Unit VII Case Study(For Dr.Michelle_KM)
Corporate Codes of Ethics, National Culture, and Earnings Discretion: International Evidence
Chu Chen1 • Giorgio Gotti2 • Tony Kang3 • Michael C. Wolfe4
Received: 10 September 2015 / Accepted: 13 May 2016 / Published online: 6 June 2016
� Springer Science+Business Media Dordrecht 2016
Abstract This study examines the role of codes of ethics
in reducing the extent to which managers act opportunis-
tically in reporting earnings. Corporate codes of ethics, by
clarifying the boundaries of ethical corporate behaviors and
making relevant social norms more salient, have the
potential to deter managers from engaging in opportunistic
financial reporting practices. In a sample of international
companies, we find that the quality of corporate codes of
ethics is associated with higher earnings quality, i.e., lower
discretionary accruals. Our results are confirmed for a
subsample of firms more likely to be engaging in oppor-
tunistic reporting behavior, i.e., firms that just meet or beat
analysts’ forecasts. Further, codes of ethics play a greater
role in reducing earnings management for firms in coun-
tries with weaker investor protection mechanisms. Our
results suggest that corporate codes of ethics can be a
viable alternative to country-level investor protection
mechanisms in curbing aggressive reporting behaviors.
Keywords Corporate ethics policy � Code of ethics � Business ethics � Earnings discretion � Accruals
JEL Classifications G300 � L210 � M140 � M410
Introduction
Generally accepted accounting principles (GAAP) impart a
variety of accounting choices and judgments on managers.
Although this discretion is a necessary component of
financial reporting, it leads to concern from investors and
regulators since it can allow opportunistic managers to
deliberately misrepresent the financial performance of the
company. Prior studies have identified various firm- and
country-level determinants of opportunistic financial
reporting by managers (e.g., Leuz et al. 2003; Bowen et al.
2008; Han et al. 2010), but the role of corporate codes of
ethics, which has a clear implication for ethical manager
behavior, has not been separately documented. The pur-
pose of this study is to fill this gap in the literature and
examine the role of codes of ethics in reducing the extent to
which managers act opportunistically in reporting earnings.
A code of ethics is a formal document that states an
organization’s primary values and the ethical rules it
expects its employees to follow (Robbins 1988). Until
recently, codes of ethics were found primarily in American
companies; however, the number of companies in other
countries with a code of ethics is increasing (Boatright
2009; McDonald 2009). A recent KPMG survey (2014)
notes that a properly implemented code is an increasingly
important instrument for today’s companies, as they
Data Availability Data used in this study are available from public sources identified in the study. We thank EIRIS for providing data on corporate ethics policy.
& Tony Kang kangt@mcmaster.ca
Chu Chen
cchen34@ewu.edu
Giorgio Gotti
ggotti@utep.edu
Michael C. Wolfe
mcwolfe@vt.edu
1 Eastern Washington University, Cheney, WA 99004, USA
2 University of Texas at El Paso, El Paso, TX 79902, USA
3 McMaster University, Hamilton, Canada
4 Virginia Tech, Blacksburg, VA 24061, USA
123
J Bus Ethics (2018) 151:141–163
https://doi.org/10.1007/s10551-016-3210-y
contribute to a company’s strategic positioning, identity
and reputation, culture and work climate, and to its finan-
cial performance. While an important goal for a code of
ethics is to impact the decision-making of individuals in the
organization (Molander 1987; Frankel 1989; Stevens 1994;
Kaptein and Wempe 1998; Lere and Gaumnitz 2003),
many commentators believe that codes are merely window-
dressing designed to improve the company’s reputation and
minimize potential legal exposure and, thus, are unlikely to
have any observable impact on behavior (e.g., Lere and
Gaumnitz 2003; Kaptein and Schwartz 2008). Despite the
increasing prevalence of corporate codes of ethics around
the world and the apparent skepticism regarding their
effectiveness, there is a paucity of empirical evidence
regarding their impact on decision-making (e.g., Somers
2001; Singh 2011; Kaptein 2015). Furthermore, the evi-
dence that does exist provides mixed and inconclusive
results (e.g., Stevens 1994; Helin and Sandstrom 2007;
Kaptein and Schwartz 2008). 1
Based on the differing
opinions and lack of empirical evidence regarding their
efficacy, it is unclear whether a code of ethics will have any
impact on opportunistic financial reporting.
Despite the lack of conclusive empirical evidence sup-
porting the impact of codes of ethics on decision-making,
extant theory predicts such a relationship. While traditional
agency models assume that opportunistic managers are
motivated solely by self-interest and extrinsic rewards,
there is also evidence that managers’ behavior is shaped by
the desire to achieve social norms such as fairness and
reciprocity (e.g., Fehr and Schmidt 1999; Fehr and Gächter
2000; Cohen et al. 2007; Bosse and Phillips 2015). For
example, managers may be more likely to act fairly if they
perceive that they are being treated fairly in return (e.g.,
Bosse and Phillips 2015). Another possibility is that con-
textual cues such as codes of ethics can ‘‘activate’’ social
norms like fairness and reciprocity that help control
opportunistic behavior (e.g., Bicchieri 2006; Davidson and
Stevens 2012). To the extent this ‘‘activation’’ leads man-
agers to incorporate these social norms into their decision-
making, we predict that firms with a higher-quality code of
ethics in place will be less likely to engage in opportunistic
earnings management.
Our focus is on the relationship between codes of ethics
and opportunistic financial reporting. Not all reporting
discretion reflects opportunistic behavior, and, while some
may regard any earnings management activity from man-
agers as unethical, others take a more nuanced approach.
For example, Ball (2009) defines earnings management as
‘‘managers intervening in the reporting of their own
financial performance’’ (p. 280). This definition encom-
passes a wide range of practices ranging from actions that
are neither illegal nor violations of accounting rules to
instances of outright fraud. For example, while the under-
statement of the provision for bad debts or drawing down
reserves might be deemed ‘‘aggressive’’ accounting, it is
not necessarily a violation of GAAP (Dechow and Skinner
2000). While all of these actions ‘‘undermine the quality of
financial reporting to some degree,’’ not all would neces-
sarily be perceived as unethical behavior (Ball 2009).
Clearly, financial reporting that explicitly violates GAAP
(i.e., fraud) is unethical. However, an action does not
necessarily need to be illegal or violate GAAP to be
deemed opportunistic reporting behavior (e.g., Dechow and
Skinner 2000). The primary distinction between oppor-
tunistic earnings management, which is more likely to be
regarded as unfair and unethical, and other instances of
accounting discretion, which are necessary and expected, is
the intent of the manager.
For example, Healy and Wahlen (1999) define earnings
management to be as follows: ‘‘earnings management
occurs when managers use judgment in financial reporting
and in structuring transactions to alter financial reports to
either mislead some stakeholders about the underlying
economic performance of the company or to influence
contractual outcomes that depend on reported earnings
numbers’’ (p. 368). Schipper (1989) defines earnings
management as a ‘‘purposeful intervention in the external
financial reporting process, with the intent of obtaining
some private gain (as opposed to, say, merely facilitating
the neutral operation of the process)’’ (p. 92). It is this type
of opportunistic reporting behavior designed to ‘‘mislead
some stakeholders’’ to ‘‘influence contractual outcomes’’
with ‘‘the intent of obtaining some private gain’’ that is
most likely to be affected by social norms like fairness and
reciprocity and, to the extent they help to activate these
norms, to be affected by a firm’s code of ethics.
Although managerial intent cannot be observed directly,
opportunistic reporting will likely take place through the
most judgmental portion of earnings, e.g., discretionary
accruals (Jones 1991). Following extant research on earn-
ings management and earnings quality (DeFond and
Subramanyam 1998; Kothari et al. 2005) we use discre-
tionary accruals as our proxy for earnings management. If
managers from companies with higher-quality codes of
ethics are less likely to engage in opportunistic earnings
management, we expect to observe higher earnings quality
(i.e., lower discretionary accruals) for these firms. As
mentioned previously, not all accounting judgments and
estimates, including those that result in discretionary
accruals, are necessarily the product of opportunistic
1 For example, in their review of empirical studies examining the
effectiveness of business codes, Kaptein and Schwartz (2008) find
that 35 % of the studies found a strong positive relationship, 16 %
reported a weak positive relationship, 33 % found no significant
relationship, 14 % reported mixed results, and one study found
evidence of a negative relationship.
142 C. Chen et al.
123
behavior. Therefore, we also examine the effect of codes of
ethics for a subsample of firms that just meet or beat
analyst earnings forecasts. Since managers have strong
incentives to beat benchmarks, firms just beating earnings
targets are more likely to be engaging in opportunistic
earnings management (Dechow and Skinner 2000).
We also examine whether the relation between codes of
ethics and discretionary accruals varies with the strength of
country-level legal institutions. Legal systems protect
investors by conferring on them rights to discipline man-
agers (e.g., replace managers), as well as by enforcing
contracts designed to limit insiders’ control benefits (La
Porta et al. 1998; Dyck and Zingales 2004). Leuz et al.
(2003) find a greater use of earnings management in
weaker investor protection countries where both the prob-
ability of being detected and the magnitude of disciplinary
action are lower. External and internal governance mech-
anisms can act as either substitutes or compliments in
constraining opportunistic behavior (e.g., Misangyi and
Acharya 2014), and we present arguments for both a
complementary and substitutive relationship between codes
of ethics and investor protection mechanisms in con-
straining opportunistic reporting behavior. Thus, we make
no ex ante prediction regarding this relationship.
To address our research question, we use an international
sample which includes firms from 19 countries represented
in the EIRIS database. This source contains firm-level cor-
porate ethics policy information from different countries and
has been used in prior studies examining business ethics
(Scholtens and Dam 2007). Our results indicate that the
existence and extent of corporate codes of ethics and their
system for implementation are associated with higher earn-
ings quality (lower discretionary accruals). However, the
relation is observed only for firms headquartered in weaker
investor protection regimes. These result are confirmed for a
subsample of firms that just meet or beat analysts’ forecasts,
which is consistent with our main premise that codes of
ethics help constrain opportunistic earnings management
behavior. These results suggest that corporate codes of ethics
have the potential to be an effective governance tool for
shareholders to mitigate opportunistic reporting behavior,
particularly in the absence of strong legal infrastructure.
This study contributes to several strands of literature.
First, it contributes to the stream of literature that examines
the role of codes of ethics in corporate decision-making.
Prior evidence, which is based almost exclusively on sur-
veys and interviews covering a small-sample of firms from
a single country, provides mixed and inconclusive results
(e.g., Stevens 1994; Helin and Sandstrom 2007; Kaptein
and Schwartz 2008). 2 Our results, which show that the
existence and implementation of corporate codes of ethics
play an effective role in deterring opportunistic reporting
behavior, are based on a larger cross-country sample,
enhancing the external validity of the documented relation.
In addition, we show that the role codes of ethics play in
financial reporting varies with the institutional environment
in which the firm operates.
Second, we analyze different dimensions of corporate
codes of ethics and document that different aspects of a
code, e.g., the presence of a code, the way it is imple-
mented, and whether the code includes a policy on cor-
ruption and human rights, are all relevant in the corporate
financial reporting setting. This finding suggests that future
research that examines the role of codes of ethics in
financial reporting decisions need not be overly concerned
about controlling for their various sub-features.
Third, this study contributes to the prior literature on the
role of ‘‘soft’’ institutions, i.e., culture and ethics, on
earnings management behaviors in international capital
markets (Chih et al. 2008; Huang et al. 2008; Han et al.
2010). An important implication of our findings to inves-
tors and regulators around the world is that corporate codes
of ethics can be a viable mechanism for deterring oppor-
tunistic reporting behavior when country-level institutions
fail to deter such behavior and when incentives or moni-
toring mechanisms are too costly or ineffective.
The remainder of the paper is organized as follows: We
first provide a literature review and develop our hypothe-
ses, followed by our research design, results, sensitivity
tests, and conclusions.
Literature Review and Hypotheses Development
Prior Evidence on Corporate Codes of Ethics
The number of companies with corporate codes of ethics
has increased significantly over the past few decades.
While Kaptein (2004) found that only 52.5 % of the largest
two hundred companies in the world had a business code in
2001, a more recent study by KPMG indicates that 86 % of
the Fortune Global 200 companies had adopted a code by
2007. 3 Prior evidence also indicates that there are cross-
country differences regarding the prevalence and quality of
codes of ethics. The KPMG study indicates that, while all
of the North American companies and 80 % of the Euro-
pean countries in the Fortune Global 200 had adopted some
form of corporate code by 2007, only half of the countries
headquartered in Asia had a code in place. Scholtens and
Dam (2007) examine cross-country differences in the
quality of codes of ethics and find that firms from the 2 For example, Choi and Pae (2011) are based on an ethical
commitment survey conducted for 252 Korean firms in 2004. 3 See ‘‘Business Codes of the Global 200’’ (KPMG 2008).
Corporate Codes of Ethics, National Culture, and Earnings Discretion 143
123
United States, Australia, and the Netherlands score best.
Firms from Luxembourg, Singapore, and Hong Kong score
the worst. Using Hofstede’s (2001) cultural values defini-
tions, they also find evidence that the cultural values of
individualism and uncertainty avoidance are positively
associated with higher ethical policies, while masculinity
and power distance are negatively associated with the
quality of ethical policies.
The prior empirical evidence examining the effect of
codes of ethics on decision-making provides mixed and
inconclusive results (e.g., Stevens 1994; Helin and Sand-
strom 2007; Kaptein and Schwartz 2008). These studies
generally utilize surveys and interviews and focus on a
small-sample of firms. For example, Singh (2006) surveys
490 Canadian corporations and finds that 68 % believe that
their code of ethics has a positive effect on the company’s
profits. Choi and Pae (2011) survey 252 Korean firms and
find evidence that companies with a higher level of ethical
commitment are associated with better quality financial
reporting. Chih et al. (2008) document that firms that exhibit
more socially responsible corporate behavior manage earn-
ings less, i.e., their earnings are less smooth, and display less
loss avoidance. Huang et al. (2008) find that firms that
commit to ethical corporate behavior, i.e., those with more
independent boards, engage in less earnings management.
Other studies find little or no relationship between codes
of ethics and behavior. For example, based on their survey
of 315 U.S. companies, McKendall et al. (2002) fail to find
evidence that firms with well-developed codes have fewer
legal violations. Similarly, Mathews (1987) examines a
sample of manufacturing firms in the U.S. and finds little
evidence of a relationship between codes of ethics and
violations of government regulations.
