Oct 14 Corporate Governance and Earnings

594
Corporate Governance: An International Review, 2009, 17(5): 594610
Corporate Governance and Earnings

Management: A Meta-Analysis
Emma Garca-Meca* and Juan P. Snchez-Ballesta
ABSTRACT
Manuscript Type: Review
Research Question/Issue: The goal of this paper is to meta-analyze the results of 35 studies that examine the effect on
earnings management of rms boards of directors and ownership structure. We examine whether differences in results are
attributable to moderating effects related to the system of corporate governance, the measurement of the governance
variable, or the particular specications of discretionary accruals models.
Research Findings/Insights: The ndings show that the variation in the results of previous studies on CEO duality and
audit committee independence are caused by sampling error. In addition, the measurement of dependent variable, discretionary accruals, and the corporate governance system moderate the association between earnings management and some
corporate governance variables.
Theoretical/Academic Implications: The measurement of variables, especially discretionary accruals, inuences the ndings found in previous studies. The ndings emphasize the need to explicitly consider the legal and institutional setting
when one analyzes the effect of mechanisms of corporate governance on discretionary accruals. Future research should
include matrix correlations, and consider detailed measures of earnings management and more attributes of boards of
directors in order to facilitate research using meta-analysis.
Practitioner/Policy Implications: The results suggest that board independence, board size, and audit committee independence can improve investor condence by constraining earnings management. Additional empirical evidence regarding
rened measures of ownership and board, specically board independence, would be very useful in gaining greater
understanding of how the different approaches to these constructs inuence earnings management.
Keywords: Corporate Governance, Audit Committee, Board of Directors, Ownership Issues, Ownership Structure,
Earnings Management, Agency Theory, Meta-Analysis
INTRODUCTION
contractual outcomes (Healey & Wahlen, 1999; Leuz, Nanda,

& Wysocki, 2003).
Accounting earnings are more reliable and more informative when managers opportunistic behavior is controlled
through a variety of monitoring systems (Dechow, Sloan, &
Sweeney, 1996; Wild, 1996). After several recent nancial
scandals, such as Enron, Xerox, or Worldcom, there has been
an international trend towards developing and implementing corporate governance mechanisms to ght against the
opportunistic behaviors that have undermined investors
credibility in nancial information. Corporate governance
attributes help investors by aligning the interests of managers with the interests of shareholders and by enhancing the
reliability of nancial information and the integrity of the
nancial reporting process (Watts & Zimmerman, 1986).
Although prior work has provided some insight into the
role of corporate governance, the results of similar studies
are frequently contradictory and there are several features
corg_753
he very nature of accounting accruals gives managers a

great deal of discretion in determining the earnings a
rm reports in any given period because of information
asymmetry between managers and owners. Managers can
manipulate earnings in order to maximize their own interests or to signal their private information, thus inuencing
the informativeness of earnings (Chung, Firth, & Kim, 2002;
Gul, Chen, & Tsui, 2003; Healy, 1985; Holthausen, Larcker, &
Sloan, 1995). Earnings management can be dened as the
alteration of a rms reported economic performance by
insiders either to mislead some stakeholders or to inuence
*Address for correspondence: Accounting and Finance Department, University of

Murcia, Campus de Espinardo 30100, Murcia, Spain. Tel:. +34-968367923; E-mail:
emmagar@um.es
594..610
2009 Blackwell Publishing Ltd

doi:10.1111/j.1467-8683.2009.00753.x
CORPORATE GOVERNANCE AND EARNINGS MANAGEMENT
595
of research that make it difcult to draw substantive conclusions (Larcker, Richardson, & Tuna, 2007). To test the
validity and generalizability of the substantial research
undertaken in this eld of research, it is necessary to
review, synthesize, and assess relevant empirical research.
Following Leonidou, Katsikeas, and Samiee (2002) such an
undertaking is important at this stage for three reasons.
First, most studies represent attempts aimed at investigating and testing only certain dimensions of corporate governance. Second, investigation efforts take place at different
moments and in varying legal contexts, with a possible
exogenous effect on the ndings. Third, research designs,
such as the measurement of some variables, may be
diverse. The conicting ndings of previous research limit
the theoretical and research development of this eld.
Taking the above into account, we have identied 35 relevant empirical studies that examine the relation between
earnings management and corporate governance. Our
objective in this paper is to integrate these results, achieve
a quantitative generalization, and nd effects or relationships that are not obvious from other ways of summarizing
research, such as narrative approaches. We will use the
meta-analysis technique, which is a quantitative review
methodology widely accepted in medical research and
other disciplines besides management. Where there are a
sufcient number of studies, most observers would be
more comfortable with conclusions drawn from a metaanalytic review rather than narrative approaches, as metaanalysis can account for sampling error and other statistical
artifacts in the data from the studies on which the analysis
relies (Hunter & Schmidt, 1990).
The benets of the meta-analysis, the recent interest in
corporate governance, together with the social importance of
the credibility in nancial information lead us to metaanalyze the relation between corporate governance and
earnings management.
Following Denis and McConnell (2003), we classify corporate governance mechanisms into two categories boards
of directors and ownership structure and we analyze the
effect on earnings management of several dimensions: (1)
Boards of directors: Board independence, board size, CEO
duality, and audit committee independence; and (2) Ownership structure: Insider ownership, concentration, and institutional ownership. Meta-analysis will allow us rst to
aggregate results across studies in order to obtain a robust
estimate of the relationship between each corporate governance variable and earnings management. The selection of
variables is based on the governance categories found in the
empirical research. Some mechanisms of corporate governance (e.g., CEO remuneration, family ownership) are not
analyzed because there are too few studies for meta-analysis
to be applied.
In addition, we also analyze whether differences in
studies are due to moderator effects such as the measurement of discretionary accruals (type and sign of the model);
the approach used to dene some corporate governance variables (ownership concentration and insider ownership); and
the system of corporate governance (Anglo-American, communitarian, or emerging system).
Accordingly, this paper addresses several research questions. What is the overall effect of the different corporate
governance attributes on earnings management? Are independent audit committees or blockholders more effective in
reducing earnings manipulation in Anglo-American countries in comparison to communitarian countries? Do the
results depend on the measurement of discretionary accruals? Are the ndings moderated by the measurement of the
governance variables?
The ndings show that in some mechanisms, such as
CEO duality and audit committee independence, the variations in results found in previous studies are due to
sampling error. The measurement of dependent variable,
discretionary accruals, is also a factor that explains differences in previous ndings. Specically, our results show
that board size and board independence only have a negative effect on earnings management with total accruals
models. This may suggest that rms with larger and more
independent boards usually have fewer discretionary
accruals choices related to asset depreciation. This suggests
that when different discretionary accruals models are used,
results can change considerably, which conrms that the
denition of variables matters, especially with constructs
such as earnings management.
The results do not support most corporate practice recommendations that strongly suggest the positions of board
chairman and CEO be held by different individuals. Yet we
do see that abnormal accruals are less pronounced in rms
with independent audit committees.
In addition, we nd signicant differences between
corporate governance systems with regard to the role of
independent directors, a mechanism that does not appear
to be efcient in constraining earnings management practices in communitarian and emerging countries. The
greater presence there of controlling shareholders and less
of a board tradition of defence against managers would
explain these results. Nevertheless, we are concerned
about the measure of board independence, overall in communitarian studies, where there are many fears that board
members are not independent of those who nominate
them.
We attempt to shed additional light for regulators, such as
the Organization for Economic and Corporate Development
Council or the Commission of the European Communities,
which are engaged in the formulation of guidelines for
improved corporate governance. The results suggest that
codes of good governance should explicitly consider the
institutional framework of a country, because the implementation of some good practices from other countries without
considering the origin of a countrys legal institution could
be ineffective. The ndings also support regulators attempts
in communitarian and emerging countries to improve the
independence of corporate boards.
The rest of the paper is as follows: in the second section,
we discuss the literature about corporate governance and
earnings management; then, we examine the possible moderators for the relationships analyzed. In the methodology
section we present the meta-analytic technique used and
the description of sample and variables. In the results
section we show the results of the meta-analyses for each
corporate governance variable, and we end with a discussion of results in a summary, discussion, and further
research section.
Volume 17
Number 5
September 2009
CORPORATE GOVERNANCE
596
CORPORATE GOVERNANCE AND

EARNINGS MANAGEMENT
We discuss earnings management from two perspectives
boards of directors and ownership structure.
Boards of Directors and Earnings Management

