Page 1
School of Accounting Seminar Series Semester 2, 2012
Earnings management decisions: The role of economics and ethics
Paul Coram The University of Melbourne
Date: Friday, 3rd August 2012 Time: 3.00pm – 4.30pm Venue: Tyree Energy Technologies Building LGO5
(Refer to campus map reference H6 here)
Australian School of Business School of Accounting
Page 2
1
Earnings Management Decisions: The Role of Economics and Ethics
Paul Coram
The University of Melbourne
James Frederickson
Melbourne Business School
Matt Pinnuck
The University of Melbourne
July 20, 2012
ABSTRACT: This study through a survey of 225 CFOs and CEOs of listed companies
attempts to better understand earnings management (EM) decisions. We extend Graham
et al. (2005) by providing participants with a case study scenario which asked them to
decide on whether to manage earnings when earnings were not expected to meet market
expectations. The survey then evaluates how they incorporate economic factors such as
costs and benefits to various stakeholders in their decision making, as well as evaluating
whether their perception of whether EM is ethical makes a difference. We find the most
significant economic factor affecting the EM decision is to avoid the costs to current
shareholders from not managing earnings to meet the market expectations. However, the
perception of whether EM is ethical and the related issue of whether EM is perceived as
lying are also significant factors that affect the decision to manage earnings. Finally, we
provide evidence that the choice by managers on the method of accrual compared to real
operational EM is primarily driven by ethical perceptions of these alternative actions.
[Draft – please do not quote without permission]
Page 3
2
I. INTRODUCTION
Earnings management (EM) is a pervasive aspect of corporate reporting which accounting
researchers attribute to contracting or market based incentives. However, despite the wealth
of archival literature on this topic there is relatively little research on the decision making
processes of managers who undertake these actions with the exception of Graham et al.
(2005). Motivated by the lack of understanding of the decision making processes by
managers we conduct a survey of 225 CFOs and CEOs of listed companies (hereafter
‗managers‘) in relation to EM to examine how they weigh up the various economics factors
in making their decisions. We also evaluate whether factors beyond economics such as
ethical considerations affect these decisions as recent economic research has questioned the
assumption of economic man (Benabou and Tirole 2011) and notes there is ample evidence
that people often behave ―morally‖.
EM refers to situations where managers apply accounting standards and/or structure
transactions to alter the company‘s financial statements with the intent of either misleading
some stakeholders about the company‘s true economic performance or influencing
contractual outcomes that are based on reported accounting numbers (Healy and Wahlen
1999). Prior archival research has documented that managers are more likely to manage
earnings in certain situations such as when unmanaged earnings are below an important
benchmark or when the firm is in danger of violating an accounting-based debt covenant and
commonly have used a version of the Jones (1991a) model to measure EM (see, e.g., Fields et
al. 2001, Dechow and Skinner 2000, or Healy and Wahlen 1999 for summaries). However,
documenting situations where managers are likely to manage earnings provides only a partial
picture of factors that influence managers. A complete picture requires that we understand the
specific factors that managers consider when deciding whether to manage earnings in a
particular situation. Through a variety of mechanisms that are based on accounting numbers
Page 4
3
(e.g., debt covenants, investment decisions, and so forth), EM redistributes wealth across a
range of stakeholders (e.g., the manager, current shareholders, current debtholders, future
shareholders). Presumably, the decision whether to manage earnings in a particular situation
is based on managers undertaking an expected cost-benefit analysis that incorporates the
effects on the various stakeholders potentially affected in that situation. Understanding which
stakeholders (and the expected costs and benefits to those stakeholders) managers consider
when deciding whether to manage earnings and how they trade off the expected costs and
benefits is critical for designing governance structures and compensation packages that will
curtail EM.
To better understand the EM decision, a prominent survey by Graham et al. (2005)
attempted to address a number of important questions from the prior literature such as which
benchmarks are most important to managers and what factors motivate managers to manage
earnings or make real operational adjustments. Their survey provided a rank ordering of
CFOs perceived importance of each motive to manage earnings and found managers were
most interested in meeting or beating earnings benchmarks to influence stock prices and their
own reputations. A surprising finding to emerge from their study is that managers prefer to
manage earnings via real operational methods rather than by accounting adjustments. In
considering this finding Graham et al. (2005) speculated that these responses may have been
affected by ethical considerations, which had been found in an earlier study by Bruns and
Merchant (1990). These findings as well as recent economics research that has questioned the
assumption of economic man (Benabou and Tirole 2011) raise the issue of whether
considerations beyond economics are factored into managers‘ decisions, which is part of
what we examine in this paper. This research project will disentangle the effects of the
various competing economic factors as well as ethical considerations that may affect
managers‘ decisions to engage in EM.
Page 5
4
Our approach to address these questions is by a survey that places managers in a
situation where there is pressure created by earnings being below market expectations and
then asking them to make a decision on whether they would manage earnings, and if so,
whether they would use real operational or accounting adjustments. The specific motives of
managers are then identified by asking questions on the magnitude of costs to various
stakeholders from the various available alternatives of: doing nothing; managing earnings
using accounting adjustments; or managing earnings using real operational adjustments.
Questions were also asked about managers‘ perceptions on whether undertaking EM is
ethical and also whether they think it is lying. This enabled us to model factors that influence
managers in deciding whether to manage earnings, and also factors that affect the approach
they would use to manage earnings (i.e., real versus accounting adjustments). We can
therefore more explicitly examine how managers consider the effects on various stakeholders
as well as the relative importance of economics and ethics in their decision making on EM.
This study is the first to try and more comprehensively understand managers‘ decision
making in this area. It provides several important new insights. First, we document the
evaluation of economic costs and benefits to stakeholders by managers in deciding on
whether to undertake EM. Second, we show managers‘ perceptions of the ethics of EM
actions and the related concept of whether they perceive EM as lying both provide significant
explanatory power in the decision to manage earnings. Finally, we examine the factors that
affect the choice between real operational and accounting EM or alternatively, why some
firms decide to do nothing. Again, economics is important but ethical considerations also
affect this choice. In summary, for both the decision to manage earnings and the method of
managing earnings we find that economics and ethical factors are important.
Page 6
5
II. LITERATURE REVIEW
Earnings Management
The question of why and when managers manage earnings has been the subject of
significant research over many years. It is an important topic that is of interest both to the
accounting profession and regulators as high profile cases of ‗earnings management‘ have
resulted in significant negative publicity for the accounting profession. The main motivations
for undertaking EM documented in the prior literature has come from either contracting or
capital markets based incentives. From a contracting perspective, studies on the effect of
being close to breaching a lending covenant have generally found little evidence of EM
(Healy and Palepu 1990; DeAngelo et al. 1994). The effect of management compensation
contracts on EM has however found evidence consistent with expectations from contracting
theory (Healy 1985; Guidry et al. 1999). DeChow and Sloan (1991) also observed real
operational EM by observing CEOs in their final years reducing R&D spending to increase
reported earnings.
More recently, the EM research has focused more on capital market incentives. Evidence
shows that managers overstate earnings prior to initial public offering of stock (Teoh, Welch
and Wong 1998). There has also been research that has observed EM to meet the expectations
of analysts (Kasznik, 1999). The focus on the importance of meeting these targets does
suggest that the decision to manage earnings is a more complex decision with a number of
possible costs to various stakeholders compared to if it is based solely on contracting factors
alone. Graham et al (2005) find that two of the most important benchmarks for managers are
meeting analysts‘ expectations and prior earnings.
The contracting and market based studies have identified situations where EM is
expected to occur and then observed whether it does. It has examined a number of separate
Page 7
6
and different factors that are associated with EM. However the situations provides no
understanding of the underlying economic motivations of the corporate manager. Positive
accounting research (Watts and Zimmerman 1986; Christie and Zimmerman 1994), under
which research on EM is typically classified, assumes that when deciding what decision to
make, managers maximize an objective function. Given that the manager‘s objective function
is unknown, the only way to fully understand the EM decision in an unbiased manner is to
specify and test an objective function that meets two conditions. The first is that the
objective function includes all key stakeholders likely to be affected by the EM decision.
Without including these stakeholders, it is impossible to identify and understand how
managers trade-off different stakeholders when deciding whether to manage earnings. Further,
if the net expected costs or benefits from EM are correlated across two or more stakeholder
groups, including all potentially relevant stakeholders is necessary to avoid a correlated
omitted variable problem. From a broad stakeholder theory perspective (Jensen 2001),
managers‘ objective functions could potentially reflect the expected costs and benefits to the
following stakeholders: the manager, current shareholders, current debtholders, other current
stakeholders (e.g., suppliers, employees, customers), and future stakeholders.
The second necessary condition is that the empirical test must reflect both the expected
costs and benefits to each stakeholder group if earnings are managed versus if they are not.
For example, assume that the benefit to the manager from managing earnings is a bonus
while the costs are job and reputation loss if the EM is detected. If the empirical test includes
only the expected benefits or only the expected costs, as is common in the existing research,
the net expected cost/benefit is measured asymmetrically, which in turn will yield
significantly biased inferences about the relative importance of each stakeholder.
Graham et al. (2005) attempted to answer some of these questions about the EM decision
through a survey. However, in doing so they raised more questions, particularly relating to
Page 8
7
other considerations such the role of ethics in these decisions and the use of real operational
EM both of which are addressed in this present study.
