2002) NYSE (2003) AMEX (2003) NASDAQ (2003)
applying
accounting
principles in
connection with
accounting for
estimates,
accruals, and
reserves used in
the company's
financial
statements,3)
experience with
internal
accounting
controls and
procedures, and
4) an
understanding of
audit committee
functions.
professional
certification in
accounting, or any
other comparable
experience or
background which
results in the
individual's
financial
sophistication,
including but not
limited to being or
having been a chief
executive officer,
chief financial
officer, other
senior officer with
financial oversight
responsibilities.
accounting,
requisite
professional
certification in
accounting, or any
other comparable
experience or
background which
results in the
individual’s
financial
sophistication,
including being or
having been a chief
executive officer,
chief financial
officer or other
senior officer with
financial oversight
responsibilities.
Audit committee
responsibilities
Oversee the
accounting and
financial
reporting process
of the company
and the audits of
the financial
statements of the
company.
Assist board
oversight of: 1)
the integrity of the
company's
financial
statements, 2) the
company's
compliance with
legal and
regulatory
requirements, 3)
the independent
auditor's
qualifications and
independence, 4)
the performance
of the company's
internal and
external audit
function and 5)
preparation of the
audit committee
report for the
company's proxy
statement.
The audit
committee of each
issuer must have
necessary
procedures relating
to: (a) registered
public accounting
firms, (b)
complaints relating
to accounting,
internal accounting
controls or
auditing matters,
(c) authority to
engage advisors,
and (d) funding as
determined by the
audit committee.
Oversee the
company's
accounting and
financial reporting
processes and the
audits of its
financial
statements,
responsible to
ensure the external
auditor's
independence and
review and
approve all related-
party transactions.
In addition
external auditor’s
accountability to
the Committee.
Audit committee
meetings
n/a Separate, periodic
meetings with
management,
internal auditors
and external
auditors.
The audit
committee of each
issuer must meet at
least on a quarterly
basis.
n/a
Acta Wasaensia 135
APPENDIX 2. Summary of studies related to the development of the research
model.
K
ey
fin
di
n
gs
N
eg
at
iv
e
re
la
tio
n
be
tw
ee
n
au
di
t c
o
m
m
itt
ee
in
de
pe
n
de
n
ce
an
d
di
sc
re
tio
n
ar
y
ac
cr
u
al
s.
A
u
di
t c
o
m
m
itt
ee
ac
tiv
ity
an
d
its
m
em
be
rs
’
fin
an
ci
al
so
ph
ist
ic
at
io
n
ar
e
as
so
ci
at
ed
w
ith
re
du
ce
d
le
v
el
s
o
f
di
sc
re
tio
n
ar
y
cu
rr
en
t a
cc
ru
al
s.
M
ea
su
re
s
o
f a
u
di
t c
o
m
m
itt
ee
ef
fe
ct
iv
en
es
s,
ex
pe
rt
ise
,
in
de
pe
n
de
n
ce
an
d
re
sp
o
n
sib
ili
tie
s,
ar
e
n
eg
at
iv
el
y
as
so
ci
at
ed
w
ith
ag
gr
es
siv
e
ea
rn
in
gs
m
an
ag
em
en
t.
N
o
ev
id
en
ce
th
at
pr
es
en
ce
/a
bs
en
ce
o
f a
u
di
t c
o
m
m
itt
ee
af
fe
ct
s
th
e
ex
te
n
t o
f i
n
co
m
e-
in
cr
ea
sin
g
ea
rn
in
gs
m
an
ip
u
la
tio
n
.
A
u
di
t c
o
m
m
itt
ee
in
de
pe
n
de
n
ce
is
re
la
te
d
to
lo
w
er
ab
n
o
rm
al
w
o
rk
in
g
ca
pi
ta
l a
cc
ru
al
s.
Co
m
pa
ra
tiv
el
y
lo
w
er
ea
rn
in
gs
m
an
ag
em
en
t f
o
r
th
e
fir
m
s
w
ith
au
di
t c
o
m
m
itt
ee
s
co
m
pa
re
d
to
th
e
fir
m
s
w
ith
o
u
t a
u
di
t c
o
m
m
itt
ee
s.
Th
e
pr
es
en
ce
o
f a
n
au
di
t c
o
m
m
itt
ee
cu
rb
s
u
pw
ar
d
ea
rn
in
gs
m
an
ag
em
en
t.
M
et
ho
d
A
rc
hi
v
al
A
rc
hi
v
al
A
rc
hi
v
al
A
rc
hi
v
al
A
rc
hi
v
al
A
rc
hi
v
al
A
rc
hi
v
al
Sa
m
pl
e
A
ll
fir
m
ye
ar
s
lis
te
d
o
n
th
e
S&
P
50
0
as
o
f M
ar
ch
31
,
19
92
an
d
19
93
.
28
2
fir
m
-
ye
ar
o
bs
er
v
at
io
n
s
o
f S
&
P
50
0
fir
m
s
fo
r
th
e
ye
ar
s
19
92
,
19
94
,
an
d
19
96
.
Sa
m
pl
e
ba
se
d
o
n
co
m
pl
et
e
se
t o
f f
irm
s
o
n
Co
m
pu
st
at
w
ith
a
D
ec
em
be
r
31
,
19
96
ye
ar
en
d.
U
K
lis
te
d
fir
m
s
w
ith
fis
ca
l y
ea
r
en
ds
be
tw
ee
n
Ju
n
e
30
,
19
93
an
d
M
ay
31
,
19
96
.
27
1
Si
n
ga
po
re
an
an
d
27
9
M
al
ay
sia
n
fir
m
s
lis
te
d
o
n
ei
th
er
o
n
th
e
Si
n
ga
po
re
o
r
th
e
K
u
al
a
Lu
m
pu
r
St
o
ck
Ex
ch
an
ge
.