Boards of directors can play a signicant role in controlling
agency problems. From an agency perspective, the ability of
the board to act as an effective monitoring mechanism
depends on its independence of management (Beasley, 1996;
Dechow et al., 1996). According to Fama and Jensen (1983),
independent directors on boards make boards more effective
in monitoring managers and exercising control on behalf of
shareholders. The Securities and Exchange Commissions
Regulation 14A, Item 6b, sets the conditions under which
directors afliation with a rm must be disclosed in proxy
materials. Directors with the following relationships must be
identied: (1) employment by the corporation or an afliate
within the last ve years; (2) any family relationship closer
than second cousin; (3) afliation in the last two years with a
concern that has had a customer, supplier, banker, or creditor
relationship with the corporation; (4) afliation with an
investment banker that has performed services for the
company within two years or will do so within one year; (5)
holding control of corporate stock; and (6) association with a
law rm engaged by the corporation. Nevertheless, previous
research that analyzes the effect of board independence on
earnings management is not so specic in the description of
board independence, which is usually referred to nonexecutive directors. This limits the analysis of board independence to the category of non-executive directors, and restricts
the possibility of a deeper examination of how these different
roles may constrain managerial discretional behavior.
Results on the association between earnings management
and board independence in previous literature are conicting. While Davidson, Goodwin-Stewart, and Kent (2005) nd
empirical support for the effective role of independent
directors in constraining earnings management in Australian rms, Bradbury, Mak, and Tan (2006) in Singapore fail to
nd any association between earnings management and
board independence.
Another characteristic that is seen to inuence a boards
ability to monitor is board size. Boards can become less
effective in controlling management as board size increases
due to problems of coordination and communication
(Jensen, 1993). Nevertheless, results regarding the effect of
board size on earnings management are not so obvious.
Some authors nd a positive association between board size
and earnings management (Chin, Firth, and Rui [2006] for
313 rms in Hong Kong), and others a negative relation
(Xie, Wallace, and Dadalt [2003] for a sample of 110 US
rms) or even no relation (Bradbury et al. [2006] for rms in
Malaysia and Singapore).
Another important characteristic of boards is whether
roles of the chairperson and the chief executive ofcer (CEO)
are vested in different people. Corporate governance guidelines assume that a board is less able to perform a monitoring role when the CEO is also the chairperson of the board.
Volume 17
Number 5
September 2009
CEO duality indicates that less control is likely to be exercised over managements activities and behavior. Empirical
evidence on the association between CEO duality and
opportunistic managerial behavior, however, seems not to
support this theory since most authors do not nd any signicant relation (Bugshan, 2005; Cornett, Marcus, Saunders,
& Tehranian, 2006; Davidson et al., 2005).
In monitoring the nancial discretion of management, it is
the audit committee that is likely to provide shareholders
with the most protection in maintaining the credibility of a
rms nancial statements. Thus, independent audit committees can potentially improve the quality and credibility of
nancial reporting. The best practice standard establishes
100 per cent of independent directors, and although in most
studies the percentage of independent directors is higher
than 50 per cent, it does not reach 100 per cent. Whereas in
the samples of Yang and Krishnan (2005) for US and
Bugshan (2005) for Australia there are 82 per cent of independent directors in the audit committee, in Garca-Osma
and Gill de Albornoz (2007) for Spain the mean is 56 per cent.
Nevertheless, results in this area are also conicting. While
Klein (2002) reports a negative relation between earnings
management and audit committee independence, other
authors nd no association between both variables (Yang
and Krishnan [2005] in the USA, and Garca-Osma and Gill
de Albornoz [2007] in Spain).
Ownership Structure and Earnings Management

Agency theory predicts that low levels of insider ownership
imply a poor alignment of interests between management
and shareholders (Jensen & Meckling, 1976); that is, managers with little ownership may have incentives to manage
accounting numbers so as to increase earnings-based
compensation, relax contractual constraints, or avoid debt
covenants (Healy, 1985; Holthausen et al., 1995). Insider ownership can be seen as a way to constrain the opportunistic
behavior of managers, so the level of discretionary accruals is
predicted to be negatively associated with insider ownership
(Wareld, Wild, & Wild, 1995). The entrenchment hypothesis
states on the other hand that high levels of insider ownership
may be ineffective in prompting insiders to make valuemaximizing decisions, which may result in an increase in
earnings management (Cornett, Marcus, & Tehranian, 2008)
Other authors do not nd any signicant association between
insider ownership and earnings management (Bowen,
Rajgopal, & Venkatachalam, 2004; Koh, 2003).
The majority of empirical research on the effect of ownership concentration on earnings management asserts that
monitoring by owners improves the quality of managerial
decisions and thus increases rm value, because the presence of substantial blockholders leads to closer monitoring
of management, implying less opportunity for accruals management or earnings manipulation. Yeo, Tan, Ho, and Chen
(2002) examine the monitoring role played by external unrelated blockholders, which results in reducing the opportunity for earnings management, with a strong positive effect
on earnings informativeness. Davidson et al. (2005) and
Snchez-Ballesta and Garca-Meca (2007b) do not nd any
signicant relation.
597
More recently there has been more focus on the role of

institutional investors in monitoring, disciplining, and inuencing corporate managers. While Rajgopal, Venkatachalam,
and Jiambalvo, (2002) and Jiraporn and Gleason (2007)
suggest institutional investors serve as monitors, mitigating
earnings management behavior, others nd no relation
between institutional investors and earnings management
(e.g. Siregar and Utama, 2008).
Another moderator variable for the relation between

corporate governance and earnings management is the
measurement of discretionary accruals. Authors have used
different models to separate observed accruals into their
non-discretionary and discretionary components. The most
common are the Jones (1991) model and Jones modied
model (Dechow, Sloan, & Sweeney, 1995). Each includes two
versions of total accruals and working capital accruals (total
accruals = working capital accruals-depreciation and amortization expenses for the period).
The main difference between versions is that total accruals
models estimate non-discretionary accruals controlling for
the component of long-term accruals, i.e., the level of gross
property, plant, and equipment. Such a calculation is subject
to considerable subjectivity in the estimates of useful life
and residual value of xed assets. The working capital model
only focuses on short-term accruals, noting that depreciation
offers limited potential as a tool for systematic earnings
management, since changes in depreciation policy cannot be
made very frequently without attracting adverse attention
from the auditor or investors. Nevertheless, some studies
use different models than Jones (1991) in order to calculate
the abnormal component of accruals, e.g., the Healy (1985)
and DeAngelo (1986) models. Meta-analysis will allow us to
evaluate if the heterogeneity in this research eld is based on
estimates from these models and if corporate governance
mechanisms are equally effective in constraining discretional behavior with working capital accruals and long-term
accruals. Hence, the second hypothesis is:
MODERATORS FOR THE META-ANALYSIS

There are striking differences across countries in corporate
governance systems for a variety of reasons, including laws,
capital market characteristics, culture, history, and industrial
organization. These differences in the business and institutional environments of countries are noted in the law and
nance approach (LaPorta, Lopez de Silanes, & Shleifer,
1999). Control mechanisms function differently, depending
on the legal and institutional setting (Lpez-de-Foronda,
Lpez-Iturriaga, & Santamara-Mariscal, 2007). Leuz et al.
(2003) note that earnings management is expected to
decrease in investor protection countries because strong
protection limits insiders ability to acquire private control
benets, which reduces their incentives to mask rm performance. Thus, we will study whether corporate governance
mechanisms actively monitor and take actions that reduce
the incidence of earnings management when the incentives
for manipulations differ with the context. Millar, Eldomiaty,
Choi, and Hilton, (2005) classify corporate governance
systems into the triad of the shareholder-interest driven
Anglo-American business system, the stakeholder driven
communitarian system and the emerging business system.
This classication takes into account the reality of the
context, non-economic forces that inuence rm capabilities,
and management discretionary behavior.
While the Anglo-American system is characterized by a
widely dispersed ownership and the basic conict of interest
is between managers and shareholders, the governance
system typical of the communitarian system is characterized
by concentrated ownership and the basic conict is between
holding companies, banks, institutions and weak shareholders. In most emerging economies, their relationship-based
institutions have led to a business system characterized by a
concentration of ownership and control of corporations and
banks by families, a lack of transparency, and weak investor
protection. The differences between the three systems
described by Millar et al. (2005), and determined by the legal
systems and investor protection, are likely to inuence the
relationship between governance mechanisms and earnings
management.
In our approach, we study whether differences in study
results are attributable to moderating effects related to
systems of corporate governance, distinguishing between
Anglo-American, communitarian, and emerging systems,
on the basis of Millar et al. (2005). Considering the corporate
governance system as a moderating variable, we propose a
rst hypothesis:
H1: The corporate governance system moderates the relation
between corporate governance variables and earnings
management.
H2: The specic model of discretionary accruals (total accruals

versus working capital accruals, and Jones versus other models)
moderates the relation between corporate governance variables
and earnings management.
Previous meta-analyses suggest that studies should be
classied according to differences in the measurement of the
dependent variable to reduce the level of variance in results
(e.g. Dalton, Daily, Ellstrand, & Johnson, 1998). Apart from
the model selected, another important difference in this variable in primary studies is related to its sign. Researchers
have used either absolute values of abnormal accruals or
signed accruals. According to Wareld et al. (1995), absolute
abnormal accruals tend to measure the extent to which
managers knowingly pursue certain techniques to adjust
reported numbers. While this measures the size of the accruals, it loses information on the sign of the accruals. Therefore,
some studies run separate regressions depending on
whether the abnormal accruals are signed (e.g., Cornett et
al., 2006; Piot & Janin, 2005) or unsigned. We test whether in
spite of their being used as measures of the same concept in
most of the primary studies the different signs in the models
may be a moderator of its relation with earnings management. Accordingly, we examine whether differences in
results are due to this specication of discretionary
accruals models (signed versus unsigned/absolute value
models). Therefore, we propose the hypothesis:
H3: The sign of the discretionary accruals model (absolute
versus signed) moderates the relation between corporate governance variables and earnings management.
Volume 17
Number 5
September 2009
598
Another possible moderating variable in this eld of

