labelle et al - 2010 - ethics, diversity management, and financial reporting quality

The framework is used to examine the relation between governance and business ethics, as proxied by diversity management DM, and financial reporting quality, as proxied by the magnitude

Ethics, Diversity Management,

and Financial Reporting Quality

Re´al Labelle Rim Makni Gargouri Claude Francoeur

ABSTRACT This article proposes and empirically tests

a theoretical framework incorporating Reidenbach and

model of corporate moral development The framework

is used to examine the relation between governance and

business ethics, as proxied by diversity management

(DM), and financial reporting quality, as proxied by the

magnitude of earnings management (EM) The level of

DM and governance quality are measured in accordance

with the ratings of Jantzi Research (JR), a leading

pro-vider of social and governance research for institutional

investors This DM score is part of an index developed by

JR that investment managers use to integrate DM criteria

into their investment decisions As expected, a negative

relation between corporate DM development and

financial reporting quality is found while controlling for

other factors known in the literatures on governance and

accounting choices to affect earnings quality Despite

some caveats presented in conclusion, this study

con-tributes to the ethics, governance, and financial reporting

literatures by studying the dynamics between governance

and ethics in the prevention of EM.

KEY WORDS: ethics, diversity, diversity management,

governance, financial reporting quality, earnings

man-agement, earnings quality

Introduction

As accounting is closely regulated, the issue of

financial reporting quality (FRQ) is often addressed

as a matter of compliance to various governance

regulations rather than ethics This legal oversight

perspective prevails as business ethics does not appear

to be sufficiently developed to prevent managers

from manipulating the information they provide to

various stakeholders about their own performance.1

However, despite this emphasis on compliance,

FRQ also presupposes unwritten rules of ethical behavior Governance complements business ethics

In other words, if managers were expected to behave

in an ethical manner, governance would be less relevant2 and, in the limit, ethics could substitute for governance These ‘‘legalistic’’ and ‘‘ethical’’ perspectives correspond to distinct stages in the conceptual model of corporate moral development (CMD) proposed by Reidenbach and Robin (1991) This theory is inspired by the study of Kolhlberg (1984) on individual moral development, and Piaget (1932/1962) on child development The objective of this article is to use the theory of CMD to study the dynamics between governance and ethics in the prevention of earnings management (EM), a proxy for FRQ

To date, relatively few studies have examined the presumed relationship between FRQ and board of directors’ characteristics other than the ones com-plying with generally accepted governance principles

as reviewed in He et al (2009).3 The objective of this article is to propose and test a theory-based framework (Fig.1) where diversity management (DM) is presented as an ethical complement to statutory governance aimed at improving FRQ and ultimately at creating value Rather than concentrate

on the concept of independence or diversity of financial interests promulgated in statutory gover-nance, diversity of knowledge, competence, and organization’s values are also considered We use the Canadian Social Investment Database (CSID) com-piled by Jantzi Research (JR) on both DM and governance corporate policies to proxy for the firms’ propensity to respectively implement and respect such stakeholder- and shareholder-oriented policies FRQ is appraised by the firms’ propensity to more

or less manage earnings Other factors known to be

DOI 10.1007/s10551-009-0225-7

associated with the quality of earnings in various

areas of the accounting choices literature are also

controlled for

Governance comprises the set of principles or rules

aimed at improving the accuracy and reliability of

corporate disclosures with which firms have to

comply to insure investors’ protection Reidenbach

and Robin4 (1991) argue that an ethical approach

mainly based on rule compliance is at the lower level

of CMD They propose five stages of organizational

moral development including the legalistic (stage 2),

the responsive (stage 3), and the emergent ethical

(stage 4) organization In this study, a firm which

adopts a DM policy thus showing concern for other

corporate stakeholders than owners (stage 3 in R &

R, pp 278, 282) understands the value of not acting

solely on a legal basis (stage 2 in R & R, pp 276,

282) ‘‘Diversity management is a voluntary

organi-zational program designed to create greater inclusion

of all individuals into informal social networks and

formal company programs’’ (Gilbert et al., 1999, p

61) In order to reach R & R’s fourth stage of CMD,

i.e., the emerging ethical organization, top

manage-ment values must become organizational values and

be integrated in the organizational culture

Our research design is aimed at examining whe-ther the degree of moral or ethical development of a corporation is related to the quality of its financial reporting We do this by isolating the marginal effect

