The Effects of Board Characteristics on Earnings Management around Australian CEO Changes

Description
Prior research has shown evidence of earnings
management in financial reports of US and
Australian firms changing chief executive
officer (CEO). This paper examines whether
corporate boards, with certain characteristics
associated with strong corporate governance,
are effective in controlling any earnings
management in the financial reports of
Australian firms that change CEOs. Since
hiring, monitoring and replacing the CEO
are key roles of the board of directors,
research in this specific context is considered
particularly appropriate.

Accounting Research Journal
The Effects of Board Characteristics on Earnings Management around Australian
CEO Changes
Paul Mather Alan Ramsay
Article information:
To cite this document:
Paul Mather Alan Ramsay, (2006),"The Effects of Board Characteristics on Earnings Management
around Australian CEO Changes", Accounting Research J ournal, Vol. 19 Iss 2 pp. 78 - 93
Permanent link to this document:http://dx.doi.org/10.1108/10309610680000680
Downloaded on: 24 January 2016, At: 20:59 (PT)
References: this document contains references to 0 other documents.
To copy this document: [email protected]
The fulltext of this document has been downloaded 945 times since 2006*
Users who downloaded this article also downloaded:
Rashidah Abdul Rahman, Fairuzana Haneem Mohamed Ali, (2006),"Board, audit committee,
culture and earnings management: Malaysian evidence", Managerial Auditing J ournal, Vol. 21 Iss
7 pp. 783-804http://dx.doi.org/10.1108/02686900610680549
Sandra Maria Geraldes Alves, (2011),"The effect of the board structure on earnings management:
evidence from Portugal", J ournal of Financial Reporting and Accounting, Vol. 9 Iss 2 pp. 141-160http://dx.doi.org/10.1108/19852511111173103
Norman Mohd Saleh, Takiah Mohd Iskandar, Mohd Mohid Rahmat, (2007),"Audit committee
characteristics and earnings management: evidence from Malaysia", Asian Review of Accounting,
Vol. 15 Iss 2 pp. 147-163http://dx.doi.org/10.1108/13217340710823369
Access to this document was granted through an Emerald subscription provided by emerald-
srm:115632 []
For Authors
If you would like to write for this, or any other Emerald publication, then please use our
Emerald for Authors service information about how to choose which publication to write for and
submission guidelines are available for all. Please visit www.emeraldinsight.com/authors for
more information.
About Emerald www.emeraldinsight.com
Emerald is a global publisher linking research and practice to the benefit of society. The
company manages a portfolio of more than 290 journals and over 2,350 books and book series
volumes, as well as providing an extensive range of online products and additional customer
resources and services.
Emerald is both COUNTER 4 and TRANSFER compliant. The organization is a partner of the
Committee on Publication Ethics (COPE) and also works with Portico and the LOCKSS initiative
for digital archive preservation.
*Related content and download information correct at time of
download.
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
0
:
5
9

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
ACCOUNTING RESEARCH JOURNAL  VOLUME 19 NO 2 (2006) 

78

The Effects of Board Characteristics on
Earnings Management around Australian
CEO Changes
Paul Mather and Alan Ramsay
Department of Accounting and Finance
Monash University

Abstract

Prior research has shown evidence of earnings
management in financial reports of US and
Australian firms changing chief executive
officer (CEO). This paper examines whether
corporate boards, with certain characteristics
associated with strong corporate governance,
are effective in controlling any earnings
management in the financial reports of
Australian firms that change CEOs. Since
hiring, monitoring and replacing the CEO
are key roles of the board of directors,
research in this specific context is considered
particularly appropriate. After controlling for
contemporaneous and lagged profitability in the
year of CEO change, we find evidence of
negative unexpected accruals in our sub-sample
of firms where the CEO resigned. For this
group, larger boards and a higher proportion
of independent directors appear to limit
observed negative earnings management. In the
case of CEO retirements there is evidence of
positive unexpected accruals in the period
of CEO change. However, none of the board
characteristics show any significant relationship
with unexpected accruals. In the period after
CEO change, we find no evidence of positive
unexpected accruals for CEO resignations and