research is the measurement of insider ownership, because
empirical studies use different measures to test the effect of
insider ownership on earnings management. Since previous
meta-analyzes suggest that the measurement of the explanatory variables may be a moderator effect (Dalton et al., 1998;
Snchez-Ballesta & Garca-Meca, 2007a), we test whether in
spite of being used as measures of the same concept in most
of the primary studies, the different operational denitions
of insider ownership may be a moderator of the relation
with earnings management. Although in their initial contribution Jensen and Meckling (1976) focused on managerial
ownership as incentive to align the interests of managers
and shareholders, this same argument of agency theory has
been extended to board members. Boards of directors are,
among other tasks, charged with monitoring and disciplining senior management. Thus, they are considered one of the
most important factors inuencing the integrity of the nancial accounting process (Anderson, Mansi, & Reeb, 2003;
Dalton, Daily, Trevis, & Roengpitya, 2003) identify different
categories of insider ownership used in empirical research
on protability: CEO equity, managerial equity, ofcer and
director equity, inside board equity, and outside board
equity. Most of the papers on earnings management analyzed only distinguish between board ownership and management ownership. The rst category includes ownership
by members of the board of directors. The second refers to
executive ofcer ownership. This separation will allow us to
test whether there are differences in earnings management
between board members and executive managers when they
are owners of the company. On the other hand, some papers
rely on ofcers and directors to capture insider equity
ownership, without distinguishing between both groups
(e.g., Gabrielsen, Gramlich, & Plenborg, 2002; Rajgopal et al.,
2002). Thus, due to this limitation, and following Dalton et al.
(2003) and Snchez-Ballesta and Garca-Meca (2007a), we
have also included this third category of insider ownership
in the meta-analysis, so we distinguish between managerial
ownership, board ownership, and ofcer and director
ownership.
H4: The measure of insider ownership (managerial versus
board versus both ofcer and director ownership) moderates the
relation between this variable and earnings management.
Finally, empirical studies use different measures to test the
effect of ownership concentration on earnings management.
Most authors use the largest shareholder as a proxy of ownership concentration (e.g., Davidson et al., 2005; Ding,
Zhang, & Zhang, 2007). Others also consider the percentage
of total shares held by the top 20 shareholders (Bugshan,
2005); the ve largest shareholders (Lpez Iturriaga &
Saona Hoffman, 2005); or ownership from the second to the
tenth shareholder (Liu & Lu, 2007). Since the results on the
effect of ownership concentration on earnings management
are heterogeneous, it is fruitful to test whether there are
differential effects across different measures of ownership
concentration. In other words, we test whether in spite of
being used as measures of the same concept, the different
operational denitions of ownership concentration may be a
moderator of the relation between concentration and earnings management. Such an investigation would identify the
Volume 17
Number 5
September 2009
measure of ownership concentration for which more pronounced effects may be found. We distinguish between the
largest shareholder and other blockholders, a category that
includes those denitions that consider more than one
shareholder.
H5: The measure of ownership concentration moderates the
relation between this variable and earnings management.
METHODOLOGY
Meta-analysis is a technique that allows rigorous integration
of the ndings of previous studies on a particular topic in
order to assess the overall effect of the studies. Literature
reviews that are simply narrative can be misleading because
different researchers may reach different conclusions about
a set of individual studies due to variations in characteristics
such as sample size, measurement of variables, and time
period (Hunter & Schmidt, 1990). The meta-analysis technique, however, allows researchers to evaluate the effect of
these different data characteristics (moderators) on the results
(Hunter & Schmidt, 1990; Rosenthal, 1991; Wolf, 1986).
Meta-analyses in corporate governance have studied the
effect of board composition and size on rm performance
(Dalton et al., 1998; Dalton, Daily, Johnson, & Ellstrand, 1999;
Rhoades, Rechner, & Sundaramurthy 2000), the relationship
between board leadership structure and performance
(Rhoades, Rechner, & Sundaramurthy, 2001) and the association between ownership structure and rm performance
(Dalton et al., 2003; Snchez-Ballesta & Garca-Meca, 2007a).
One research topic of interest in recent years has been the
effect of corporate governance mechanisms on earnings
management.
For the objectives sought here we perform several independent meta-analyses. Each one examines the relationship
between earnings management and one corporate governance variable.
Literature Search
We rst used different combinations of keywords to search
for articles that report ndings on the relation between earnings management and boards of directors and ownership
structure mechanisms of corporate governance. Keywords
used included earnings management, discretionary
accruals, nancial reporting quality, corporate governance, ownership structure, board of directors,
CEO, board independence, insider ownership, ownership concentration, audit committee, board size, and
institutional ownership to search databases and editorial
sources including the ISI web of Science, ScienceDirect, EJS
Ebsco, Blackwell, Emerald, ABI Inform, and SSRN. We also
consulted the major journals of accounting and nance that
typically publish this kind of research (The Accounting
Review; Contemporary Accounting Research; Journal of Accounting and Economics; Corporate Governance: An International
Review; Journal of Financial Economics; Journal of Business,
Finance and Accounting.). References in the most recent
articles were also examined to identify other sources. These
searches yielded a total of 66 published and unpublished
599
TABLE 1
Selection of Earnings Management Studies for Meta-Analysis
Initial sample
Criteria leading to exclusion of studies
CG variables not included in our meta-analysis (CEO compensation, Audit quality and
nancial expertise, wedge between control and ownership)
Other attributes of earnings (total accruals, value relevance, timeliness, persistence)
Results non transformable into r
Composite measures of corporate governance variables or accruals
Non English
Studies on specic events and rms
Final sample
studies with quantitative data on corporate governance variables and discretionary accruals.
As detailed in Table 1, the difference between the number of studies and the number of usable samples is due
to different reasons: corporate governance variables not
included, such as CEO compensation (Balsam, 1998; Bergstresser & Philippon, 2006), composite measures of accruals
or corporate governance variables (Dhaliwal, Naiker, &
Navissi, 2006; Larcker et al., 2007; Leuz et al., 2003); results
not transformable into r (Jaggi & Tsui, 2007; Peasnell, Pope
& Young, 2000, 2005). This reduced the initial sample to 35
studies and 81 individual correlations from 1995 to 2008
that examine the effect of corporate governance mechanisms on discretionary accruals:
Variables Analyzed
Most of the articles we identied use discretionary accruals
as a proxy for earnings management. Thus, we focus our
research on this measurement, and exclude other attributes
of earnings such as total accruals, the value relevance of
earnings, earnings timeliness, or earnings persistence (e.g.,
Beekes, Pope, & Young, 2004; Oei, Ramsay, & Mather, 2008).
Discretionary accruals, our focus of interest, are the abnormal component of accruals, i.e., the portion of accruals not
explained by different factors such as (in the Jones [1991]
original model) change in revenues, and the level of gross
property, plant, and equipment (the last component is
included only in total accruals models).
Two articles analyzing the effect of insider ownership on
discretionary accruals use as estimates of discretionary
accruals: (1) the difference between accruals and accruals
expected based on the average of the previous four
years (Gabrielsen et al., 2002); and (2) the Healy (1985) and
DeAngelo (1986) models (Wareld et al., 1995). All
the others calculate the abnormal component of accruals
using the original version of the Jones (1991) model or its
various modications, such as that of Dechow et al. (1995)
(see Table 2). On the other hand, some primary studies
use models of total accruals, while others use working
Number of
studies
Percentage
66
100%
(11)
16.67%
(8)
(5)
(3)
(2)
(2)
35
12.12%
7.58%
4.55%
3.03%
3.03%
53.03%
capital accruals, and some measure discretionary

accruals in absolute value while others use signed
accruals.
We analyzed the moderator effect played by these different operational denitions of discretionary accruals. This
moderator factor was only studied in those cases where
there were sufcient studies.
For the independent variables in primary studies, we
looked for board of directors and ownership mechanisms
with enough previous research to be meta-analyzed. CEO
remuneration and other mechanisms of corporate governance are not analyzed because there are too few studies to
apply meta-analysis. This led us to focus on particular
dimensions of boards and ownership board independence,
CEO duality, board size, and audit committee independence
for board of directors; and insider ownership, ownership
concentration, and institutional ownership for ownership
structure.
We also found that some of these constructs have different
operational denitions, as explained in the moderators
section. Insider ownership can be measured in terms of
ofcer and director ownership, board ownership, and
management ownership. Ownership concentration in most
articles is measured as the ownership held by the largest
shareholder. Others measure ownership concentration as
ownership held by the 20 top owners, the Herndahl index
from the second to the tenth largest shareholders, or ownership by other groups of shareholders. We focus on the
moderator effect played by these different operational denitions of independent variables in their association with
discretionary accruals, grouping the largest shareholder
against denitions that include more shareholders. Finally,
CEO duality in some studies is coded 1 if CEO = Chairman
(non-independence between the two roles), and 0 otherwise,
while in other studies it is coded 1 if the Chairman is not the
CEO (independence of the two roles). To be consistent and
to focus all studies on CEO duality, we change the signs of
the associations in the cases that code 1 if the Chairman is
not the CEO. In this way we meta-analyze the effect of CEO
duality on discretionary accruals.
Volume 17
Number 5
September 2009
Volume 17
Number 5
19881990
Period
September 2009
19931998
19911997
19972000
19932000
2000
19931997
19912001
19992001
Bowen et al. (2004)

Park and Shin (2004)
Bugshan (2005)
Cheng and Wareld (2005)
Davidson et al. (2005)
Hsu and Koh (2005)

Lpez and Saona (2005)
Piot and Janin (2005)
Jones Modied
Jones
Jones
Jones Modied with CF