Hypothesis Development: Corporate Codes of Ethics
and Earnings Discretion
In traditional agency models, individuals are assumed to be
motivated solely by self-interest. When both the incentive
and opportunity (through information asymmetry) are
presented, a rational individual will choose to act oppor-
tunistically (Baiman 1982). Principals can attempt to con-
trol this opportunistic behavior through either increased
monitoring or by more closely aligning the agent’s material
interests with their own. While traditional agency theory is
pervasive and provides a framework for explaining many
dimensions of the contracting environment, its assumption
that narrow self-interest is the sole determinant of man-
agement behavior has been criticized for its self-fulfilling
effect on social norms (e.g., Ferraro et al. 2005) and its lack
of empirical validity (Eisenhart 1989).
One of the potential issues with the typical agency
model is its focus solely on extrinsic monetary rewards
while ignoring intrinsic-based rewards such as personal
satisfaction and ethical engagements with contracting
partners (Cohen et al. 2007). Several studies have shown
that managers’ behavior is also shaped by the desire to
achieve social norms such as fairness and reciprocity (e.g.,
Fehr and Schmidt 1999; Fehr and Gächter 2000). Cohen
et al. (2007) find evidence consistent with managers being
less likely to undertake a potentially opportunistic action
when they perceive the action to be unfair. Bosse and
Phillips (2015) present an agency-theory based model that
assumes agents are ‘‘boundedly’’ self-interested rather than
‘‘narrowly’’ self-interested and argue that ‘‘boundedly’’
self-interested actors will seek to maximize their own self-
interest but only as long as perceived norms of fairness and
reciprocity are not violated.
Previous studies have also identified mechanisms that
could affect the likelihood that managers will move beyond
‘‘narrow’’ self-interest and incorporate these types of social
norms into their decision-making. For example, managers
may be more likely to act fairly if they perceive that they
are being treated fairly in return (e.g., Bosse and Phillips
2015). Another possibility is the use of external cues to
facilitate desirable ethical behavior. For example, Mazar
et al. (2008) find that experimental participants are less
likely to behave dishonestly when they pay attention to
honesty standards such as The Ten Commandments and
honor codes. In a principal/agent setting, codes of ethics
can also serve as a contextual cue to induce ethical
behavior by the agent.
A code of ethics is a formal document that states an
organization’s primary values and the ethical rules it
expects its employees to follow (Robbins 1988). It sets out
the values that underpin the code, describe the company’s
obligation to its stakeholders, and provide details of how
the company plans to implement its values and vision, as
well as guidance to staff on ethical standards and how to
achieve them (Ladd 1991). Such codes help corporate
stakeholders understand the difference between ‘right’ and
‘wrong’ and induce (deter) right (wrong) behaviors. As a
result, a code of ethics can effect a change in action by
changing the individual’s beliefs as to whether an action is
ethical or not.
Utilizing Bicchieri’s (2006) model of social norm acti-
vation, Davidson and Stevens (2012) assert that a code of
ethics can affect this change in beliefs by activating social
norms (such as fairness and reciprocity) that help control
opportunistic behavior. The activation of the relevant
norms can occur if the code of ethics makes relevant
behavioral rules more salient, increases the belief that a
large subset of the population conforms to these behavioral
rules, and increases the belief that the behavioral rules are
valid or reasonable (Bicchieri 2006; Davidson and Stevens
2012). To the extent these norms become internalized as a
144 C. Chen et al.
123
result of this activation, they create a benchmark from
which managers can evaluate the ethical implications of
their actions.
When managers face the decision to behave oppor-
tunistically, they likely weigh the expected benefits and
costs of their actions. If managers are completely self-in-
terested, their focus will likely rest solely on the magnitude
of the rewards, the probability of being caught, and the
magnitude of the punishment (e.g., Mazar et al. 2008).
However, it is also possible that managers will consider
other issues like fairness and reciprocity to the extent they
have become salient factors in the decision-making pro-
cess. If contextual cues can facilitate this process by
helping to activate social norms, we predict that the exis-
tence of a high-quality code of ethics within the organi-
zation will play a role in constraining potential
opportunistic behavior.
The quality of a firm’s code of ethics is determined by
both the comprehensiveness of the code and the compre-
hensiveness of the system for implementing the code, both
of which are important determinants of a code’s potential
effectiveness. Regarding the comprehensiveness of the
code itself, Paine et al. (2005) examine several sources of
conduct guidelines for multinational companies including
business sector codes, the Sarbanes–Oxley Act, SEC reg-
ulations, and the NYSE and Nasdaq corporate governance
rules and find that nearly all enjoin companies to obey the
law, protect the environment, avoid bribery, and conduct
business in a truthful manner. Other consistent themes
include the disclosure of relevant information in a timely
manner, keeping accurate records, honoring agreements,
respecting human dignity and human rights, protecting
health and safety, and contributing to society through
innovation. 4 A comprehensive, and thus high-quality, code
is likely to significantly address several, if not all, of these
issues. Similarly, a comprehensive system for implement-
ing a code likely includes an extensive program for com-
municating the code and educating employees, a system for
code enforcement through appropriate disciplinary mea-
sures, appropriate monitoring, auditing, and whistleblow-
ing systems, and provisions for changing the code as new
situations and challenges arise (Boatright 2009). A high-
quality implementation system is likely to strongly address
these issues as well as many others.
This discussion leads us to the following set of
hypotheses:
H1a Firms with a higher-quality code of ethics are less
likely to engage in opportunistic earnings management.
H1b Firms with a higher-quality code of ethics imple-
mentation system are less likely to engage in opportunistic
earnings management.
Hypothesis Development: The Mediating Role
of Investor Protection
We next examine whether the posited relationship between
codes of ethics and opportunistic financial reporting varies
based on the level of investor protection. The prior litera-
ture has introduced several empirical measures designed to
capture differences in investor protection around the world.
For example, anti-director rights and anti-self-dealing
measures are designed to capture the strength of legal
protections for minority shareholders against corporate
insiders’ expropriation and self-dealing transactions (e.g.,
La Porta et al. 1998; Djankov et al. 2008). Previous studies
have also examined differences relating to the country’s
legal system (code law vs. common law) and found evi-
dence consistent with common law countries being asso-
ciated with stronger investor protection mechanisms (La
Porta et al. 1998).
Several studies have utilized these proxies to investigate
the effect of investor protection mechanisms on accounting
outcomes. For example, Leuz et al. (2003) find that earn-
ings management is negatively related to the strength of
investor protection mechanisms. The authors argue that
insiders use financial reporting discretion to overstate
earnings and conceal unfavorable earnings realizations in
an attempt to conceal private control benefits and reduce
the likelihood of outside intervention. Since strong legal
systems protect investors by conferring on them rights to
discipline managers (e.g., replace managers), as well as by
enforcing contracts designed to limit insiders’ control
benefits (Dyck and Zingales 2004; La Porta et al. 1998), the
incentive to manage earnings is much stronger in countries
where the legal protection of outside investors is weak.
Han et al. (2010) use Gray’s (1988) model to examine
the effects of both culture and investor protections on
earnings management. Gray’s (1988) model posits that
accounting outcomes are a product of culture and the
interaction of culture and legal institutions. Utilizing this
model, Han et al. (2010) predict that earnings management
will be greater in countries where individualism is the
dominant culture and lower in countries where uncertainty
avoidance in the dominant culture. They find support for
both predictions. They also find evidence that managers
from both high individualistic and strong uncertainty-
avoidant cultures are more likely to manage earnings when
they are located in stronger investor protection regimes.
As a result of these prior studies indicating that earnings
management is mitigated by strong investor protection
mechanisms, we predict that the association between codes
4 The authors identify eight basic governing principles: fiduciary,
property, reliability, transparency, dignity, fairness, citizenship, and
responsiveness.
Corporate Codes of Ethics, National Culture, and Earnings Discretion 145
123
of ethics and earnings management will vary based on the
strength of the country’s legal environment. However, we
make no ex ante prediction regarding this relationship. The
previous literature on corporate governance has examined
several mechanisms that reduce management’s propensity
to act opportunistically. This includes mechanisms which
are external to the organization such as institutional
shareholders and the strength of the legal system as well as
internal mechanisms such as board monitoring and codes
of ethics. However, the evidence regarding the interaction
between these external and internal mechanisms is incon-
clusive, as prior studies have found evidence of both a
complementary effect and substitutive effect (e.g., Mis-
angyi and Acharya 2014). Thus, it is unclear whether codes
of ethics will play a greater role in constraining oppor-
tunistic behavior in stronger (complementary effect) or in
weaker (substitution effect) investor protection countries.
Regarding the complement view, in the Bicchieri
(2006) model of social norm activation, conformity to a
social norm is dependent on the expectations of others’
behaviors and/or beliefs regarding the norm. Furthermore,
although legal norms and social norms are separate con-
cepts, social norms can be made more explicit in the
presence of supporting laws (Bicchieri 2006). Thus, social
norms may be more likely to be activated in countries
where there are already laws and other systems in place to
impact expectations and beliefs. In addition, enforcement
provisions potentially play a large role in the effectiveness
of a firm’s code of ethics (Lere and Gaumnitz 2003). To
the extent that enforcement mechanisms are more likely to
be in place and/or enforced in stronger legal environ-
ments, this would also result in the effect of codes of
ethics being observed in countries with stronger legal
protections mechanisms.
On the other hand, if the probability of managers being
caught and punished is lower in countries with weaker
investor protection mechanisms, affecting the manager’s
behavior through the activation of social norms such as
fairness and reciprocity may play a greater role in reducing
opportunistic behavior. Put simply, if external mechanisms
are not present to deter opportunistic behavior, internal
mechanisms like codes of ethics may become much more
salient. Also, a code of ethics must have a significant effect
on a manager’s beliefs in order to induce a change in
behavior (Lere and Gaumnitz 2003). Such a significant
impact may be more likely to occur where other potential
cues, such as laws, are not present to impact the manager’s
prior ethical beliefs.
To summarize, the complement (substitution) view
suggests that codes of ethics will play a greater role in
constraining opportunistic earnings management in stron-
ger (weaker) investor protection countries. Since it is not
clear ex ante which of these two competing views is more
likely to be observed, we state our next hypothesis in a
non-directional manner:
H2 The association between codes of ethics and oppor-
tunistic earnings management varies with the strength of
investor protection.
Research Design
Data and Sample Selection
This section of our paper introduces the data we gather
about codes of ethics, cultural values, and accounting
information that we use in our analysis. Following prior
research (Scholtens and Dam 2007), we derive data about
codes of ethics from the Ethical Investment Research
Service (EIRIS). EIRIS is a foundation established in the
UK in 1983 with the mission to provide independent
assessments of environmental, social, and governance
performance. EIRIS research covers, from 2003 to 2012,
about 3500 firms from 40 different countries on more than
110 different environment, social, and governance areas. It
gathers data annually through questionnaires and surveys
across six different areas: environment, governance, human
rights, positive products and services, stakeholders’ issues,
and ethical concerns. EIRISs main research areas include
environmental issues, social issues, human rights, animal
welfare, and other traditional negative issues such as
gambling and tobacco. 5 Following Scholtens and Dam
(2007), we focus on the existence and comprehensiveness
of a firm’s code of ethics, the existence and comprehen-
siveness of the firm’s system for implementing its code of
ethics, whether the firm has policies and procedures on
bribery and corruption, and the extent of the firm’s policy
addressing human rights issues. 6
We also compute a measure of corporate governance
using EIRIS data. Since EIRIS did not begin compiling the
corporate governance data for U.S. firms until 2004, we
utilize corporate governance data from 2004 for U.S.
observations occurring in 2003. To perform our empirical
tests, we require accounting data from Compustat Global
Vantage for the period from 2002 to 2012. We gather
Hofstede’s cultural values scores from Hofstede (2001) and
utilize measures of inflation and GDP growth from the
World Bank. Our final sample includes data from 2003 to
2012 for 9826 firm-year observations from 19 countries.
5 Further information on EIRIS can be obtained at their website:
http://www.eiris.org. 6 The area related to the firm’s communication of its code of ethics
cannot be used in our study because EIRIS dropped the questionnaire
inquiring about this issue in 2006.
146 C. Chen et al.
123
Measurement of Variables
Discretionary Accruals
Following extant research on earnings management
(Kothari et al. 2005; DeFond and Subramanyam 1998), we
use discretionary accruals as our proxy for earnings man-
agement. We measure discretionary accruals using the
performance-controlled Jones model (1991) introduced by
Kothari et al. (2005), which has been used extensively in
the prior literature.
We first estimate the following model using an ordinary
least square regression (firm subscripts omitted):
TACCt
TAt�1 ¼ a0
1
TAt�1
� � þ a1
DREVt TAt�1
� �
þ a2 PPEt
TAt�1
� � þ a3ROAt þ e
ð1Þ
where TACC is total accruals in period t, DREV is the change in revenue for period t, PPE is the level of property,
plant, and equipment in period t, and ROA is return on
assets (net income deflated by the beginning of the year
total assets) which controls for the effect of performance on
discretionary accruals. Following prior literature (e.g., Han
et al. 2010), we deflate all variables in the model by the
lagged book value of total assets to address the potential
effect of heteroskedasticity in the error term. We use the
residuals from this model as our measure of discretionary
accruals. In our analysis, we use the absolute value of
discretionary accruals, positive discretionary accruals (in-
come-increasing), and negative discretionary accruals (in-
come-decreasing).
Measures of Codes of Ethics
We derive our measures of the quality of a firm’s code of
ethics from the EIRIS database (Scholtens and Dam 2007).
EIRIS gathers information from a variety of sources. These
includes:
– Reading and analyzing company public documents and
public filings, including annual reports, websites,
sustainability reports, and other publications;
– Collecting responses to annual surveys and profile
mailings that companies send to EIRIS;
– Using independent regulatory sources, including rele-
vant regulatory data;
– Analyzing media coverage and press releases;
– Reading reports and interacting with organization and
NGOs;
– Interaction with company representatives, without
compromising the independence and autonomy of
EIRIS.
To assess each firm, EIRIS has a scoring table that we
use to code the variables, i.e., EIRIS assigns grades on
specific attributes. As Scholtens and Dam (2007) point out,
the procedure is not free from subjectivity. 7 EIRIS classi-
fies firms based on their answers to the following questions:
1. Systems Does the firm have a code of ethics and, if so,
how comprehensive is it? The possible answers are no
(coded = 0), limited (coded = 1), basic (coded = 2),
intermediate (coded = 3), and advanced (coded = 4).
2. Implementation of the firm’s code of ethics: Does the
firm have a system for implementing a code of ethics
and, if so, how comprehensive is it? The possible
answers are no (coded = 0), limited (coded = 1),
basic (coded = 2), intermediate (coded = 3), and
advanced (coded = 4).