Acknowledgements: The research assistance of Calvin Tan
and Clifton Phua is gratefully acknowledged. The paper has
benefited from the helpful comments of Nicholas Carline,
Robert Faff, Matt Hall, Mike Jones, Graham Peirson and
an anonymous referee. Participants at the AFAANZ
conference, Brisbane, July 2003, the Financial Reporting
Conference, Cardiff, July 2003, and research seminars at
Monash University and the University of Melbourne are also
acknowledged. Nevertheless, the usual caveat applies. The
Monash University Research Fund provided financial
support for this project.
none of the board characteristics show any
significant relationship with unexpected
accruals. For CEO retirements, our analysis
indicates that a higher proportion of executive
and affiliated director shareholding goes some
way towards counteracting the observed
positive unexpected accruals. When lagged
unexpected accruals are included in the
regression equation to control for accrual
reversals, CEO duality significantly increases
the already positive earnings management
found in CEO retirements in the period
following CEO change.
1. Introduction
The context of chief executive officer (CEO)
change provides CEOs with incentives (human
capital development, remuneration) to
opportunistically manage reported earnings
(Pourciau, 1993). Recent empirical evidence
provides support for these arguments in the
context of CEO changes by Australian firms
(Godfrey, Mather and Ramsay, 2003; Wells,
2002). This paper investigates whether certain
characteristics of a firm’s board of directors
constrain such earnings management in the
financial reports of Australian firms that change
CEOs. The specific research questions
addressed are: (a) do board characteristics
constrain downward earnings management in
financial reports in the year of a CEO change?
and (b) do these board characteristics constrain
upward earnings management in the year after
CEO change?
Several interrelated factors motivate this
study. First, issues involving corporate
governance, especially for listed companies,
have regularly been the focus of much attention
and the subject of international debate in the
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
0
:
5
9

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
The Effects of Board Characteristics on Earnings Management around Australian CEO Changes

79

past decade. As a result of the recent instances
of corporate failures and accounting scandals,
national regulators have established corporate
governance codes, such as the Sarbanes-Oxley
Act (2002) in the United States and the ASX
Corporate Governance Council (2003) in
Australia. These regulators believe that
improving corporate governance structures
within firms will compel managers to act in the
shareholders’ best interests and will thus ensure
that resources are more optimally allocated. As
the recommendations of the corporate
governance codes may impose implementation
costs on companies, an empirical examination
of the efficacy of some of the regulators’
recommendations is considered useful.
Second, corporate governance is concerned
with (amongst other things) ensuring that a
firm’s resources are optimally employed for the
benefit of shareholders (Fama and Jensen,
1983). A growing body of literature examines
whether corporate governance mechanisms,
such as the size and structure of the firm’s board
of directors mitigate opportunistic earnings
management (for example, Dechow, Sloan and
Sweeney, 1996; Peasnell, Pope and Young,
2005; Xie, Davidson and DaDalt, 2002). This
paper aims to extend this literature by testing
whether certain board characteristics are
effective in controlling perceived opportunistic
earnings management around CEO changes.
Since hiring, monitoring and, if necessary,
replacing the CEO are key roles of the board of
directors (see, Weisbach, 1988), research in this
specific context is considered particularly
appropriate.
Our results show that in the year of CEO
change, after controlling for contemporaneous
and lagged profitability, negative earnings
management is only observed where the CEO
resigns. In this situation, larger boards and a
higher proportion of independent directors
significantly reduce the negative unexpected
accruals suggesting that these board
characteristics provide more effective
monitoring in this context. In the case of CEO
retirements and in an interesting contrast to
those who resign, there is evidence of positive
unexpected accruals in the period of CEO
change. However, for CEO retirements, none of
the board characteristics examined show any
significant relationship with unexpected
accruals. In the period after CEO change, we
find no evidence of positive unexpected
accruals for CEO resignations. Our analysis
indicates that none of the board characteristics
examined show any significant relationship
with unexpected accruals. For CEO retirements,
we find evidence of significant positive
unexpected accruals. The board characteristic of
a higher proportion of executive and affiliated
director shareholding goes some way towards
counteracting the observed positive unexpected
accruals.
To counter possible accrual reversal effects
between the period of and the period after CEO
change, we include lagged unexpected accruals
in our regression model for period t+1. For
CEO resignations, lagged unexpected accruals
are insignificant, current period unexpected
accruals are insignificant, as are all board
characteristic variables. For CEO retirements,
there is evidence of accrual reversals as lagged
unexpected accruals are significantly negative
(5% level) and current period unexpected
accruals are significantly positive (5% level). Of
the board characteristic variables, CEO duality
is positive and significant (5% level). That is,
following a CEO retirement, the already
significantly positive unexpected accruals are
further increased when the CEO is also
chairman of the board.
2. Hypothesis Development
2.1 Theoretical framework: CEO change
and earnings management
An agency framework may be used to show that
new CEOs have incentives to increase their
welfare by establishing tenure, enhancing their
human capital (Fama, 1980) and improving
their compensation by taking a ‘big bath’ in the
year of CEO change and managing earnings
upward in the following year (Pourciau 1993;
Godfrey et al. 2003).
Vancil (1987) argues that CEOs face two
critical tasks on appointment. First, they have to
manage the expectations of the board and senior
staff. Second, they need to earn the confidence
of all stakeholders by achieving an initial set of
performance targets in their first year or two as
CEO. It is argued that to manage expectations
and set (and reach) achievable performance
goals, the new CEO has strong incentives to
engage in initial downward earnings
management followed by subsequent upward
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
0
:
5
9