Ethics
Defining ethics is difficult, although it relates to a basic question of ―What ought one to
do?‖1 In examining the effect of ethics on EM, Merchant and Rockness (1994) noted that
most EM acts were legal, however the ethical perspective raised the question of whether they
were the ―right things to do‖ (p.81). The difficulty in determining what one ought to do or the
right thing to do comes from the fact that people bring morality and values to their ethical
decision making, which comes from traditions or theories – which will vary between
individuals. Further complexity in evaluating an ethical dilemma can be due to important
values held by individuals may conflict. For example, two important values that many
cultures and traditions agree upon as morally right would be to tell the truth and to not cause
others harm. What happens when these two values conflict? The answer to this might then
depend on the ethical theory that an individual ascribes to. A teleological approach assesses
the rightness or wrongness of something by the consequences, therefore they would choose
lying over harming others. A deontological approach relates to universal laws, therefore if
they think that lying is always wrong, they would tell the truth under all circumstances. A
recent experimental paper by Johnson et al. (2012) provides some evidence that for many
managers, the ends may justify the means in relation to EM, thereby suggesting that
managers may adopt a teleological approach to this decision.
There is limited research that has examined ethical considerations on managers‘ decision
making relating to accounting. This may be partially due to the assumption underpinning
positive accounting theory that managers actions will only be in their own self interest (Watts
1 This question was raised by Socrates in the fifth century BC.
Page 9
8
and Zimmerman 1986). Erhard and Jensen (2012) have proposed a positive model of integrity
which relates to ―honoring ones word‖. They assert their definition of integrity has nothing to
do with morality and ethics, which are normative concepts. While it is true that they are
normative concepts, there has been significant interest in ethics in recent years due to the
perceived lack of ethics in some of the high profile corporate collapses in the early 2000s, as
well as in the current global financial crisis. In addition, some have recognized that in order
for economic settings to function smoothly, ethics and trust are necessary (e.g., Holmstrom
2005; Noreen 1988).
In discussing their finding that US CFOs prefer to manage earnings via real operational
methods, Graham et al. (2005) speculate—but do not test—that this effect could be due to
ethical considerations. In this discussion, Graham et al. make reference to the study by Bruns
and Merchant (1990) who found that managers perceive managing earnings via real
operational methods to be more ethical than using accounting methods. A number of studies
have examined the ethical dimension in decision making and Loe et al. (2000) provide a good
review of this literature. However, from this review, it is apparent that the majority of this
research has focused on the role of awareness, individual and organizational factors in ethical
decision making. An exception was Merchant and Rockness (1994) who examined the
morality of EM with a focus on the moral issue itself. They found some evidence that the
acceptability of various EM practices varied with the type, size, timing, and purpose of
actions. Their reasoning for applying an ethical perspective is because many types of EM
behaviors are not obviously acceptable or unacceptable. Therefore the concept of whether the
actions are the ‗right thing to do‘ can become important.
In the evaluation of an ethical decision, it is a determination of what is acceptable, or
therefore what is ‗right‘. Our study will be the first to evaluate whether the decision making
of managers goes beyond the traditional economic considerations to explain EM. It will
Page 10
9
address the question of how managers see these decisions. Are these purely economic
decisions in that the actions by managers are solely in their self interest? Due to contract
design, this self interest is also usually linked to the short term effect on the company.
Alternatively, does ethics in determining what is ‗right‘ beyond personal self interest come
into these decisions. Jones (1991b, p.367) defined an ethical decision as a ―decision that is
both legal and morally acceptable to the larger community‖, and a moral issue as being
present when ―a person‘s actions, when freely performed, may harm or benefit others‖ (from
Velasquez and Rostankowski, 1985). Jones developed the concept of moral intensity as a
framework to evaluate the moral issue itself itself over six dimensions, which are: magnitude
of consequences; social consensus; probability of effect; temporal immediacy; proximity; and
concentration of effect. Most of these are specific considerations on the effect of the decision
on others – although social consensus might include other aspects, e.g., whether the action is
perceived as lying or not. Our study takes into account some of these other considerations
(which we define as economic factors in the survey) to provide a comprehensive
understanding of the factors affecting managers decision making in relation to EM.
In determining what is ‗right‘ a deontological approach to ethical decision making would
expect that individuals should tell the truth under all circumstances whereas a teleological
approach would take the view that the consequences were most important.2 Both of these
approaches contrast with a purely economic perspective where it would be assumed that a lie
would be told whenever it benefits the liar and irrespective of the effect on any other party.3
Gneezy (2005) experimentally evaluated whether people do behave in this way or whether
the propensity to lie is affected by differing gains to the individual or consequences to others.
He found that people are sensitive to the amount of their gain when deciding to lie and are
2 Erhard and Jensen‘s (2012) concept of integrity would seem to be much more focused on the former rather
than the latter. 3 The economic approach is concerned with some consequences – that is, those related to the individuals self
interest.
Page 11
10
also sensitive to the level of harm to others from the lie. In relating this to EM, this would
suggest that whether the manager thinks that EM is lying will have a significant effect on his
or her behavior. We will test these findings of Gneezy (2005) in this study by evaluating
whether those who think that EM is lying will still undertake these actions based on
variations in the benefits and consequences to others.
Recent accounting studies have shown that managers‘ decisions on EM can be affected
by more than the traditional motivations as would be expected based on economic theory.
McGuire et al. (2012) found that greater levels of religiosity are associated with less EM.
Although not explicit in that paper, for many religious people ‗lying‘ is perceived to be
wrong in an absolute sense, which may the reason for this finding. A recent survey by
Abernathy et al. (2012) found an ethical work climate that focuses on ‗self‘ yields agents who
are more likely to manipulate accounting earnings. Hunton et al. (2011) found an association
between perceived tone at the top and accruals quality. Dikolli et al. (2012) found a positive
association between the managerial trait of integrity as defined by Jensen (2009) and accruals
quality. In summary, these studies provide some evidence to suggest exploring factors that
might explain decision making beyond economics can be important.
As well as consideration of the ethics of deciding to manage earnings or not, the choice
of EM method also appears to be one partly driven by ethical considerations and we consider
this issue in the next section.
Real Earnings Management
Real earnings management (REM) is when managers undertake actions that ―change the
timing or structuring of an operation, investment, and/or financial transaction in an effort to
influence the output of the accounting system‖ (Gunny, p.855, 2010). Studies have shown
that REM exists, such as Roychowdhury (2006) who developed empirical proxies for REM
Page 12
11
and found that managers avoid reporting losses by undertaking REM. There is also evidence
of an increase in REM in recent years as the method by which to undertake EM (Bartov and
Cohen 2009). This is consistent with the findings by Graham et al. (2005), who found 80
percent of their sample of CFOs said they would undertake REM to meet an earnings target.
Graham et al. expressed surprise at these findings as they reasonably assert that these
decisions are ones that sacrifice economic value. However, interestingly, recent archival
research seems to suggest that firms undertaking REM perform at least as well or better than
those who do not in the longer term (Gunny 2010). It should be noted that REM studies still
suffer from the measurement issues associated with studies on accrual EM (see Fields et al.
2001) and also the problem of determining whether the changes in actual decisions are
undertaken solely for financial reporting reasons. Archival studies are not able to evaluate
why the decision to make the real operational adjustments occurred.
Irrespective of some of the recent changes in preferences for managers to use real rather
than accrual EM, the key questions about the factors that affect their decisions of this nature
are largely unexplored. Some experiments have examined these questions relating to
motivations for REM, such as Bhojraj and Libby (2005) who showed evidence of managerial
myopia to meet earnings benchmarks. Further, Seybert (2010) found that allowing
capitalization of research and development would lead to REM through overinvestment in
continuing projects so managers could avoid reputation damage from asset impairment. As
noted earlier in this paper, the survey by Graham et al. (2005) found a preference by
managers to use REM. McGuire et al. (2012) found religiosity is negatively associated with
abnormal accruals, but positively associated with proxies for REM. For religious people
values that inform ethical decision making can be guided by prescribed moral rules (such as
lying is always wrong), this therefore provides some evidence consistent with prior research
Page 13
12
that suggests managers‘ perceive REM to be more ethical than accrual EM (Bruns and
Merchant 1990; Graham et al 2005).
III. RESEARCH QUESTIONS
Our study examines a decision by managers on whether to manage earnings and how
they will manage earnings when they are below market expectations. In evaluating the
economic considerations that affect manager‘s EM decisions the first order effect that we
examine is the economic considerations of the likelihood of success of the action and the
direct effect on the manager. We then explore whether the direct effect on the manager is due
to consideration of shareholders and other stakeholders and the relative importance of these
groups in the decision making process. This has not been done before and will provide insight
into some important questions on why managers make these decisions.