3. Corruption Does the company have policies and
procedures on bribery and corruption? Here, the
possible responses are the firm has no policy disclosed
(coded = 0), it has adopted a policy (coded = 1), or it
has a clear policy and procedures (coded = 2). 8
4. Human rights On this issue, EIRIS includes answers to
two different questions: (a) What is the extent of policy
addressing human rights issues? And (b) what is the
extent of policy and systems addressing human rights
issues? In this study, we use the more comprehensive
answer to question (b), but results using question
(a) are qualitatively similar to the results reported
using question (b).
EIRIS launched a new grading system on human rights
in the fourth quarter of 2007. The new system consists of
five levels of responses (up from four, as used by Scholtens
and Dam (2007). The ratings for the old measure are still
available until 2010; for comparability, we modified the
new measure for 2011 and 2012 to match the former
ranking method. (For the new measure, we coded the
answer ‘‘No evidence of’’ equal to 0; ‘‘Limited’’ as equal to
1; ‘‘Intermediate’’ as equal to 2; ‘‘Advanced’’ or ‘‘Good’’ as
equal to 3. For the old measure, we coded similarly). 9
7 While it is not clear how such bias affects our inference, to the
extent the biases/errors are randomly distributed across firms, they
will bias against documenting our findings. 8 EIRIS refers to this variable as ‘‘governance.’’ We label it as
‘‘corruption’’ as the former is a bit too broad of a concept and the
latter is more descriptive of what the variable measures. 9 The data are no longer available to replicate the fifth measure
illustrated by Scholtens and Dam (2007), Communication of the code
of ethics (whether the company has adopted a code of ethics or
business principles by which it communicates to all employees). The
answer was either no evidence of, has adopted, or clearly commu-
nicates. For this measure (whatID = 224 in the access file), the data
are only available until 2006. We cannot find this question in the
‘‘Guide to EIRIS Research’’ handbook published in the summer of
2013.
Corporate Codes of Ethics, National Culture, and Earnings Discretion 147
123
We also perform a factor analysis to construct a vari-
able, EthicalSystem, which combines information from the
individual variables. The factor analysis suggests that there
is only one factor with an eigenvalue larger than one (value
of 1.94). Systems and Corruption are the two variables
from our main analysis that contribute the most to this
factor (uniqueness of 0.20 and 0.42, respectively). For this
part of the analysis, we initially exclude Human Rights, as
it restricts our sample size by approximately 37 % (from
8649 to 5479). However, with or without Human Rights in
the analysis, we obtain only one factor with an eigenvalue
over one. The variable with the highest factor loading is
Systems in both cases. The main factor analysis is based on
the three variables excluding Human Rights, as we are able
to preserve a larger number of observations under that
approach. 10
Measures of Investor Protection
To test for differences between firms from stronger and
weaker investor protection legal regimes, we split our
sample into two subsamples using three different investor
protection measures (Behn et al. 2013). Our first measure
of investor protection is based on the anti-director rights
index developed in La Porta et al. (1998). This index
measures the legal protection of minority shareholders
against corporate insiders’ expropriation activities around
the world. The observation is classified as high protection if
the score is equal to or larger than the median, and low
otherwise.
The second measure of investor protection is based on
the anti-self-dealing index developed in Djankov et al.
(2008). This index measures the legal protection of
minority shareholders against corporate insiders’ expro-
priation activities by focusing on legal rules applied around
the world, and includes private enforcement mechanisms
(disclosures, approval, and litigation) that govern self-
dealing transactions. As the authors confirm, this index
generally performs better than the previous anti-director
rights index developed by the same group of authors (La
Porta et al. 1998). Again, the observation is classified as
high protection if the score is equal to or larger than the
median, and low otherwise. The third measure is based on
the country’s legal system (code law vs. common law).
Previous research indicates that common law countries are
associated with stronger investor protection regimes (La
Porta et al. 1998).
Empirical Models
To test our hypotheses we adopt the following model:
Disc:accrualst ¼ Codes of ethics measure þ Cultural factors þ Controls þ error
ð2Þ
where disc. Accruals (DACC) is our measure of earnings
management and is measured as the residuals of the ordi-
nary least square regression from Model (1) which is
described in Sect. 3.2.1 above. Codes of Ethics Measures
refer to the variables described in Sect. 3.2.2: Systems,
Implementation, Corruption, and Human Rights. Based on
Han et al. (2010), which shows that national culture is a
significant explanatory variable for earnings management
in international capital markets, and Scholtens and Dam
(2007), which shows that culture is associated with a firm’s
ethical policies, we also control for the following cultural
factors from Hofstede (2001): PDI: Power Distance score,
IDV: Individualism score, MAS: Masculinity score, and
UAI: Uncertainty Avoidance score.
Following Leuz et al. (2003), we also control for infla-
tion and GDP growth. INFLATION is measured by the
annual growth rate of the GDP implicit deflator showing
the rate of price change in the economy as a whole. The
GDP implicit deflator is the ratio of GDP in the current
local currency to GDP in the constant local currency.
GDP_GROWTH is measured as the annual percentage
growth rate of GDP at market prices based on the constant
local currency.
We also include several firm-level controls. First of all,
since corporate governance mechanisms could be associ-
ated with both a higher-quality code of ethics and less
opportunistic behavior from management, we control for
the strength of the firm’s corporate governance. Using
EIRIS data, we measure Corporate Governance with a
score from zero to three based on the following survey
questions: (1) Does the company separate the roles of
chairman and chief executive? (2) Does the company have
a board comprising more than 33 % independent directors?
(3) Does the company have an audit committee comprising
a majority of independent directors? and (4) Does the
company disclose director remuneration? Companies’
answers could be ‘‘none,’’ ‘‘one,’’ ‘‘some,’’ or ‘‘all’’.
Therefore, the range of this measure is 0–3.
We also include in our model controls for the size and
growth opportunities of the firm, as previous research by
Roychowdhury (2006) indicates that these two factors can
explain variation in earnings management. We measure
size as the natural logarithm of market value of equity
(LNSIZE) and growth opportunities with the book-to-
market ratio (BM). Because earnings management can also
be connected to an equity offering, we include in our model
10 The results are unchanged when Human Rights is included in the
principal component analysis.
148 C. Chen et al.
123
Table 1 Summary statistics
Country Number of
firm-year obs.
Average system Average
implementation
Average
corruption
Average human
Panel A: Observations by country
Australia 417 3.16 2.44 1.22 0.48
Austria 64 2.16 1.77 1.04 1.13
Belgium 76 2.76 2.13 1.07 1.24
Denmark 68 2.76 2.34 1.25 1.44
Finland 132 2.79 1.89 1.01 1.58
France 476 2.94 2.19 1.34 1.62
Germany 482 2.30 1.78 1.05 1.08
Hong Kong 121 1.28 0.93 0.51 0.19
Italy 121 3.75 2.78 1.79 1.21
Japan 3172 2.28 2.38 0.64 0.59
The Netherlands 149 3.54 2.85 1.51 1.65
Norway 67 3.07 1.96 1.16 2.08
Singapore 138 1.12 1.01 0.96 0.07
South Korea 252 2.97 2.52 1.01 0.54
Spain 135 2.61 1.75 1.00 1.62
Sweden 185 3.16 2.49 1.44 1.79
Switzerland 157 2.89 2.13 1.38 1.08
UK 853 1.97 1.74 0.70 0.96
United States 2761 3.32 2.55 1.29 0.73
Total 9826
N Mean SD 25 % Median 75 %
Panel B: Descriptive statistics of regression variables
|DACC| 9826 0.187 0.377 0.023 0.059 0.179
Systems 9826 2.688 1.310 2.000 3.000 4.000
Implementation 9824 2.289 1.259 2.000 2.000 3.000
Corruption 8651 0.984 0.960 0.000 1.000 2.000
Human rights 6322 0.836 0.953 0.000 0.000 2.000
EthicalSystem 8649 0.000 1.000 -0.634 0.131 1.008
PDI 9826 46.120 11.237 38.000 40.000 54.000
IDV 9826 67.782 22.195 46.000 71.000 91.000
MAS 9826 68.415 22.461 62.000 62.000 95.000
UAI 9826 64.689 23.915 46.000 58.000 92.000
GDP growth 9826 0.014 0.026 0.001 0.018 0.028
INFLATION 9826 0.015 0.015 0.000 0.016 0.028
CORPORATE_GOVERNANCE 9826 2.068 0.730 2.000 2.000 3.000
ROA 9776 0.059 0.092 0.021 0.051 0.092
LNSIZE 9826 11.703 4.743 8.150 9.698 16.759
BM 9826 0.653 0.525 0.311 0.530 0.840
LEV 9826 0.235 0.178 0.099 0.217 0.335
ISSUE 9826 0.391 0.488 0.000 0.000 1.000
LOSS 9826 0.745 0.436 0.000 1.000 1.000
Corporate Codes of Ethics, National Culture, and Earnings Discretion 149
123
Table 1 continued
|DACC| Systems Implem Corruption Human EthicalSys PDI IDV MAS
Panel C: correlation table
|DACC| 1
Systems 0.0564* 1
Implem -0.0516* 0.7193* 1
Corruption 0.0674* 0.7389* 0.5385* 1
Human -0.0122 0.4105* 0.3573* 0.4453* 1
EthicalSys 0.0343* 0.9355* 0.8521* 0.8589* 0.4571* 1
PDI -0.2356* -0.1493* -0.0227* -0.1563* -0.1048* -0.1219* 1
IDV 0.3952* 0.2704* 0.0650* 0.2465* 0.1492* 0.2255* -0.6840* 1
MAS -0.1159* -0.2010* 0.0276* -0.2221* -0.2833* -0.1532* 0.3050* -0.3936* 1
UAI -0.3334* -0.1093* 0.0682* -0.1363* -0.0395* -0.0690* 0.6113* -0.6706* 0.6039*
GDP GR 0.0817* -0.0646* -0.1243* -0.0054 -0.0371* -0.0862* 0.0101 -0.0014 -0.1430*
INFL 0.2695* 0.1595* -0.0349* 0.1739* 0.0812* 0.1217* -0.3543* 0.5290* -0.5581*
CORP_G 0.1281* 0.2022* 0.0734* 0.1714* 0.1827* 0.1729* -0.3537* 0.4594* -0.4776*
ROA 0.1238* 0.0194 -0.0289* 0.0201 0.0262* 0.0117 -0.1219* 0.1572* -0.1541*
LNSIZE -0.2466* -0.0795* 0.1341* -0.1309* -0.1182* -0.0313* 0.5326* -0.7745* 0.5526*
BM -0.1937* -0.0599* 0.0436* -0.0842* -0.0166 -0.0452* 0.2117* -0.3190* 0.1266*
LEV 0.0486* 0.0753* 0.0496* 0.0863* 0.1151* 0.0766* -0.0367* 0.0702* -0.1196*
ISSUE 0.0914* -0.0260* -0.0836* 0.0240* 0.0136 -0.0298* -0.0733* 0.0903* -0.1187*
LOSS -0.4476* -0.2738* -0.1199* -0.1855* 0.0401* -0.2191* 0.3186* -0.6119* 0.1671*
UAI GDP GR INFL CORP_G ROA LNSIZE BM LEV ISSUE LOSS
Panel C: correlation table
|DACC|
Systems
Implem
Corruption
Human
EthicalSys
PDI
IDV
MAS
UAI 1
GDP GR -0.1992* 1
INFL -0.6168* 0.2915* 1
CORP_G -0.5296* 0.0714* 0.4292* 1
ROA -0.1931* 0.1560* 0.1543* 0.1156* 1
LNSIZE 0.7324* -0.0489* -0.5950* -0.5381* -0.0810* 1
BM 0.2509* -0.1238* -0.1021* -0.1141* -0.3254* 0.1486* 1
LEV -0.0856* 0.0091 0.0977* 0.0744* -0.1073* -0.0984* -0.0411* 1
ISSUE -0.1161* 0.1023* 0.0763* 0.0637* 0.2813* -0.0620* -0.2196* 0.0701* 1
LOSS 0.4571* -0.0862* -0.3637* -0.1249* -0.2592* 0.2984* 0.3164* -0.0064 -0.1256* 1
The table reports Pearson product moment correlations. * denotes significance at the 5 % level (two-tailed)
Please refer to the Appendix for variable definitions
150 C. Chen et al.
123
Table 2 Test of H1: Regression of discretionary
accruals on ethics policy
variables
Dependent variable
|DACC| [?]DACC [-]DACC
Panel A: Code of ethics system
Systems -0.006*** (-3.651) -0.008*** (-3.582) 0.004* (1.680)
PDI 0.001*** (4.316) 0.001*** (3.520) -0.001*** (-3.063)
IDV 0.002*** (11.200) 0.002*** (7.405) -0.002*** (-9.010)
MAS 0.001*** (10.663) 0.001*** (5.649) -0.001*** (-8.427)
UAI -0.001*** (-9.615) -0.001*** (-6.903) 0.001*** (6.575)
GDP growth 0.494*** (5.027) 0.555*** (3.225) -0.470*** (-4.015)
Inflation 1.137*** (7.113) 1.085*** (4.464) -1.240*** (-6.168)
Corporate governance 0.004 (1.542) 0.002 (0.479) -0.004 (-0.982)
ROA 0.117*** (2.591) 0.123 (1.127) -0.119** (-2.529)
LNSIZE 0.003*** (5.936) 0.004*** (4.462) -0.003*** (-5.119)
BM -0.003 (-0.792) -0.002 (-0.550) 0.002 (0.395)
LEV 0.042*** (3.080) 0.080*** (3.059) -0.016 (-1.011)
ISSUE 0.007* (1.782) -0.006 (-0.661) -0.015*** (-3.295)
LOSS -0.320*** (-26.182) -0.519*** (-15.567) 0.289*** (22.279)
Constant 0.212*** (3.739) 0.177*** (3.617) -0.251*** (v3.879)
Year dummies Yes Yes Yes
Industry dummies Yes Yes Yes
N 9776 2446 7330
Adj. R 2
0.269 0.407 0.266
Panel B: Code of ethics implementation
Implementation -0.006*** (-4.124) -0.007*** (-3.520) 0.005*** (2.807)
PDI 0.001*** (4.490) 0.001*** (3.690) -0.001*** (-3.180)
IDV 0.002*** (11.157) 0.002*** (7.258) -0.002*** (-9.098)
MAS 0.001*** (11.497) 0.001*** (6.184) -0.001*** (-8.959)
UAI -0.001*** (-10.394) -0.001*** (-7.286) 0.001*** (7.190)
GDP growth 0.484*** (4.902) 0.563*** (3.238) -0.450*** (-3.852)
Inflation 1.125*** (6.998) 1.066*** (4.377) -1.217*** (-6.046)
Corporate governance 0.004 (1.443) 0.001 (0.206) -0.004 (-1.173)
ROA 0.121*** (2.687) 0.126 (1.163) -0.121*** (-2.616)
LNSIZE 0.003*** (6.132) 0.004*** (4.422) -0.003*** (-5.407)
BM -0.002 (-0.577) -0.002 (-0.392) 0.001 (0.270)
LEV 0.041*** (3.045) 0.082*** (3.172) -0.014 (-0.946)
ISSUE 0.007* (1.805) -0.005 (-0.620) -0.015*** (-3.306)
LOSS -0.317*** (-26.160) -0.516*** (-15.457) 0.288*** (22.452)
Constant 0.201*** (3.546) 0.162*** (3.209) -0.244*** (-3.771)
Year dummies Yes Yes Yes
Industry dummies Yes Yes Yes
N 9774 2445 7329
Adj. R 2
0.269 0.406 0.267
Panel C: Corruption policy
Corruption -0.008*** (-3.288) -0.011*** (-3.262) 0.005* (1.728)
PDI 0.001*** (4.216) 0.001*** (3.550) -0.001*** (-2.802)
IDV 0.002*** (11.726) 0.003*** (7.654) -0.002*** (-9.402)
MAS 0.001*** (11.578) 0.001*** (6.546) -0.001*** (-8.978)
UAI -0.001*** (-10.509) -0.001*** (-7.839) 0.001*** (7.077)
GDP growth 0.684*** (6.017) 0.643*** (3.150) -0.718*** (-5.239)
Inflation 1.173*** (6.473) 1.213*** (4.286) -1.223*** (-5.267)
Corporate Codes of Ethics, National Culture, and Earnings Discretion 151
123
the leverage ratio (LEV), measured by debt to total assets,
and a dummy variable (ISSUE) equal to 1 if shareholders’
equity increases by more than 10 % compared to the pre-
vious year and 0 otherwise. Since discretionary accruals
vary with the financial performance of the firm, we also
include return on assets (ROA) and a dummy variable
(LOSS) equal to 1 if the firm reported a loss during the year
and 0 otherwise. Since Corporate Governance and Ethics
policies can change during the same year, we also run our
main model with a lag measure (previous year) of
corporate governance. All results and inferences are robust
to this different model specification.