Jones Modied
Jones Modied
Jones
Jones Modied
Healy (1985), De Angelo

(1986): accruals expected
(accruals 5 prev. years)
Jones Modied
Jones Modied
Accrualst-Accrualst-1
(average prev. 4 years)
Jones
Jones Modied
Jones Modied
Jones Modied
Jones
Jones Modied
Version
Model
USA (A)
USA (A)
Denmark (C)
USA (A)
Board
Insider, Institutional
Concentration
Insider
Insider
Concentration
Concentration, Insider
Insider
Concentration
Insider
Insider
Concentration
Institutional
Institutional
Insider
Insider
Ownership structure
Internal CG Variables
USA (A)
Board_indep, AC_indep
USA (A)
Singapore (A)
Australia (A)
Australia (A)
USA (A)
Size, Board_indep,
AC_indep
TA
USA (A)
CEO_dual
WCA Canada (A)
Board_indep
TA
Australia (A) Size, Board_indep
TA
USA (A)
TA
Australia (A) Board_indep, CEO_dual,
AC_indep
TA
Australia (A)
TA
Chile (E)
TA
France (C)
TA
TA
TA
TA
TA
WCA
TA
TA
TA
TA
Term
Country and
CG System
Ta: total accruals; wca: working capital accruals; AC_indep: audit committee independence; Size: board size; Board_indep: board independence; CEO_dual: CEO duality;
Institutional: institutional ownership; Insider: insider ownership; concentration: ownership concentration; CF: cash ow; BM: book to market. A: Anglo-American; C:
Communitarian; E: Emerging counries.
19921993
19891995
19901992
1993
19931997
1992, 1994, 1996
Klein (2002)
Rajgopal et al. (2002)
Yeo et al. (2002)
Gul et al. (2003)
Koh (2003)
Xie et al. (2003)
Cheng and Reitenga (2001) 19871996

Chung et al. (2002)
19881996
Gabrielsen et al. (2002)
19911995
Wareld et al. (1995)
Study
TABLE 2
Sample Studies
600
19961999
19942003
20002004
Setia-Atmaja et al. (2008)
Volume 17
Number 5
Jones Modied
with CF
Jones Modied
Jones Modied
Jones/Jones
CF/Marginal
Jones Modied
Jones Modied
USA (A)
USA (A)
Indonesia (E)
Japan (C)
TA
TA
TA/WCA Australia (A)
WCA
TA
CEO_dual, Size,
Board_indep,
Size, Board_indep,
CEO_dual
Board_indep
China (E)
Hong Kong (A) Size, Board_indep
Australia (A)
China (E)
CEO_dual, Board_indep
Spain (C)
WCA/TA Spain (C)
TA
TA
TA
TA
TA
TA
AC_indep
Board
Institutional
Ownership structure
Internal CG Variables
Insider
Institutional
Institutional
Concentration
Insider
Institutional
Insider, Concentration
Singapore and CEO_dual, Size,

Insider, Concentration
Malasia (E)
Institutional
Hong Kong (A) Size, Board_indep
Concentration
USA (A)
Size, Board_indep,
CEO_dual
Taiwn (E)
Board_indep
Insider
USA (A)
USA (A)
Country
TA: total accruals; WCA: working capital accruals; AC_indep: audit committee independence; Size: board size; Board_indep: board independence; CEO_dual: CEO
duality; Institutional: institutional ownership; Insider: insider ownership; concentration: ownership concentration; CF: cash ow; BM: book to market. A: AngloAmerican; C: Communitarian; E: Emerging counries.
19951996/
20002002
Teshima and Shuto (2008) 19911999
Siregar and Utama (2008)
19992001
2002
19992000
19951998
19992005
19992002
Ding et al. (2007)

Jaggi and Leung (2007)
Koh (2007)
Liu and Lu (2007)
Snchez-Ballesta and
Garca-Meca (2007b)
Garca-Osma and Gill de
Albornoz (2007)
Jiraporn et al. (2007)
Cornett et al. (2008)
Jones Modied
controlling
performance
Jones
Jones Modied
Jones Modied
Jones Modied
Jones Modied
WCA
TA
20002003
Jones Modied
Jones Modied
Chen et al. (2007)
Bradbury et al. (2006)
19931998
19932000
Term
Chin et al. (2006)

Cornett et al. (2006)
Version
Model
19962000
Jones
TA
20002001/ Jones modied with CF TA
20022004
and BM
19931998
Jones Modied with CF TA
Period
Yang and Krishnan (2005)

Ahmed et al. (2006)
Study
TABLE 2
Sample Studies (Continued)

601
September 2009
602
Effect Size and Study Selection

In meta-analysis the index used to represent and standardize
the ndings of primary studies is called effect size (Lipsey
and Wilson, 2001:34). We use the Pearson correlation coefcient (r) between corporate governance measures and discretionary accruals as effect size to integrate the results of
different studies. We need sample articles report the correlation coefcient between discretionary accruals and one of
the corporate governance variables analyzed. When r statistics were not reported, but other statistics transformable into
r statistics were, we used formulas given by Wolf (1986),
Rosenthal (1991), and Lipsey and Wilson (2001) to transform
the statistic into an r statistic. We obtained 39 correlations for
board of directors, of which 16 refer to board independence,
eight to board size, seven to CEO duality, and eight to audit
committee independence. For ownership we obtained 38
correlations overall, of which 19 correspond to insider ownership, nine to ownership concentration, and 10 to institutional ownership. The basic characteristics of the studies
included in our analysis are detailed in Table 2.
Meta-Analytic Technique
We use the meta-analytic technique developed by Hunter
and Schmidt (Hunter & Schmidt, 1990; Hunter, Schmidt, &
Jackson, 1982), which is commonly used in economics and in
other studies on corporate governance (Dalton et al., 1998,
1999, 2003; Rhoades et al., 2000, 2001). Thus, for each association between corporate governance variable and discretionary accruals we calculate the weighted mean correlation
coefcient as an estimate of the population mean correlation
(r), the total observed variance, the sampling error variance
and the population variance estimate.
When a study offered various correlations between discretionary accruals and one corporate governance variable
(correlations, for example, due to various measures of discretionary accruals: total accruals against working capital
accruals, or signed accruals against absolute accruals), we
rst used one correlation coefcient per study (the mean
correlation coefcient) in the overall meta-analysis in order
to maintain independence between observations (Hunter
and Schmidt, 1990). Then, in the subgroup moderator analyses of the measurement of variables, we use the original
correlation coefcients, maintaining one correlation per
study.1
Second, to evaluate whether the empirical correlations are
homogeneous, we use two tests: (1) The observed variance
explained by sampling errors, according to which if between
50 and 75 per cent of the observed variance across studies
(which are corrected only for sampling error) can be
explained by sampling error, we can conclude that there is
no true variance in the studies and thus the association is
unmoderated and homogeneous; and (2) the Q statistic of
homogeneity which follows a chi-square distribution whose
signicance would indicate rejection of the null hypothesis
of homogeneity.
The hypothesis of homogeneity will be rejected in many
cases, so in order to limit Type I error rates we use a random
effects model (Hunter & Schmidt, 2000; Overton, 1998;
Shadish & Haddock, 1994), a more conservative approach
Volume 17
Number 5
September 2009
than the xed effects one, and one which provides wider
condence intervals around the mean correlation.
In our rst analysis we do not correct for statistical artifacts that are different from the sampling error, such as
range restriction and measurement unreliability, because
this information was not provided in primary studies. Brierley (1999), Tosi, Werner, Katz, and Gmez-Meja (2000),
and Rhoades et al. (2001), among others, do not carry out
this correction either, since accounting data are supposed
to suffer less from reliability issues than psychological
constructs. Nevertheless, as argued by Sundaramurthy,
Rhoades, and Rechner (2005), since most governance
research is conducted on large rms, it is likely that some
range restriction exists. Therefore, in a second analysis, following Dalton et al. (1998, 1999, 2003), and Sundarmurthy et
al. (2005), we explore this issue by repeating the analysis at
different levels of reliability (.8 and .9), and do not nd
major changes in the results.
RESULTS
We summarize the results for each corporate governance
variable separately in Tables 3 to 9. In each table we offer rst
the results of the overall meta-analysis, and then, if the
homogeneity tests are rejected and indicate the presence
of heterogeneity, we search for moderators, splitting the
sample according to the measures and models of discretionary accruals, the measures of independent variables and the
corporate governance system. The individual meta-analyzes
of specic subgroups in each corporate governance variable
allow us to test the hypotheses we have posed.
Board Mechanisms
Table 3 shows the results of the meta-analysis of the effect of
board independence on discretionary accruals. An overall
meta-analysis is conducted for the 16 studies that examine
the association between board independence and discretionary accruals. The results show a weak negative association
(z = 1.80, p < .10) between both variables, which suggests
that greater board independence may constrain earnings
management.
Since the homogeneity tests are rejected, we deepen our
analysis in the search for moderators. First, we examine
whether heterogeneity in independence is due to the
measure of accruals. The results (working capital against
total accruals and signed accruals against absolute accruals)
show a weak negative association when discretionary
accruals are measured using the total accruals models
(z = 1.79, p < .10), which would weakly support H2, and
non-signicant associations for the rest of the measures of
discretionary accruals, although.there is still heterogeneity
in the correlations.
Second, we study the moderator effect played by the corporate governance system, and nd that in Anglo-American
countries independent directors are effective in constraining
earnings management (z = 3.06, p < .01). In communitarian
and emerging countries, however, we do not nd a signicant association with earnings management, although the
number of correlations in communitarians is small. These
603
TABLE 3
Meta-Analysis of the Impact of Board Independence on Earnings Management
Variable
Sample
Number of
correlations
Mean
correlation
(r)
%
S e/S2r
2
Condence
interval (95%)
Min
BOARD INDEPENDENCE
(1)
(2)
Board independence overall