of the stakeholder-oriented policy of promoting diversity and employment equity from the gover-nance-presumed effect on FRQ as proxied by EM The diversity of boards, other than the one re-quired in governance codes to insure that the financial interests of board members are aligned with shareholders rather than managers, is also starting to permeate governance regulation around the world Many propositions for governance reforms have explicitly emphasized the importance of diversity in the boardroom (Adams and Ferreira,2008) In Great Britain, the Higgs report (Higgs, 2003), commis-sioned by the British Department of Trade and Industry, maintains that diversity can improve board efficiency Diversity of boards has also attracted the interest

of institutional investors for several years For in-stance, according to Carter et al (2003), the Teachers Insurance and Annuity Association-College Retirement Equities Fund has adopted a policy statement on cor-porate governance (CG), which stipulates that the board must be made up of qualified individuals that

Fig 1 Ethics, diversity management, and financial reporting quality.

reflect a diversity of experience, gender, race, and

age (TIAA-CREF,1997) TIAA uses diversity as an

investment criterion and considers this board

char-acteristic as fundamental in limiting executives’

dis-cretionary behavior (Carter et al., 2003)

Despite the presumed role of ethics or CMD as a

complement to governance regulation in controlling

opportunistic EM behavior, there is a near vacuum

of empirical literature on this subject This is most

likely due to the fact that business ethics is more

difficult to observe and measure than compliance

with governance regulation We propose to use DM

as defined above to proxy for the level of business

ethics or CMD To our knowledge, one exploratory

(Krishnan and Parsons, 2008) and two unpublished

studies (Francis et al., 2009; Gul et al., 2007) have

started to test the link between gender diversity as

measured by the proportion of women on boards

and FRQ We contend that the concept of DM is

much larger than gender diversity which may be

closer to affirmative action (Gilbert et al.,1999) than

to business ethics Again, because gender diversity is

easier to observe than DM, there is a growing

lit-erature investigating the impact of gender on various

corporate decisions such as capital structure decisions

(Huang and Kisgen, 2008), merger and acquisition

(Levi et al., 2008), and going public (Mohan and

Chen,2004)

In order to examine whether the degree to which

firms, instead of solely relying on compliance, also

stress ethics and diversity in the workplace have

better FRQ, we use data from the CSID compiled

by JR, a leading provider of research on governance

and social research for institutional investors The

CSID score for CG measures compliance with

governance regulations The CSID score for DM

incorporates several aspects of diversity in addition to

gender diversity on the board which was the sole

aspect examined in previous research Our sample is

composed of all the companies appraised by JR in

2004 and 2005 As hypothesized, our findings show

a significant negative association between DM, our

proxy for CMD, and EM, our proxy for FRQ

There is no statistically significant relation between

CG and FRQ We interpret these results as

con-firming that FRQ is related to a higher level of

CMD Indeed, the degree of implementation and

respect of an ethical stakeholder-oriented policy

such as the promotion of diversity on the board and

among management and employees appears to en-hance earnings quality This is a significant result as earnings quality is commonly used as an important governance tool in measuring management’s trans-parency and performance

The remainder of the article is organized as fol-lows The next section presents the theoretical framework of the study followed by a review of the literature in ‘‘Review of empirical literature’’ sec-tion ‘‘Model development, data collection, and variable measurement’’ section describes the meth-odology and empirical models used Results are presented and discussed in ‘‘Analysis and interpre-tation of results’’ section The last section presents the conclusions and caveats and proposes some avenues for future research

Theoretical framework Figure 1proposes a theory-based framework linking ethics or CMD and diversity in governance with the prevention of EM and ultimately value creation Governance is presented as both advisory (left-hand side) and fiduciary (right-hand side) in nature Both roles may affect or be affected by business ethics or CMD in a different manner