Th
e
da
ta
-
sa
m
pl
in
g
pe
rio
d
is
ye
ar
20
00
.
52
3
fir
m
-
ye
ar
o
bs
er
v
at
io
n
s
o
f
H
o
n
g
K
o
n
g
lis
te
d
co
m
pa
n
ie
s
du
rin
g
19
99
-
20
00
.
M
ai
n
co
m
pa
n
ie
s
o
n
th
e
Fr
en
ch
st
o
ck
m
ar
ke
t,
sp
ec
ifi
ca
lly
th
e
SB
F
12
0
In
de
x
co
m
pa
n
ie
s.
Th
re
e
fin
an
ci
al
ye
ar
s:
19
99
,
20
00
an
d
20
01
.
A
u
di
t c
o
m
m
itt
ee
ef
fe
ct
iv
en
es
s
a
n
d
fin
a
n
ci
a
l r
ep
o
rt
in
g
qu
a
lit
y
A
u
th
o
r(s
)
K
le
in
(20
02
a)
X
ie
,
D
av
id
so
n
&
D
aD
al
t
(20
03
)
B
éd
ar
d,
Ch
to
u
ro
u
&
Co
u
rt
ea
u
(20
04
)
Pe
as
n
el
l,
Po
pe
&
Y
o
u
n
g
(20
05
)
B
ra
db
u
ry
,
M
ak
&
Ta
n
(20
06
)
Ja
gg
i &
Le
u
n
g
(20
07
)
Pi
o
t &
Ja
n
in
(20
07
)
136 Acta Wasaensia
K
ey
fin
di
n
gs
Po
sit
iv
e
re
la
tio
n
sh
ip
be
tw
ee
n
siz
e-
re
la
te
d
au
di
t f
ee
s
an
d
th
e
pr
es
en
ce
o
f a
n
au
di
t c
o
m
m
itt
ee
.
A
u
di
t c
o
m
m
itt
ee
in
de
pe
n
de
n
ce
an
d
fin
an
ci
al
ex
pe
rt
ise
ar
e
po
sit
iv
el
y
as
so
ci
at
ed
w
ith
au
di
t f
ee
s.
M
ee
tin
g
fre
qu
en
cy
w
as
n
o
t a
ss
o
ci
at
ed
w
ith
au
di
t f
ee
s.
A
u
di
t c
o
m
m
itt
ee
in
de
pe
n
de
n
ce
,
ac
tiv
ity
le
v
el
an
d
ex
pe
rt
ise
ar
e
po
sit
iv
el
y
as
so
ci
at
ed
w
ith
au
di
t f
ee
s.
Ex
ist
en
ce
o
f a
n
au
di
t c
o
m
m
itt
ee
an
d
m
o
re
fre
qu
en
t
au
di
t c
o
m
m
itt
ee
m
ee
tin
gs
ar
e
as
so
ci
at
ed
w
ith
hi
gh
er
au
di
t f
ee
s.
A
po
sit
iv
e
re
la
tio
n
sh
ip
be
tw
ee
n
pr
es
en
ce
o
f a
n
au
di
t
co
m
m
itt
ee
an
d
au
di
t f
ee
s.
A
po
sit
iv
e
as
so
ci
at
io
n
be
tw
ee
n
au
di
t c
o
m
m
itt
ee
in
de
pe
n
de
n
ce
an
d
au
di
t f
ee
s.
A
u
di
t c
o
m
m
itt
ee
siz
e,
au
di
t c
o
m
m
itt
ee
m
em
be
r
ex
pe
rt
ise
,
an
d
au
di
t c
o
m
m
itt
ee
m
em
be
r
in
de
pe
n
de
n
ce
ar
e
po
sit
iv
el
y
as
so
ci
at
ed
w
ith
au
di
t f
ee
le
v
el
s.
Ex
ist
en
ce
o
f a
n
au
di
t c
o
m
m
itt
ee
is
po
sit
iv
el
y
as
so
ci
at
ed
w
ith
au
di
t f
ee
s.
M
et
ho
d
A
rc
hi
v
al
A
rc
hi
v
al
A
rc
hi
v
al
A
rc
hi
v
al
A
rc
hi
v
al
A
rc
hi
v
al
A
rc
hi
v
al
A
rc
hi
v
al
Sa
m
pl
e
U
K
co
m
pa
n
ie
s
w
hi
ch
w
er
e
m
em
be
rs
o
f t
he
Fi
n
an
ci
al
Ti
m
es
A
ll
Sh
ar
e
In
de
x
(F
TA
SI
) i
n
D
ec
em
be
r
19
91
.
A
sa
m
pl
e
o
f 4
92
n
o
n
re
gu
la
te
d,
B
ig
5-
au
di
te
d
co
m
pa
n
ie
s
th
at
fil
ed
pr
o
x
y
st
at
em
en
ts
w
ith
th
e
SE
C
in
20
01
.
79
2
U
S
co
m
pa
n
ie
s
(S
&
P
15
00
fir
m
s
o
r
o
th
er
la
rg
e
fir
m
s).
40
1
A
u
st
ra
lia
n
pu
bl
ic
ly
lis
te
d
co
m
pa
n
ie
s.
Sa
m
pl
e
o
f l
ist
ed
co
m
pa
n
ie
s
in
B
el
gi
u
m
.
35
8
fir
m
s
lis
te
d
o
n
N
Y
SE
w
ith
fis
ca
l y
ea
r-
en
d
D
ec
em
be
r
31
,
20
00
.
Fo
rt
u
n
e
50
0
fir
m
s
in
20
01
.