15,155
Subgroups of Independence by measurement of
Working capital accruals
2,210
Total accruals
13,550
Signed Accruals
12,263
Absolute Accruals
6,130
Subgroups of Board independence by countries
Anglo-American
4,671
Comunitarian
410
Emerging
10,074
16
accruals
6
12
10
9
8
2
6
c2k-1
Max
(3)
(4)
(5)
(6)
-.031
21.635
-.066
.003
73.955***
.004
-.034
-.019
-.030
26.550
20.473
24.834
15.462
-.077
-.071
-.054
-.094
.085
.003
.017
.034
22.599***
58.613***
40.267***
58.208***
-.081**
.029
-.011
30.077
29.622
25.584
-.133
-.149
-.049
-.029
.207
.028
26.598***
6.752**
23.452***
p < .10; *p < .05; **p < .01; ***p < .001.
TABLE 4
Meta-Analysis of the Impact of CEO Duality on Earnings Management
Variable
Sample
Number of
correlations
Mean
correlation
(r)
%
S e/S2r
2
Min
CEO DUALITY
CEO DUALITY overall
Condence
interval (95%)
(1)
(2)
(3)
(4)
12,364
-.001
83.762
c2k-1
Max
(5)
-.019
(6)
.020
8.357
p < .10; *p < .05; **p < .01; ***p < .001.
different results by type of country would support H1 for

this variable. The ndings in Anglo-American countries
could be due to the higher tradition of a market of independent directors and the superior investor protection of these
countries. The ndings are in line with Kim, KitsabunnaratChatjuthamard, and Nofsinge (2007) results, which show
that shareholder protection rights and rms board independence are positively related. When a countrys minority
shareholder rights are strong, then minority shareholders
should have the legal power to affect board composition.
In Table 4 we display the results of the overall metaanalysis of the seven studies that examine the association
between CEO duality and earnings management. We do not
nd evidence of a true non-zero correlation. Thus, we
cannot support the hypothesis that rms with CEO duality
roles are involved in more earnings management than rms
whose CEO is independent of the chairman. As the variance explained by sampling error is 84 per cent, and the
homogeneity test is not rejected, we do not search for moderators, since the size of any moderator analysis would be
too small.
In Table 5 we show the results of the meta-analysis of the
association between board size and earnings management.
The results of the overall meta-analysis (eight studies) show
a negative and signicant effect (z = 2.81, p < .01) of board
size on discretionary accruals. When we divide the sample
according to the measures of accruals in order to search for
moderators, we nd this negative effect (z = 2.90, p < .01) in
total accruals models, but not in working capital models,
which supports H2 for this variable. Both absolute and
signed accruals maintain this negative association, but the
association for signed accruals is weak (z = 1.67, p < .10). The
system of corporate governance also inuences the association between board size and discretionary accruals (H1),
since the negative association (z = 2.62, p < .01) is maintained in Anglo-American. but not in emerging. countries.
Volume 17
Number 5
September 2009
604
TABLE 5
Meta-Analysis of the Impact of Board Size on Earnings Management
Variable
Sample
Number of
correlations
Mean
correlation
(r)
%
S e/S2r
2
Condence
interval (95%)
Min
BOARD SIZE
(1)
(2)
Board size overall

4,313
8
Subgroups of Board size by measurement of accruals
Working Capital Accruals
1,516
4
Total Accruals
3,247
5
Signed Accruals
2,136
5
Absolute Accruals
3,049
5
Subgroups of Board size by countries
Anglo-American
3,006
5
Emerging
1,307
3
c2k-1
Max
(3)
(4)
(5)
(6)
-.054**
62.358
-.092
-.016
12.829
-.051
-.077**
-.049
-.052**
64.368
42.913
53.135
93.518
-.113
-.129
-.108
-.088
.012
-.025
.009
-.015
6.214
11.651*
9.410
5.347
-.060**
-.042
64.253
63.168
-.104
-.110
-.015
.027
7.782
4.749
p < .10; *p < .05; **p < .01; ***p < .001.
TABLE 6
Meta-Analysis of the Impact of Audit committee on Earnings Management
Variable
Sample
Number of
correlations
Mean
correlation
(r)
%
S e/S2r
2
Condence
interval (95%)
Min
AUDIT COMMITTEE
INDEPENDENCE
Audit committee independence
(1)
(2)
(3)
(4)
3,662
-.058***
100.000
Max
(5)
-.087
c2k-1
(6)
-.029
6.391
p < .10; *p < .05; **p < .01; ***p < .001.
Table 6 displays the results of the association between

audit committee independence and discretionary accruals.
The ndings of the overall meta-analysis of eight studies are
strongly signicant (z = 3.92, p < .001), which shows that
independent audit committees are effective mechanisms in
limiting earnings management. Moreover, since the variance
explained by sampling error is 100 per cent and the
chi-square test is non-signicant, we do not search for
moderators.
Ownership Structure
Table 7 provides the results of the meta-analysis of the association between insider ownership and discretionary accruals. The overall meta-analysis of 19 studies does not show a
signicant association between the variables. In order to
reduce heterogeneity, we rst search for moderators according to the operational denition of insider ownership, ofcer
and director ownership, board ownership, and management
Volume 17
Number 5
September 2009
ownership. Only in board ownership do we nd a signicant

and negative effect on discretionary accruals (z = 2.93,
p < .01). The other two categories show non-signicant associations with discretionary accruals. This suggests that directors ownership is more effective than managers ownership
in limiting earnings management, and conrms H4. Moreover, the homogeneity test is not rejected in this category.
Second, we strengthen our analysis of the categories
where homogeneity tests are rejected ofcer and director
ownership and management ownership. For ofcer and
director ownership, since there are two estimation models
other than the Jones models, we looked to see if these introduced any bias in the results. Neither this subgroup nor the
Jones group show signicant associations with discretionary
accruals, as in the overall meta-analysis of ofcers and directors, and the same non-signicant ndings are obtained
when subgroups of signed and absolute accruals are considered. For management ownership, we could divide the
sample into signed accruals, which show a positive associa-
605
TABLE 7
Meta-Analysis of the Impact of Insider Ownership on Earnings Management
Variable
Sample
Number of
correlations
Mean
correlation
(r)
%
S e/S2r
2
Condence
interval (95%)
Min
INSIDER OWNERSHIP
(1)
(2)
(3)
Insider ownership overall

53,339
19
.001
Subgroups of insider ownership
Ofcer and director ownership
14,327
6
.011
Board ownership
19,587
6
-.025**
Management ownership
19,425
7
.018
Subgroups of ofcers and directors by Accruals models
Jones models
10,779
4
.020
Others models
3,548
2
-.015
Subgroups of ofcers and directors by measurement of accruals
Signed Accruals
4,134
2
-.0126
Absolute Accruals
14,327
6
.014
Subgroups of management ownership by measurement of accruals
Signed Accruals
15,548
5
.018*
Absolute Accruals
6,369
4
.022
Subgroups of management ownership by countries
Anglo-American
10,956
4
.030
Emerging
8,214
2
.002
(4)
c2k-1
Max
(5)
(6)
31.149
-.015
.016
60.997***
46.871
70.690
36.442
-.013
-.042
-.005
.035
-.008
.042
12.801*
8.488
19.209**
53.363
94.727
-.006
-.049
.046
.019
7.496
2.111
27.343
57.844
-.071
-.007
.046
.036
7.314**
10.373
98.099
27.091
.002
-.025
.033
.069
5.097
14.765**
27.951
100.000
-.005
-.014
.066
.019
14.311**
1.143
p < .10; *p < .05; **p < .01; ***p < .001.
tion (z = 2.18, p < .05), and absolute accruals, which is not

signicant. This would conrm H3 for insider ownership.
The study of the association between management ownership and discretionary accruals shows a weak positive association (z = 1.69, p < .10) in Anglo-American countries (H1).
Results of the meta-analysis for ownership concentration
are shown in Table 8. The overall meta-analysis of nine
studies shows a non-signicant association, but heterogeneity in correlations. We rst search for moderators according
to the operational denition of ownership concentration (the
largest shareholder against other blockholders), and the
results maintained the non-signicant association. Second,
we analyze differences in the denitions of accruals. Signed
and absolute accruals were also non-signicantly associated
with ownership concentration. While total accruals models
are non-signicant, working capital models show a positive
effect on ownership concentration (z = 8.71, p < .001),
although the research evidence is skimpy. Third, we divide
the sample according to the system of corporate governance.
In Anglo-American and emerging countries the association
between ownership concentration and discretionary accruals is not signicant.
In Table 9, we analyze the association between discretionary accruals and institutional ownership. The results of the
overall meta-analysis of ten studies are non-signicant. We
divide the sample according to the sign of accruals (signed
and absolute). The association of signed accruals with insti-
tutional ownership is weakly positive (z = 1.88, p < .1), but

the effect of the absolute value of accruals is negative (z = 4.56,
p < .001), conrming H3 for institutional ownership.
SUMMARY, DISCUSSION, AND FURTHER

RESEARCH
Codes of best practices around the world seek to move
towards a greater presence of independent directors on their
boards. Our results show that independent boards are more
effective in preventing managerial discretionary behavior,
although the effects of compliance with best practices affect
earnings management in some but not all countries. The
results show independent directors appear to be less effective in carrying out this theoretical role of constraining
earnings management in communitarian and emerging
economies. The monitoring role of independent directors
suggested by agency theory does not take place in those
corporate governance systems. Their weaker tradition of the
board-monitoring role may be the reason that board independence is not such a useful control mechanism in preventing earnings management there.
The results also suggest that countries with poor legal
environments might not have rms with desirable and
effective boards of directors, noting that the effect of board
independence on earnings management depends on inves-
Volume 17
Number 5
September 2009
606
TABLE 8
Meta-Analysis of the Impact of Ownership concentration on Earnings Management
Variable
Sample
Number of
correlations
Mean
correlation
(r)
%
S e/S2r
2
Condence
interval (95%)
Min
OWNERSHIP
CONCENTRATION
(1)
(2)
(3)
Ownership concentration overall