We also ask managers their perceptions on the ethics of EM to see whether it plays a part
in their decision making process. In discussing the fact that CFOs stated they were more
willing to report taking real decisions rather than accounting decisions, Graham et al (2005)
suggest that this effect could be due to ethical considerations. This present study will also
explore the effect of ethical considerations on managers‘ decision making process. Merchant
and Rockness (1994) examined the ethics of EM but took a very different approach to our
study.4 We will evaluate whether the harm to others affects decision making and whether
managers see this effect as an ethical consideration. Due to contracts, sometimes the harm to
others has a direct effect on the manager. We will evaluate whether some considerations of
4 Merchant and Rockness used a questionnaire that consisted of 13 potentially questionable earnings
management activities, these varied by: type of action; consistency with GAAP; the direction of the effect on
earnings; materiality; the period of effect; and the purpose in mind. They then asked participants the ethical
acceptability of each case presented and they found that ethical judgments were affected by the type of earnings
management and that using real compared to accounting methods to adjust earnings was much more ethically
acceptable.
Page 14
13
the harm to others takes on an ethical dimension, beyond managers‘ self interest. As there
have been very few studies to try to understand the decision making processes of managers
associated with EM, we will frame our study in terms of research questions.
The first research questions therefore relates to the question of why managers manage
earnings:
RQ 1: What economic factors are important in the decision to manage earnings?
RQ 2: What is the relative importance of economics and ethics on the decision to
manage earnings?
By determining whether ethics makes a difference beyond the potential harm to
stakeholders, we are moving towards exploring values that inform ethical decision making
that are not related to economics at all such as whether the action seems the ―right thing to do‖
or whether it involves lying or not. As noted by Gneezy (2005), extreme economic theory
would suggest that ―lies will be told whenever it is beneficial to the liar‖(p.384). This would
seem to be contrary to the positive notion of integrity which means to ―honor ones word‖ as
proposed by Erhard and Jensen (2012). By definition as per Healy and Wahlen (1999) earlier
in this paper, EM involves deception therefore technically it could be construed as lying. We
therefore raise the question that has not been previously considered in the literature: whether
managers (a) view EM actions as ‗lying‘, and if so (b) will they still do it?
RQ 3a: Is earnings management perceived as lying?
RQ 3b: Does perceiving earnings management as lying affect the decision to manage
earnings?
Page 15
14
Gneezy (2005), showed that inconsistent with an economic theory, individuals are
sensitive to the harm that lying may cause to others. The third research question therefore
examines whether this finding applies to the EM decision to see whether the decision to
manage earnings for those who think EM is ‗lying‘ is affected by the costs and benefits to
other parties. We address this as follows:
RQ 4: Where EM is viewed as lying, is the decision to undertake EM affected by the
costs and benefits to other parties?
The final issue that we examine is how managers decide whether to adjust earnings – by
real or accrual earnings adjustments? Graham et al. (2005) provided the interesting finding
that managers prefer real over accrual EM techniques. In evaluating their findings, they
thought that the response of managers to their survey question on this issue may have been
affected by whether they perceived the two approaches as ethically different. Studies in
recent years have examined REM and found some evidence of it being associated with future
firm value (Gunny, 2010). However, even if the decision is made for financial reporting
reasons, for many firms it is quite possible that reductions in expenditure for the purposes of
managing earnings (REM) might also be beneficial to the efficiency of the firm in the longer
term anyway. That is, the ‗requirement‘ to reduce real expenses to manage earnings may act
as an impetus to improve efficiencies in the longer term. This argument would support the
findings of Gunny (2010). We will be able to evaluate whether managers choosing to
undertake REM are doing it with consideration of future shareholders. Our study will be the
first to examine perceived costs and benefits to stakeholders and whether ethics are
considered by managers when actually making a choice of real versus accrual EM.
Page 16
15
We evaluate whether ethics affects this choice as follows:
RQ 5: What is the relative importance of economics and ethics on the choice of
method to manage earnings?
IV. RESEARCH DESIGN
A difficulty inherent in the archival studies of EM that examine specific incentives to
manage earnings is due to endogeneity and correlated omitted variables problems, as a result
this makes it difficult to draw strong causal influences to explain this behavior (e.g., Fields et
al. 2001). As pointed out by Libby and Seybert (2008) in a review of behavioral studies of
EM, the advantage of experiments and surveys is that they can address unanswered questions
from prior archival research. They further state that behavioral studies of EM are able to
isolate specific motives for EM and causally link them to EM attempts. How managers
actually consider these economic and ethical factors and make these tradeoffs are empirical
questions which will be examined by a survey in this research paper.
Participants
A survey was sent to all the CFOs and CEOs of approximately 1200 of the largest public
companies in Australia.5 This was done by obtaining a mailing list as well as significant hand
collection of data. For all of the CEOs we attempted to find the related CFO if we did not
have them already, similarly for all of the CFOs we attempted to find the related CEO. This
process ultimately resulted in a list of 1044 CFOs and 1180 CEOs who were sent the surveys.
5 Australia has a similar corporate governance framework and institutional environment to the United States (see
Leuz et al. (2003)). Prior research has also shown that earnings benchmark beating is important in the Australian
environment (Carey and Simnett 2006). Although SOX was obviously not introduced in Australia, there were
significant regulatory changes imposed at about the same time through CLERP 9. There is no evidence to
suggest that they CFOs and CEOs would respond in any differently from an equivalent group in the US to a
survey of this type.
Page 17
16
The first survey was mailed out in December 2009. To encourage responses a donation
would be made on their behalf on completion of the survey. The number of responses
(response rate) from the first mail out was 82 CFOs (7.9%) and 59 CEOs (5.0%). A second
request was sent in April 2010, this resulted in the sample increasing to 146 CFOs (14%) and
90 CEOs (7.6%) to give a total sample size of 236. The final number analyzed in this paper
was 225 as we excluded any respondents who did not answer the majority of questions. For
this type of participant group the response rate is reasonable, for example, Graham et al.
(2005) achieved an overall average of 10.4% in their survey of financial executives.
Of the respondents, 62% were CFOs and 38% were CEOs. As would be expected they
were very experienced with an average of 23 years of professional work experience,
including an average of 4.8 years in their current positions. A significant number (65%) had
professional accounting qualifications, with 20% CPAs and 45% Chartered Accountants.
Ninety-two percent of respondents were male.
Survey Task
The participants were provided with information about a hypothetical company facing a
situation where the earnings per share is not going to meet the markets‘ expectations by 4%.6
Some extracts from the survey are presented in the Appendix. They were then told there were
two possible actions that were available under these circumstances. First, take no special
action, or second, try to improve reported performance by either accounting or operational
adjustments. They were then asked in Question 1 the likelihood that their firm would improve
reported performance using accounting or operational adjustments on a scale of 0 (certain not
to happen) to 100% (certain to happen). This was then followed by a direct question (Qu. 2)
6 Considerable feedback from practitioners and academics was incorporated in determining this amount. Four
percent was decided on because: (a) it was a significant amount; (b) achievable to be adjusted by real or
accounting means in the time frame outlined in the case study materials; and (c) it was less than 5%, which is an
important materiality benchmark for auditors. The responses to the survey would suggest that the percentage
chosen was reasonable.
Page 18
17
on the decision their firm would actually make in this situation to either take no special
actions or try to improve reported performance. If they selected the latter, they were further
asked whether they would rely solely on accounting adjustments (0%) or operational
adjustments (100%). In the final part of this section in Question 3 they were asked their
perceived likelihood that their firm could increase reported earnings by each of the possible
options of: only accounting adjustments; only operational adjustments; or a combination of
both.
The next part of the task was to evaluate the magnitude and likelihood of costs for each of
the stakeholder groups that might be affected by each of the three possible alternatives
(Questions 4, 5 & 6). The stakeholder groups comprised: CFO and CEO; company in the
short term; company in the long term; company‘s current shareholders; company‘s future
shareholders; company‘s debtholders; and other stakeholders (e.g., employees, customers, or
general public).7
To assess the ‗ethics‘ associated with the options available, we asked how ethically
questionable the participant assessed each course of action (Qu. 13), which was to get an
overall sense of the participant‘s perceived ‗rightness‘ or ‗wrongness‘ of the action. Second,
we asked whether the participant thought that manipulating earnings through accounting or
operational adjustments was ‗lying‘ (Other Questions 2 & 3), which is an important value
many people bring to their ethical decision making.
Extensive time was spent on developing the survey with reference to the prior literature
on EM. An important criteria in this development was that the participants had to make a
‗decision‘ so that modeling of the factors that were associated with that decision could occur.
The draft survey was circulated to a number of experts for comment before it was finalized.
7 Other questions were asked about corporate governance. However, these responses are not included in this
paper.
Page 19
18
The experts who perused the materials and provided us with comments, included: three
academics, two regulators, two business professional association members, one technical
audit partner, and one retired CFO of a top 100 Australian company. The feedback from the
experts was considered by the research team and incorporated where appropriate into the case
materials.
V. RESULTS
Descriptive Statistics
Table 1 reports descriptive statistics relating the various questions from our survey about
EM. Question 1 asked managers the likelihood that they would manage earnings and the
mean response of 54% indicates it is more likely on average that managers would choose to
manage earnings. In Question 2 they were asked to actually make a decision and 68% said
that they would manage earnings. For those who would manage earnings, the clear preference
was to use operational adjustments with an overall average of 74%. Question 3 asked
perceptions on the likelihood of success of each of the methods to manage earnings and using
a combination of accounting and operational adjustments is most likely to be successful
(62%), followed by operational (51%), and then accounting adjustments (38%).