While our test variable is a firm-level variable, we
include several country-level control variables in the firm-
level regression. This will not only bias the country-level
coefficients for more representative countries but could
also induce serial correlation in the error term, which can
affect the entire model. Given that the sample countries
have unequal numbers of observations, we run this model
using a Weighted Least Squares (WLS) regression to
Table 2 continued Dependent variable
|DACC| [?]DACC [-]DACC
Corporate governance 0.000 (0.160) -0.003 (-0.644) -0.000 (-0.097)
ROA 0.119** (2.524) 0.139 (1.290) -0.115** (-2.313)
LNSIZE 0.003*** (4.945) 0.003*** (3.491) -0.003*** (-4.736)
BM -0.002 (-0.602) -0.000 (-0.105) 0.001 (0.268)
LEV 0.049*** (3.281) 0.101*** (3.672) -0.020 (-1.194)
ISSUE 0.009** (2.036) -0.005 (-0.547) -0.018*** (-3.400)
LOSS -0.335*** (-24.772) -0.506*** (-14.755) 0.307*** (21.309)
Constant 0.199*** (3.259) 0.127** (2.284) -0.237*** (-3.373)
Year dummies Yes Yes Yes
Industry dummies Yes Yes Yes
N 8606 2130 6476
Adj. R 2
0.279 0.443 0.274
Panel D: Human rights ethical policy
Human rights -0.002 (-1.274) -0.013*** (-4.280) -0.001 (-0.453)
PDI 0.000*** (4.270) 0.000** (2.178) -0.000*** (-2.984)
IDV 0.001*** (7.805) 0.001*** (4.535) -0.001*** (-6.760)
MAS 0.001*** (7.769) 0.001*** (3.347) -0.001*** (-6.446)
UAI -0.001*** (-7.284) -0.001*** (-4.399) 0.001*** (5.459)
GDP growth 0.352*** (3.637) 0.483*** (3.090) -0.299** (-2.540)
Inflation 1.186*** (7.150) 0.500** (1.972) -1.511*** (-7.451)
Corporate governance -0.001 (-0.372) -0.003 (-0.853) 0.001 (0.453)
ROA 0.063* (1.822) 0.070 (0.980) -0.053 (-1.314)
LNSIZE 0.003*** (5.994) 0.002*** (2.717) -0.003*** (-5.449)
BM -0.005** (-2.139) 0.000 (0.058) 0.006 (1.548)
LEV 0.004 (0.334) 0.049** (2.396) 0.026* (1.798)
ISSUE 0.010*** (2.808) -0.002 (-0.463) -0.018*** (-4.049)
LOSS -0.307*** (-20.147) -0.552*** (-13.462) 0.272*** (16.855)
Constant 0.209*** (2.916) 0.286*** (5.687) -0.213*** (-2.625)
Year dummies Yes Yes Yes
Industry dummies Yes Yes Yes
N 6287 1635 4652
Adj. R 2
0.270 0.409 0.264
Please refer to the Appendix for variable definitions. *, **, and *** indicate statistical significance at the
10, 5, and 1 % levels (two-tailed), respectively
152 C. Chen et al.
123
T a b le
3 T e st
o f H 2 : T h e in v e st o r p ro te c ti o n h y p o th e si s
T e st
v a ri a b le s
L o w
a n ti -d ir e c to r ri g h ts
H ig h a n ti -d ir e c to r ri g h ts
P a n e l A : P a rt it io n in g b a se d o n a n ti -d ir e c to r ri g h ts
in d e x (d e p e n d e n t v a ri a b le : |D A C C |)
S y st e m s
- 0 .0 0 8 * * * (-
3 .2 6 0 )
- 0 .0 0 1 (-
0 .2 8 5 )
Im p le m e n ta ti o n
- 0 .0 0 6 * * * (-
2 .9 4 7 )
- 0 .0 0 0 (-
0 .0 6 7 )
C o rr u p ti o n
- 0 .0 0 7 * * (-
2 .0 9 4 )
0 .0 0 4 (0 .9 4 7 )
H u m a n ri g h ts
0 .0 0 0 (0 .1 1 4 )
- 0 .0 0 1 (-
0 .3 9 1 )
C o n tr o l v a ri a b le s
P D I
0 .0 0 0 (0 .2 8 2 )
0 .0 0 0 (0 .0 5 4 )
0 .0 0 0 (0 .1 3 6 )
0 .0 0 1 * (1 .8 4 1 )
0 .0 0 2 * * (2 .1 4 1 )
0 .0 0 2 * * (2 .0 4 5 )
0 .0 0 1 (1 .2 6 1 )
0 .0 0 4 * * * (4 .2 9 6 )
ID V
0 .0 0 4 * * * (5 .7 3 1 )
0 .0 0 4 * * * (5 .7 4 7 )
0 .0 0 4 * * * (5 .6 4 7 )
0 .0 0 2 * * * (4 .0 2 1 )
0 .0 0 2 * * * (3 .3 6 4 )
0 .0 0 2 * * * (3 .2 7 5 )
0 .0 0 2 * * (2 .4 9 3 )
0 .0 0 4 * * * (5 .4 0 9 )
M A S
0 .0 0 1 * * * (4 .5 3 5 )
0 .0 0 1 * * * (4 .6 5 9 )
0 .0 0 1 * * * (4 .5 9 3 )
0 .0 0 1 * * * (4 .8 9 2 )
- 0 .0 0 0 (-
0 .1 0 7 )
0 .0 0 0 (0 .0 1 4 )
0 .0 0 0 (1 .3 0 4 )
- 0 .0 0 1 * * (-
2 .4 1 5 )
U A I
- 0 .0 0 2 * * * (-
3 .5 2 0 )
- 0 .0 0 1 * * * (-
3 .2 7 3 )
- 0 .0 0 2 * * * (-
3 .2 2 9 )
- 0 .0 0 2 * * * (-
4 .5 5 1 )
- 0 .0 0 0 (-
1 .4 1 8 )
- 0 .0 0 0 (-
1 .4 1 2 )
- 0 .0 0 0 * (-
1 .6 4 6 )
- 0 .0 0 1 * * * (-
2 .5 9 0 )
G D P g ro w th
0 .2 7 3 (1 .4 8 1 )
0 .2 6 6 (1 .4 3 8 )
0 .6 7 7 * * * (3 .1 2 4 )
0 .1 3 8 (0 .7 8 1 )
0 .2 0 5 (1 .4 1 3 )
0 .2 1 0 (1 .4 4 2 )
0 .3 3 1 * * (1 .9 9 6 )
0 .1 8 7 (1 .3 9 5 )
In fl a ti o n
0 .4 1 0 (1 .1 1 5 )
0 .3 9 6 (1 .0 8 6 )
0 .2 1 5 (0 .4 9 3 )
0 .7 1 9 * * (2 .4 4 8 )
0 .3 8 5 (1 .4 5 5 )
0 .4 0 0 (1 .5 0 4 )
0 .4 9 2 * (1 .6 7 5 )
0 .4 2 8 * (1 .7 0 4 )
C o rp o ra te
g o v e rn a n c e
- 0 .0 0 7 (-
1 .4 9 1 )
- 0 .0 0 8 * (-
1 .7 6 6 )
- 0 .0 1 2 * * (-
2 .3 0 0 )
- 0 .0 1 3 * * * (-
3 .2 3 6 )
0 .0 2 1 * * * (4 .2 9 7 )
0 .0 2 1 * * * (4 .2 9 4 )
0 .0 2 0 * * * (3 .7 8 6 )
0 .0 0 8 * (1 .8 9 1 )
R O A
- 0 .0 2 1 (-
0 .3 8 6 )
- 0 .0 1 5 (-
0 .2 7 4 )
- 0 .0 3 2 (-
0 .5 3 4 )
0 .0 1 9 (0 .3 5 3 )
0 .1 9 4 * * * (3 .4 9 0 )
0 .1 9 4 * * * (3 .4 8 9 )
0 .2 0 8 * * * (3 .6 6 2 )
0 .1 0 3 * * (2 .2 6 1 )
B M
- 0 .0 0 3 * (-
1 .8 5 3 )
- 0 .0 0 3 (-
1 .5 7 9 )
- 0 .0 0 5 * * (-
2 .1 8 1 )
- 0 .0 0 5 * * * (-
2 .8 0 6 )
0 .0 0 2 (0 .8 8 5 )
0 .0 0 1 (0 .7 7 7 )
- 0 .0 0 0 (-
0 .0 4 2 )
0 .0 0 6 * * * (3 .4 7 2 )
L E V
- 0 .0 0 6 - (-
0 .9 4 5 )
- 0 .0 0 6 (-
0 .8 4 3 )
- 0 .0 0 6 (-
0 .7 8 3 )
- 0 .0 1 1 * * * (-
2 .5 9 4 )
- 0 .0 0 4 (-
0 .9 7 8 )
- 0 .0 0 4 (-
0 .9 8 6 )
- 0 .0 0 6 (-
1 .2 7 2 )
0 .0 0 1 (0 .2 5 4 )
IS S U E
0 .0 2 2 (1 .1 3 8 )
0 .0 2 1 (1 .1 0 4 )
0 .0 3 7 * (1 .6 8 5 )
- 0 .0 2 7 * (-
1 .7 1 3 )
0 .0 4 5 * * (2 .1 1 9 )
0 .0 4 5 * * (2 .1 0 9 )
0 .0 4 9 * * (2 .1 7 8 )
0 .0 1 8 (0 .9 8 2 )
L O S S
0 .0 1 9 * * * (3 .8 4 1 )
0 .0 1 9 * * * (3 .8 4 6 )
0 .0 2 2 * * * (4 .0 6 2 )
0 .0 1 8 * * * (3 .8 2 7 )
0 .0 0 2 (0 .3 8 0 )
0 .0 0 2 (0 .3 7 3 )
0 .0 0 5 (0 .6 8 0 )
- 0 .0 0 1 (-
0 .1 7 4 )
C o n st a n t
- 0 .2 9 5 * * * (-
1 5 .2 6 5 )
- 0 .2 9 4 * * * (-
1 5 .3 4 0 )
- 0 .3 1 6 * * * (-
1 4 .4 0 4 )
- 0 .2 8 2 * * * (-
1 3 .9 7 1 )
0 .0 0 0 (0 .0 0 0 )
0 .0 0 0 (0 .0 0 0 )
0 .0 0 0 (0 .0 0 0 )
0 .0 0 0 (0 .0 0 0 )
Y e a r d u m m ie s
Y e s
Y e s
Y e s
Y e s
Y e s
Y e s
Y e s
Y e s
In d u st ry
d u m m ie s
Y e s
Y e s
Y e s
Y e s
Y e s
Y e s
Y e s
Y e s
N 4 6 5 4
4 6 5 4
4 0 8 4
3 0 3 1
5 1 2 2
5 1 2 0
4 5 2 2
3 2 5 6
A d j. R 2
0 .3 6 5
0 .3 6 4
0 .3 7 7
0 .3 2 9
0 .2 3 5
0 .2 3 5
0 .2 4 5
0 .1 8 3
Corporate Codes of Ethics, National Culture, and Earnings Discretion 153
123
T a b le
3 c o n ti n u e d
T e st
v a ri a b le s
L o w
a n ti -s e lf -d e a li n g in d e x
H ig h a n ti -s e lf -d e a li n g in d e x
P a n e l B : P a rt it io n in g b a se d o n a n ti -s e lf -d e a li n g in d e x (d e p e n d e n t v a ri a b le : |D A C C |
S y st e m s
- 0 .0 0 3 * * * (-
3 .0 9 0 )
0 .0 0 1 (0 .1 4 7 )
Im p le m e n ta ti o n
- 0 .0 0 3 * * (-
2 .5 7 3 )
- 0 .0 0 2 (-
0 .5 3 1 )
C o rr u p ti o n
- 0 .0 0 3 * (-
1 .7 6 2 )
0 .0 1 1 (1 .5 7 1 )
H u m a n ri g h ts
- 0 .0 0 2 * * (-
1 .9 8 6 )
0 .0 1 3 (1 .1 6 9 )
C o n tr o l v a ri a b le s
P D I
0 .0 0 0 (0 .4 9 7 )
0 .0 0 0 (0 .3 1 0 )
0 .0 0 0 (0 .5 2 9 )
0 .0 0 0 (1 .3 3 5 )
0 .0 2 4 * * * (7 .4 5 3 )
0 .0 2 4 * * * (7 .4 6 4 )
0 .0 2 8 * * * (7 .7 1 1 )
0 .0 2 5 * * * (6 .0 8 5 )
ID V