13
0
co
m
pa
n
ie
s
lis
te
d
o
n
th
e
N
ew
Ze
al
an
d
St
o
ck
Ex
ch
an
ge
in
19
95
an
d
20
05
.
A
u
di
t c
o
m
m
itt
ee
ef
fe
ct
iv
en
es
s
a
n
d
a
u
di
t q
u
a
lit
y
A
u
th
o
r(s
)
Co
lli
er
&
G
re
go
ry
(19
96
)
A
bb
o
tt,
Pa
rk
er
,
Pe
te
rs
&
R
ag
hu
n
an
da
n
(20
03
a)
Le
e
&
M
an
de
(20
05
)
G
o
o
dw
in
-
St
ew
ar
t &
K
en
t (
20
06
)
K
n
ec
he
l &
W
ill
ek
en
s
(20
06
)
M
itr
a,
H
o
ss
ai
n
&
D
ei
s
(20
07
)
V
af
ea
s
&
W
ae
ge
le
in
(20
07
)
H
ay
,
K
n
ec
he
l &
Li
n
g
(20
08
)
Acta Wasaensia 137
K
ey
fin
di
n
gs
Cl
ie
n
ts
o
f n
o
n
-
B
ig
Si
x
au
di
to
rs
re
po
rt
di
sc
re
tio
n
ar
y
ac
cr
u
al
s
th
at
ar
e
hi
gh
er
th
an
th
e
di
sc
re
tio
n
ar
y
ac
cr
u
al
s
re
po
rt
ed
by
cl
ie
n
ts
o
f B
ig
Si
x
au
di
to
rs
.
B
ig
Si
x
au
di
te
d
co
m
pa
n
ie
s
ha
v
e
lo
w
er
le
v
el
o
f
es
tim
at
ed
di
sc
re
tio
n
ar
y
ac
cr
u
al
s
th
an
n
o
n
-
Bi
g
Si
x
au
di
te
d
co
m
pa
n
ie
s.
A
u
di
t f
ee
s
ar
e
n
eg
at
iv
el
y
as
so
ci
at
ed
w
ith
sm
al
l e
ar
n
in
gs
su
rp
ris
es
an
d
th
e
m
ag
n
itu
de
o
f d
isc
re
tio
n
ar
y
ac
cr
u
al
s.
Cl
ie
n
ts
o
f i
n
du
st
ry
sp
ec
ia
lis
t a
u
di
to
rs
ha
v
e
lo
w
er
di
sc
re
tio
n
ar
y
ac
cr
u
al
s
an
d
hi
gh
er
ea
rn
in
gs
re
sp
o
n
se
co
ef
fic
ie
n
ts
th
an
fir
m
s
n
o
t a
u
di
te
d
by
in
du
st
ry
sp
ec
ia
lis
t a
u
di
to
rs
.
R
es
u
lts
su
gg
es
t h
ig
he
r
ea
rn
in
gs
qu
al
ity
w
ith
lo
n
ge
r
au
di
to
r
te
n
u
re
.
Th
e
ea
rn
in
gs
o
f c
lie
n
ts
o
f s
pe
ci
al
ist
au
di
to
rs
ar
e
m
o
re
tim
el
y
in
re
fle
ct
in
g
ba
d
n
ew
s
th
an
ea
rn
in
gs
o
f c
lie
n
ts
o
f
n
o
n
-
sp
ec
ia
lis
t a
u
di
to
rs
H
ig
he
r
fe
es
pa
id
to
au
di
to
rs
ar
e
as
so
ci
at
ed
w
ith
sm
al
le
r
ac
cr
u
al
s.
W
ea
k
ev
id
en
ce
th
at
lo
n
ge
r
au
di
to
r
te
n
u
re
in
cr
ea
se
s
th
e
qu
al
ity
o
f r
ep
o
rt
ed
ea
rn
in
gs
.
A
u
di
t
fe
es
re
su
lt
in
hi
gh
er
ac
cr
u
al
qu
al
ity
.
M
et
ho
d
A
rc
hi
v
al
A
rc
hi
v
al
A
rc
hi
v
al
A
rc
hi
v
al
A
rc
hi
v
al
A
rc
hi
v
al
A
rc
hi
v
al
A
rc
hi
v
al
A
rc
hi
v
al
Sa
m
pl
e
10
39
7
fir
m
ye
ar
o
bs
er
v
at
io
n
s
au
di
te
d
by
B
ig
Si
x
au
di
to
rs
an
d
21
79
fir
m
ye
ar
o
bs
er
v
at
io
n
s
au
di
te
d
by
n
o
n
-
B
ig
Si
x
au
di
to
rs
.
A
sa
m
pl
e
o
f N
A
SD
A
Q
co
m
pa
n
ie
s
30
74
co
m
pa
n
ie
s
w
hi
ch
iss
u
ed
pr
o
x
y
st
at
em
en
ts
w
ith
th
e
SE
C
in
20
01
.
A
sa
m
pl
e
o
f U
S
co
m
pa
n
ie
s
au
di
te
d
by
B
ig
Si
x
(F
iv
e)
au
di
to
rs
fro
m
19
91
to
19
99
.
A
sa
m
pl
e
o
f c
o
m
pa
n
ie
s
o
n
th
e
20
01
Co
m
pu
st
at
an
n
u
al
in
du
st
ria
l a
n
d
re
se
ar
ch
fil
es
.
20
64
6
fir
m
ye
ar
o
bs
er
v
at
io
n
s
re
pr
es
en
tin
g
35
50
U
S
co
m
pa
n
ie
s.