From the fiduciary perspective, qualified directors have to be independent and even specialists in monitoring, as per the stipulation of the current regulation and generally accepted governance prin-ciples concerning the composition of the board and its audit committee.5In theory and in corporate law, primacy is given to shareholders’ interests Agency theory, the main theory providing explanations for governance, predicts management interests are dif-ferent and even in conflict with those of share-holders Thus, the emphasis is on the diversity of opinions and financial interests of board members This kind of diversity is required from them so they are incited to keep on questioning or monitoring whether management stays in compliance with governance regulation and meets its fiduciary duty From the advisory perspective of governance (left-hand side of Fig 1), board effectiveness requires a diversity of knowledge, competences, and organiza-tional values to guide and contribute to organizaorganiza-tional learning and strategic decision making The emphasis

is on counseling and mentoring management, and not

on the statutory characteristics such as the diversity of

interests or the financial literacy of the directors

Theories based on resources, competences, and

organizational learning explain the more direct link

with value creation shown in Fig 1 Although

investors’ protection still matters, under that

per-spective, there is room to develop an organizational

culture or corporate policies which integrate a wider

range of stakeholders instead of having to focus by law

on shareholders This time, diversity is necessary for

board members to be able to ask knowledgeable

questions to shape the managerial decision-making

process and the organization’s culture For R & R

(1991), the moral development of a corporation is

determined by the organization’s culture and, in

reciprocal fashion, helps define that culture In

essence, it is the organization’s culture under the

impetus of top management and advisory

gover-nance, which undergoes moral development

Nowadays, society is demanding that the

eco-nomic development of corporations be equated with

their moral development The moral development

of a corporation can be classified according to the

degree to which its social responsibility is recognized

and blended with its economic mission R & R

(1991) propose a model of CMD comprised of 5

stages based on the following types of behavior

The first and the lowest stage of CMD is the

‘‘amoral organization’’ with no set of value other

than greed For this kind of organization,

gover-nance is just another set of rules to circumvent Stage

two is the ‘‘legalistic corporation’’ so named because

of the preoccupation it exhibits for compliance with

the letter of the law as opposed to its spirit The

principal emphasis is still on profitability, but the

difference between stage 2 and 1 is that the latter is

concerned with the legality of profits, not necessarily

with their morality Owners are still the principal

stakeholders These organizations are followers

and not social leaders Society can expect, for the

most part, that they adhere to the rules of statutory

governance

Contrary to their legalistic counterparts in stage 2,

the stage 3 ‘‘responsive corporations’’ begin to develop

cultures that contain values other than productivity

and a sense of legality Management understands the

value of not acting solely on a legal basis This is the case

when a firm adopts DM policies thus showing concern

for other corporate stakeholders than owners According to Gilbert et al (1999, p 61), ‘‘diversity management is a voluntary organizational program designed to create greater inclusion of all individuals into informal social networks and formal company programs.’’ In order to reach R & R’s fourth stage of CMD, i.e., ‘‘emerging ethical organization,’’ top management and board values must become organi-zational values or part of the organiorgani-zational culture This is the first stage to exhibit an active concern for ethical outcomes In R & R’s conceptual model of CMD, the final stage is the ‘‘ethical organization.’’ According to R & R, it is difficult to find organizations which have reached this level of development where ethics is fully integrated in the firm’s mission and organization’s culture and where there is a balanced concern for ethical and economic outcomes

This article examines whether the degree of moral

or ethical development of a corporation is related to the quality of its financial reporting In order to achieve this objective, we have to isolate the mar-ginal effect of CMD over the governance presumed effect on FRQ We use the stakeholder-oriented policy of promoting diversity and employment equity to proxy for CMD In the next section, we review the empirical literature on diversity and EM

Review of empirical literature Despite the presumed role of ethics or CMD as a complement to governance regulation in controlling opportunistic EM behavior, there is a near vacuum of empirical literature on the subject This may be due

to the fact that business ethics is more difficult to observe and measure than compliance with gover-nance regulation as proxied for instance by the per-centage of unrelated directors We use DM as defined above by Gilbert et al (1999, p 61) to proxy for the level of business ethics or CMD For these authors,

‘‘diversity management (DM) is a voluntary organi-zational program ….’’ Thus, DM results from a voluntary management decision which does not de-pend solely on laws, since sheer compliance to affirmative action quotas or targets represents only minimum acceptable standards of behavior ‘‘Ethical behavior focused on diversity management takes knowledge, commitment and work beyond the law’’