- 0 .0 0 0 * * (-
2 .0 8 0 )
- 0 .0 0 0 * (-
1 .8 2 8 )
- 0 .0 0 1 * * (-
2 .3 7 8 )
- 0 .0 0 1 * * * (-
2 .7 1 3 )
0 .0 1 5 * * * (7 .1 5 2 )
0 .0 1 5 * * * (7 .2 2 1 )
0 .0 1 7 * * * (7 .4 2 3 )
0 .0 1 5 * * * (6 .1 3 0 )
M A S
- 0 .0 0 0 (-
0 .4 5 8 )
- 0 .0 0 0 (-
0 .3 8 4 )
- 0 .0 0 0 (-
0 .2 3 5 )
0 .0 0 0 (0 .3 0 2 )
0 .0 2 2 * * * (7 .9 4 3 )
0 .0 2 1 * * * (7 .8 4 7 )
0 .0 2 5 * * * (8 .3 1 5 )
0 .0 2 6 * * * (6 .1 6 1 )
U A I
- 0 .0 0 0 (-
0 .7 5 5 )
- 0 .0 0 0 (-
0 .6 4 3 )
- 0 .0 0 0 (-
0 .9 4 7 )
- 0 .0 0 0 * (-
1 .6 5 1 )
- 0 .0 0 8 * * * (-
7 .6 6 3 )
- 0 .0 0 8 * * * (-
7 .7 2 5 )
- 0 .0 0 9 * * * (-
7 .9 4 5 )
- 0 .0 0 9 * * * (-
6 .6 8 6 )
G D P g ro w th
- 0 .0 7 3 (-
0 .8 1 1 )
- 0 .0 8 4 (-
0 .9 2 5 )
- 0 .1 5 2 (-
1 .6 0 4 )
- 0 .0 6 6 (-
0 .7 3 0 )
0 .5 6 7 * * (2 .1 3 9 )
0 .5 2 9 * * (1 .9 8 8 )
1 .0 0 7 * * * (3 .4 0 0 )
0 .6 5 8 * * (2 .4 8 4 )
In fl a ti o n
- 0 .0 3 4 (-
0 .1 8 6 )
- 0 .0 4 1 (-
0 .2 2 7 )
0 .0 5 7 (0 .2 7 3 )
0 .0 1 3 (0 .0 7 2 )
2 .1 5 1 * * * (3 .7 7 2 )
2 .0 8 3 * * * (3 .6 2 6 )
1 .6 5 0 * * * (2 .7 5 3 )
3 .2 7 5 * * * (5 .0 6 2 )
C o rp o ra te
g o v e rn a n c e
0 .0 0 1 (0 .5 7 3 )
0 .0 0 1 (0 .3 8 1 )
- 0 .0 0 0 (-
0 .1 6 1 )
0 .0 0 2 (0 .8 0 0 )
- 0 .0 2 0 * * (-
2 .4 2 8 )
- 0 .0 2 0 * * (-
2 .4 6 7 )
- 0 .0 2 5 * * * (-
2 .7 8 7 )
- 0 .0 3 8 * * * (v 3 .4 0 1 )
R O A
0 .0 9 0 * * (2 .0 1 0 )
0 .0 9 3 * * (2 .0 6 3 )
0 .0 8 9 * (1 .8 1 4 )
0 .1 1 3 * * * (2 .6 8 8 )
0 .1 8 6 * * * (2 .8 6 0 )
0 .1 8 8 * * * (2 .9 0 6 )
0 .2 0 5 * * * (3 .1 3 5 )
0 .0 2 7 (0 .4 3 2 )
B M
- 0 .0 0 1 * (-
1 .8 2 4 )
- 0 .0 0 1 (-
1 .3 5 5 )
- 0 .0 0 1 * * (-
2 .0 5 5 )
- 0 .0 0 1 * (-
1 .9 4 1 )
0 .0 0 7 (1 .5 7 3 )
0 .0 0 8 * (1 .8 5 8 )
0 .0 0 7 (1 .5 0 6 )
0 .0 0 3 (0 .5 2 6 )
L E V
- 0 .0 0 0 (-
0 .0 6 3 )
0 .0 0 0 (0 .0 1 6 )
0 .0 0 3 (0 .6 1 8 )
- 0 .0 0 2 (-
1 .0 8 3 )
- 0 .0 1 1 (-
1 .3 0 9 )
- 0 .0 1 0 (-
1 .1 8 1 )
- 0 .0 0 9 (-
0 .9 9 8 )
- 0 .0 0 7 (-
0 .7 0 5 )
IS S U E
0 .0 1 5 (1 .3 4 5 )
0 .0 1 5 (1 .3 0 2 )
0 .0 1 7 (1 .3 6 6 )
- 0 .0 0 9 (-
0 .9 5 4 )
0 .0 6 2 * (1 .7 6 0 )
0 .0 6 0 * (1 .6 9 0 )
0 .0 7 5 * * (1 .9 7 2 )
0 .0 6 3 (1 .4 7 1 )
L O S S
0 .0 0 3 (1 .1 4 1 )
0 .0 0 3 (1 .1 4 4 )
0 .0 0 3 (1 .2 1 5 )
0 .0 0 3 (1 .1 7 1 )
0 .0 2 6 * * (2 .5 3 9 )
0 .0 2 7 * * * (2 .6 6 0 )
0 .0 3 0 * * * (2 .7 2 4 )
0 .0 3 1 * * * (2 .6 2 4 )
C o n st a n t
0 .0 0 0 (0 .0 0 0 )
0 .0 0 0 (0 .0 0 0 )
0 .0 0 0 (0 .0 0 0 )
0 .0 0 0 (0 .0 0 0 )
- 0 .1 8 7 * * * (-
7 .7 5 7 )
- 0 .1 8 4 * * * (-
7 .6 5 7 )
- 0 .1 8 8 * * * (-
7 .0 4 1 )
- 0 .1 4 9 * * * (-
4 .5 9 9 )
Y e a r d u m m ie s
Y e s
Y e s
Y e s
Y e s
Y e s
Y e s
Y e s
Y e s
In d u st ry
d u m m ie s
Y e s
Y e s
Y e s
Y e s
Y e s
Y e s
Y e s
Y e s
N 5 4 4 1
5 4 4 0
4 7 6 6
4 1 8 7
4 3 3 5
4 3 3 4
3 8 4 0
2 1 0 0
A d j. R 2
0 .0 7 9
0 .0 7 9
0 .0 7 4
0 .0 9 3
0 .3 0 0
0 .3 0 1
0 .3 0 4
0 .3 2 4
T e st
v a ri a b le s
C o d e L a w
C o m m o n L a w
P a n e l C : P a rt it io n in g b a se d o n le g a l o ri g in
(d e p e n d e n t v a ri a b le : |D A C C |)
S y st e m s
- 0 .0 0 3 * * * (-
2 .9 1 9 )
0 .0 0 1 (0 .0 9 7 )
Im p le m e n ta ti o n
- 0 .0 0 3 * * (-
2 .3 2 8 )
- 0 .0 0 1 (-
0 .2 0 9 )
C o rr u p ti o n
- 0 .0 0 3 (-
1 .5 7 0 )
0 .0 0 9 (1 .2 5 9 )
H u m a n ri g h ts
- 0 .0 0 2 (-
1 .5 0 5 )
0 .0 2 3 * * (2 .0 2 5 )
C o n tr o l v a ri a b le s
P D I
0 .0 0 0 (0 .6 2 0 )
0 .0 0 0 (0 .4 5 9 )
0 .0 0 0 (0 .6 2 5 )
0 .0 0 0 * (1 .6 7 0 )
0 .0 9 3 * * * (1 1 .8 1 5 )
0 .0 9 3 * * * (1 1 .8 3 6 )
0 .1 0 2 * * * (1 1 .5 5 0 )
0 .0 8 5 * * * (8 .4 3 5 )
ID V
- 0 .0 0 1 * * * (-
2 .6 8 0 )
- 0 .0 0 1 * * (-
2 .4 5 7 )
- 0 .0 0 1 * * * (-
3 .1 3 3 )
- 0 .0 0 1 * * * (-
3 .3 7 1 )
0 .0 4 3 * * * (1 1 .9 1 2 )
0 .0 4 3 * * * (1 1 .9 9 6 )
0 .0 4 7 * * * (1 1 .7 9 1 )
0 .0 3 9 * * * (8 .8 8 6 )
M A S
- 0 .0 0 0 (-
0 .3 4 9 )
- 0 .0 0 0 (-
0 .2 6 9 )
- 0 .0 0 0 (-
0 .1 3 3 )
0 .0 0 0 (0 .6 7 2 )
0 .0 4 8 * * * (1 2 .2 5 0 )
0 .0 4 7 * * * (1 2 .1 2 5 )
0 .0 5 3 * * * (1 2 .2 1 1 )
0 .0 4 6 * * * (8 .9 9 7 )
U A I
- 0 .0 0 0 (-
1 .2 3 5 )
- 0 .0 0 0 (-
1 .1 3 7 )
- 0 .0 0 0 (-
1 .4 7 1 )
- 0 .0 0 1 * * (-
2 .3 0 0 )
- 0 .0 0 1 (-
1 .0 4 9 )
- 0 .0 0 1 (-
1 .1 6 2 )
- 0 .0 0 2 * (-
1 .7 2 7 )
- 0 .0 0 2 (-
1 .1 0 0 )
G D P g ro w th
- 0 .1 0 7 (-
1 .2 2 6 )
- 0 .1 1 7 (-
1 .3 2 0 )
- 0 .1 9 1 * * (-
2 .0 5 8 )
- 0 .0 7 1 (-
0 .7 9 7 )
0 .4 0 1 (1 .3 7 6 )
0 .3 5 3 (1 .1 9 6 )
0 .7 4 6 * * (2 .3 2 2 )
0 .6 6 0 * * (2 .2 9 6 )
In fl a ti o n
- 0 .0 4 1 (-
0 .2 3 0 )
- 0 .0 4 8 (-
0 .2 7 0 )
0 .0 3 0 (0 .1 5 0 )
0 .0 1 3 (0 .0 8 0 )
3 .4 2 9 * * * (5 .2 2 6 )
3 .3 5 2 * * * (5 .0 3 3 )
3 .1 8 8 * * * (4 .6 6 1 )
4 .6 7 7 * * * (6 .2 7 4 )
C o rp o ra te
g o v e rn a n c e
- 0 .0 0 0 (-
0 .0 8 1 )
- 0 .0 0 1 (-
0 .3 6 2 )
- 0 .0 0 2 (-
0 .7 6 2 )
0 .0 0 1 (0 .2 7 3 )
0 .0 0 4 (0 .4 5 6 )
0 .0 0 4 (0 .3 9 7 )
0 .0 0 0 (0 .0 3 5 )
- 0 .0 1 7 (-
1 .5 0 9 )
R O A
0 .0 9 3 * * (2 .1 1 9 )
0 .0 9 5 * * (2 .1 8 0 )
0 .0 9 2 * (1 .9 1 8 )
0 .1 1 4 * * * (2 .8 1 8 )
0 .2 3 9 * * * (4 .0 8 7 )
0 .2 4 1 * * * (4 .1 1 2 )
0 .2 5 7 * * * (4 .3 1 5 )
0 .1 7 6 * * * (2 .6 3 0 )
B M
- 0 .0 0 1 * * (-
2 .0 2 7 )
- 0 .0 0 1 (-
1 .6 1 8 )
- 0 .0 0 1 * * (-
2 .2 9 6 )
- 0 .0 0 1 * * (-
2 .3 0 7 )
- 0 .0 0 9 * (-
1 .9 4 6 )
- 0 .0 0 8 * (-
1 .7 6 9 )
- 0 .0 0 9 * (-
1 .6 8 5 )
- 0 .0 1 4 * * (-
2 .2 1 9 )
154 C. Chen et al.
123
address this issue. The weight is inversely proportional to
the number of observations in each country. 11
Results
Panel A of Table 1 provides the number of observations
included in our sample by country of origin. To be
included in the sample, a firm must have code of ethics
data available in EIRIS and firm-level control variables
available in Compustat Global Vantage. The final sample
comprises firms from 19 countries and consists of a total
of 9826 firm-year observations that meet the data
requirements. 12
Japan (n = 3172) and the United States
(n = 2761) are most heavily represented in the sample.
We also include the mean ethic codes scores for each
country in our sample. Similar to Scholtens and Dam
(2007), American, Australian, and Dutch firms score high
while firms from Singapore and Hong Kong score poorly.
For our sample of firms, Italy and Sweden also score well
while the United Kingdom and Austria fall near the
bottom.
In Panel B, we present descriptive statistics for all the
variables included in our study. The Human Rights
variable has fewer observations in the EIRIS dataset than
the other ethics code measures (6322 vs. 9826 for Sys-
tems). The distribution of our test variables is comparable
to the numbers reported in prior studies. Panel C provides
univariate correlations. As expected, the code of ethics
variables is highly correlated, with the Pearson correla-
tion ranging from 0.3573 (between Implementation
and Human Rights) to 0.7389 (between Systems and
Corruption).