U
S
co
m
pa
n
ie
s
w
hi
ch
re
po
rt
ed
da
ta
o
n
au
di
t a
n
d
n
o
n
-
.
au
di
t s
er
v
ic
es
fe
es
fo
r
fis
ca
l y
ea
rs
20
00
an
d
20
01
.
.
M
ai
n
co
m
pa
n
ie
s
o
n
th
e
Fr
en
ch
st
o
ck
m
ar
ke
t (
th
e
SB
F
12
0
In
de
x
co
m
pa
n
ie
s).
Co
m
pa
n
ie
s
lis
te
d
o
n
N
Y
SE
,
A
M
EX
o
r
N
A
SD
A
Q.
62
84
fir
m
-
ye
ar
o
bs
er
v
at
io
n
s
fro
m
ye
ar
s
20
00
an
d
20
01
.
A
u
di
t
qu
a
lit
y
a
n
d
fin
a
n
ci
a
l r
ep
o
rt
in
g
qu
a
lit
y
A
u
th
o
r(s
)
B
ec
ke
r,
D
eF
o
n
d,
Jia
m
ba
lv
o
&
Su
br
am
an
ya
m
(19
98
)
Fr
an
ci
s,
M
ay
de
w
&
Sp
ar
ks
(19
99
)
Fr
an
ke
l,
Jo
hn
so
n
&
N
el
so
n
(20
02
)
B
al
sa
m
,
K
ris
hn
an
&
Y
an
g
(20
03
)
M
ye
rs
,
M
ye
rs
&
O
m
er
(20
03
)
K
ris
hn
an
(20
05
)
La
rc
ke
l &
R
ic
ha
rd
so
n
(20
04
)
Pi
o
t &
Ja
n
in
(20
07
)
Sr
in
id
hi
&
G
u
l (
20
07
)
138 Acta Wasaensia
APPENDIX 3. Definitions for variables used in the analyses.
Variable name Description of the variable
Audit committee effectiveness
ACSIZE Number of audit committee members
ACEXP Ratio of financial experts on the audit committee
ACMEET Number of audit committee meetings during fiscal
year
Audit quality
AUDITFEE Natural logarithm of audit fees paid to the incumbent
auditor
Financial reporting quality
ACC Discretionary accruals (Residual from the regression
model 6 scaled by 100)
Control variables for financial reporting quality
TA Natural logarithm of total assets
OPCYCLE Operating cycle (Calculated as presented in
Equations 7-8)
SALEG Growth rate in net sales
LOSS Indicator variable, 1 if the net income of the fiscal
year is negative, otherwise 0
Control variables for audit quality
TA Natural logarithm of total assets
INVREC Total inventories and total receivables to total assets
FOROPR Foreign sales to total sales
QR Quick ratio
LOSS Indicator variable, 1 if the net income of the fiscal
year is negative, otherwise 0
BSIZE Number of board members
BIND Board independence %
BMEET Number of board meetings during fiscal year
Acta Wasaensia 139
APPENDIX 4. Sample selection criteria.
N
15
00
24
0
26
0
10
00
In
iti
a
l s
a
m
pl
e
S&
P
15
00
fir
m
s
(co
m
pr
ise
s
o
f S
&
P
50
0,
S&
P
M
id
Ca
p
40
0
an
d
S&
P
Sm
al
lC
ap
60
0
fir
m
s)
Le
ss
Fi
n
an
ci
al
in
st
itu
tio
n
s
(S
IC
co
de
s
60
00
-
69
99
)
Fi
rm
s
in
in
du
st
rie
s
w
ith
in
su
ffi
ci
en
t d
at
a
fo
r
es
tim
at
in
g
ac
cr
u
al
s
(le
ss
th
an
20
o
bs
er
v
at
io
n
)
Fi
n
a
l s
a
m
pl
e
140 Acta Wasaensia
APPENDIX 5. Descriptive statistics of variables used in statistical analyses.
Table 1. Descriptive statistics for variables used in main analysis.
Minimum Maximum Mean Std.Dev. Skewness Kurtosis
ACSIZE 2 8 3.74 .928 1.051 1.428
ACEXP 0 1.00 .4871 .28409 .737 -.757
ACMEET 0 31 9.02 3.616 1.328 4.756
AUDITFEE 12.00 18.27 14.6051 .97259 .416 .164
ACC -49.04 48.68 .000 11.39802 .142 2.020
TA 17.55 27.27 21.4201 1.50657 .416 -.005
OPCYCLE 4.58 664.33 120.7825 75.64084 2.107 7.296
SALESG -.87 6.47 .1405 .27582 12.487 281.326
LOSS 0 1 .1004 .30069 2.663 5.103
INVREC 0 .89 .2483 .15784 .903 .671
FOROPR 0 1.02 .2972 .26144 .510 -.727
QR .077 24.470 1.74010 2.022967 4.730 35.100
BSIZE 4 17 9.11 2.119 .357 .031
BIND 0 100.000 72.91053 13.753116 -.810 .906
BMEET 2 31 7.80 3.184 1.866 7.006
Notes:
The variables are as defined in Appendix 3
Table 2. Descriptive statistics of winsorized variables.
Minimum Maximum Mean Std.Dev. Skewness Kurtosis
ACSIZE 3 6 3.74 .852 .905 -.038
ACEXP .17 1.00 .4903 .27950 .812 -.787
ACMEET 4 17 8.93 3.178 .478 -.144
AUDITFEE 12.93 16.74 14.6043 .92984 .372 -.435
ACC -21.31 26.39 .0603 10.30859 .267 .128
TA 18.80 24.50 21.4099 1.44328 .298 -.617
OPCYCLE 22.37 327.43 118.5289 65.76321 1.209 1.536
SALESG -.19 .62 .1356 .16311 .942 1.253
LOSS .00 1.00 .1004 .30069 2.663 5.103
INVREC .04 .64 .2468 .15237 .728 -.090
FOROPR .00 .82 .2945 .25536 .412 -1.007
QR .257 7.309 1.65395 1.513498 2.168 4.637
BSIZE 5 13 9.07 2.024 .140 -.656
BIND 42.857 90.909 73.10669 13.020966 -.487 -.607
BMEET 4 15 7.70 2.775 .834 .197
Notes:
The variables are as defined in Appendix 3
Acta Wasaensia 141
APPENDIX 6. Correlation matrix.