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
ACCOUNTING RESEARCH JOURNAL  VOLUME 19 NO 2 (2006) 

80

earnings management (Pourciau, 1993). Such a
strategy sets an initial low benchmark against
which performance is judged and, at the same
time, increases the ability of the new CEO to
subsequently manage earnings upwards.
Godfrey et al. (2003) provide anecdotal
insights into the Australian institutional context
which suggest that earnings management at the
time of CEO change is normal and expected
and that such earnings management takes place
even where the new CEO is an internal
appointee. However, the likelihood of such
earnings management appears to be
significantly reduced where the departing CEO
continues to serve the firm in some capacity,
especially as chairman of the board. Similarly,
anecdotal evidence in the financial press
suggests that such earnings management around
CEO changes is common practice in Australia.
1

Related to the above, the institutional context
of CEO change may play an important role in
mitigating or accentuating opportunistic
behaviour by the incoming CEO. Where the
succession process is orderly and planned
(‘retirement’), there are relatively fewer
incentives and opportunities for earnings
management, as compared to the case of an
unplanned CEO change (‘resignation’), where
incentives and opportunities for earnings
management are likely to be greater (Pourciau,
1993). This is consistent with the empirical
evidence of Wells (2002) and Godfrey et al.,
(2003), who find evidence of earnings
management mainly for CEO resignations
and/or non-routine changes in CEO.
Earnings management around senior
management changes has been extensively
studied in the US (DeAngelo, 1988; Elliott and
Shaw, 1988; Strong and Meyer, 1987; Pourciau,
1993). Evidence generally supports the finding
of negative unexpected accruals in the period of
CEO change and positive unexpected accruals
in the following period. In Australia, two
studies have investigated earnings management
behaviour around CEO changes. Wells (2002)
finds weak evidence of negative unexpected
accruals in the period of CEO change, but
stronger evidence of earnings management
through abnormal and extraordinary items.

1 For example, see the Business Australian, 24 April 2000,
P32.
Godfrey et al. (2003) find evidence of negative
unexpected accruals in the period of CEO
change, followed by positive unexpected
accruals in the period after change.
An alternative view is that, rather than such
behaviour being opportunistic, the new CEO
may be dealing with problems previously
ignored by the old management. In other words,
the new CEO is recording write-offs that should
have been made by the previous management.
In fact, Pourciau states that the evidence she
reports (and previously reported) is consistent
with both these hypotheses and that she cannot
differentiate between the two. We argue,
however, that by examining the effect of board
characteristics in this context should enable us
to differentiate between these two alternative
explanations, as the accurate recognition of
necessary write-offs should not be significantly
reduced by board characteristics associated with
strong corporate governance.
2

It may also be argued that, if faced with the
threat of job termination, the outgoing CEO
may have incentives to oversee income
increasing accruals in the period before the
CEO change. Thus if accruals return to ‘normal’
levels in the period of CEO change, it will
appear that the incoming CEO has engaged in
downward earnings management. However, the
focus of this paper is to examine the effects of
corporate governance factors on earnings
management around the CEO change.
Irrespective of whether it was undertaken by the
outgoing or incoming CEO, we gain insights
into the board’s ability to constrain such
earnings management
3
.
2.2 Theoretical framework: corporate
governance
There are many definitions of corporate
governance in the literature. Presenting a
financial accounting and financial economics
perspective, Sloan defines corporate governance
as “the mechanisms that have evolved to
mitigate incentive problems created by the