(Gilbert et al., 1999, p 73) As underlined in our

theoretical framework, corporate DM development

may run parallel to the responsive (R & R’s stage 3)

and emerging ethical (R & R’s stage 3) organizations’

moral development stages

To our knowledge, one exploratory (Krishnan and

Parsons,2008) and two unpublished studies (Francis

et al.,2009; Gul et al.,2007) have started to test the

link between a subset of DM, i.e., gender diversity as

measured by the proportion of women on boards,

and FRQ We contend that the concept of DM is

much larger than gender diversity which may be

closer to affirmative action (Gilbert et al.,1999) than

to business ethics Conceptually, DM is closely

re-lated to organizational culture and encompasses the

equitable management of several aspects of human

resources including gender, race, ethnicity, age,

personality, and cognitive style However, as most

studies to date have focused on gender diversity, our

literature review will mostly focus on that partial

view of diversity

Much attention has been paid to the diversity of

interests between directors and top management and

its indirect effect on performance (Francoeur et al.,

2008; Levi et al., 2008) This study focuses on both

the potential direct and indirect relations between

DM and agency costs where managers may take

advantage of information asymmetry to fiddle with

earnings (right hand side of Fig.1) The board of

directors generally oversees at least four important

functions (Mallin,2004; Monks and Minow,2004)

Managerial control and compliance with laws and

regulations are mostly related to fiduciary governance

(stage 2 of R & R) while questioning, informing, and

advising managers, and ensuring good relations with

the external environment are mostly related to the

advisory role of governance

Diversified boards can potentially not only improve

their oversight capacity but also the quality of their

decisions (Dallas, 2002) through their enlightened

questions and advices One of the most significant

governance challenges that managers, administrators,

and shareholders of the modern company face is

establishing an optimal mix within the board of

directors in terms of gender, race, and culture (Carter

et al.,2003) The business case for DM is closely linked

to agency as well as to resource-based theories

It suggests that a diversified board increases its

independence and activism thus ensuring better control over managers’ behavior (Carter et al.,2003,

2008) Adams (2008) asserts that women directors have an impact similar to the one of independent directors, particularly regarding control over mana-gerial discretion Smith et al (2006) argue that the heterogeneous makeup of top management is an as-pect of good CG that helps constrain EM practices Boards of directors that include women members are purportedly better able to improve control and con-strain opportunistic EM behaviors (Gul et al.,2007) Furthermore, according to Cohen et al (1998), Kle-nke (2003), and Trinidad and Normore (2005), wo-men employ a more democratic, trust-based leadership style They also exhibit greater risk aversion

in financial decision making (Hinz et al.,1997; Powell and Ansic,1997; Riley and Chow,1992; Sunden and Surette,1998) Betz et al (1989), Mason and Mudrack (1996), and Clikeman et al (2001) find that women exhibit higher ethical values in their decisions than men do

Earnings may be managed while complying with the rules, but some manipulations may be unethical (Bruns and Merchant, 1990; Gaa, 2007) Opportu-nistic EM is likely to attract meticulous examination

by investors, analysts, and regulators, which increases the risks of litigation and loss of reputation Several studies have shown a strong association between EM and the risk of litigation (DuCharme et al., 2004; Heninger, 2001; Kasznik, 1999) and the loss of corporate reputation (Kaplan and Ravenscroft,2004; Hunton et al., 2006) Owing to their more pro-nounced ethical values and greater risk aversion, women directors might be more averse to EM and the above consequences, compared with their male counterparts (Gul et al., 2007)

Diversity management may thus affect EM in var-ious ways Diversified, trust-based as opposed to compliance- or obedience-oriented management style implies greater sharing of information among directors, and between directors, top management, and employees This reduction in information asym-metry putatively improves control over the corporate financial reporting function thus minimizing oppor-tunistic EM to camouflage mediocre performance and expropriate shareholders’ wealth

Beyond diversity of the board of directors and top management, DM systems and programs related to

equity and equality in employment are an important

dimension of diversity Armstrong et al (2008) posit

that diversity and employment equality management

policy is part of a High Performance Work System,

which improves the quality of work and

organiza-tional productivity In 2006, the Internaorganiza-tional Labour

Organization (ILO) mentioned that employment

equality and employee diversity are some of the

characteristics of high performance workplaces In

this context, the employees’ perceptions of their

organization’s values of ethics and equity motivate

them to put forth additional efforts (Lambert,2000)

The increase in employee diversity is a

contem-porary human resources management strategy whose

objective is to generate a strategic advantage for the

company (Cox,1991; Thomas,1991) and a potential

source of organizational efficiency (Miller, 1998)