SA
LE
SG
-
.
05
6*
-
.
03
3
-
.
05
4
-
.
10
4*
*
*
-
.
06
0*
-
.
02
5
-
.
06
7*
*
1
O
PC
Y
CL
E
-
.
07
8*
*
*
.
06
4*
.
03
2
.
03
8
.
27
3*
*
*
-
.
03
7 1
TA
.
36
3*
*
*
.
12
4*
*
*
.
11
1*
*
*
.
79
2*
*
*
-
.
22
3*
*
*
1
A
CC
-
.
01
5
.
01
3
-
.
11
1*
*
*
-
.
21
1*
*
*
1
A
U
D
IT
FE
E
.
27
9*
*
*
.
11
7*
*
*
.
21
4*
*
*
1
A
CM
EE
T
-
.
03
1
.
08
3*
*
*
1
A
CE
X
P
-
.
10
0*
*
*
1
A
CS
IZ
E
1
A
CS
IZ
E
A
CE
X
P
A
CM
EE
T
A
U
D
IT
FE
E
A
CC
TA
O
PC
Y
CL
E
SA
LE
SG
LO
SS
IN
V
R
EC
FO
R
O
PR
QR
B
SI
ZE
B
IN
D
B
M
EE
T
N
o
te
s:
*
*
*
,
*
*
,
*
de
n
o
te
sig
n
ifi
ca
n
ce
at
1%
,
5%
an
d
10
%
le
v
el
re
sp
ec
tiv
el
y
Th
e
va
ri
a
bl
es
a
re
a
s
de
fin
ed
in
Ap
pe
n
di
x
3
142 Acta Wasaensia
B
M
EE
T
.
04
9
.
02
4
.
26
1*
*
*
.
12
5*
*
*
-
.
14
0*
*
*
.
11
7*
*
*
.
05
7*
-
.
05
6*
.
12
7*
*
*
-
.
15
6*
*
*
-
.
01
0
.
05
2
.
04
2
.
10
3*
*
*
1
B
IN
D
.
22
5*
*
*
.
03
2
.
07
9*
*
.
18
9*
*
*
-
.
07
7*
*
.
15
8*
*
*
-
.
01
1
-
.
04
5
-
.
03
0
-
.
03
5
.
06
7*
-
.
06
3*
.
09
9*
*
*
1
B
SI
ZE
.
43
7*
*
*
.
05
9*
.
05
7*
.
50
0*
*
*
-
.
12
1*
*
*
.
59
2*
*
*
-
.
05
1
-
.
07
8*
*
-
.
01
9
-
.
03
8
-
.
01
2
-
.
28
1*
*
*
1
QR
-
.
21
1*
*
*
-
.
04
1
-
.
06
7*
*
-
.
31
4*
*
*
.
06
3*
*
-
.
33
7*
*
*
.
18
7*
*
*
.
04
4
.
07
4*
*
-
.
14
0*
*
*
.
09
7*
*
*
1
FO
R
O
PR
-
.
04
7
.
03
8
.
05
5
.
33
1*
*
*
-
.
08
0*
*
.
08
4*
*
.
23
4*
*
*
.
00
2
.
07
5*
*
.
13
8*
*
*
1
IN
V
R
EC
.
02
0
-
.
01
7
.
00
8
.
02
2
.
49
2*
*
*
-
.
18
6*
*
*
.
19
4*
*
*
-
.
01
0
-
.
02
5 1
LO
SS
-
.
07
6*
*
-
.
01
0
.
06
5*
.
00
3
-
.
10
3*
*
*
-
.
10
7*
*
*
.
16
5*
*
*
-
.
08
8*
*
1
(C
o
n
tin
u
ed
)
A
CS
IZ
E
A
CE
X
P
A
CM
EE
T
A
U
D
IT
FE
E
A
CC
TA
O
PC
Y
CL
E
SA
LE
SG
LO
SS
IN
V
R
EC
FO
R
O
PR
QR
B
SI
ZE
B
IN
D
B
M
EE
T
N
o
te
s:
*
*
*
,
*
*
,
*
de
n
o
te
sig
n
ifi
ca
n
ce
at
1%
,
5%
an
d
10
%
le
v
el
re
sp
ec
tiv
el
y
Th
e
va
ri
a
bl
es
a
re
a
s
de
fin
ed
in
Ap
pe
n
di
x
3
Acta Wasaensia 143
APPENDIX 7. Main relationships as scatterplots.
144 Acta Wasaensia
Acta Wasaensia 145
APPENDIX 8. Companies grouped by industries.
%
in
tt
he
sa
m
pl
e 4.
4
3.
7
2.
1
9.
5
2.
3
2.
3
7.
1
11
.
4
3.
3
9.
5
3.
1
8.
6
3.
7
2.
1
2.
8
2.
8
2.
2
14
.
0
2.
9
2.
2
10
0.