10,632
9
.013
Subgroups of ownership concentration
Largest shareholder
9,082
6
.026
Others blockholders
7,527
4
.011
Subgroups of ownership concentration by measurement of accruals
Signed accruals
8,727
5
.024
Absolute accruals
1,415
3
.009
Working capital accruals
821
2
.111***
Total accruals
9,811
7
.005
Subgroups of ownership concentration by countries
Anglo-American
2,033
4
.049
Emerging
8,396
4
.004
(4)
c2k-1
Max
(5)
(6)
26.768
-.024
.050
33.622***
35.768
21.548
-.009
-.037
.060
.060
16.775**
18.563***
36.205
100.000
100.000
28.097
-.011
-.038
.086
-.032
.059
.056
.135
.042
13.810**
2.432
0.271
24.914***
50.338
17.814
-.012
-.046
.110
.055
7.946*
22.455***
p < .10; *p < .05; **p < .01; ***p < .001.
TABLE 9
Meta-Analysis of the Impact of Institutional Ownership on Earnings Management
Variable
Sample
Number of
correlations
Mean
correlation
(r)
%
S e/S2r
2
Condence
interval (95%)
Min
INSTITUTIONAL OWNERSHIP
(1)
(2)
(3)
Institutional ownership overall

37,729
10
Subgroups of Inst. Own. by measurement of Accruals
Signed Accruals
25,917
6
Absolute value of Accruals
10,001
3
.005
.031
-.070***
(4)
c2k-1
Max
(5)
(6)
8.969
-.029
.039
111.500***
14.563
41.964
-.001
-.100
.062
-.040
41.199***
7.149*
p < .10; *p < .05; **p < .01; ***p < .001.
tor protection rights. Countries with better shareholder protection rights and enforcement should have empowered
minority shareholders who should be able to affect board
composition. This result contributes to the literature on the
association between corporate governance and disclosure by
testing how effective the governance recommendations
introduced by codes of best practice are at constraining earnings manipulation in Anglo-American countries.
For the rest of the countries, these ndings support the
importance of establishing a nomination process that guarantees that directors are selected using independent and
professional procedures. The ndings also encourage the
attempts of regulators in communitarian and emerging
Volume 17
Number 5
September 2009
economies to improve the independence of their corporate

boards.
Nevertheless, these results should be interpreted with
caution since there are few studies in the communitarian
system. This would be well worth investigation in future
research so as to conrm or reject these ndings with more
empirical evidence.
Another relevant limitation regarding the results of this
corporate governance attribute is its measure. Regulators,
commentators, and courts have all used independence to
mean different things at different times for different reasons.
In corporate governance literature most papers measure
board independence as the proportion of independent direc-
607
tors divided by the total number of directors of the board.

However, the denition of the independent director term
varies across theoretical literature. Some papers distinguish
between non-executive and executive director; other consider outsider versus insider directors. Although there is
limited evidence that denition of outside/independent
makes a difference (Rhoades et al., 2000), we were unable to
determine if there were differences between the denitions
used in the earnings management literature. We consider
that the term of board independence usually employed in
this literature refers to non-executive directors, although it
may depend on the country analyzed.
Apart from the fact that most of papers in the earnings
management eld do not give much information regarding
the measure of board independence, the studies analyzed
do not go deeply into the different approaches to measuring
board independence. This may conrm previous concerns
(Larcker et al., 2007) that in corporate governance literature
single indicators are used as measures for ill-dened and
complex corporate governance constructs. Additional
empirical evidence regarding rened measures of board
independence in terms of SEC regulation or those used in
the literature (Dalton et al., 1998; Rhoades et al., 2000)
would be very useful to gain greater the understanding of
how these different measures of board composition inuence earnings management. For example, some kinds of
directors (gray directors) have additional characteristics that
call into question the actual level of independence they
bring to the monitoring of managements nancial reporting decisions. This category is not usually considered in
most studies, and we are concerned about the possible
inclusion of these directors in the category of board independence, overall in communitarian studies, where there
are many fears that board members are not independent of
those who nominate them. A report made in 2007 by the
Financial Studies Association in Spain notes that only onethird of independent directors reported by Spanish listed
rms are objectively independent. This serves as a timely
reminder to regulators that if they wish to protect investors
it is their duty to ensure that the board of directors governing the companies are capable of independent supervision
of management.
The results conrm previous meta-analyses in other areas
that suggest that the measurement of variables is an important moderator effect that justies the heterogeneity found
in a research eld (Dalton et al., 1998). Measurement consequences occur not only with corporate governance variables,
but also with earnings management models. The negative
effect of board independence and board size on earnings
management occurs only with total accruals models, with
the implication that rms with independent and larger
boards usually have fewer discretionary accruals choices
related to asset depreciation, which would show a controlling role for these mechanisms in earnings manipulation.
This result may have occurred because depreciation is a very
visible tool for systematic earnings management, and
changes in depreciation policy cannot be made without
attracting adverse attention from a large and independent
board of directors. Our results suggest that accruals models
that do not consider long-term discretionary accruals choices
might lead us to make misleading inferences about the role
of these corporate governance mechanisms in earnings management behavior.

The ndings also provide evidence that the insider
ownership/earnings management relation is moderated by
the particular measure of the insider ownership variable.
When we focus on management ownership, we nd that
high levels of ownership are not signicant in aligning
owner and manager interests in avoiding manipulating of
results. When we focus only on board ownership, the
convergence-of-interests hypothesis suggested by agency
theory is conrmed. Thus, although in their initial development of agency theory Jensen and Meckling (1976) focus on
managerial ownership as an incentive to align the interests
of managers and shareholders, our ndings suggest that this
effect on earnings management is conrmed only in the case
of board members, suggesting that the two groups (directors
and managers) have different objectives. The meta-analytic
results suggest that board ownership can be seen as a
mechanism to constrain the opportunistic behavior of managers, and highlights the importance of considering the type
of insider ownership when one investigates its effect on
earnings management.
The ndings of the overall meta-analysis of the relation
between board size and earnings management suggest that
this governance mechanism limits managerial discretion;
that is, larger boards are more effective in preventing managerial discretionary behavior. Although this nding could be
contrary to what is to be expected according to Jensen (1993),
one reason behind these results may be that larger boards are
more likely to devolve responsibilities to board committee
members than are smaller boards. The formation of subcommittees of larger boards is likely to provide greater monitoring benets than smaller boards themselves (Klein, 2002).
Moreover, as boards become larger, they are likely to
include more independent directors with valuable experience (Xie et al., 2003).
In terms of CEO duality, our results show that in spite of
differences in the ndings of previous studies, these are due
to sampling error, and the real correlation between CEO
duality and discretionary accruals in the population is not
signicantly different from zero. Although most corporate
practice recommendations strongly suggest separating the
roles of board chairman and CEO, the evidence until now
does not support the view that independence of roles
favours control over managers discretionary accruals activities and behavior.
Audit Committee independence is found to be one of the
major corporate governance mechanisms in constraining
earnings management. This governance attribute is likely to
provide shareholders with the greatest protection in maintaining the credibility of a rms nancial statements. Since
the audit committee samples examined in our study are
comprised, on average, of a majority of independent directors, this nding supports both the published literature and
governance recommendations which suggest that audit
committees should consist exclusively of non-executive or
independent directors in order to constrain discretionary
accruals and to enhance the credibility of nancial statements (e.g., BRC, 1999; Menon & Williams, 1994).
One limitation of our results is the number of studies
available for some tests. Additional empirical evidence
Volume 17
Number 5
September 2009
608
would be most useful to conrm the relationships analyzed.