INSERT TABLE 1 ABOUT HERE
Questions 4, 5 and 6 asked about the costs to various stakeholders of the options
available on a scale of 0 = absolutely no costs to 100 = very large costs. Question 4 related to
the costs of taking no special action rather than managing earnings. This question by
addressing the costs of taking no action provides a measure on the benefits of EM. The three
questions where the costs are perceived as most significant (>40) are on the CFO and CEO,
company in the short term, and the company‘s current shareholders. Question 5 asked about
Page 20
19
the costs of accounting adjustment. Costs to the company in the long term are highest with a
mean of 41, which is consistent with accrual reversal of these types of adjustments in the
longer term. The effect on the CFO and CEO were also high with a mean of 38, which is
consistent with the potential consequences of these types of actions if they are detected.
Question 6 asked about costs of operational adjustments and generally the perceived costs are
lower than those for accounting adjustments. The highest cost is for the company in the long
term, with a mean of 34. This is consistent with the proposition that real operational EM
sacrifices economic value.
Table 2 presents correlations between the some of the variables reported in Table 1.
There is a high level of correlation between the various costs associated with doing nothing
(Qu 4). There is also a correlation between perceptions of EM actions as unethical (Q13_2)
and perceptions of whether they are lying for both accounting adjustments (OQ2, 0.33) and
operational adjustments (OQ3, 0.77). Another correlation of note is that if participants
thought that operational earnings adjustments were unethical or lying they were less likely to
adjust earnings and also much less likely to use it as a method of managing earnings.
INSERT TABLE 2 ABOUT HERE
Why Do Corporate Managers Manage Earnings?
A conceptual model of the EM decision was developed that included all economic and
ethical reasons for the decision. OLS regressions were then estimated that included all of the
economic and ethical responses of participants as independent variables. We estimated two
regressions with: (i) the decision to manage earnings as the dependent variable; and (ii) the
decision on the method of managing earnings as dependent variable. The estimated
regressions are as follows:
Page 21
20
(a) EM (Q1) = (Economic Consequences, Ethics and Lying)
(b) EM (Q2-2) = (Economic Consequences, Ethics and Lying)
To evaluate the first and second research questions five regressions were run with the
decision to manage earnings as the dependent variable and they are reported in Table 3 below.
These regressions are to try and explain the tradeoffs between stakeholders and other
considerations by managers in their decisions on whether to manage earnings.
INSERT TABLE 3 ABOUT HERE
The first regression (a) in Table 3, includes the likelihood of success of the various EM
actions (Q3) as well as the benefits and costs of taking the actions on the CEO/CFO (Q4A_1,
Q5A_1, Q6A_1). The overall R2 for this first regression is 0.24. The likelihood of success
through accounting adjustments (Q3_1, p=0.077), operational adjustments (Q3_2, p=0.001),
or and a combination of operational and accounting adjustments (Q3_3, p=0.002) are all
significantly associated with the decision to manage earnings. If managers perceive there is a
cost to themselves to doing nothing they are more likely to manage earnings (Q4A_1,
p=0.002). However, they also are more likely to manage earnings if they think there is a low
cost to themselves from accounting adjustments (Q5A_1, p=0.10). The second regression (b)
incorporates the benefits and costs to shareholders and other stakeholders from the decision
to manage earnings and the R2 increases to 0.31. The interesting finding from this regression
is that once these other factors are included, the significant effect for self interest to
CEO/CFOs goes away. The most significant benefit of EM to emerge from the second
regression is the avoidance of costs to current shareholders by managing earnings (Q4A_4,
p<0.001). This suggests that current shareholders are the mediating factor in manager‘s
consideration of the effect on themselves. The other costs to emerge as significant are if there
are perceived low costs to future shareholders (Q5A_5, p=0.102) or debt-holders (Q5A_6,
Page 22
21
p=0.046) from accounting adjustments, then managers will be more likely to undertake EM.
Therefore in addressing Research Question 1, the high R2 does confirm economic factors are
important, further this analysis shows the overriding economic factor of importance in
deciding on whether to undertake EM is through consideration of the effect on current
shareholders from doing nothing when earnings are not going to meet market expectations.
The next two regressions address Research Question 2, which evaluates whether ethical
factors also affect managers‘ decisions to undertake EM. The third regression (c) in Table 3,
shows the first regression (a) plus inclusion of the perceived ethics of the various EM options
(Q13). As can be seen the R2 increased from 0.24 to 0.41. When ethics are included, the
affect of benefits and costs to CEO/CFOs themselves becomes insignificant. If managers
think that taking no special action is ethically questionable, they are more likely to manage
earnings (Q13_1, p<0.001). If they think that operational adjustments are ethically
questionable, they are less likely to manage earnings (Q13_3, p<0.001), which could be
driven in part by the fact that the majority who choose to manage earnings do so by
operational methods. The fourth (d) regression, shows the first regression (a) with the
inclusion of whether managers perceive the EM methods as lying or not (OQ2&3). Again,
this provides an increase in explanatory power over the first regression (R2 increased from
0.24 to 0.33), although it is not as much of an increase as from inclusion of the ethics
questions. If managers think that accounting (OQ2, p<0.001) or operational adjustments
(OQ3, p=0.015) are lying they are less likely to manage earnings. In this regression, unlike
the third one, if the managers perceive a cost to themselves to doing nothing, it is still a
significant factor in their decision to manage earnings (Q4A_1, p=0.042), indicating that this
consideration has an ethical component but does not relate to lying.
The fifth regression (e) includes all questions from the previous four regressions. The
overall R2 increases to 0.49. This shows that there is incremental explanatory power from all
Page 23
22
of the various components measured by this survey to understand the EM decision. In this
overall model, the effect on current shareholders from doing nothing remains an important
consideration in deciding whether to undertake EM or not (Q4A_4, p<0.001). If there are low
costs to future shareholders and debt-holders from accounting adjustments, managers are
more likely to undertake EM. A significant effect comes through only in the complete model
for the effect on current shareholders from operational adjustments. That is, if there are low
costs perceived to current shareholders from operational adjustments, managers are more
likely to manage earnings (Q6A_4, p=0.057).
In relation to the ethical factors in the complete regression (e), the perceived ethics of
taking no action or operational adjustments continues to be a significant factor relating to the
EM decision. However, the ethical perception of accounting adjustments was not significant
in the third regression (c) and becomes even more insignificant in the complete regression (e).
Whether accounting adjustments are perceived as lying remains very significant (OQ2,
p=0.003), however whether operational adjustments are perceived as lying becomes
insignificant with inclusion of all questions in the complete regression (OQ3, p=0.302).
A regression was also run (not reported) that only examined the effect of ethics on the
EM decision. This provided an R2 of 0.35, however, 0.25 of that explanatory power was
correlated with economic factors, indicating that an important part of managers‘ ethical
evaluation is the harm to others. However, there was still 0.10 of ‗pure ethics‘ beyond any
economic considerations, which helps explain the significant effect from perceptions of lying
on the EM decision found in this study.
Is EM Lying and Does Perceiving EM as Lying Affect the Decision to Manage Earnings?
Research Question 3a addresses whether EM is perceived as lying, which has not been
examined in the literature before. In Table 1, the descriptive responses to these questions is
Page 24
23
presented. Other Question (OQ) 2 shows a mean of 60% who think accounting adjustments
are lying compared to OQ 3 with a mean of 27% who think that operational adjustments are
lying. Question 13 relates to perceived ethics of the available options, as can be seen, taking
no action is the most ethical with a mean of 24 (scale of 0 = not (ethically) questionable at all
to 100 = Extremely (ethically) questionable), followed by operational adjustments (34), then
accounting adjustments (67) as least ethical. From the correlations in Table 2, it is shown that
a major reason accounting adjustments are seen as unethical is because they are seen as lying
(0.33). In relation to RQ3b, which relates to whether if perceiving EM as lying affects the
decision to manage earnings, results presented on Table 3 showed that perceiving accounting
adjustments as lying (OQ2) means that managers will be less likely to manage earnings
(p=0.003).
Research Question 4 is an evaluation of whether for those who think EM is lying, the
economic costs and benefits of the EM makes a difference to their decision on whether to
manage earnings. Table 4 provides an analysis by splitting the sample into whether they think
that accounting EM is lying or not.8 Table 4, Panel A provides details of the economic
benefits of EM (through avoidance of the costs of doing nothing on stakeholders – Question
Set 4). As can be seen the R2 is relatively low for those that think accounting EM is lying (R
2
= 0.05), however it increases significantly for those who do not think accounting EM is lying
in Panel B (R2 = 0.16) indicating economic benefits are more important for the group who do
not think that accounting EM is lying. For this group the significant individual factor is the
effect on current shareholders as a reason for undertaking EM (Q4A_4, p=0.038). The second
regression in Table 4, Panel A, incorporates the economic benefits and costs (costs to
stakeholders of accounting adjustments - Question Set 5) of accounting adjustments. For
those who think accounting EM is lying, the inclusion of consequences increases the R2 to
8 The focus is only on accounting EM because for the majority of managers they do not perceive operational EM
as lying.