Table 2 presents the main results of our study. Panel A
presents results for the Systems variable, Panel B presents
results for Implementation, Panel C presents results for
Corruption, and Panel D presents results for Human
Rights. Given the high correlations among the four vari-
ables, we examine each measure separately to avoid issues
relating to multicollinearity. In the models with the Sys-
tems, Implementation, and Corruption variables (Panels A,
B, and C), we find a significant negative association
T a b le
3 c o n ti n u e d
T e st
v a ri a b le s
C o d e L a w
C o m m o n L a w
L E V
- 0 .0 0 1 (-
0 .2 1 7 )
- 0 .0 0 0 (-
0 .1 2 5 )
0 .0 0 2 (0 .4 5 9 )
- 0 .0 0 3 (-
1 .1 9 2 )
- 0 .0 3 5 * * * (-
3 .8 8 1 )
- 0 .0 3 4 * * * (-
3 .7 4 9 )
- 0 .0 4 0 * * * (-
4 .0 9 3 )
- 0 .0 2 5 * * (-
2 .4 9 3 )
IS S U E
0 .0 1 5 (1 .3 4 5 )
0 .0 1 4 (1 .2 9 5 )
0 .0 1 7 (1 .3 9 8 )
- 0 .0 0 9 (-
1 .0 2 2 )
0 .0 8 5 * * (2 .3 2 7 )
0 .0 8 1 * * (2 .2 2 0 )
0 .0 9 5 * * (2 .3 8 5 )
0 .0 7 5 * (1 .6 6 0 )
L O S S
0 .0 0 3 (1 .3 1 1 )
0 .0 0 3 (1 .3 3 1 )
0 .0 0 4 (1 .3 4 1 )
0 .0 0 3 (1 .2 9 0 )
0 .0 3 4 * * * (3 .1 0 5 )
0 .0 3 6 * * * (3 .3 0 4 )
0 .0 3 8 * * * (3 .1 2 1 )
0 .0 3 8 * * * (2 .9 5 2 )
C o n st a n t
0 .0 0 0 (0 .0 0 0 )
0 .0 0 0 (0 .0 0 0 )
0 .0 0 0 (0 .0 0 0 )
0 .0 0 0 (0 .0 0 0 )
0 .0 8 7 * * (2 .3 5 8 )
0 .0 8 8 * * (2 .4 0 5 )
0 .1 0 6 * * * (2 .5 8 6 )
0 .0 5 6 (1 .2 7 8 )
Y e a r d u m m ie s
Y e s
Y e s
Y e s
Y e s
Y e s
Y e s
Y e s
Y e s
In d u st ry
d u m m ie s
Y e s
Y e s
Y e s
Y e s
Y e s
Y e s
Y e s
Y e s
N 5 5 1 6
5 5 1 5
4 8 3 2
4 2 2 4
4 2 6 0
4 2 5 9
3 7 7 4
2 0 6 3
A d j. R 2
0 .0 8 2
0 .0 8 1
0 .0 8 0
0 .0 9 4
0 .3 1 2
0 .3 1 4
0 .3 1 1
0 .3 4 5
P le a se
re fe r to
th e A p p e n d ix
fo r v a ri a b le
d e fi n it io n s. * , * * , a n d * * * in d ic a te
st a ti st ic a l si g n ifi c a n c e a t th e 1 0 , 5 , a n d 1 %
le v e ls
(t w o -t a il e d ), re sp e c ti v e ly
11 The results are qualitatively similar when unweighted OLS is
estimated. The ethics variables have the predicted sign, although the
significance levels are slightly lower (p \ 0.10), presumably due to the econometric issues discussed above. 12
For comparison, Scholtens and Dam (2007), which also use the
EIRIS data, have 24 countries represented in their sample. We have a
slightly lower number of countries represented because we require
extensive firm-level control variables which were not required in
Scholtens and Dam (2007).
Corporate Codes of Ethics, National Culture, and Earnings Discretion 155
123
between the ethics code variable and the absolute value of
discretionary accruals (p \ 0.01). Although the coefficient in Panel D (Human Rights) is negative, it is not significant.
These results appears to be driven by a reduction in
income-increasing discretionary accruals, as the coefficient
for [?]DACC is negative and significant (p \ 0.01) for all four variables. Although the coefficients for income-de-
creasing accruals ([-]DACC) are negative for Systems,
Implementation, and Corruption, the results are generally
weaker (p \ 0.10 for both Systems and Corruption). Overall, these results are consistent with earnings quality
being higher (i.e., lower discretionary accruals) for firms
with more effective codes of ethics and provide evidence in
support of H1a and H1b.
Table 3 introduces our results testing H2. To test for
differences between firms from stronger and weaker
investor protection regimes, we split our sample into two
subsamples using the three different investor protection
measures discussed in Sect. 3.2.3. Our first measure of
investor protection is based on the anti-director rights index
developed in La Porta et al. (1998). The observation is
classified as high protection if the anti-director rights index
score is equal to or larger than the median, and low
otherwise. The results utilizing this measure are provided
in Panel A. With the exception of Human Rights, the
coefficients for the ethics code variables are all negative
and significant for the low anti-director rights countries.
None of the variables are significant for the high anti-di-
rector rights subsample.
The second measure of investor protection is based on
the anti-self-dealing index developed in Djankov et al.
(2008). Again, the observation is classified as high pro-
tection if the score is equal to or larger than the median,
and low otherwise. These results (provided in Panel B) are
consistent with those reported in Panel A as the ethics code
coefficients for the subsample of firms from the low anti-
self-dealing countries are all negative and significant while
none of the variables are significant for the high anti-self-
dealing subsample. Results for the third measure, which is
based on the country’s legal system (code law vs. common
law) are provided in Panel C. For the code law countries,
although all four ethics codes coefficients are negative,
only Systems and Implementation are significantly differ-
ent from zero. Overall, these results are consistent with
codes of ethics playing a greater role in determining
earnings quality when investor protection mechanisms are
weak and fail to prevent opportunistic behavior (i.e., the
substitution view). Table 4 Regressions with all four ethics policy variables
Dependent variable |DACC|
Systems -0.003 (-1.417)
Implementation -0.004* (-1.830)
Governance 0.003 (0.930)
Human rights 0.001 (0.669)
PDI 0.001*** (4.874)
IDV 0.002*** (8.316)
MAS 0.001*** (8.513)
UAI -0.001*** (-7.509)
GDP growth 0.431*** (3.936)
Inflation 1.429*** (7.618)
Corporate governance -0.001 (-0.374)
ROA 0.079** (2.052)
LNSIZE 0.004*** (6.309)
BM -0.004 (-1.557)
LEV 0.011 (0.810)
ISSUE 0.011*** (2.872)
LOSS -0.336*** (-19.109)
Constant 0.179** (2.252)
Year dummies Yes
Industry dummies Yes
N 5446
Adj. R 2
0.287
Please refer to the Appendix for variable definitions. *, **, and ***
indicate statistical significance at the 10, 5, and 1 % levels (two-
tailed), respectively
Table 5 Test of H1: Regression of discretionary accruals on Ethi- calSystem factor
|DACC|
EthicalSystem -0.009*** (-3.972)
PDI 0.001*** (4.523)
IDV 0.002*** (11.867)
MAS 0.001*** (11.635)
UAI -0.001*** (-10.379)
GDP growth 0.674*** (5.943)
Inflation 1.156*** (6.369)
Corporate governance 0.002 (0.719)
ROA 0.121** (2.570)
LNSIZE 0.003*** (5.522)
BM -0.002 (-0.558)
LEV 0.049*** (3.301)
ISSUE 0.009** (2.056)
LOSS -0.336*** (-24.829)
Constant 0.174*** (2.833)
Year dummies Yes
Industry dummies Yes
N 8604
Adj. R 2
0.280
Please refer to the Appendix for variable definitions. *, **, and ***
indicate statistical significance at the 10, 5, and 1 % levels (two--
tailed), respectively
156 C. Chen et al.
123
Next, we include all four code of ethics variables in the
same model and rerun the regression for the full sample of
firms. Since the four variables are highly correlated, the
results are subject to strong multicollinearity. The presence
of such multicollinearity induces a bias toward insignifi-
cance (Maddala 1983); thus, the ethics variables will most
likely turn out to be either insignificant or less significant
when all are examined concurrently. However, significant
variables in this model can also provide additional insight
regarding the most relevant dimensions of codes of ethics
in deterring opportunistic behavior. These results are
reported in Table 4. As expected, the significance of the
four variables decreases substantially with only Imple-
mentation remaining significant, although at a much lower
level (p \ 0.10). The insignificance/lower significance is consistent with multicollinearity among the four ethics
code variables biasing the t-statistics toward insignificance
(Maddala 1983).
While it is clear that the four variables are highly cor-
related, it is also likely that each variable captures a dif-
ferent dimension of a high-quality code of ethics. We next
provide results using EthicalSystem, the variable con-
structed using factor analysis that combines the different
information provided in the four code of ethics variables.
These results are provided in Table 5. Consistent with the
main analysis of the paper, we find a significant negative
association between |DACC| and EthicalSystem (p \ 0.01).
Table 6 presents the results for EthicalSystem separately
for stronger and weaker investor protection countries.
Again consistent with our prior results, we find that the
negative association between |DACC| and EthicalSystem is
driven by firms in weaker investor protection countries as
the coefficients for the low protection subsamples are
negative and significant (p \ 0.01) for all three measures. None of the coefficients are significant for the high-pro-
tection subsamples.
We next examine a subsample of firms that just meet or
beat analysts’ earnings forecasts. Since managers have
strong incentives to beat benchmarks, firms just beating
earnings targets are more likely to be engaging in oppor-
tunistic earnings management (Dechow and Skinner 2000).
We first measure the difference between the firm’s actual
earnings per share (EPS) and the median analyst forecast
for each company. We then create two subsamples: firms
that just beat the median analyst forecast (i.e., the actual
EPS is greater than the median analyst forecast by 10 % or
less) and firms that just miss the median forecast (i.e., the
actual EPS is less than the median analyst forecast by 10 %
or less). The results for these two subsamples are provided
in Table 7. For the subsample of firms that just beat ana-
lysts’ forecasts, the negative association between the code
of ethics variables and |DACC| is significant (p \ 0.01) for all of the ethics code variables other than Human Rights.
Only the coefficient for Implementation is significant for
Table 6 Test of H2: The investor protection hypothesis (dependent variable: |DACC|)
Low anti-dir
right
High anti-dir
right
Low anti-self-
dealing
High anti-self-
dealing
Code law Common law
EthicalSystem -0.010*** (-2.895) 0.000 (0.034) -0.004*** (-2.671) 0.004 (0.500) -0.004*** (-2.453) 0.003 (0.396)
PDI -0.000 (-0.005) 0.002 (1.403) 0.000 (0.514) 0.028*** (7.700) 0.000 (0.622) 0.101*** (11.533)
IDV 0.004*** (5.934) 0.002*** (2.673) -0.000** (-2.064) 0.017*** (7.421) -0.001*** (-2.818) 0.047*** (11.740)
MAS 0.001*** (4.549) 0.000 (1.058) -0.000 (-0.276) 0.025*** (8.227) -0.000 (-0.172) 0.053*** (12.139)
UAI -0.001*** (-3.042) -0.000 (-1.604) -0.000 (-0.824) -0.009*** (-7.953) -0.000 (-1.367) -0.003* (-1.774)
GDP Growth 0.663*** (3.046) 0.322* (1.949) -0.163* (-1.715) 0.968*** (3.261) -0.202** (-2.173) 0.688** (2.129)
Inflation 0.206 (0.474) 0.454 (1.531) 0.060 (0.289) 1.570*** (2.589) 0.029 (0.143) 3.114*** (4.479)
Corporate
governance
-0.010* (-1.905) 0.020*** (3.829) 0.001 (0.250) -0.025*** (-2.765) -0.001 (-0.348) 0.000 (0.016)
ROA -0.031 (-0.521) 0.207*** (3.646) 0.090* (1.842) 0.206*** (3.167) 0.092* (1.942) 0.257*** (4.333)
BM -0.004* (-1.764) 0.000 (0.160) -0.001 (-1.609) 0.008* (1.704) -0.001* (-1.834) -0.008 (-1.429)
LEV -0.005 (-0.700) -0.005 (-1.154) 0.003 (0.650) -0.008 (-0.857) 0.002 (0.472) -0.039*** (-3.942)
ISSUE 0.037* (1.702) 0.049** (2.176) 0.017 (1.375) 0.072* (1.873) 0.017 (1.415) 0.089** (2.252)
LOSS 0.022*** (4.061) 0.005 (0.656) 0.003 (1.255) 0.032*** (2.818) 0.004 (1.383) 0.040*** (3.295)
Constant -0.314*** (-14.349) 0.000 (.) 0.000 (.) -0.185*** (-6.958) 0.000 (.) 0.108*** (2.641)
Year dummies Yes Yes Yes Yes Yes Yes
Industry dummies Yes Yes Yes Yes Yes Yes
N 4084 4520 4765 3839 4831 3773
Adj. R 2
0.378 0.244 0.078 0.305 0.083 0.312
Please refer to the Appendix for variable definitions. *, **, and *** indicate statistical significance at the 10, 5, and 1 % levels (two-tailed),
respectively
Corporate Codes of Ethics, National Culture, and Earnings Discretion 157
123
T a b le
7 R e g re ss io n o f d is c re ti o n a ry
a c c ru a ls
a n d m e e t o r b e a t a n a ly st s’
fo re c a st s
V a ri a b le s
(1 )
(2 )
(3 )
(4 )
(5 )
(6 )
(7 )
(8 )
(9 )
(1 0 )
B e a t e a rn in g s fo re c a st
b y \ 1 0 %
S y st e m s
- 0 .0 1 4 * * *
(- 5 .7 1 0 )
- 0 .0 0 3
(- 1 .3 5 4 )
Im p le m e n ta ti o n
- 0 .0 1 3 * * *
(- 6 .0 8 0 )
- 0 .0 0 6 * * * (-
2 .5 8 1 )
C o rr u p ti o n
- 0 .0 1 3 * * *
(- 4 .2 1 7 )
- 0 .0 0 3
(- 0 .9 3 7 )
H u m a n ri g h ts
0 .0 0 1
(0 .3 0 5 )
- 0 .0 0 3
(- 1 .1 2 1 )
E th ic a lS y st e m
- 0 .0 1 8 * * *
(- 5 .6 3 9 )
- 0 .0 0 5
(- 1 .3 5 0 )
P D I
0 .0 0 1 * * *
(6 .8 2 4 )
0 .0 0 1 * * * (6 .7 3 9 )
0 .0 0 1 * * *
(7 .5 1 9 )
0 .0 0 1 * * *
(4 .2 6 7 )
0 .0 0 1 * * *
(7 .7 6 9 )
0 .0 0 1 * * *
(4 .0 2 3 )
0 .0 0 1 * * *
(4 .2 0 7 )
0 .0 0 1 * * *
(4 .0 6 3 )
0 .0 0 1 * * *
(3 .9 2 2 )
0 .0 0 1 * * *
(4 .0 3 6 )
ID V
0 .0 0 2 * * *
(7 .6 3 3 )
0 .0 0 1 * * *
(7 .3 2 8 )
0 .0 0 2 * * *
(7 .5 2 9 )
0 .0 0 0 * *
(2 .1 5 7 )
0 .0 0 2 * * *
(7 .9 1 7 )
0 .0 0 1 * * *
(5 .9 1 0 )
0 .0 0 1 * * *
(5 .9 2 4 )
0 .0 0 1 * * *
(5 .8 8 3 )
0 .0 0 1 * * *
(3 .8 2 6 )
0 .0 0 1 * * *
(5 .8 9 2 )
M A S
0 .0 0 1 * * *
(8 .8 1 4 )
0 .0 0 1 * * *
(9 .7 9 6 )
0 .0 0 1 * * *
(9 .8 7 4 )
0 .0 0 1 * * *
(5 .5 2 9 )
0 .0 0 1 * * *
(9 .4 4 6 )
0 .0 0 1 * * *
(8 .3 6 5 )
0 .0 0 1 * * *
(8 .6 9 5 )
0 .0 0 1 * * *
(8 .2 8 0 )
0 .0 0 1 * * *
(5 .3 3 3 )
0 .0 0 1 * * *
(8 .3 4 3 )
U A I
- 0 .0 0 1 * * *
(- 8 .6 6 2 )
- 0 .0 0 1 * * *
(- 9 .4 5 4 )
- 0 .0 0 1 * * *
(- 9 .8 9 4 )
- 0 .0 0 1 * * *
(- 6 .1 6 7 )
- 0 .0 0 1 * * *
(- 9 .3 7 2 )
- 0 .0 0 1 * * *
(- 5 .9 3 6 )
- 0 .0 0 1 * * *
(- 5 .9 8 3 )
- 0 .0 0 1 * * *
(- 6 .4 2 0 )
- 0 .0 0 1 * * *
(- 4 .5 3 2 )
- 0 .0 0 1 * * *
(- 6 .1 6 3 )
G D P G ro w th
0 .0 0 4
(0 .0 4 5 )
- 0 .0 1 9
(- 0 .2 0 8 )
0 .0 5 2
(0 .5 4 2 )
- 0 .1 1 8
(- 1 .3 3 4 )
0 .0 1 6
(0 .1 6 1 )
0 .0 8 2
(0 .8 5 8 )
0 .0 7 5
(0 .7 8 7 )
0 .0 5 7
(0 .5 2 1 )
0 .1 6 2 *
(1 .8 1 3 )
0 .0 5 4
(0 .4 8 7 )
In fl a ti o n
0 .6 3 9 * * *
(3 .1 0 4 )
0 .6 6 7 * * *
(3 .2 3 9 )
0 .4 8 0 * *
(2 .1 2 1 )
0 .9 5 8 * * *
(3 .9 8 1 )
0 .5 4 7 * *
(2 .4 0 8 )
0 .6 2 7 * * *
(3 .5 1 2 )
0 .6 4 9 * * *
(3 .5 6 6 )
0 .6 3 2 * * *
(3 .2 3 4 )
0 .5 7 9 * * *
(3 .1 9 2 )
0 .6 5 1 * * *
(3 .3 1 5 )
C o rp o ra te
G o v .