2 In fact it may be argued that the opposite is likely to
be the case. Effective corporate governance should
increase the likelihood of economically necessary write
offs being made.
3 In any event, Wells (2002) specifically examines and
finds no evidence of earnings management by outgoing
CEOs.
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
0
:
5
9

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
The Effects of Board Characteristics on Earnings Management around Australian CEO Changes

81

separation of management and financing of
business entities” (Sloan, 2001, pp.336). This
takes a classic agency approach with corporate
governance viewed as a series of contractual
and control mechanisms to monitor and control
management’s behaviour, particularly in public
corporations where there is separation of
ownership and control.
The board of directors is central to corporate
governance and helps mitigate agency problems
associated with managing corporations (Fama
and Jensen, 1983). One of the board’s important
roles is to monitor senior management including
the CEO. As the inside (or executive) directors
are by definition part of the aforementioned
senior management team, the monitoring role is
likely to primarily be the responsibility of the
independent directors. Empirical evidence in
support of this monitoring role is provided by
Mace (1971) and Wiesbach (1988) who find
that outside directors were proactive in
replacing CEOs who were not performing. The
economics of the managerial labour market may
also provide incentives for independent
directors to be effective monitors as to do
otherwise would result in reputation losses that
diminish the value of their human capital
(Fama, 1983).
4
For example, Gilson (1990)
finds that directors who leave distressed firms
hold fewer directorships in the future. A further
issue is that executive directors owning equity
may reduce agency problems by more closely
aligning their interests to those of the other
shareholders (Jensen and Meckling, 1976).
5

It is often argued that large boards are less
effective because coordination and process
problems outweigh the advantages of having
access to wider expertise and/or greater skills
(for example, see Jensen, 1993). There is a
growing body of research into board size and
corporate performance but the results are mixed.
For example, Yermack (1996) and Eisenberg,
Sundgren and Wells (1998) provide evidence
that smaller boards are associated with stronger

4 Notwithstanding the above, some critics maintain that
independent directors perform little or no real monitoring
role because they lack the necessary time and expertise
to minimize the existing information asymmetries and
challenge management effectively (Hart, 1995).
5 This is likely to be particularly so where the value of the
equity forms a significant part of the inside directors’
personal wealth.
financial performance. In contrast, Dalton,
Daily, Johnson and Ellstrand (1999) show a
positive and significant relationship between
board size and financial performance. The latter
was a meta-analysis of numerous prior research
samples with an aggregated sample size of
20,620 observations. There is far less research
that focuses on the relationship between board
size and its monitoring role, but Xie, et al.
(2002) and Chtourou et al. (2001) investigate
and find that larger boards are associated with a
lower level of earnings management. An
explanation suggested is that a larger board is
more likely to have independent directors with
financial accounting expertise and is therefore
likely to be better at preventing earnings
management.
It may also be argued that there is an inherent
conflict between the CEO and the independent
directors on the board. The CEO has incentives
to control the board to maximize his or her own
welfare (for example, tenure and compensation)
but independent directors have incentives to
maintain their independence, monitor the CEO
and replace him or her where appropriate
(Hermalin and Weisbach, 2003). Arguably, a
CEO who also chairs the board (CEO duality) is
going to have an impact on its independence
and reduce the board’s ability to effectively
monitor the CEO.
6
On the other hand, splitting
the roles of CEO and Chairman may introduce
agency costs associated with controlling the
behaviour of the non-CEO chairman. Such costs
are avoided where there is CEO duality
(Brickley, Jarrell and Coles, 1996).
2.3 Prior research into the impact of
corporate governance on earnings
management
Dechow, et al. (1996), investigates earnings
manipulation behaviour by firms subject to SEC
enforcement actions in the US. Results showed
that firms manipulating earnings are more likely

6 It should be noted that the recently released ASX Best
Practice Recommendations (ASX Corporate Governance
Council, 2003) in Australia included recommendations
that a majority of the board should be independent
directors and that the roles of chairperson and CEO
should be separated. However, empirical evidence from
the US suggests that CEO duality is the norm and that
there is no financial performance differential or share
market reaction to combining or separating these roles
(Brickley et al., 1996, p.67).
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
0
:
5
9

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
ACCOUNTING RESEARCH JOURNAL  VOLUME 19 NO 2 (2006) 