Effective management of diversity could foster

cre-ativity, innovation, and problem-solving capacity

(Thomas,1992) DM practices and policies also

im-prove the quality of decision making (Flood et al.,

2005) and reinforce the organizational commitment

of employees (O’Connell and Russel,2005)

In order to summarize, DM policies purportedly

generate a sense of equity and satisfaction in the

workplace that provides incentives to act in the

company’s interests In that environment,

opportu-nistic EM behavior posing as a risk to undermine the

success of the organization is likely to be minimized

Thus, we hypothesize that the level of CMD as

proxied by DM putatively constrains managerial

discretion in EM We, therefore, expect a negative

relation between DM and EM while controlling for

CG and other FRQ determinants often found to be

relevant in the various areas of the accounting choices

literature

Model development, data collection,

and variable measurement

In this section, we first present the model and

sample, and then explain why the variables were

chosen and how they are measured In order to

examine the association between DM, our proxy for

CMD, and the magnitude of EM, our proxy for the

firm’s FRQ, we ran the following panel probit

model using Stata software:

EMit¼ a1þ a2DMit
1þ a3CGit
1þ a4CFPit
1

þ a5SIZEit
1þ a6DEBTit
1

þ a7EXCOMPit
1þ a8RISKit
1

þ a9MBRATIOit
1þ a10BLOCKit
1

where EM = 1 for firms with high absolute discre-tionary accruals (DA), and 0 for low absolute DA;

DM = JR diversity management score; CG = JR corporate governance score; CFP = corporate financial performance; SIZE = natural logarithm of total assets; DEBT = level of debt; RISK = sys-tematic risk b of the market model; MBRA-TIO = market to book ratio; EXCOMP = average ratio of bonus to total pay for all executives; BLOCK = cumulative percentage of shares owned

by blockholders (>10%); Industry: GOLD = Gold and precious minerals; OIL = Oil and gas; CONS = consumer products; IND = industrial products; COM = communication and media; MERCH = merchandizing

The dependent variable (EM) is measured for the period 2005–2006, while the explanatory and con-trol variables are measured for 2004–2005 This is done to avoid any potential endogeneity problem The initial sample consists of all the 160 compa-nies of the CSID for which DM and CG scores are compiled by JR Owing to their particular accounting standards and rules, 41 firms operating in the public and financial services sectors are excluded

An additional 41 firms are excluded due to missing data on accounting or stock market The final sample consists of 78 companies or 156 observations over 2 years

Dependent variable – financial reporting quality

We use EM through DA to proxy for FRQ DA are estimated using the cross-sectional version of the Jones model modified by Dechow et al (1995) Dechow et al (1995) eliminated the tendency of the Jones’ model to incorrectly measure DA when dis-cretion is exerted on the operating cash flow They adjusted the variations in operating cash flow to take into account the change in accounts receivable dur-ing the examination period Jones’ original model

presumes that managerial discretion is not exerted on

operating cash flow in the estimation period or in the

examination period In contrast, the modified model

presumes that all variations in credit sales in the

examination period are associated with EM This

modification is based mainly on the fact that it is

easier to manage earnings by manipulating the

rec-ognition of credit sales than that of cash sales The

modified version of Jones’ model is therefore

ex-pressed as follows:

TAt ¼ a1

1

At
1

þ a2ðDREVt
DRECtÞ

þ a3PPEtþ eit

TA = total accruals = net income (item 75 of the

Stock Guide database) less operating cash flow

(item 110 of the Stock Guide database); At-1 =

total assets (item 94 of the Stock Guide database)

at t - 1; DREV = revenues (item 61 of the Stock

Guide database) in year t less revenues in year t - 1

scaled by total assets at t - 1; DREC = net

receiv-ables (item 85 of the Stock Guide database) in year

t less net receivables in year t - 1 scaled by total

assets at t - 1; PPE = gross property plant and

equipment (item 89 of the Stock Guide database)

in year t scaled by total assets at t – 1

All variables are standardized by the amount of

total assets at t - 1 to reduce heteroskedasticity

Based on the estimated coefficients, we calculate

the nondiscretionary accruals (NDA) of the sample

as follows:

NDAt ¼ ^a1

1

At
1

þ ^a2ðDREVt
DRECtÞ

þ ^a3PPEt

where ^a1, ^a2 and ^a3 are the estimated coefficients

of a1, a2, and a3, respectively

The DA of the sample are measured by the

dif-ference between the total accruals and the

nondis-cretionary accruals:

DAt ¼ TAt
NDAt

As proposed by Huddart and Louis (2006), the

propensity of a firm to manage earnings (EM) is

measured by a dichotomous variable that takes the

value of 1 for firms with high absolute DA, and 0 for

low absolute DA firms We use the absolute value of

accruals as earnings that may be managed upward or

downward depending on the incentives In order to distinguish high and low DA firms, the absolute values of DA are first sorted in decreasing order for

2005 and 2006 Then, for each year, the sample is divided into two groups of 39 firms whose absolute values of DA fall on either side of the median The group with values higher than the median consists of companies whose magnitude of EM is high, whereas the group whose absolute DA values fall below the median consists of companies exhibiting less EM

Independent variable of interest – diversity management

As outlined earlier, the concept of DM is harder to observe and measure than gender diversity which is the only element of diversity examined in previous research in relation to EM We thus use the CSID compiled by JR6for 2004 and 2005 JR is a leading provider of governance and social, such as DM, research for institutional investors JR’s rating of DM

is part of an index that investment managers use to integrate DM criteria into their investment deci-sions As shown in Appendix 1, 75% of the DM rating on 10 concerns the firms’ stakeholder-ori-ented policies and efforts to recruit and integrate minorities and women in the workplace It includes the firm’s policy on diversity and employment equity, the managerial structure and responsibility, the public reporting on diversity issues, employee training and communication, recruitment, retention and promotion programs, parental benefits, and other diversity initiatives or benefits The percent-ages of women on the board of directors and among senior officers accounts for 12.5% of the total score each

Control variables Other factors in the governance and accounting choices literatures are known to constitute deterrents (governance, size, and to a lesser extent ownership concentration) or incentives for EM, namely, cor-porate financial performance (CFP), debt and com-pensation contracts, growth opportunities, and risk

In order to gauge the indirect advisory and direct fiduciary role the board of directors might play in preventing EM (Fig 1), we first include the quality

of the firm’s governance in the model (Niu,2006).

For that purpose, we use the CSID CG score

compiled by JR for the years 2004 and 2005 CG is a

composite index, as in the case of DM, measuring

corporate fiduciary governance such as the

inde-pendence of board members from management or

the separation of the chairperson and CEO positions,

the existence of formal CG principles, of a code of

business conduct and concerns about non-voting or

multivoting common shares, excessive

compensa-tion, and stock option plan dilution

The sign of the relation between the firm’s

ownership structure and its propensity to more or

less manage earnings is uncertain In governance,

ownership concentration in the hands of one or

more blocks of shareholders (BLOCK) may be seen

as another means of controlling executives’

discre-tion with regard to opportunistic EM behavior

(Bushee, 1998; Dempsy et al.,1993) On the other

hand, closely held firms may be less concerned with

earnings quality than widely held ones, as their

shareholders can obtain information directly from

the firm (Bushee et al., 2003; Gelb, 2000)

There-fore, we control for the firm’s ownership

concen-tration without qualifying its relation with EM We

use the Stock Guide database measure, i.e., the

cumulative percentage of shares held by company

directors and other individuals or institutions

hold-ing more than 10% of the share equity This

pro-portion corresponds to the disclosure requirement

threshold with regard to ownership

In the voluntary disclosure literature, Clarkson

et al (1999) and Lang and Lundholm (1993) show

that larger firms are generally more forthcoming

than smaller ones in terms of financial reporting Size

(SIZE) is measured by the natural logarithm of total

assets

Guay et al (1996) distinguish three perspectives

on the relation between performance and managerial

discretionary behaviors affecting accruals From a

performance perspective, management may use

accruals to reduce earnings (-) to enhance their

ability to reflect the true underlying performance of

the firm From an opportunism perspective,

man-agement may use accruals to subvert the ability of

earnings to reflect true underlying performance

of-ten but not always to portray performance in a more

favorable way (+) Most empirical evidence indeed

generally indicates that poorly performing firms are

more likely to manage earnings (Burgstahler and Dichev,1997; DeAngelo et al.,1994; Mard, 2004) Finally, from a noise component perspective, accruals may introduce noise into the performance signal (?/+) In order to control for CFP, we use stock market return (STOCKRET) As a sensitivity analysis, we also use two accounting proxies often used in previous research to measure performance: return on assets (ROA) and on shareholders’ equity (ROE)