In this sense, the addition of new studies with different
correlation signs could inuence the estimates in some
groups with few correlations. New research would also help
us to evaluate the moderating role played by other variables
that could inuence the effectiveness of governance mechanisms in constraining earnings management, such as
changes in governance regulation (e.g., the Sabarnes-Oxley
legislation in the US). Another limitation is the bias of the
sample to Anglo-American countries (68.57 per cent), particularly the US (37.14 per cent), whereas the percentage of
studies in emerging and communitarian is, respectively,
17.15 per cent and 14.28 per cent. Differences in accounting
systems, rather than governance systems, could also have
explained some differences in results.
Further research in the area of ownership identity (ownership by different types of large shareholders other corporations, families, and government) would be fruitful, and
may provide more insight into blockholders and earnings
management relationships. Moreover, as many of the critical
decisions of boards are driven by committees, further
research on their roles and status would be rewarding to
provide evidence on the effect of board committees on rms
earnings management; examples might relate to their nomination and compensation, and to strategy, audit, and ethics
committees.
Finally, although much of the attention in earnings management has been directed towards large corporations, they
represent only a small part of the total business world. Consideration of small- and medium-sized enterprises in the
corporate governance-earnings management area may also
be fruitful.
Despite these limitations, we believe that this study makes
an important theoretical and practical contribution. First, the
paper shows that the measurement of variables, especially
discretionary accruals, inuences the ndings found in
previous studies on the association of this variable with
corporate governance factors. Consequently, if theoretical
differences and managerial motivations comprised in accruals models are not taken into account and looked at in more
depth in corporate governance studies, research faces a
severe limitation to precisely ascertain the theoretical knowledge on factors associated to discretionary accounting
choices. Second, the ndings emphasize the need to explicitly
consider the legal and institutional setting when one analyzes
the effect of mechanisms of corporate governance on discretionary accruals, suggesting that the implementation of some
good practices from other countries without considering the
origin of a countrys legal institution could be ineffective.
Finally, another important theoretical implication is the need
to consider matrix correlations and more attributes of board
(e.g., CEO compensation, audit expertise) in future research
in order to be included in future meta-analyses.
Regarding the practical implications, the results suggest
that some board independence, board size, and audit committee independence can improve investor condence by
constraining earnings management. This is especially important as corporate governance may be used as a key to help
restore investor condence in markets that have experienced
nancial problems, which, at the present, is a real problem in
most of the world economies. Additional empirical evidence
Volume 17
Number 5
September 2009
regarding rened measures of ownership and board, specically board independence, would be very useful to gain
greater understanding of how the different approaches to
these constructs inuence earnings management.
The ndings also support regulators attempts in communitarian and emerging countries to improve the independence of corporate boards, through the use of professional
nomination processes that guarantee that directors are
selected via independent procedures.
ACKNOWLEDGEMENTS
The authors gratefully acknowledge the helpful comments
and suggestions received from two anonymous referees and
from the Associate Editor during the review process. We also
thank the Research Agency of the Spanish Government for
nancial support (Project SEJ2007-61450/ECON).
NOTE
1. This is why the number of studies of the overall meta-analysis
does not agree with the sum of the correlations of the subgroups.
REFERENCES*
*Asterisk denotes studies that were examined within this
meta-analysis.
*Ahmed, K., Duellman, S., & Abdel-Meguid, A. 2006. The Sarbanes
Oxley Act, auditor independence and accounting accruals: An
empirical analysis. Working Paper SSRN, Syracuse University.
Anderson, R. C., Mansi, S. A., & Reeb, D. M. 2003. Founding family
ownership and the agency cost of debt. Journal of Financial
Economics, 68: 263285.
Balsam, S. 1998. Discretionary accounting choices and CEO compensation. Contemporary Accounting Research, 15: 229252.
Beasley, M. S. 1996. An empirical analysis of the relation between
the board of directors and nancial statement fraud. The
Accounting Review, 71: 443465.
Beekes, W., Pope, P., & Young, S. 2004. The link between earnings
timeliness, earnings conservatism and board composition: Evidence from the UK. Corporate Governance: An International
Review, 12: 4759.
Bergstresser, D. & Philippon, T. 2006. CEO incentives and earnings
management. Journal of Financial Economics, 80: 511529.
*Bowen, R., Rajgopal, S., & Venkatachalam, M. 2004. Accounting
discretion, corporate governance and rm performance. Working
Paper SSRN, University of Washington.
*Bradbury, M., Mak, Y., & Tan, S. 2006. Board characteristics, audit
committee characteristics, and abnormal accruals. Pacic
Accounting Review, 18: 4768.
BRC (Blue Ribbon Committee on Improving the Effectiveness of
Corporate Audit Committees). 1999. Report and recommendations of the Blue Ribbon Committee on Improving the Effectiveness of Corporate Audit Committees. New York: NYSE.
Brierley, J.A. 1999. Accountants job satisfaction: A meta-analysis.
British Accounting Review, 31: 6384.
*Bugshan, T. 2005. Corporate governance, earnings management,
and the information content of accounting earnings: Theoretical
model and empirical tests. Ph.D. Dissertation, Bond University,
Australia.
*Chen, K., Elder, R., & Hsieh, Y. 2007. Corporate governance and
earnings management: The implications of corporate governance
best practice principles for Taiwanese listed companies. Working
609
paper SSRN. Journal of Contemporary Accounting and Economics, National Cheng Kung University.
*Cheng, A. & Reitenga, A. 2001. Characteristics of institutional
investors and discretionary accruals. Working Paper, Bauer
Collage of Business, University of Houston. Journal of Applied
Business Research (forthcoming).
*Cheng, Q. & Wareld, T. 2005. Equity incentives and earnings
management. The Accounting Review, 80: 441477.
*Chin, K., Firth, M., & Rui, O. 2006. Earnings management, corporate governance, and the market performance of seasoned equity
offerings. Journal of Contemporary Accounting and Economics,
7398.
*Chung, R., Firth, M., & Kim, J. 2002. Institutional monitoring and
opportunistic rarnings management. Journal of Corporate
Finance, 8: 2948.
*Cornett, M., Marcus, A., Saunders, A., & Tehranian, H. 2006. Earnings management, corporate governance, and true nancial
performance. Working Paper, Southern Illinois University, Carbondale, IL.
*Cornett, M., Marcus, A., & Tehranian, H. 2008. Corporate governance and pay-for-performance: The impact of earnings management. Journal of Financial Economics, 87: 357373.
Dalton, D. R., Daily, C. M., Ellstrand, A. E., & Johnson, J. L. 1998.
Meta-analytic reviews of board composition, leadership structure, and nancial performance. Strategic Management Journal,
19: 269290.
Dalton, D. R., Daily, C. M., Johnson, J. L., & Ellstrand, A. E. 1999.
Number of directors and nancial performance: A meta-analysis,
Academy of Management Journal, 42: 674686.
Dalton, D., Daily, C., Trevis, S., & Roengpitya, R. 2003. Metaanalyses of nancial performance and equity: Fusion or confusion? Academy of Management Journal, 46: 1326.
*Davidson, R., Goodwin-Stewart, J., & Kent, P. 2005. Internal governance structures and earnings management. Accounting and
Finance, 45: 241267.
DeAngelo, L. 1986. Accounting numbers as market valuation substitutes: A study of management buy-outs of public stockholders. The Accounting Review, 61: 400420.
Dechow, P. M., Sloan, R. G., & Sweeney, A. P. 1995. Detecting
earnings management. The Accounting Review, 70: 193
225.
Dechow, P. M., Sloan, R. G., & Sweeney, A. P. 1996. Causes and
consequences of earnings manipulation: An analysis of rms
subject to enforcement actions by the SEC. Contemporary
Accounting Review 13(1): 136.
Denis, D. & McConnell, J. 2003. International corporate governance.
Journal of Financial and Quantitative Analysis, 38: 136.
Dhaliwal, D., Naiker, V., & Navissi, F. 2006. Audit committee nancial expertise, corporate governance, and accruals quality: An
empirical analysis. SSRN working paper series.
*Ding, Y., Zhang, H., & Zhang, J. 2007. Private vs. state ownership
and earnings management: Evidence from Chinese listed companies. Corporate Governance: An International Review, 15:
223238.
Fama, E. & Jensen, M., 1983. Separation of ownership and control.
Journal of Law and Economics, 26: 301325.
*Gabrielsen, G., Gramlich, J., & Plenborg, T. 2002. Managerial ownership, information content of earnings, and discretionary accruals in a non-US setting. Journal of Business, Finance and
Accounting, 29: 967988.
*Garca Osma, B. & Gill de Albornoz, B. 2007. The effect of the
board composition and its monitoring committees on earnings
management: Evidence from Spain. Corporate Governance: An
International Review, 15: 14131428.
*Gul, F.A., Chen, C. J. P., & Tsui, J. S. L. 2003. Discretionary accounting accruals, managers incentives, and audit fees. Contemporary
Accounting Research, 20: 441464.
Healy, J. 1985. The effect of bonus schemes on accounting decisions.