0 .0 0 1
(0 .2 9 7 )
- 0 .0 0 0
(- 0 .0 3 8 )
- 0 .0 0 6
(- 1 .4 4 3 )
- 0 .0 0 4
(- 0 .9 8 8 )
- 0 .0 0 2
(- 0 .3 4 4 )
0 .0 0 4
(1 .0 2 2 )
0 .0 0 6
(1 .4 6 8 )
0 .0 0 2
(0 .4 0 1 )
0 .0 0 2
(0 .5 5 9 )
0 .0 0 3
(0 .6 7 4 )
R O A
0 .1 6 0 * * *
(3 .0 5 9 )
0 .1 7 3 * * *
(3 .4 8 2 )
0 .1 8 6 * * *
(3 .5 1 7 )
0 .0 2 5
(0 .4 9 3 )
0 .1 8 7 * * *
(3 .6 4 4 )
0 .2 1 5 * *
(2 .3 4 1 )
0 .2 1 6 * *
(2 .3 6 4 )
0 .2 0 6 * *
(2 .0 9 7 )
0 .0 8 2 *
(1 .8 3 0 )
0 .2 0 7 * *
(2 .1 1 2 )
L N S IZ E
0 .0 0 2 *
(1 .9 4 6 )
0 .0 0 2 * *
(2 .1 8 1 )
0 .0 0 1
(0 .6 3 0 )
0 .0 0 1
(1 .0 2 8 )
0 .0 0 2 *
(1 .8 7 4 )
0 .0 0 2 * * *
(2 .9 9 9 )
0 .0 0 3 * * *
(3 .3 2 1 )
0 .0 0 2 * * *
(2 .6 8 9 )
0 .0 0 2 * * *
(2 .7 8 5 )
0 .0 0 2 * * *
(2 .8 4 8 )
B M
- 0 .0 1 6 * * *
(- 3 .7 1 0 )
- 0 .0 1 5 * * *
(- 3 .5 7 7 )
- 0 .0 1 9 * * *
(- 4 .1 1 4 )
- 0 .0 1 9 * * *
(- 3 .3 0 9 )
- 0 .0 1 8 * * *
(- 3 .9 0 9 )
- 0 .0 0 6 *
(- 1 .7 5 1 )
- 0 .0 0 5
(- 1 .4 0 3 )
- 0 .0 0 5
(- 1 .2 1 0 )
- 0 .0 0 6
(- 1 .6 4 4 )
- 0 .0 0 4
(- 1 .1 3 3 )
L E V
0 .0 3 6 *
(1 .9 2 2 )
0 .0 3 5 *
(1 .8 4 3 )
0 .0 3 9 *
(1 .9 5 2 )
0 .0 2 0
(0 .8 7 6 )
0 .0 4 6 * *
(2 .2 5 6 )
0 .0 1 6
(0 .7 5 4 )
0 .0 1 4
(0 .6 5 3 )
0 .0 2 4
(1 .0 5 4 )
- 0 .0 0 9
(- 0 .4 4 7 )
0 .0 2 3
(1 .0 4 3 )
IS S U E
0 .0 1 0 *
(1 .9 5 0 )
0 .0 1 0 *
(1 .9 5 8 )
0 .0 1 0 *
(1 .7 2 3 )
0 .0 1 6 * * *
(3 .1 6 3 )
0 .0 1 0 *
(1 .7 5 6 )
0 .0 0 1
(0 .1 4 5 )
0 .0 0 0
(0 .0 4 9 )
- 0 .0 0 0
(- 0 .0 1 3 )
0 .0 1 1 * *
(2 .0 5 1 )
- 0 .0 0 1
(- 0 .0 8 9 )
158 C. Chen et al.
123
the subsample of firms that just miss the forecast. These
results are consistent with our main premise that codes of
ethics help constrain opportunistic earnings management
behavior and provide additional evidence in support of H1a
and H1b.
Sensitivity Tests
We next include a control in our model for the general
ethical behavior of the firm. Prior evidence indicates that
Corporate Social Responsibility (CSR) is negatively rela-
ted to earnings management (e.g., Kim et al. 2012). One
possibility is that the presence and quality of a code of
ethics are related to the already established ethical behavior
of the firm, and, as a result, unlikely to induce changes in
decision-making. Following Brammer and Pavelin (2006),
we utilize data from EIRIS to construct a measure of CRS
and include this variable in our model along with the
EthicalSystem variable. Although the correlation between
the CSR variable and EthicalSystem is high (0.5146), our
results (provided in Table 8) are robust to the inclusion of
the CSR variable.
Since a few of the control variables are not changing
over time (for instance, the culture variables), a cross-
sectional dependence of data across time has the potential
to bias our results. For this reason, as a sensitivity test, we
estimate for each year/industry a separate regression of
the absolute value of abnormal accruals on the Ethi-
calSystem variable and control variables. Then, in the
spirit of Fama and MacBeth (1973), we test for the sig-
nificance of the estimated coefficients. The results (pro-
vided in Table 9) indicate that the coefficient for
EthicalSystem is, as in our main analysis, negative and
significant, providing further evidence for our first set of
hypotheses. We also confirm our results using an ordinary
least squares (OLS) regression with Newey–West stan-
dard errors.
In order to mitigate potential endogeneity, we also use a
change specification to examine the relationship between
codes of ethics and discretionary accruals. These results are
provided in Table 10. The national culture variables drop
out as they are constant over the sample period. We find
significant negative coefficients for the changes in Systems,
Corruption, and EthicalSystem when the change in |DACC|
is the dependent variable. Although the coefficients for
Implementation and Human Rights are negative, they are
not significant. These results alleviate the concern of
endogeneity due to omitted variables and reverse causality.
They are consistent with an increase in the strength of the
firm’s code of ethics being associated with a decrease in
opportunistic earnings management and help strengthen
our earlier inferences.
T a b le
7 c o n ti n u e d
V a ri a b le s
(1 )
(2 )
(3 )
(4 )
(5 )
(6 )
(7 )
(8 )
(9 )
(1 0 )
B e a t e a rn in g s fo re c a st
b y \ 1 0 %
L O S S
- 0 .3 3 8 * * *
(- 2 2 .5 7 1 )
- 0 .3 3 3 * * *
(- 2 2 .5 3 3 )
- 0 .3 5 2 * * *
(- 2 1 .4 9 8 )
- 0 .3 2 9 * * *
(- 1 8 .8 5 9 )
- 0 .3 5 5 * * *
(- 2 1 .6 0 1 )
- 0 .3 7 5 * * *
(- 1 4 .0 5 6 )
- 0 .3 7 5 * * *
(v 1 4 .1 0 9 )
- 0 .3 9 7 * * *
(- 1 3 .4 7 8 )
- 0 .3 6 3 * * *
(- 1 0 .7 7 1 )
- 0 .3 9 8 * * *
(- 1 3 .4 6 5 )
C o n st a n t
0 .2 8 4 * * *
(1 0 .1 7 3 )
0 .2 7 3 * * *
(9 .6 9 9 )
0 .2 9 0 * * *
(9 .6 4 4 )
0 .3 3 9 * * *
(1 0 .9 9 6 )
0 .2 4 1 * * *
(7 .6 8 1 )
0 .2 8 3 * * *
(7 .0 6 4 )
0 .2 7 3 * * *
(6 .8 2 5 )
0 .2 9 6 * * *
(6 .7 5 4 )
0 .3 0 4 * * *
(6 .7 0 1 )
0 .2 8 5 * * *
(6 .5 3 8 )
O b se rv a ti o n s
4 4 1 5
4 4 1 4
3 9 1 3
2 6 6 3
3 9 1 2
2 1 9 2
2 1 9 1
1 9 1 3
1 4 8 1
1 9 1 2
A d j. R 2
0 .2 8 3
0 .2 8 3
0 .2 9 1
0 .2 9 7
0 .2 9 4
0 .2 8 8
0 .2 8 9
0 .3 0 0
0 .2 7 4
0 .3 0 1
R o b u st
t- st a ti st ic s in
p a re n th e se s
* * * p \
0 .0 1 , * * p \
0 .0 5 , * p \
0 .1
Corporate Codes of Ethics, National Culture, and Earnings Discretion 159
123
Finally, as another way to control for uneven country
representations in the sample, we use the country-year
medians of the regression variables, and rerun the analysis
(e.g., Hail and Leuz 2006). Untabulated results confirm our
earlier findings. 13
Conclusions
We examine the role of codes of ethics in reducing the
extent to which managers act opportunistically in reporting
earnings. We measure the quality of a firm’s code of ethics
based on its existence and comprehensiveness, how it is
implemented, the existence of a human rights policy, and
the existence of policies and procedures related to bribery
and corruption. We rely on the firm’s level of discretionary
accruals as a proxy for earnings management. We find that
companies with higher-quality codes of ethics are associ-
ated with higher earnings quality (lower earnings man-
agement) and that these results are driven by firms from
countries characterized by weaker investor protection legal
systems. Our results are also confirmed for a sample of
firms that just meet or beat analysts’ forecasts, which are at
greater risk for opportunistic earnings reporting.
An important implication of our findings is that corpo-
rate codes of ethics can be a viable mechanism for deter-
ring opportunistic reporting behavior when country-level
institutions fail to deter such behavior and when incentives
or monitoring mechanisms are too costly or ineffective.
Our evidence has important implications for firms, regu-
lators, and investors as they continue to evaluate the suf-
ficiency and impact of codes of ethics around the world.
For example, shareholders in emerging countries with
weak corporate governance may consider demanding that
the firm adopt a code of ethics to ensure less opportunistic
reporting behavior. Regulators (especially those in weak
investor protection regimes) should also consider mandat-
ing the adoption of codes of ethics.
We close with some caveats and suggestions for
future research. While we provide evidence consistent
with high-quality codes of ethics influencing the oppor-
tunistic earnings behavior of managers, causality is dif-
ficult to prove. Although we take efforts to address
Table 8 Test of H1: Regression of discretionary accruals on Ethi- calSystem factor with CSR
|DACC|
EthicalSystem -0.009*** (-3.674)
PDI 0.001*** (4.513)
IDV 0.002*** (11.820)
MAS 0.001*** (11.365)
UAI -0.001*** (-10.217)
GDP growth 0.677*** (5.931)
Inflation 1.164*** (6.354)
Corporate governance 0.002 (0.656)
Corporate social responsibility 0.000 (0.271)
ROA 0.129*** (2.796)
LNSIZE 0.003*** (5.453)
BM -0.002 (-0.519)
LEV 0.046*** (3.188)
ISSUE 0.008* (1.956)
LOSS -0.337*** (-24.709)
Constant 0.173*** (2.808)
Year dummies Yes
Industry dummies Yes
N 8593
Adj. R 2
0.280
Please refer to the Appendix for variable definitions. *, **, and ***
indicate statistical significance at the 10, 5, and 1 % levels (two-
tailed), respectively
Table 9 Regression of discretionary accruals on EthicalSystem FactorA Fama–MacBeth type estimation
|DACC|
EthicalSystem -0.173*** (-3.296)
PDI 0.012 (1.319)
IDV 0.036** (2.258)
MAS 0.006 (0.848)
UAI 0.005 (0.521)
GDP Growth 0.917** (2.176)
Inflation -1.589 (-1.105)
Corporate governance -0.051 (-0.930)
ROA -0.546 (-0.542)
LNSIZE -0.015 (-0.182)
BM 0.125 (1.319)
LEV 0.331** (2.319)
ISSUE -0.027 (-0.569)
LOSS -0.175*** (-2.948)
Constant -3.601** (-2.060)
N. 8604
N of groups 53
R 2
0.149
In the first step, for each year-industry group we regress the absolute
value of abnormal accruals on the factor identified in the factor
analysis (EthicalSystem) and control variables. Then, in the second
step, we aggregate the coefficient estimates from the first step
Please refer to the Appendix for variable definitions. *, **, and ***
indicate statistical significance at the 10, 5, and 1 % levels (two-
tailed), respectively
13 We obtain 163 country-year medians of the regression variables
under this procedure. The EthicalSystem variable remains negative
and significant at p \ 0.10.
160 C. Chen et al.
123
alternative explanations, the strength of our inferences is
dependent on how well we have addressed other mech-
anisms that could explain our results. Secondly, we
utilize data from EIRIS to construct our code of ethics
variables. Since EIRIS data are based on surveys
administered to firms, we cannot eliminate the possibility
that some companies either mismeasure or bias their
responses.
Additional research is needed to better understand the
specific mechanisms through which codes of ethics impact
opportunistic reporting. We also believe that our under-
standing of the effectiveness of codes of ethics in miti-
gating opportunistic behavior will be enhanced by gaining
a better understanding of cross-country differences in the
content and implementation of codes. Finally, our data
indicate that although codes have become more prevalent
worldwide, significant variation remains regarding the
quality of the codes and their implementation. We leave it
to future research to determine if these gaps in quality
are reduced over time and what impact this has on the
relationship between codes of ethics and opportunistic
behavior.