From their extensive review of the literature on accounting choices, Watts and Zimmerman (1986) also assert that firms use DA to avoid violation of accounting-based-debt covenant Consistent with prior empirical work (e.g., Bowen et al 2008), we control for the level of debt We measure this variable

as the ratio of long-term debt to total assets (DEBT) Healy (1985), Holthausen et al (1995) and Gaver et al (1995) also identify executive compensation (EX-COMP) as another contractual motivation for EM In this case, earnings could be managed upward to im-prove an impending bonus, or downward if the manager believes that the firm is not in a position to meet the contractual earnings’ target This is referred

to as the ‘‘big bath’’ hypothesis While we have to control for that motivation which is frequently examined in the accounting choice literature, we do not specify the direction of the effect of executive compensation on EM In order to proxy for com-pensation, we use the measure introduced by Park and Shin (2004), namely, the mean ratio (premium/total compensation) of the five most highly paid managers This allows capturing the weight of compensation rather than simply the existence or not of a com-pensation plan Total comcom-pensation is the sum of the salary, premium, present value of stock option plans, value of restricted share grants, long-term incentive compensation, and any other annual components of compensation

Given the political costs linked to the company’s risk level, firms with very high risk have a greater incentive to manage earnings (Zmijewski and Ha-german, 1981) In fact, firms with higher risks face more political visibility and use accounting discretion

to reduce the risk perception (Warfield et al.,1995)

We expect a positive association between risk and

EM Following Warfield et al (1995) and Riahi-Belkaoui (2003), this variable (RISK) is measured by the systematic risk b of the market model

We also include the Market to Book ratio

(MBRATIO) to measure growth opportunities

According to Park and Shin (2004, p 443), ‘‘it is

easier for fast-growing firms to engage in EM than

slow-growing or stagnant firms because it is generally

harder to see through the business activities of

fast-growing firms.’’

Lastly, we use industries (IND) as dichotomous

variables As there exists mimetism in accounting

choices within industries, to examine firms’ EM

behavior, we have to control for any differences that

may exist between industries The breakdown of

observations by industrial sector, as defined by the

Toronto Stock Exchange, is presented in TableI

Analysis and interpretation of results

This section presents the univariate and multivariate

statistical analyses of the data and their interpretation

Descriptive statistics and univariate analysis

TableII presents the descriptive statistics of the

continuous independent variables for the total sample

(Part A), for highly manipulative (Part B) and weakly

manipulative firms (Part C) The results of univariate

means comparison tests (Part D) bring to light

significant differences between the two groups of firms

The DM variable reveals a significant difference

between high and low EM firms As expected, the

difference in CMD as proxied by DM is in favor of

the firms who are less prone to EM It is statistically

significant at the 1% level There is no significant difference as far as the quality of their fiduciary governance is concerned

As for other control variables, differences are not significant except in the case of firm performance as measured by stock return (at the 1% level) and growth opportunities as measured by MBRATIO (at the 5% level) In the case of firm performance, the difference is no longer significant when measured in accounting terms (ROA and ROE) When perfor-mance is measured in market terms (STOCKRET), high EM firms exhibit on average a significantly stronger financial performance (1%) than low EM ones As inferred by Guay et al (1996), in our sample, management seems to use accruals to reduce earnings

to enhance their ability to reflect the true underlying performance of the firm In contrast, high EM firms have on average more growth opportunities (5%) as measured by the MBRATIO than do companies that engage in less EM This is consistent with Park and Shin’s (2004) findings about fast-growing firms tak-ing advantage of greater information asymmetry to engage in more EM

There is no significant statistical difference be-tween the two groups of companies in terms of size, debt and compensation contracts, risk and existence

of a shareholder block

TableIII presents contingency analyses that integrates dichotomous variables The results indi-cate that the proportion of companies engaged in

EM is significantly higher (lower) in the metals and minerals, oil and gas (consumer products, industrial products and communication, and media sectors) than in other industrial sectors

TABLE I Distribution across industries

Gold and precious minerals (GOLD) 20 12.8

TABLE