0
D
es
cr
ip
tio
n
o
f t
he
in
du
st
ry
O
il
an
d
ga
s
ex
tr
ac
tio
n
Fo
o
d
an
d
ki
n
dr
ed
pr
o
du
ct
s
Pr
in
tin
g,
pu
bl
ish
in
g
an
d
al
lie
d
in
du
st
rie
s
Ch
em
ic
al
s
an
d
al
lie
d
pr
o
du
ct
s
Pr
im
ar
y
m
et
al
in
du
st
rie
s
Fa
br
ic
at
ed
m
et
al
pr
o
du
ct
s,
ex
ce
pt
m
ac
hi
n
er
y
an
d
tr
an
sp
o
rt
at
io
n
eq
u
ip
m
en
t
In
du
st
ria
l a
n
d
co
m
m
er
ci
al
m
ac
hi
n
er
y
an
d
co
m
pu
te
r
eq
u
ip
m
en
t
El
ec
tr
o
n
ic
an
d
o
th
er
el
ec
tr
ic
al
eq
u
ip
m
en
t
an
d
co
m
po
n
en
ts
,
ex
ce
pt
co
m
pu
te
r
eq
u
ip
m
en
t
Tr
an
sp
o
rt
at
io
n
eq
u
ip
m
en
t
M
ea
su
rin
g,
an
al
yz
in
g
an
d
co
n
tr
o
lli
n
g
in
st
ru
m
en
ts
Co
m
m
u
n
ic
at
io
n
s
El
ec
tr
ic
,
ga
s
an
d
sa
n
ita
ry
se
rv
ic
es
W
ho
le
sa
le
tr
ad
e-
du
ra
bl
e
go
o
ds
W
ho
le
sa
le
tr
ad
e-
n
o
n
-
du
ra
bl
e
go
o
ds
A
pp
ar
el
an
d
ac
ce
ss
o
ry
st
o
re
s
Ea
tin
g
an
d
dr
in
ki
n
g
pl
ac
es
M
isc
el
la
n
eo
u
s
re
ta
il
B
us
in
es
s
se
rv
ic
es
H
ea
lth
se
rv
ic
es
En
gi
n
ee
rin
g,
A
cc
o
u
n
tin
g,
R
es
ea
rc
h,
M
an
ag
em
en
t a
n
d
re
la
te
d
se
rv
ic
es
N
u
m
be
r
o
f f
ir
m
s 4
4 37
21
95
23
23
71
11
4 33
95
31
86
37
21
28
28
22
14
0 29
22
10
00
Tw
o
di
gi
t S
IC
co
de
13
20
27
28
33
34
35
36
37
38
48
49
50
51
56
58
59
73
80
87
To
ta
l
146 Acta Wasaensia
APPENDIX 9. Results for model ACSIZEAUDITFEEACC.
Variable Condition 1
ACC
Variable Condition 2
AUDITFEE
Variable Conditions 3
and 4
ACC
Intercept
t
se
TA
t
se
OPCYCLE
t
se
SALESG
t
se
LOSS
t
se
Audit committee
effectiveness
ACSIZE
t
se
34.307***
7.046
4.869
-1.987***
-8.358
.238
.045***
9.822
.005
-2.423**
-1.987
1.219
-6.766***
-5.949
1.137
1.039***
2.609
.398
Intercept
t
se
TA
t
se
INVREC
t
se
FOROPR
t
se
QR
t
se
LOSS
t
se
BSIZE
t
se
BIND
t
se
BMEET
t
se
Audit committee
effectiveness
ACSIZE
t
se
3.619***
12.111
.299
.468***
31.263
.015
.812***
7.016
.116
.914***
12.728
.072
-.029***
-3.143
.009
.217***
3.775
.058
.028***
2.634
.011
.004***
2.883
.001
.014**
2.519
.006
-.031
-1.423
.022
Intercept
t
se
TA
t
se
OPCYCLE
t
se
SALESG
t
se
LOSS
t
se
Audit committee
effectiveness
ACSIZE
t
se
Audit quality
AUDITFEE
t
se
39.347***
7.431
5.295
-1.321***
-3.613
.366
.045***
10.007
.005
-2.611**
-2.142
1.219
-6.447***
-5.643
1.142
1.040***
2.617
.397
-1.329**
-2.392
.556
R2 .157 .695 .162
Sobel Test -
Notes:
***, **, *denote significance at 1%, 5% and 10% level respectively
The variables are as defined in Appendix 3
Acta Wasaensia 147
APPENDIX 10. Results for model ACEXPAUDITFEEACC.
Variable Condition 1
ACC
Variable Condition 2
AUDITFEE
Variable Conditions 3
and 4
ACC
Intercept
t
se
TA
t
se
OPCYCLE
t
se
SALESG
t
se
LOSS
t
se
Audit committee
effectiveness
ACEXP
t
se
33.805***
6.926
4.881
-1.793***
-7.939
.226
.044***
9.584
.005
-2.582**
-2.113
1.222
-6.858***
-6.014
1.140
.793
.645
1.230
Intercept
t
se
TA
t
se
INVREC
t
se
FOROPR
t
se
QR
t
se
LOSS
t
se
BSIZE
t
se
BIND
t
se
BMEET
t
se
Audit committee
effectiveness
ACEXP
t
se
3.623***
12.112
.299
.464***
31.088
.015
.801***
6.926
.116
.920***
12.821
.072
-.029***
-3.080
.009
.221***
3.845
.058
.023**
2.314
.010
.003***
2.649
.001
.014**
2.480
.006
.041
.655
.062
Intercept
t
se
TA
t
se
OPCYCLE
t
se
SALESG
t
se
LOSS
t
se
Audit committee
effectiveness
ACEXP
t
se
Audit quality
AUDITFEE
t
se
38.874***
7.323
5.308
-1.125***
-3.139
.358
.045***
9.767
.005
-2.770**
-2.268
1.221
-6.537***
-5.707
1.146
.860
.701
1.228
-1.337**
-2.398
.558
R2 .151 .694 .156
Sobel Test -
Notes:
***, **, *denote significance at 1%, 5% and 10% level respectively
The variables are as defined in Appendix 3
148 Acta Wasaensia
APPENDIX 11. Results for model ACMEETAUDITFEEACC.