Page 25
24
0.10. One of the specific consequences of accounting adjustments relating to the company in
the short term is marginally significant (Q5A_2, p = 0.08), which indicates this negative
consequence of accounting adjustments is associated with managers choosing to engage in
EM. This finding may come about because most managers who choose to engage in EM use
operational not accounting adjustments. For the group who think that accounting EM is not
lying, the explanatory power hardly changes with the inclusion of consequences in Panel B
(R2 from 0.16 to 0.18), and none of the individual consequences are significant.
Overall, our findings show that for those who think EM is lying, that economic factors
do not affect their propensity to undertake these EM actions, which is not consistent with
Gneezy (2005). However, it does also illustrate there are a significant group of managers
whose actions seem to be primarily determined by whether they think accounting EM is lying.
This does seem to indicate that the view by management on this issue is an important
management trait that has economic consequences, which is explored further in the next
research question.
INSERT TABLE 4 ABOUT HERE
Table 5 controls for ethics to evaluate the decision of the groups who think that
accounting is lying compared to those who do not. In Table 5, Panel A, for those who think
that accounting is lying, the ethical factors (Q15) have significant explanatory power in the
regression (R2 of 0.40). When economic considerations are introduced into this regression the
explanatory power barely changes (R2 of 0.41), which is consistent with findings in Table 4.
If accounting EM is perceived as lying, the ethics of taking no action or operational
adjustments makes a difference to the EM decision but nothing else does. Specifically, for
this group, if they think that taking no action is unethical, they are more likely to manage
Page 26
25
earnings, and if they think that operational adjustments are unethical they are less likely to
manage earnings.
INSERT TABLE 5 ABOUT HERE
Table 5, Panel B shows the group who do not think that accounting is lying and also
controlling for ethics. In this case, ethics alone has a much lower explanatory power for those
who think that accounting is lying (R2 of 0.25). However, inclusion of economics has a much
more significant effect on the decisions of this group (R2 of 0.37). In evaluating specific costs
and benefits that are considered in deciding to manage earnings, the benefits are to current
shareholders (by avoiding costs of doing nothing) from EM actions (Q4A_4, p=0.001) and
the potential costs are to the CEO or CFO (Q5A_1, p=0.076).
Therefore in addressing RQ 4, where EM is viewed as lying, the economic costs and
benefits to other parties do not matter. However, if it is not viewed as lying economic costs
and benefits do affect managers‘ decisions.
Choice of Earnings Management Technique
Research Question 5 evaluates the relative importance of economics and ethical factors
in the choice of EM technique (i.e., for those who have decided to manage earnings). Three
regressions are run as shown in Table 6.
INSERT TABLE 6 ABOUT HERE
The first regression in Table 6 shows economic factors that affect this choice. The main
economic determinant on this choice is the likelihood of success of a particular EM technique.
If managers do not think that accounting EM can achieve the required increase in earnings
then they are more likely to use operational adjustments (Q3_1, p=0.002). Similarly, if
managers do not think that operational EM could achieve the required increase in earnings
Page 27
26
then they are more likely to use accounting adjustments (Q3_2, p<0.001). The only other
economic variable of significance is perceptions of the cost of accounting adjustments to the
company in the long term. If managers think these costs are high, they are more likely to use
operational adjustments (Q5A_3, p=0.020). However, this consideration of the longer term
effect on the company appears to have an ethical component because it becomes insignificant
when ethics variables are added in the second regression.
The second regression in Table 6 adds the ethics questions and this significantly
increases the explanatory power of the model to an R2 = 0.48 (from R
2 = 0.29 for the model
with economics factors only). If managers think that accounting adjustments are ethically
questionable, they are more likely to use operational adjustments to manage earnings (Q13_2,
p<0.001). Similarly, if managers think that operational adjustments are ethically questionable,
they are more likely to use accounting adjustments to manage earnings (Q13_3, p=0.001).
The final regression includes the questions asked on whether the EM actions were perceived
as lying or not. If managers think that accounting adjustments are lying, they are more likely
to use operational adjustments to manage earnings (OQ2, p=0.058). However, whether
managers think operational adjustments are lying or not does not affect their choice between
the two types of EM techniques (OQ3, p=0.301). These findings suggest that managers do
not see operational adjustments as lying or deception but they are seen as by some as
unethical. This is consistent with the adjustments being ‗real‘ decisions which could therefore
more easily be justified to oneself as not lying. From this third regression it seems the
primary determinants on the choice of method relates to ethical perceptions of the actions,
and for accounting adjustments it also relates to whether managers perceive the actions as
lying. As reported earlier, many more managers think accounting adjustments are unethical
and lying. This therefore translates into the much higher use of operational EM methods as
has been shown in this survey and the survey of Graham et al. (2005).
Page 28
27
VI. DISCUSSION AND CONCLUSIONS
This study examines economic factors and tradeoffs between stakeholders that affect the
decision to manage earnings and is the first to show in a comprehensive model of factors
associated with EM decisions that ethical considerations are important in the decision and the
choice of the method of EM.
The first research question evaluated what economic factors were important in the
decision to manage earnings. The most important factor affecting this decision is the
likelihood of success of the EM. The most significant perceived cost relates to the costs of
doing nothing, of which the most important aspect are the costs to the current shareholders
from taking no action. The second research question examined the relative importance of
economic and ethical factors for managers in deciding on whether to manage earnings. We
find that economic factors, ethics and lying all have a statistically significant effect on the
decision to undertake EM. The most important factors are economics and ethics, and they are
highly correlated, showing that ethical considerations do often relate to economic factors.
Which is consistent with a teleological approach to ethical decision making and the moral
intensity model proposed by Jones (1991b).
From the detailed analysis performed addressing the second research question, the
‗ethical acceptability‘ of taking no action or operational adjustments are both associated with
the decision to manage earnings. Those who viewed taking no action as ethically
questionable were more likely to manage earnings. We are not aware of not managing
earnings as being documented as ethically questionable in any previous study. It does also
illustrate the importance of benchmarks to managers, but suggests they think there are ethical
implications to these decisions rather than the traditional view of meeting these targets being
Page 29
28
purely associated with short-term self-interest. This finding is consistent with the recent
experiment by Johnson et al. (2012), where they found that if the organizational
consequences were positive from undertaking EM, they perceived the EM actions as less
unethical. Those who viewed operational adjustments as unethical were less likely to manage
earnings. This is not surprising as we also find that operational adjustments were the most
‗popular‘ method of managing earnings by those who chose to do so, with 74% of our sample
preferring this method.
The next research questions explored a part of ethical decision making (not previously
explored in the accounting literature) on whether managers see EM as ‗lying‘ and if so,
whether this affects their decisions. From our descriptive data, the majority of managers
perceived undertaking accounting EM adjustments as lying, although this was not found for
operational EM. This flows through to the decision on EM, if accounting adjustments were
perceived as lying, managers were less likely to manage earnings. This does show that
perceptions of ‗lying‘ of these actions can make difference to behavior of managers. We also
tested the findings by Gneezy (2005) on whether the decision to lie is affected by costs and
benefits to other parties and we found that for those who think accounting EM is lying their
decisions are not affected by economic considerations.
The final research question related to the relative importance of economic and ethical
factors for managers in deciding on the method of managing earnings. In terms of economic
factors affecting the method chosen, the perceived likelihood of success of the particular
method is again the most important factor. However, ethical considerations are more
important from managers‘ perspective in making this choice as the ethical acceptability of
accounting compared to operational adjustments was associated with the choice of the
method of managing earnings. That is, if managers thought accounting adjustments were
more unethical they were more likely to use operational adjustments to manage, whereas
Page 30
29
those who thought that operational adjustments were more unethical were more likely to use
accounting adjustments to manage. The other consideration of importance was whether they
thought making accounting adjustments is ‗lying‘, if they did, they were less likely to use
accounting adjustments to manage earnings.
This study shows that the decision to manage earnings and the method to manage
earnings are complex decisions. Some economic factors clearly affect managers‘ decisions on
EM such as the likelihood of success of particular methods of EM. The affect on current
shareholders is also an important consideration, which quite reasonably shows that managers
see their duty directly related to current shareholders, which not surprising as it is usually
aligned to their economic ‗self-interest‘. However, a significant new finding from this
research is that decisions on EM are not solely affected by direct economic self interest and
that there is a consideration of what is ‗right‘ or ‗ethical‘ in a determination of EM decisions.
Associated with this, whether they believe the actions to be lying or not also has an impact.
Further, ethical perceptions of both real operational and accounting EM and the perceptions
of whether accounting adjustments are lying significantly effects the choice between real
operational and accounting EM. Therefore this study provides evidence on the importance of
ethics in these decisions on EM as was alluded to by Graham et al. (2005), and is consistent
with recent economic research that has questioned the assumption of ‗economic man‘
(Benabou and Tirole 2011).
In summary, this study is the first to comprehensively evaluate the spectrum of factors
that has been shown to affect EM decisions from the prior literature and shows that beyond
the importance of these factors, ethical considerations make a difference. Therefore, a
broader perspective in evaluating EM that encompasses the ethical decision making
framework of managers is important to properly understand the decision to manage or to not
manage earnings.
Page 31
30
APPENDIX
General Background Information
Assume ―the Company‖ is listed on the Australian Stock Exchange, and it has a large, diversified
shareholder base.
The Company has an independent Audit Committee, and a Corporate Code of Ethics for all directors,
officers, and employees.