Acknowledgments We thank seminar participants at the 2013 European Accounting Association Annual Congress and 2013
American Accounting Association Annual Meeting
Appendix: Variable Definitions
Test variables (from EIRIS):
1. Code of Ethics Systems The first question about the
firm’s code of ethics is whether the company has a
code of ethics and, if so, how comprehensive is it. The
answer is either no, limited, basic, intermediate, or
advanced.
2. Code of Ethics Implementation The second question is
whether the company has a system for implementing a
code of ethics and, if so, how comprehensive is it. The
answer is either no, limited, basic, intermediate, or
advanced.
3. Corruption Does the company have policies and
procedures on bribery and corruption? Here, the firm
can have a clear policy and procedures, an adopted
policy, or no policy disclosed.
4. Human Rights Policy and System Overall Rating What
is the extent of policy addressing human rights issues?
The answer is either no evidence of, has adopted, or
clearly communicates.
Culture-related variables from Hofstede (2001)
PDI: Power distance score
IDV: Individualism score
Table 10 Regressions of change in discretionary accruals
Variables (1) (2) (3) (4) (5)
DSystems -0.019*** (-2.949)
DImplementation -0.011 (-1.619)
DCorruption -0.024** (-2.504)
DHuman Rights -0.011 (-1.027)
DEthicalSystem -0.028*** (-2.684)
DGDP Growth -0.170 (-1.362) -0.167 (-1.333) -0.242* (-1.761) -0.138 (-1.146) -0.242* (-1.758)
DInflation 3.192*** (9.837) 3.160*** (9.740) 3.559*** (9.800) 3.444*** (10.411) 3.559*** (9.802)
DCorp_Governance 0.016 (1.418) 0.015 (1.330) 0.017 (1.338) -0.004 (-0.356) 0.018 (1.428)
DROA 0.084 (1.520) 0.084 (1.518) 0.104* (1.717) 0.023 (0.318) 0.106* (1.747)
DLNSIZE -0.039*** (-2.850) -0.040*** (-2.914) -0.048*** (-3.178) -0.027* (-1.855) -0.048*** (-3.151)
DBM -0.036* (-1.923) -0.037** (-1.971) -0.042** (-1.991) -0.032 (-1.539) -0.041* (-1.956)
DLEV 0.324*** (8.225) 0.323*** (8.183) 0.344*** (7.765) 0.154*** (3.357) 0.343*** (7.760)
DISSUE -0.004 (-0.518) -0.003 (-0.470) -0.005 (-0.621) 0.014* (1.793) -0.006 (-0.656)
DLOSS -0.016 (-0.627) -0.017 (-0.647) -0.015 (-0.502) -0.080*** (-2.756) -0.014 (-0.485)
Constant -0.044*** (-10.122) -0.045*** (-10.156) -0.056*** (-11.414) -0.032*** (-7.268) -0.053*** (-10.496)
Year dummies No No No No No
Industry dummies No No No No No
Observations 8391 8389 7289 5284 7288
Adj. R 2
0.025 0.024 0.028 0.030 0.028
t Statistics in parentheses
*** p \ 0.01, ** p \ 0.05, * p \ 0.1
Corporate Codes of Ethics, National Culture, and Earnings Discretion 161
123
MAS: Masculinity score
UAI: Uncertainty avoidance score
Country-level control variables
INFLATION: Measured by the annual growth rate of the
GDP implicit deflator, showing the rate of price change in
the economy as a whole. The GDP implicit deflator is the
ratio of GDP in current local currency to GDP in constant
local currency from the World Bank.
GDP_GROWTH: Growth of gross domestic product,
measured as the annual percentage growth rate of GDP at
market prices based on constant local currency also from
the World Bank
Firm-level control variables
CORPORATE GOVERNANCE: We measure Corporate
Governance with a score from zero to three based on the
survey question ‘‘Does the Company: (1) separate the roles
of chairman and chief executive (2) have a board com-
prising more than 33 % independent directors, (3) have an
audit committee comprising a majority of independent
directors, and (4) disclose director remuneration. Compa-
nies’ answers could be ‘‘none,’’ ‘‘one,’’ ‘‘some,’’ or ‘‘all.’’
Therefore, the range of this measure is 0-3.
ROA: Return on Assets, measured as net income
deflated by the beginning of the year total assets.
LNSIZE: Natural logarithm of market value of equity
BM: Book-to-market ratio
LEV: Leverage ratio, measured by debt to total assets
ISSUE: 1 if shareholders’ equity increases by more than
10 %; 0 otherwise
LOSS: 1 if the firm reported a loss during the year; 0
otherwise
References
Baiman, S. (1982). Agency research in managerial accounting.
Journal of Accounting Literature, 1, 154–213.
Ball, R. (2009). Market and political/regulatory perspectives on the
recent accounting scandals. Journal of Accounting Research,
47(2), 277–323.
Behn, B. K., Gotti, G., Herrmann, D., & Kang, T. (2013).
Classification shifting in an international setting: Investor
protection and financial analysts monitoring. Journal of Inter-
national Accounting Research, 12(2), 27–50.
Bicchieri, C. (2006). The grammar of society: The nature and
dynamics of social norms. New York: Cambridge University
Press.
Boatright, J. (2009). Ethics and the conduct of business (6th ed.).
Upper Saddle River, NJ: Prentice Hall.
Bosse, D., & Phillips, R. (2016). Agency theory and bounded self-
interest. Academy of Management Review, 41, 276–297.
Bowen, R. M., Rajgopal, S., & Venkatachalam, M. (2008). Account-
ing discretion, corporate governance, and firm performance.
Contemporary Accounting Research, 25(2), 351–405.
Brammer, S., & Pavelin, S. (2006). Corporate reputation and social
performance : The importance of fit. Journal of Management
Studies, 43(3), 435–455.
Chih, H.-L., Shen, C.-H., & Kang, F.-C. (2008). Corporate social
responsibility, investor protection, and earnings management:
Some international evidence. Journal of Business Ethics,
79(1–2), 179–198.
Choi, T. H., & Pae, J. (2011). Business ethics and financial reporting
quality: Evidence from Korea. Journal of Business Ethics,
103(3), 403–427.
Cohen, J. R., Holder-Webb, L., Sharp, D. J., & Pant, L. W. (2007).
The effects of perceived fairness on opportunistic behavior.
Contemporary Accounting Research, 24(4), 1119–1138.
Davidson, B. I., & Stevens, D. E. (2012). Can a code of ethics
improve manager behavior and investor confidence? An exper-
imental study. The Accounting Review, 88(1), 51–74.
Dechow, P., & Skinner, D. (2000). Earnings management: reconciling
the views of accounting academics, practitioners, and regulators.
Accounting Horizons, 14(2), 235–250.
DeFond, M. L., & Subramanyam, K. (1998). Auditor changes and
discretionary accruals. Journal of Accounting and Economics,
25(1), 35–67.
Djankov, S., La Porta, R., Lopez-de-Silanes, F., & Shleifer, A. (2008).
The law and cconomics of self-dealing. Journal of Financial
Economics, 88(3), 430–465.
Dyck, A., & Zingales, L. (2004). Private benefits of control: An
international comparison. Journal of Finance, 59(2), 537–600.
Eisenhart, K. (1989). Agency theory: An assessment and review.
Academy of Management Review, 14(1), 57–74.
Fama, E. F., & MacBeth, J. D. (1973). Risk, return, and equilibrium:
Empirical tests. The Journal of Political Economy, 81(3), 607–636.
Fehr, E., & Gächter, S. (2000). Fairness and retaliation: The
economics of reciprocity. The Journal of Economic Perspectives,
14(3), 159–181.
Fehr, E., & Schmidt, K. M. (1999). A theory of fairness, competition,
and cooperation. Quarterly Journal of Economics, 114(3),
817–868.
Ferraro, F., Pfeffer, J., & Sutton, R. (2005). Economics language and
assumptions: How theories can become self-fulfilling. Academy
of Management Review, 30(1), 8–24.
Frankel, M. S. (1989). Professional codes: Why, how, and with what
impact? Journal of Business Ethics, 8(2–3), 109–115.
Gray,S.J.(1988). Towards atheoryofculturalinfluence onthe development
of accounting systems internationally. Abacus, 24(1), 1–15.
Hail, L., & Leuz, C. (2006). International differences in the cost of
equity capital: Do legal institutions and security regulation
matter? Journal of Accounting Research, 44(3), 485–531.
Han, S., Kang, K., Salter, S., & Yoo, Y. K. (2010). A cross-country
study on the effects of national culture on earnings management.
Journal of International Business Studies, 41(1), 123–141.
Healy, P., & Wahlen, J. (1999). A review of the earnings management
literature and its implications for standard setting. Accounting
Horizons, 13(4), 365–383.
Helin, S., & Sandstrom, J. (2007). An inquiry into the study of
corporate codes of ethics. Journal of Business Ethics, 75(3),
253–271.
Hofstede, G. (2001). Culture’s consequences: Comparing values,
behaviors, institutions, and organizations across nations. Thou-
sand Oaks, CA: Sage.
Huang, P., Louwers, T. J., Moffitt, J. S., & Zhang, Y. (2008). Ethical
management, corporate governance, and abnormal accruals.
Journal of Business Ethics, 83(3), 469–487.
162 C. Chen et al.
123
Jones, J. J. (1991). Earnings management during import relief
investigations. Journal of Accounting Research, 29(2), 193–228.
Kaptein, M. (2004). Business codes of multinational firms: What do
they say? Journal of Business Ethics, 50(1), 13–31.
Kaptein, M. (2015). The effectiveness of ethics programs: The role of
scope, composition, and sequence. Journal of Business Ethics,
132(2), 415–431.
Kaptein, M., & Schwartz, M. S. (2008). The effectiveness of business
codes: A critical examination of existing studies and the
development of an integrated research model. Journal of
Business Ethics, 77(2), 111–127.
Kaptein, M., & Wempe, J. (1998). Twelve Gordian knots when
developing an organizational code of ethics. Journal of Business
Ethics, 17(8), 853–869.
Kim, Y., Park, M. S., & Wier, B. (2012). Is earnings quality
associated with corporate social responsibility? The Accounting
Review, 87(3), 761–796.
Kothari, S., Leone, A. J., & Wasley, C. E. (2005). Performance
matched discretionary accrual measures. Journal of Accounting
and Economics, 39(1), 163–197.
KPMG. (2008). Business codes of the Global 200: Their prevalence,
content and embedding. The Netherlands: KPMG.
KPMG. (2014). The business codes of the Fortune Global 200: What
the largest companies in the world say and do. Amstelveen:
KPMG.
La Porta, R., Lopez-de-Silanes, F., Schleifer, A., & Vishny, R. W.
(1998). Law and finance. Journal of Political Economy, 106,
1113–1155.
Ladd, J. (1991). The quest for a code of professional ethics: An
intellectual and moral confusion. In D. G. Johnson (Ed.), Ethical
issues in engineering. Englewood Cliffs, NJ: Prentice Hall.
Lere, J. C., & Gaumnitz, B. R. (2003). The impact of codes of ethics
on decision making: Some insights from information economics.
Journal of Business Ethics, 48(4), 365–379.
Leuz, C., Nanda, D., & Wysocki, P. D. (2003). Earnings management
and investor protection: An international comparison. Journal of
Financial Economics, 69(3), 505–527.
Maddala, G. S. (1983). Limited-dependent and qualitative variables
in econometrics. Cambridge: Cambridge University Press.
Mathews, C. (1987). Codes of ethics: Organizational behavior and
misbehavior. In W. Frederick (Ed.), Research in corporate social
performance and policy 9 (pp. 107–130). Greenwich: JAI Press.
Mazar, N., Amir, O., & Ariely, D. (2008). The dishonesty of honest
people: A theory of self-concept maintenance. Journal of
Marketing Research, 45(6), 633–644.
McDonald, G. M. (2009). An anthology of codes of ethics. European
Business Review, 21(4), 344–372.
McKendall, M., DeMarr, B., & Jones-Rikkers, C. (2002). Ethical
compliance programs and corporate illegality: Testing the
assumptions of the corporate sentencing guidelines. Journal of
Business Ethics, 37(4), 367–383.
Misangyi, V. F., & Acharya, A. G. (2014). Substitutes or complements?
A configurational examination of corporate governance mecha-
nisms. Academy of Management Journal, 57(6), 1681–1705.
Molander, E. A. (1987). A paradigm for design, promulgation and
enforcement of ethical codes. Journal of Business Ethics, 6(8),
619–631.
Paine, L., Deshpande, R., Margolis, J. D., & Bettcher, K. E. (2005).
Up to code: Does your company’s conduct meet world-class
standards? Harvard Business Review, 38(12), 122–133.
Robbins, S. (1988). Management. Englewood Cliffs, NJ: Prentice Hall. Roychowdhury, S. (2006). Earnings management through real
activities manipulation. Journal of Accounting and Economics,
42(3), 335–370.
Schipper, K. (1989). Commentary on earnings management. Ac-
counting Horizons, 3(4), 91–102.
Scholtens, B., & Dam, L. (2007). Cultural values and international
differences in business ethics. Journal of Business Ethics, 75(3),
273–284.
Singh, J. B. (2006). Ethics programs in Canada’s largest corporations.
Business and Society Review, 111(2), 119–136.
Singh, J. B. (2011). Determinants of the effectiveness of corporate
codes of ethics: An empirical study. Journal of Business Ethics,
101(3), 385–395.
Somers, M. J. (2001). Ethical codes of conduct and organizational
context: A study of the relationship between codes of conduct,
employee behavior and organizational values. Journal of Busi-
ness Ethics, 30(2), 185–195.
Stevens, B. (1994). An analysis of corporate ethical code studies:
‘‘Where do we go from here?’’. Journal of Business Ethics,
13(1), 63–69.
Corporate Codes of Ethics, National Culture, and Earnings Discretion 163
123
Journal of Business Ethics is a copyright of Springer, 2018. All Rights Reserved.
- Corporate Codes of Ethics, National Culture, and Earnings Discretion: International Evidence
- Abstract
- Introduction
- Literature Review and Hypotheses Development
- Prior Evidence on Corporate Codes of Ethics
- Hypothesis Development: Corporate Codes of Ethics and Earnings Discretion
- Hypothesis Development: The Mediating Role of Investor Protection
- Research Design
- Data and Sample Selection
- Measurement of Variables
- Discretionary Accruals
- Measures of Codes of Ethics
- Measures of Investor Protection
- Empirical Models
- Results
- Sensitivity Tests
- Conclusions
- Acknowledgments
- Appendix: Variable Definitions
- Country-level control variables
- Firm-level control variables
- References