82

to have boards that are dominated by insiders,
more likely to have CEO duality, less likely to
have an audit committee and less likely to have
an outside block holder. Similar research looks
at the relation between board composition and
financial statement fraud (Beasley, 1996).
Beasley finds that firms experiencing fraud
have a significantly lower proportion of
independent board members, but the existence
of an audit committee does not significantly
affect the likelihood of fraud.
These early papers investigated extreme
cases of earnings manipulation, beyond the
more subtle accruals-based earnings
management permitted by accounting standards.
More recent research has investigated the
broader relationship between earnings
management and corporate governance
structures. Xie, et al. (2002), investigate
earnings management and corporate governance
in US firms, focusing on the role of the board
and the audit committee. They find that larger
boards and greater independent representation
on the board are associated with a lower level of
earnings management. However, the mere
existence of an audit committee does not
mitigate earnings management behaviour.
Similarly, Peasnell, et al. (2005) investigate
corporate governance and earnings management
in UK firms. Their results show that the
likelihood of managers making earnings-
increasing abnormal accruals to avoid falling
below important earnings thresholds is
negatively related to the proportion of
independent directors on the board. They too
found no evidence that the existence of an audit
committee influences the extent of earnings
management.
In further US research, Klein, (2002),
investigates whether board and audit committee
characteristics are related to earnings
management. Contrary to the tenor of the Blue
Ribbon committee recommendations, she finds
a non-linear, negative relation between audit
committee independence and earnings
management. Finally, Chtourou, Bedard and
Courteau, (2001) investigate corporate
governance and earnings management for two
samples of US companies, one selected because
of their higher than normal discretionary
accruals and the other selected because of their
lower than expected discretionary accruals.
They find that larger boards are associated with
lower earnings management and the existence
of large shareholders seems to decrease the
likelihood of positive discretionary accruals. No
prior research has investigated the effects of
corporate governance factors on earnings
management in the context of CEO change.
2.4 Hypotheses
Based on the above discussion, we expect that a
new CEO has incentives to engage in earnings
management that allows him or her to attribute
poor performance to his or her predecessor and
support, ex post, the claim of being a better
manager. This is especially so when the CEO
change is unplanned (resignations). Hiring,
monitoring and, if necessary, firing the CEO are
key roles of the Board. Thus theory as well as
prior research supports the view that
characteristics such as larger boards, greater
extent of independent directors, larger
proportion of inside director shareholdings and
CEO non-duality (“board characteristics
associated with strong corporate governance”)
are likely to mitigate such earnings management
behaviour around CEO changes. Accordingly,
we predict that these board characteristics
associated with strong corporate governance
will constrain opportunistic earnings
management in the reporting periods
surrounding CEO change.
H1: the board characteristics associated with
strong corporate governance will constrain
negative earnings management in the
reporting period of the CEO change
(period t).
H2: the board characteristics associated with
strong corporate governance will constrain
positive earnings management in the
reporting period following the CEO change
(period t+1).
3. Research Method
3.1 Regression model
To test the hypotheses, we use ordinary least
squares regression of equation (1).
UA = ? + ?
1
CFO + ?
2
LEV + ?
3
SIZE
+ ?
4
CONTEMPNP
+ ?
5
LAGGEDNP + ?
6
BOARDSIZE
+ ?
7
PROPIND + ?
8
DUAL
+ ?
9
EXECSHAR (1)
The dependent variable is UA—that is,
unexpected accruals for all sample firms. The
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
0
:
5
9

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
The Effects of Board Characteristics on Earnings Management around Australian CEO Changes

83

independent variables in equation (1) are the
board characteristic variables and the control
variables. These are defined in section 3.2.
Equation (1) is tested separately on data for
period t and period t+1,
7
as we expect the sign
of the relationship between UA and the various
explanatory variables to reverse between the
period of CEO turnover (period t) and the
period after CEO turnover (period t+1). To
support H1 in period t, UA must be significantly
negative and the coefficient on one or more of
the board characteristic variables must be
significantly positive. To support H2 in period
t+1, UA must be significantly positive and the
coefficient on one or more of the board
characteristic variables must be significantly
negative.
3.2 Variable definition
3.2.1 Measuring earnings management ? UA
Accruals reflect the impact of changes in
accounting estimates that underlie much of the
accrual accounting process. We identify
unexpected accounting accruals (discretionary
accruals) reflected in earnings released by firms
during the period of a CEO change (t) and
after a CEO change (t+1) as a means of
detecting possible earnings management. Such
unexpected accruals are a proxy for the
discretionary component of reported earnings or
the extent to which earnings management has
occurred.
We follow DeAngelo (1988), Eddey and
Taylor (1999) and Godfrey et al. (2003) in
measuring unexpected accruals using each firm
as its own control and assuming that expected
accruals remain constant.
More formally, unexpected accruals in year t
may be expressed as follows:
AC
t
= NP
t
– CFO
t
(2)
UA
t
= AC
t
– AC
t-1
(3)
where AC
t
= the accrual component of
earnings in year t,