Financial Background Information
The Company has always reported earnings per share (EPS) that has met or beaten the market‘s
expectations.
The Company has long-term borrowings with a debt covenant.
The Company‘s year-end is 30 June.
Assume that today is 15 May 2010
Based on the Company‘s reported EPS for the six months ended 31 December 2009, the Company was
on track (as of 31 December 2009) to generate EPS of $1.75 for the year ended 30 June 2010, which is
equal to the market‘s expectations.
Today (i.e., 15 May 2010), the Company re-assessed its financial results for 2009/10. The relevant
information is summarised below.
Due to a series of independent, minor, company-specific setbacks unrelated to economic
conditions, the Company is now certain (as of 15 May 2010) that actual EPS for 2009/10 will fall
in the range of $1.67 to $1.69, with $1.68 being the most likely outcome. Thus, the Company now
expects to underperform the market expectation of $1.75 by 4%.
Although the decrease in EPS will move the Company closer to the debt covenant‘s required
benchmarks, the Company‘s expected EPS will be sufficient to maintain a more than acceptable
margin to avoid violating the debt covenant.
Potential Courses of Action
The Company has two general options regarding its reported performance, as the figure below indicates.
Take no special actions to improve reported performance, and thus underperform market expectations by
4%.
Use accounting and/or operational adjustments to improve reported performance by 4% to meet market
expectations.
1. In this situation, where the Company will underperform market expectations by 4% if it does nothing, what
is the likelihood that your firm would try to improve reported performance using accounting adjustments
and/or operational adjustments?
Certain
not to
happen
Certain
to happen
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
Page 32
31
2. What decision do you think your firm would actually make in this situation (please tick the one box that
corresponds to your answer)?
Take no special actions.
Try to improve reported performance, using
accounting and/or operational adjustments Indicate the proportions below.
Would rely
solely on
accounting
adjustments
Would rely
50% on
accounting
adjustments
and 50% on
operational
adjustments
Would rely
solely on
operational
adjustments
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
3. What is the likelihood that your firm would be able to increase reported earnings by 4% in this situation to meet
market expectations if it:
Certain
not to
happen
Certain
to
happen
0% 10
%
20
%
30
%
40
%
50
%
60
%
70
%
80
%
90
%
100%
• Used ONLY accounting adjustments
• Used ONLY operational adjustments
• Used a combination of accounting and
operational adjustments
Page 33
32
4. Assume your firm takes no special action [Qu 5 using only accounting adjustments; Qu6 using only
operational adjustments] to improve reported performance and underperforms market expectations by 4%.
A. How large are the potential costs that taking no special action [Qu 5 accounting; Qu 6 operational] (and thus
underperforming market expectations by 4%) will impose on each of the following stakeholder groups?
Absolutely no costs
Very large costs
0 10 20 30 40 50 60 70 80 90 100
• CFO and CEO
• Company in the short term
• Company in the long term
• Company’s current shareholders
• Company’s future shareholders
• Company’s debtholders
• Other stakeholders (e.g., employees,
customers, or general public)
B. What is the likelihood that a particular stakeholder group will actually incur the potential costs you identified
above?
Certain
not to happen
Certain
to happen
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
• CFO and CEO
• Company in the short term
• Company in the long term
• Company’s current shareholders
• Company’s future shareholders
• Company’s debtholders
• Other stakeholders (e.g., employees,
customers, or general public)
Page 34
33
Ethical Questions
11. Over what period of time do you expect any costs from each course of action to occur?
Immediately
< 6
months
6 months
to 1 year
1 to 2
years
2 to 3
years
3 to 4
years
4 to 5
years
> 5 years
• Take no special actions
• Use accounting adjustments
• Use operational adjustments
12. How many people do you think would be affected by the costs of each course of action?
None 1 2 to 10 10 to 50 50 to 1000 1000+
• Take no special actions
• Use accounting adjustments
• Use operational adjustments
13. How ethically questionable do you believe each course of action is? Not
questionable
at all
Extremely
questionable 0 10 20 30 40 50 60 70 80 90 100
• Take no special actions
• Use accounting adjustments
• Use operational adjustments
14. What percentage of people in your same senior management position do you believe would consider the course of
action to be ethically questionable?
None 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
• Take no special actions
• Use accounting adjustments
• Use operational adjustments
15. What percentage of people in your same senior management position do you think would select each of the following
actions when faced with underperforming market expectations by 4%? The percentages allocated across the actions
must add up to 100.
Take no special actions.
Rely solely on accounting adjustments to improve reported performance
Rely solely on operational adjustments to improve reported performance
Use some combination of accounting and operational adjustments to improve
reported performance
Total (sum of your responses must equal 100%) 100%
Page 35
34
Other Questions
1. Indicate your belief about the degree to which
attempts to improve reported performance via
accounting and/or operational adjustments are
initiated by a company‘s CEO versus its CFO.
CEO CFO
Initiated Initiated
0 10 20 30 40 50 60 70 80 90 100
2. Indicate the degree to which you think that
improving reported performance by using
accounting adjustments is lying?
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
3. Indicate the degree to which you think that
improving reported performance by using
operational adjustments is lying?
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
Page 36
35
REFERENCES
Abernathy, M., J. Bouwens, and L. Van Lent. 2012. Ethics, agent‘s choices, and incentive
contract design. Working Paper, University of Melbourne and Tilburg University.
Bartov, E., and D. Cohen. 2009. The 'numbers game' in the pre- and post-Sarbanes-Oxley
eras. Journal of Accounting, Auditing and Finance 24 (4): 505-534.
Benebou, R., and J. Tirole. 2011. Identity, morals, and taboos: Beliefs as assets. The
Quarterly Journal of Economics 126: 805-855.
Bhojraj, S., and R. Libby. 2005. Capital market pressure, disclosure frequency-induced
earnings/cash flow conflict, and managerial myopia. The Accounting Review 80 (1): 1-20.
Bruns, W., and K. Merchant. 1990. The dangerous morality of managing earnings.
Management Accounting 72 (2): 22–25.
Carey, P., and R. Simnett. 2006. Audit partner tenure and audit quality. The Accounting
Review 81 (3): 653-676.
Christie, A., and J. Zimmerman. 1994. Efficient and opportunistic choices of accounting
procedures: Corporate control contests. The Accounting Review 69 (4): 539-566.
DeAngelo, H., L. DeAngelo, and D. Skinner. 1994. Accounting choice in troubled companies.
Journal of Accounting and Economics 17: 113–143.
Dechow, P., and D. Skinner. 2000. Earnings management: Reconciling the views of
accounting academics, practitioners, and regulators. Accounting Horizons 14 (2-3): 235-
250.
Dechow, P., and R. Sloan. 1991. Executive incentives and the horizon problem: An empirical
investigation. Journal of Accounting and Economics 14 (1): 51-98.
Dikolli, S., W. Mayhew, and T. Steffen. 2012. Honoring one‘s word: CEO integrity and
accruals quality. Working paper, Duke University.
Erhard, W. and Jensen, M. 2012. Putting integrity into finance: A purely positive approach.
Available at: http://ssrn.com/abstract=1985594
Page 37
36
Fields, T., T. Lys, and L. Vincent. 2001. Empirical research on accounting choice. Journal of
Accounting and Economics 31 (1-3): 255-307.
Gneezy, U. 2005. Deception: The role of consequences. The American Economic Review 95
(1): 384-394.
Graham, J., C. Harvey, and S. Rajgopal. 2005. The economic implications of corporate
financial reporting. Journal of Accounting and Economics 40 (1-3): 3-73.
Guidry, F., A. Leone and S.Rock. 1999. Earnings-based bonus plans and earnings
management by business-unit managers. Journal of Accounting and Economics 26: 113–
142.
Gunny, K. 2010. The relation between earnings management using real activities
manipulation and future performance: Evidence from meeting earnings benchmarks.
Contemporary Accounting Research 27 (3): 855-888.
Healy, P. 1985. The impact of bonus schemes on the selection of accounting principles.
Journal of Accounting and Economics 7: 85–107.
Healy, P., and K. Palepu. 1990. Effectiveness of accounting-based dividend covenants.
Journal of Accounting and Economics 12: 97–123.
Healy, P., and J. Wahlen. 1999. A review of the earnings management literature and its
implications for standard setting. Accounting Horizons 13 (4): 365-383.
Hunton, J., R. Hoitash, and J. Thibodeau. 2011. The relationship between perceived tone at
the top and earnings quality. Contemporary Accounting Research 28 (4): 1190–1224.
Holmstrom, B. 2005. Pay without performance and the managerial power hypothesis: A
comment. Journal of Corporation Law 30: 703-713.
Jensen, M. 2001. Value maximisation, stakeholder theory, and the corporate objective
function. European Financial Management 7 (3): 297-317.
Jensen, M. 2009. Integrity: Without it nothing works. Available at:
http://ssrn.com/abstract=1511274
Page 38
37
Johnson, E., G. Fleischman, S. Valentine, and K. Walker. 2012. Managers‘ ethical
evaluations of earnings management and its consequences. Contemporary Accounting
Research, forthcoming.
Jones, J. 1991a. Earnings management during import relief investigations. Journal of
Accounting Research 29 (2): 193-228.