Variable Condition 1
ACC
Variable Condition 2
AUDITFEE
Variable Conditions
3 and 4
ACC
Intercept
t
se
TA
t
se
OPCYCLE
t
se
SALESG
t
se
LOSS
t
se
Audit committee
effectiveness
ACMEET
t
se
34.758***
7.133
4.873
-1.705***
-7.584
.225
.044**
9.741
.005
-2.746***
-2.255
1.217
-6.650***
-5.843
1.138
-.279***
-2.901
.096
Intercept
t
se
TA
t
se
INVREC
t
se
FOROPR
t
se
QR
t
se
LOSS
t
se
BSIZE
t
se
BIND
t
se
BMEET
t
se
Audit committee
effectiveness
ACMEET
t
se
3.557***
12.043
.295
.461***
31.382
.015
.774***
6.776
.114
.906***
12.801
.071
-.026***
-2.867
.009
.209***
3.680
.057
.023**
2.363
.010
.003**
2.474
.001
.007
1.143
.006
.026***
5.208
.005
Intercept
t
se
TA
t
se
OPCYCLE
t
se
SALESG
t
se
LOSS
t
se
Audit committee
effectiveness
ACMEET
t
se
Audit quality
AUDITFEE
t
se
38.711***
7.314
5.293
-1.176***
-3.286
.358
.045***
9.872
.005
-2.880**
-2.365
1.218
-6.419***
-5.615
1.143
-.246**
-2.517
.098
-1.073*
-1.899
.565
R2 .158 .702 .161
Sobel Test
-1.784*
Notes:
***, **, *denote significance at 1%, 5% and 10% level respectively
The variables are as defined in Appendix 3
Acta Wasaensia 149
APPENDIX 12. Results for model ACMEETAUDITFEEACC (winsorized
data).
Variable Condition 1
ACC
Variable Condition 2
AUDITFEE
Variable Conditions
3 and 4
ACC
Intercept
t
se
TA
t
se
OPCYCLE
t
se
SALESG
t
se
LOSS
t
se
Audit committee
effectiveness
ACMEET
t
se
36.085***
7.788
4.633
-1.755***
-8.271
.212
.047***
10.065
.005
-6.262***
-3.367
1.859
-6.420***
-6.282
1.022
-.282***
-2.865
.098
Intercept
t
se
TA
t
se
INVREC
t
se
FOROPR
t
se
QR
t
se
LOSS
t
se
BSIZE
t
se
BIND
t
se
BMEET
t
se
Audit committee
effectiveness
ACMEET
t
se
3.718***
12.070
.308
.454***
30.224
.015
.779***
6.821
.114
.947***
13.499
.070
-.033***
-2.649
.012
.162***
2.973
.054
.022**
2.210
.010
.003**
2.372
.001
.008
1.293
.006
.027***
4.999
.005
Intercept
t
se
TA
t
se
OPCYCLE
t
se
SALESG
t
se
LOSS
t
se
Audit committee
effectiveness
ACMEET
t
se
Audit quality
AUDITFEE
t
se
40.304***
8.028
5.021
-1.201***
-3.608
.333
.048***
10.261
.005
-6.354***
-3.422
1.857
-6.229***
-6.083
1.024
-.244**
-2.445
.100
-1.135**
-2.160
.525
R2 .175 .702 .179
Sobel Test -2.01**
Notes:
Data is winsorized by 2.5% from both tails.
***, **, *denote significance at 1%, 5% and 10% level respectively
The variables are as defined in Appendix 3
150 Acta Wasaensia
APPENDIX 13. All measures for audit committee effectiveness included.
Variable
Condition 1
ACC
Variable Condition 2
AUDITFEE
Variable Conditions 3
and 4
ACC
Intercept
t
se
TA
t
se
OPCYCLE
t
se
SALESG
t
se
LOSS
t
se
Audit committee
effectiveness
ACSIZE
t
se
ACEXP
t
se
ACMEET
t
se
35.225***
7.244
4.863
-1.949***
-8.065
.242
.045***
9.817
.005
-2.525**
-2.075
1.217
-6.545***
-5.762
1.136
1.033***
2.570
.402
1.507
1.217
1.238
-.270***
-2.798
.096
Intercept
t
se
TA
t
se
INVREC
t
se
FOROPR
t
se
QR
t
se
LOSS
t
se
BSIZE
t
se
BIND
t
se
BMEET
t
se
Audit committee
effectiveness
ACSIZE
t
se
ACEXP
t
se
ACMEET
t
se
3.557***
12.033
.296
.463***
31.080
.015
.782***
6.821
.115
.902***
12.708
.071
-.027***
-2.906
.009
.207***
3.634
.057
.026**
2.528
.010
.003***
2.601
.001
.007
1.183
.006
-.020
-.907
.022
.010
.156
.062
.025***
5.074
.005
Intercept
t
se
TA
t
se
OPCYCLE
t
se
SALESG
t
se
LOSS
t
se
Audit committee
effectiveness
ACSIZE
t
se
ACEXP
t
se
ACMEET
t
se
Audit quality
AUDITFEE
t
se
39.288***
7.437
5.282
-1.410***
-3.843
.367
.045***
9.954
.005
-2.660**
-2.186
1.217
-6.306***
-5.528
1.141
1.046***
2.604
.402
1.538
1.244
1.237
-.236**
-2.407
.098
-1.101*
-1.953
.564
R2 .164 .702 .168
Sobel Test
ACMEETAUDITFEEACC
-1,818*
Notes:
***, **, *denote significance at 1%, 5% and 10% level respectively
The variables are as defined in Appendix 3
Acta Wasaensia 151
APPENDIX 14. All measures for audit committee effectiveness included and
control variables excluded.