Journal of Accounting and Economics, 7: 85107.
Healey, P. M. & Wahlen, J. M. 1999. A review of the earnings
management literature and its implication for standard setting.
Accounting Horizons, 13: 365383.
Holthausen, R., Larcker, D., & Sloan, R. 1995. Annual bonus
schemes and the manipulation of earnings. Journal of Accounting and Economics, 19: 2974.
*Hsu, G. & Koh, P. 2005. Does the presence of institutional
investors inuence accruals management? Evidence form Australia. Corporate Governance: An International Review, 13: 809
825.
Hunter, J. E. & Schmidt, F. L. 1990. Methods of meta-analysis:
Correcting error and bias in research ndings. Beverly Hills, CA:
Sage.
Hunter, J. E. & Schmidt, F. L. 2000. Fixed effects vs. random effects
meta-analysis models: Implications for cumulative research
knowledge. International Journal of Selection and Assessment,
8: 275292.
Hunter, J. E., Schmidt, F. L., & Jackson, G. B. 1982. Meta-analysis:
Cumulating research ndings across studies. Beverly Hills, CA:
Sage.
*Jaggi, B. & Leung, S. 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: 2750.
Jaggi, B. & Tsui, J. 2007. Insider trading, earnings management, and
corporate governance: Empirical evidence based on Hong Kong
rms. Journal of International Financial Management and
Accounting, 18: 192222.
Jensen, M. 1993. The modern industrial revolution: Exit and
failure of internal control systems. Journal of Finance, 48: 831
880.
Jensen, M. & Meckling, W. 1976. Theory of the rm: Managerial
behavior, agency costs, and ownership structure. Journal of
Financial Economics, 3: 305360.
*Jiraporn, P. & Gleason, K. C. 2007. Delaware incorporation and
earnings management: An empirical analysis. Journal of Applied
Finance, 17: 4051.
Jones, J. 1991. Earnings management during import relief investigations. Journal of Accounting Research, 29: 193228.
Kim, K., Kitsabunnarat-Chatjuthamard, P., & Nofsinge, J. R. 2007.
Large shareholders, board independence, and minority shareholder rights: Evidence from Europe. Journal of Corporate
Finance, 13: 859880.
*Klein, A. 2002. Audit committee, board of director characteristics,
and earnings management. Journal of Accounting and Economics, 33: 375400.
*Koh, P. 2003. On the association between institutional ownership
and aggressive corporate earnings management in Australia.
*Koh, P. 2007. Institutional investor type, earnings management,
and benchmark beaters. Journal of Accounting and Public
Policy, 26: 267299.
La Porta, R., Lopez de Silanes, F., & Shleifer, A. 1999. Corporate
ownership around the world. Journal of Finance, 54: 471
517.
Larcker, D. F., Richardson, S. A., & Tuna, I. 2007. How important is
corporate governance? Working Paper, SSRN.
Leonidou, L. C, Katsikeas, C. S, & Samiee, S. 2002. Marketing
strategy determinants of export performance: A meta-analysis.
Journal of Business Research, 55: 5167.
Leuz, C., Nanda, D., & Wysocki, P. D. 2003. Earnings management
and investor protection: An international comparison. Journal of
Financial Economics, 69: 505527.
Lipsey, M. W. & Wilson, D. B. 2001. Practical meta-analysis. Applied
Social Research Methods, 49.
Volume 17
Number 5
September 2009
610
*Liu, Q. & Lu, Z. 2007. Corporate governance and earnings management in the Chinese listed companies: A tunnelling perspective. Journal of Corporate Finance, 13: 881906.
Lpez-de-Foronda, O., Lpez-Iturriaga, F., & Santamara-Mariscal,
M. 2007. Ownership structure, sharing of control, and legal
framework: International evidence. Corporate Governance: An
*Lpez Iturriaga, F. & Saona Hoffman, P. 2005. Earnings management and internal mechanisms of corporate governance: Empirical evidence from Chilean rms. Corporate Ownership and
Control, 3(1): 1729.
Menon, K. & Williams, J. D. 1994. The use of audit committees for
monitoring. Journal of Accounting and Public Policy, 13: 121
139.
Millar, C., Eldomiaty, T., Choi, C., & Hilton, B. 2005. Corporate
governance and institutional transparency in emerging markets.
Journal of Business Ethics, 59: 163174.
Oei, R., Ramsay, A., & Mather, P. 2008. Earnings persistence, accruals, and managerial share ownership. Accounting and Finance,
47: 475502.
Overton, R. C. 1998. A comparison of xed-effects and mixed
(random-effects) models for meta-analysis tests of moderator
variable effects. Psychological Bulletin, 97: 134147.
*Park, Y. W. & Shin, H. H. 2004. Board composition and earnings
management in Canada. Journal of Corporate Finance, 10: 431
457.
Peasnell, K. V., Pope, P. F., & Young, S. 2000. Accrual management
to meet earnings targets: UK evidence pre- and post-Cadbury.
Peasnell, K. V., Pope, P. F., & Young, S. 2005. Board monitoring and
earnings management: Do outside directors inuence abnormal
accruals? Journal of Business, Finance and Accounting, 32: 1311
1346.
*Piot, C. & Janin, R. 2005. Audit quality and earnings management
in France. Working Paper.
*Rajgopal, S., Venkatachalam, M., & Jiambalvo, J. 2002. Is institutional ownership associated with earnings management and the
extent to which stock prices reect future earnings? Contemporary Accounting Research, 19: 117136.
Rhoades, D. L., Rechner, P. L., & Sundaramurthy, C. 2000. Board
composition and nancial performance: A meta-analysis of the
inuence of outside directors. Journal of Managerial Issues, 12:
7691.
Rhoades, D. L., Rechner, P. L., & Sundaramurthy, C. 2001. A metaanalysis of board leadership structure and nancial performance:
Are two heads better than one? Corporate Governance: An
Rosenthal, R. 1991. Meta-analytic procedures for social research.
Newbury Park, CA: Sage.
Snchez-Ballesta, J. P. & Garca-Meca, E. 2007a. A meta-analytic
vision of the effect of ownership structure on rm performance.
Corporate Governance: An International Review, 15: 879893.
*Snchez-Ballesta, J. P. & Garca-Meca, E. 2007b. Ownership structure, discretionary accruals and the informativeness of earnings.
Corporate Governance: An International Review, 15: 677
691.
Volume 17
Number 5
September 2009
*Setia-Atmaja, L., Haman, J., & Tanewski, G. 2008. Earnings management, board structure and family ownership. Working Paper
SSRN, Monash University.
Shadish, W. R. & Haddock, C. K. 1994. Combining estimates
of effect size. In H. Cooper & L. V. Hedges (Eds.), The handbook
of research synthesis: 261281. New York: Russell Sage
Foundation.
*Siregar, S. & Utama, S. 2008. Type of earnings management and
the effect of ownership structure, rm size, and corporategovernance practices: Evidence from Indonesia. The International Journal of Accounting, 43: 127.
Sundaramurthy, C., Rhoades, D. L., & Rechner, P. L. 2005. A metaanalysis of the effects of executive and institutional ownership on
rm performance. Journal of Managerial Issues, 17: 494510.
*Teshima, N. & Shuto, A. 2008. Managerial ownership and earnings management: Theory and empirical evidence from Japan.
Journal of International Financial Management and Accounting, 19: 107132.
Tosi, H. L., Werner, S., Katz, K., & Gmez-Meja, L. R. 2000. How
does performance matter? A meta-analysis of CEO pay studies.
Journal of Management, 26: 301309.
*Wareld, T., Wild, J., & Wild, K. 1995. Managerial ownership,
accounting choices, and informativeness of earnings. Journal of
Accounting and Economics, 20: 6191.
Watts, R. L. & Zimmerman, J. L. 1986. Positive accounting theory.
New York: Prentice Hall.
Wild, J. 1996. The audit committee and earnings quality. Journal of
Accounting, Auditing and Finance, 11: 247276
Wolf, F. M. 1986. Meta-analysis: Quantitative methods for
research synthesis. Beverly Hills, CA: Sage.
*Xie, B., Wallace, N., & Dadalt, P. 2003. Earnings management and
corporate governance: The role of the board and the audit committee. Journal of Corporate Finance, 9: 295316.
*Yang, J. & Krishnan, J. 2005. Audit committees and quarterly earnings management. International Journal of Auditing, 9: 201219.
*Yeo, G. H. H., Tan, P., Ho, K. W., & Chen, S. S. 2002. Corporate
ownership structure and the informativeness of earnings.
Journal of Business, Finance and Accounting, 29: 10231046.
Emma Garca-Meca is a lecturer in the Department of

Accounting and Finance at the University of Murcia, Spain.
Her research focuses on nancial analysts, ownership structure, boards of directors, voluntary disclosure, intangibles,
and earnings management. She has published articles in
European Accounting Review, International Journal of Accounting, International Business Review, and some papers in Corporate Governance: An International Review.
Juan Pedro Snchez-Ballesta is a lecturer in the Department
of Accounting and Finance at the University of Murcia,
Spain. His research focuses on meta-analysis, ownership
structure, boards of directors, and earnings management. He
has published articles in International Business Review and
some papers in Corporate Governance: An International Review.

Oct 14 Corporate Governance and Earnings

Uploaded by

Copyright:

Available Formats

Oct 14 Corporate Governance and Earnings

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Oct 14 Corporate Governance and Earnings

Uploaded by

Copyright:

Available Formats

594

Corporate Governance: An International Review, 2009, 17(5): 594610

Corporate Governance and Earnings

contractual outcomes (Healey & Wahlen, 1999; Leuz, Nanda,

he very nature of accounting accruals gives managers a

*Address for correspondence: Accounting and Finance Department, University of

2009 Blackwell Publishing Ltd

CORPORATE GOVERNANCE AND EARNINGS MANAGEMENT

2009 Blackwell Publishing Ltd

CORPORATE GOVERNANCE AND

Boards of Directors and Earnings Management

Ownership Structure and Earnings Management

2009 Blackwell Publishing Ltd

CORPORATE GOVERNANCE AND EARNINGS MANAGEMENT

More recently there has been more focus on the role of

Another moderator variable for the relation between

MODERATORS FOR THE META-ANALYSIS

2009 Blackwell Publishing Ltd

H2: The specic model of discretionary accruals (total accruals

Another possible moderating variable in this eld of

2009 Blackwell Publishing Ltd

CORPORATE GOVERNANCE AND EARNINGS MANAGEMENT

2009 Blackwell Publishing Ltd

capital accruals, and some measure discretionary

Bowen et al. (2004)

Hsu and Koh (2005)

Jones Modied with CF

Healy (1985), De Angelo

Cheng and Reitenga (2001) 19871996

Wareld et al. (1995)

2009 Blackwell Publishing Ltd

2009 Blackwell Publishing Ltd

Setia-Atmaja et al. (2008)

TA/WCA Australia (A)

WCA/TA Spain (C)

Singapore and CEO_dual, Size,

Siregar and Utama (2008)

Ding et al. (2007)

Chen et al. (2007)

Bradbury et al. (2006)

Chin et al. (2006)

Yang and Krishnan (2005)

CORPORATE GOVERNANCE AND EARNINGS MANAGEMENT

Effect Size and Study Selection

2009 Blackwell Publishing Ltd

CORPORATE GOVERNANCE AND EARNINGS MANAGEMENT

Board independence overall

CEO DUALITY overall

different results by type of country would support H1 for

2009 Blackwell Publishing Ltd

Board size overall

Table 6 displays the results of the association between

ownership. Only in board ownership do we nd a signicant

2009 Blackwell Publishing Ltd

CORPORATE GOVERNANCE AND EARNINGS MANAGEMENT

Insider ownership overall

tion (z = 2.18, p < .05), and absolute accruals, which is not

2009 Blackwell Publishing Ltd

tutional ownership is weakly positive (z = 1.88, p < .1), but

SUMMARY, DISCUSSION, AND FURTHER