Jones, T. M. 1991b. Ethical decision-making by individuals in organizations: An issue-
contingent model. Academy of Management Review 16: 366-395.
Kasznik, R. 1999. On the association between voluntary disclosure and earnings
management. Journal of Accounting Research 37 (1): 57-81.
Leuz, C., D. Nanda, and P. Wysocki. 2003. Earnings management and investor protection:
An international comparison. Journal of Financial Economics 69 (3): 505-527.
Libby, R., and N. Seybert. 2008. Behavioral studies of the effects of regulation on earnings
management and accounting choice. Working paper, Cornell University.
Loe, T., L. Ferrell, and P. Mansfield. 2000. A review of empirical studies assessing ethical
decision making in business. Journal of Business Ethics 25 (3): 185-204.
McGuire, S., T. Omer, and N. Sharp. 2012. The impact of religion on financial reporting
irregularities. The Accounting Review. Accepted manuscript. doi: 10.2308/accr-10206.
Merchant, K., and J. Rockness. 1994. The ethics of managing earnings: An empirical
investigation. Journal of Accounting and Public Policy 13 (1): 79-94.
Noreen, E. 1988. The economics of ethics: A new perspective on agency theory. Accounting,
Organizations and Society 15 (4): 359-369.
Rowchowdhury, S. 2006. Earnings management through real activities manipulation. Journal
of Accounting and Economics 42: 335-370.
Seybert, N. 2010. R&D capitalization and reputation-driven real earnings management. The
Accounting Review 85 (2): 671-694
Page 39
38
Teoh, S., I. Welch, and T. Wong, T. 1998. Earnings management and the long-run market
performance of initial public offerings. Journal of Finance 53 (6): 1935–1974.
Velasquez, M. and C. Rostankowski. 1985. Ethics: Theory and Practice. Englewood Cliffs,
NJ: Prentice-Hall.
Watts, R., and J. Zimmerman. 1986. Positive Accounting Theory. London: Prentice-Hall.
Page 40
39
_STAT_ Q1 Q2_1 Q2_2 Q3_1acc Q3_2op Q3_3com Q4a_1ceocfo Q4a_2cost Q4a_3colt Q4a_4csh Q4a_5fst Q4a_6dh Q4a_7other
N 224 225 153 210 217 214 225 225 225 225 224 223 224
MIN 0 0 0.1 0 0 0 0 0 0 0 0 0 0
MAX 1 1 1 1 1 1 100 100 100 100 100 100 100
MEAN 0.54 0.68 0.74 0.38 0.51 0.62 43.20 46.09 23.02 43.47 20.09 17.27 20.40
MEDIAN 0.6 1 0.8 0.3 0.5 0.7 50 50 20 50 10 10 15
STD 0.34 0.468 0.202 0.335 0.28 0.326 28.229 27.071 22.712 26.364 22.31 21.957 20.03
Q5a_1ceocfo Q5a_2cost Q5a_3colt Q5a_4csh Q5a_5fsh Q5a_6dh Q5a_7other Q6a_1 Q6a_2 Q6a_3 Q6a_4 Q6a_5 Q6a_6 Q6a_7
N 224 225 225 225 225 224 225 225 225 225 225 225 224 225
MIN 0 0 0 0 0 0 0 0 0 0 0 0 0 0
MAX 100 100 100 100 100 100 100 100 100 100 100 100 70 100
MEAN 38.21 29.51 41.24 26.22 33.29 22.50 22.31 29.16 30.67 33.51 23.64 29.16 17.19 24.62
MEDIAN 30 20 40 20 20 10 20 20 20 30 20 20 10 20
STD 33.77 27.71 29.40 25.94 28.83 25.22 22.56 25.66 24.18 25.73 21.38 25.03 19.03 23.30
Q13_1noaction Q13_2acc Q13_3opr OQ2 OQ3
N 224 224 224 222 222
MIN 0 0 0 0 0
MAX 100 100 100 1 1
MEAN 24.78 67.01 33.53 0.60 0.27
MEDIAN 10 75 30 0.7 0.2
STD 32.90 28.98 28.72 0.33 0.27
TABLE 1 Descriptive Statistics
Page 41
40
Q1 Q2_2 Q3_1ACC Q3_2OP Q3_3COM Q4A_1CEOCFOQ4A_2COSTQ4A_3COLTQ4A_4CSHQ4A_5FST Q4A_6DH Q4A_7OTHERQ13_1NOACTIONQ13_2ACCQ13_3OPROQ2 OQ3
Q1 1.00 0.17 -0.22 0.01 0.13 -0.07 0.05 0.09 0.02 -0.09 -0.05 -0.06 0.12 -0.01 -0.23 0.07 -0.25
Q2_2 1.00 -0.02 0.53 -0.06 -0.31 -0.18 -0.01 -0.03 -0.18 -0.21 -0.05 0.06 0.10 -0.53 0.10 -0.60
Q3_1ACC 1.00 0.13 0.65 0.16 0.21 -0.04 -0.04 0.18 0.04 0.13 -0.03 -0.21 0.20 -0.43 0.23
Q3_2OP 1.00 0.01 -0.25 -0.10 0.05 -0.08 -0.12 -0.06 -0.01 0.03 -0.03 -0.24 0.00 -0.19
Q3_3COM 1.00 0.31 0.36 0.01 0.07 0.14 -0.07 0.22 0.11 -0.10 0.23 -0.26 0.10
Q4A_1CEOCFO 1.00 0.83 0.60 0.59 0.46 0.54 0.63 0.37 0.07 0.17 0.00 0.06
Q4A_2COST 1.00 0.65 0.65 0.44 0.45 0.54 0.39 0.09 0.15 -0.19 0.05
Q4A_3COLT 1.00 0.55 0.53 0.60 0.59 0.53 0.01 0.06 0.04 0.03
Q4A_4CSH 1.00 0.18 0.18 0.47 0.20 0.07 0.13 0.02 -0.03
Q4A_5FST 1.00 0.64 0.48 0.41 -0.09 0.14 -0.16 0.14
Q4A_6DH 1.00 0.51 0.29 0.06 0.03 0.03 0.16
Q4A_7OTHER 1.00 0.13 -0.02 0.05 -0.07 0.05
Q13_1NOACTION 1.00 0.06 0.11 0.06 -0.11
Q13_2ACC 1.00 0.10 0.33 -0.20
Q13_3OPR 1.00 0.05 0.77
OQ2 1.00 -0.01
OQ3 1.00
TABLE 2
Correlation Matrix
Page 42
41
Table 3
Regression of Q1 on Economics, Ethics and Lying
(a) (b) (c) (d) (e)
Variable Coefficient Prob. Coefficient Prob. Coefficient Prob. Coefficient Prob. Coefficient Prob.
C 0.2077 0.0003 0.1871 0.0017 0.4482 0 0.4432 0 0.4639 0
Q3_1ACC -0.1497 0.0766 -0.1083 0.1971 -0.1111 0.1435 -0.1627 0.044 -0.0874 0.2389
Q3_2OP 0.3100 0.0007 0.2588 0.0063 0.1840 0.0283 0.2833 0.0013 0.1802 0.0304
Q3_3COM 0.3174 0.0015 0.2609 0.0121 0.2499 0.0047 0.2887 0.0025 0.1871 0.0426
Q4A_1CEOCFO 0.0026 0.0022 0.0010 0.2697 0.0005 0.5631 0.0017 0.0416 -0.0009 0.3135
Q5A_1CEOCFO -0.0012 0.0999 -0.0003 0.7436 0.0000 0.9732 -0.0004 0.6043 0.0006 0.4534
Q6A_1 -0.0011 0.2794 -0.0018 0.1183 -0.0004 0.6215 -0.0006 0.5317 -0.0012 0.2097
Q4A_4CSH 0.0037 0.0004 0.0032 0.0006
Q4A_5FST 0.0006 0.6422 -0.0003 0.8059
Q4A_6DH 0.0011 0.4387 0.0001 0.9099
Q4A_7OTHER 0.0007 0.6226 -0.0002 0.9055
Q5A_4CSH 0.0012 0.2959 0.0010 0.3041
Q5A_5FSH -0.0018 0.102 -0.0016 0.0995
Q5A_6DH -0.0031 0.0463 -0.0026 0.0603
Q5A_7OTHER 0.0011 0.5064 0.0007 0.6448
Q6A_4 -0.0020 0.1578 -0.0024 0.0567
Q6A_5 -0.0001 0.9453 0.0011 0.3287
Q6A_6 0.0004 0.8041 0.0015 0.3615
Q6A_7 0.0011 0.3883 0.0014 0.212
Q13_1NOACTION 0.0029 0 0.0029 0
Q13_2ACC -0.0011 0.1365 0.0004 0.6368
Q13_3OPR -0.0034 0 -0.0043 0
OQ2 -0.2425 0.0005 -0.2333 0.0028
OQ3 -0.2186 0.015 0.1159 0.3021
Adjusted R-squared 0.2375 0.3063 0.4074 0.3280 0.4879
Page 43
42
Variable Coeff t-Statistic Prob. Coeff t-Statistic Prob.