Variable Condition 1
ACC
Variable Condition 2
AUDITFEE
Variable Conditions 3
and 4
ACC
Intercept
t
se
Audit committee
effectiveness
ACSIZE
t
se
ACEXP
t
se
ACMEET
t
se
3.574*
1.819
1.965
-.201
-.496
.404
.826
.624
1.325
-.358***
-3.454
.104
Intercept
t
se
Audit committee
effectiveness
ACSIZE
t
se
ACEXP
t
se
ACMEET
t
se
12.772***
82.487
.155
.303***
9.516
.032
.426***
4.083
.104
.055***
6.692
.008
Intercept
t
se
Audit committee
effectiveness
ACSIZE
t
se
ACEXP
t
se
ACMEET
t
se
Audit quality
AUDITFEE
t
se
35.912***
6.661
5.391
.567
1.370
.414
1.905
1.455
1.310
-.219**
-2.114
.104
-2.532***
-6.422
.394
R2 .012 .128 .051
Sobel Test
ACMEETAUDITFEEACC
-4.694***
Notes:
***, **, *denote significance at 1%, 5% and 10% level respectively
The variables are as defined in Appendix 3
152 Acta Wasaensia
APPENDIX 15. Results for model ACMEETINDFEEACC.
Variable Condition 1
ACC
Variable Condition 2
INDFEE
Variable Conditions 3
and 4
ACC
Intercept
t
se
TA
t
se
OPCYCLE
t
se
SALESG
t
se
LOSS
t
se
Audit committee
effectiveness
ACMEET
t
se
34.758***
7.133
4.873
-1.705***
-7.584
.225
.044***
9.741
.005
-2.746**
-2.255
1.217
-6.650***
-5.843
1.138
-.279***
-2.901
.096
Intercept
t
se
TA
t
se
INVREC
t
se
FOROPR
t
se
QR
t
se
LOSS
t
se
BSIZE
t
se
BIND
t
se
BMEET
t
se
Audit committee
effectiveness
ACMEET
t
se
-11.005***
-21.130
.521
.496***
19.152
.026
.863***
4.285
.201
.643***
5.147
.125
.020
1.243
.016
.303***
3.023
.100
.025
1.406
.018
.003
1.306
.002
.011
1.121
.010
.041***
4.715
.009
Intercept
t
se
TA
t
se
OPCYCLE
t
se
SALESG
t
se
LOSS
t
se
Audit committee
effectiveness
ACMEET
t
se
Audit quality
INDFEE
t
se
27.394***
4.539
6.036
-1.342***
-4.706
.285
.044***
9.760
.005
-2.834**
-2.330
1.216
-6.409***
-5.611
1.142
-.246**
-2.521
.097
-.721**
-2.061
.350
R2 .158 .440 .162
Sobel Test -1.877*
Notes:
***, **, *denote significance at 1%, 5% and 10% level respectively
The variables are defined as follows:
INDFEE = Ratio of audit fees relative to the industry mean
Other variables are as defined in Appendix 3
Acta Wasaensia 153
APPENDIX 16. Results for path analysis.
Figure 1. AMOS results χ2(1)= 0.204 (P=.652)
Paths Standardized
estimates
Unstandardized
estimates
Standard
error
Sig.
ACSUM AUDITFEE .365 .208 .018 ***
AUDITFEE ACC -.209 -2.450 .367 ***
Notes:
***, **, *denote significance at 1%, 5% and 10% level respectively
The variables are defined as follows:
ACSUM=Composite measure of audit committee effectiveness
Other variables are as defined in Appendix 3
ACSUM AUDITFEE ACC
-.21
e1 e2
.36
154 Acta Wasaensia
APPENDIX 17. Results for unexpected fees.
Variable
Model 1
ACC
Variable
Model 2
ACC
Intercept
t
Control variables
TA
t
OPCYCLE
t
SALESG
t
LOSS
t
Audit quality
UNEXPFEE_1
t
33.938***
7.001
-1.778***
-7.983
.043***
9.620
-2.873**
-2.364
-6.872***
-6.068
-2.392***
-3.767
Intercept
t
Control variables
TA
t
OPCYCLE
t
SALESG
t
LOSS
t
Audit quality
UNEXPFEE_2
t
33.942***
7.001
-1.778***
-7.982
.043***
9.612
-2.874**
-2.365
-6.871***
-6.066
-2.367***
-3.724
R2 .163 .163
Notes:
***, **, * denote significance at 1%, 5% and 10% level respectively
The variables are defined as follows:
UNEXPFEE_1=Unexpected audit fees (Residual from regression model 33)
UNEXPFEE_2=Unexpected audit fees (Residual from regression model 34)
Other variables are as defined in Appendix 3
Acta Wasaensia 155
APPENDIX 18. Results for moderation effect.
Variable
ACC
Intercept
t
Control variables
TA
t
OPCYCLE
t
SALESG
t
LOSS
t
Audit committee effectiveness
ACMEET
t
Audit quality
AUDITFEE
t
Interaction term
AUDITFEE*ACMEET
t
35.485***
5.601
-1.178***
-3.293
.045***
9.886
-2.834**
-2.325
-6.385***
-5.582
.113
.283
-.839
-1.355
-.026
-.927
R2 .162
Notes:
***, **, * denote significance at 1%, 5% and 10% level respectively
The variables are defined as follows:
AUDITFEE*ACMEET= Interaction term of ACMEET and AUDITFEE
Other variables are as defined in Appendix 3