C 0.3095 4.68 0 0.3505 4.87 0
Q4A_1CEOCFO -0.0019 -0.96 0.3416 -0.0015 -0.76 0.4516
Q4A_2COST 0.0017 0.85 0.3995 0.0012 0.57 0.5677
Q4A_3COLT 0.0019 0.83 0.4075 0.0012 0.52 0.6048
Q4A_4CSH 0.0012 0.71 0.4819 0.0010 0.59 0.5573
Q4A_5FST 0.0020 0.97 0.3339 0.0020 0.92 0.3584
Q4A_6DH -0.0009 -0.44 0.6604 0.0002 0.09 0.9305
Q4A_7OTHER 0.0016 0.71 0.4765 0.0013 0.59 0.5597
Q5A_1CEOCFO -0.0015 -1.11 0.2683
Q5A_2COST 0.0029 1.79 0.0766
Q5A_3COLT 0.0005 0.26 0.7947
Q5A_4CSH 0.0014 0.84 0.4058
Q5A_5FSH 0.0000 0.02 0.9802
Q5A_6DH -0.0020 -1.02 0.3104
Q5A_7OTHER -0.0030 -1.32 0.1888
Adjusted R-squared 0.0498 0.0951
F-statistic 1.8166 1.8183
Prob(F-statistic) 0.0919 0.0467
1.7290 0.1114
Variable Coeff t-Statistic Prob. Coeff t-Statistic Prob.
C 0.3814 6.09 0 0.4061 6.42 0
Q4A_1CEOCFO 0.0012 0.80 0.4272 0.0013 0.83 0.4105
Q4A_2COST 0.0009 0.54 0.5872 0.0015 0.85 0.3962
Q4A_3COLT 0.0015 0.92 0.3576 0.0026 1.50 0.1377
Q4A_4CSH 0.0032 2.10 0.0379 0.0041 2.57 0.0116
Q4A_5FST -0.0020 -1.19 0.2355 -0.0035 -1.90 0.0607
Q4A_6DH -0.0015 -0.95 0.3467 -0.0014 -0.71 0.4812
Q4A_7OTHER 0.0015 0.86 0.3896 0.0024 1.24 0.2178
Q5A_1CEOCFO 0.0012 0.85 0.3949
Q5A_2COST -0.0006 -0.33 0.7425
Q5A_3COLT -0.0023 -1.13 0.2628
Q5A_4CSH 0.0003 0.13 0.8941
Q5A_5FSH -0.0027 -1.45 0.1498
Q5A_6DH -0.0004 -0.16 0.8759
Q5A_7OTHER 0.0006 0.28 0.7804
Adjusted R-squared 0.1648 0.1816
F-statistic 4.0721 2.7280
Prob(F-statistic) 0.0006 0.0020
1.3000 0.2588
Included observations: 110
Panel A OQ2 Accounting is Lying
Panel B OQ2 Accounting is not Lying
Wald Test: Q5 (F-stat)
Wald Test: Q5 (F-stat)
TABLE 4
Anlysis of Benefits and consequences Conditional on Lying or Not Lying
Page 44
43
Robustness Measure to Lying based on a 50% cuttoff
Ethics Adjusted R-squared 0.412165
Ethics and Benefits/Consequences Adjusted R-squared 0.440128
F-statistic 1.560189 0.1025
Sample 128
Variable Coefficient t-Statistic Prob. Coefficient t-Statistic Prob.
C 0.4681 4.82 0 0.4548 3.99 0.0001
Q13_1NOACTION 0.0038 4.36 0 0.0032 2.91 0.0045
Q13_2ACC 0.0012 1.12 0.2646 0.0009 0.76 0.4477
Q13_3OPR -0.0053 -6.10 0 -0.0052 -5.78 0
Q4A_1CEOCFO -0.0036 -2.23 0.028
Q4A_2COST 0.0023 1.38 0.1714
Q4A_3COLT -0.0018 -0.88 0.3828
Q4A_4CSH 0.0015 1.07 0.2861
Q4A_5FST 0.0013 0.73 0.4689
Q4A_6DH 0.0002 0.11 0.9165
Q4A_7OTHER 0.0022 1.17 0.2441
Q5A_1CEOCFO -0.0006 -0.55 0.5813
Q5A_2COST 0.0019 1.43 0.1569
Q5A_3COLT 0.0012 0.72 0.4762
Q5A_4CSH 0.0003 0.19 0.8469
Q5A_5FSH -0.0005 -0.33 0.7405
Q5A_6DH -0.0015 -0.94 0.3522
Q5A_7OTHER -0.0016 -0.85 0.3996
Adjusted R-squared 0.3995 0.4067
F-statistic 25.6133 0.0000 5.3951 0.0000
1.1894 0.2968 Wald test F-statistic Q4 and Q5
TABLE 5
Analysis of Benefits/Consequences of Lying or Not Lying after Controlling for Ethics
Panel A OQ2 Accounting is Lying
Page 45
44
Variable Coeff t-Statistic Prob. Coeff t-Statistic Prob.
C 0.6753 10.90 0 0.5321 7.36 0
Q13_1NOACTION 0.0029 3.78 0.0003 0.0025 3.17 0.0021
Q13_2ACC 0.0001 0.11 0.9128 -0.0006 -0.50 0.6217
Q13_3OPR -0.0043 -3.79 0.0003 -0.0039 -3.56 0.0006
Q4A_1CEOCFO -0.0004 -0.30 0.7617
Q4A_2COST 0.0018 1.18 0.2414
Q4A_3COLT 0.0014 0.91 0.3649
Q4A_4CSH 0.0048 3.37 0.0011
Q4A_5FST -0.0031 -1.86 0.0654
Q4A_6DH -0.0018 -1.04 0.3009
Q4A_7OTHER 0.0009 0.54 0.5926
Q5A_1CEOCFO 0.0023 1.79 0.076
Q5A_2COST -0.0014 -0.92 0.362
Q5A_3COLT -0.0032 -1.61 0.111
Q5A_4CSH -0.0003 -0.13 0.8934
Q5A_5FSH -0.0017 -1.00 0.3211
Q5A_6DH 0.0004 0.20 0.839
Q5A_7OTHER 0.0018 0.87 0.3875
Adjusted R-squared 0.2453 0.3720
F-statistic 12.9149 0.0000 4.7627 0.0000
3.6871 0.0015
1.3075 0.2558
Panel B OQ2 Accounting is not Lying
Wald test F-statistic Q4
Wald test F-statistic Q5
TABLE 5
Page 46
45
TABLE 6
Regression of Q2_2 on Economics, Ethics and Lying
Coeff Prob. Coeff Prob. Coeff Prob.
C 0.6235 0
0.4695 0
0.4603 0
Q3_1ACC -0.1876 0.0015
-0.1159 0.0251
-0.0991 0.059
Q3_2OP 0.3243 0
0.2570 0.0001
0.2571 0.0001
Q3_3COM 0.0367 0.6139
0.0174 0.7861
-0.0028 0.9671
Q4A_1CEOCFO -0.0009 0.2982
-0.0007 0.3741
-0.0008 0.3018
Q4A_2COST 0.0002 0.8248
0.0003 0.6703
0.0003 0.7003
Q4A_3COLT 0.0000 0.9697
0.0001 0.866
0.0003 0.7586
Q4A_4CSH -0.0006 0.4243
-0.0001 0.8423
0.0001 0.8416
Q4A_5FST 0.0007 0.5043
0.0009 0.3218
0.0008 0.3826
Q4A_6DH -0.0007 0.5196
-0.0010 0.2848
-0.0011 0.2558
Q4A_7OTHER -0.0006 0.6155
-0.0009 0.3691
-0.0007 0.4748
Q5A_1CEOCFO 0.0002 0.8014
-0.0001 0.9299
0.0000 0.9449
Q5A_2COST -0.0003 0.7907
-0.0006 0.4881
-0.0004 0.6622
Q5A_3COLT 0.0027 0.0199
0.0015 0.1706
0.0011 0.3382
Q5A_4CSH 0.0001 0.898
-0.0006 0.4874
-0.0005 0.5822
Q5A_5FSH -0.0006 0.5452
-0.0001 0.8778
-0.0001 0.9313
Q5A_6DH 0.0020 0.118
0.0016 0.1548
0.0015 0.1717
Q5A_7OTHER -0.0005 0.6901
-0.0005 0.6543
-0.0004 0.7258
Q6A_1 -0.0016 0.0803
-0.0010 0.238
-0.0010 0.2458
Q6A_2 -0.0001 0.8809
-0.0002 0.8228
-0.0004 0.6224
Q6A_3 0.0006 0.5727
0.0008 0.3903
0.0008 0.3996
Q6A_4 -0.0007 0.5211
-0.0009 0.314
-0.0009 0.3092
Q6A_5 -0.0005 0.6422
0.0001 0.9544
0.0004 0.6828
Q6A_6 -0.0014 0.3572
0.0000 0.9907
-0.0001 0.9528
Q6A_7 0.0004 0.6807
-0.0008 0.3508
-0.0007 0.4169
Q13_1NOACTION
0.0005 0.2419
0.0005 0.2217
Q13_2ACC
0.0037 0
0.0028 0.0003
Q13_3OPR
-0.0023 0.001
-0.0017 0.0663
OQ2
0.1218 0.058
OQ3
-0.0960 0.3011
Adjusted R-squared 0.2887
0.4811
0.4822
Included observations: 138