7 The year of change (t) is the first year the incoming CEO
could realistically control the preparation of the
accounting numbers in the annual report. Consistent with
Pourciau (1993), we use a date three months after the
financial reporting year-end date as the cut off point. For
instance, in the case of a firm with a 31 December year-
end, 1997 would be considered to be the year of change
if the new CEO commenced duties prior to 1 April,
1998.
NP
t
= net operating profit after
interest and tax in year t,
CFO
t
= cash flow from operations in
year t, and
UA
t
= unexpected accruals in year t
Unexpected accruals in year t is the
difference between the change in net operating
profit after interest and tax and the change in
cash flow from operations, comparing year t to
year t-1. Unexpected accruals are divided by
beginning of period total assets to give a
standardised measure.
This measure of unexpected accruals has
some limitations (Dechow et al. 1995; Eddey
and Taylor, 1999). First, if unexpected accruals
vary with firm performance and firm
performance varies with the probability of CEO
change, then the ensuing tests will be biased.
Extensive empirical research in the US (Warner,
Watts and Wruck, 1988), the UK (Conyon and
Florou, 2002,) and Australia (Wells, 2002, Lau,
Sinnadurai and Wright, 2005) documents a
negative relationship between corporate
performance and CEO turnover. To overcome
this problem, we incorporate into our
regressions, control variables capturing current
and lagged profitability. This is intended to
reduce, if not eliminate, possible bias in the
measurement of unexpected accruals. Second,
there is the possibility of measurement error, so
the tests may lack sufficient power to detect
earnings management.
The benefits of this method include requiring
far less data in comparison to the common
alternatives and the underlying assumptions
being less restrictive.
8

8 An alternative method of measuring unexpected accruals
is the regression method developed by Jones (1991).
This method uses potential determinants of accounting
accruals (i.e., sales, non-current assets) to estimate
expected accruals and the resulting estimate of
unexpected accruals. However, this approach requires a
substantial time-series or cross-section of data as well as
assumptions of time-series or cross sectional stationarity.
Moreover, Dechow et al. (1995) find no evidence that
the regression method systematically outperforms the
first-differences method. However, Dechow et al. (1995)
do point out that the DeAngelo method suffers from low
power. Additionally, Guay et al. (1996) suggests that this
lack of power may not be such a problem in non-random
samples, such as ours, where management incentives are
strong.
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
0
:
5
9

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
ACCOUNTING RESEARCH JOURNAL  VOLUME 19 NO 2 (2006) 

84

3.2.2 Board characteristics
Based on the theory developed in this paper and
prior research, the following board
characteristic variables are included in the
model.
Board size (BOARDSIZE): the number of
members on the company’s board of directors,
as disclosed in the company’s annual report.
Proportion of independent directors
(PROPIND): the number of directors classified
as independent based on the Investment and
Financial Services Association definition,
divided by board size. The classification was
based on information supplied in the corporate
governance disclosures in the company’s annual
report.
CEO is also Chairperson (DUAL): a dummy
variable set to 1 if there was CEO duality, zero
otherwise.
Executive shareholdings (EXECSHAR): the
number of shares held by executive and
affiliated directors, divided by the total number
of shares outstanding. This is based on
information supplied in the company’s annual
report.
3.2.3 Control variables
Consistent with prior literature we included the
following control variables:
Cash flow from operations (CFO)
9
: cash
flow from operations in the current year. As
disclosed in the cash flow statement in the
company’s annual report (Dechow et al. 1995;
Guay et al. 1996).
Leverage (LEV): amount of interest-bearing
debt divided by year-end total assets (Peasnell
et al. 2005).
Size (SIZE): the amount of total assets as at
the end of the reporting period (Xie et al. 2002;
Peasnell et al. 2005).
Given the prior evidence of a negative
relationship between CEO turnover and
corporate performance, (Warner, Watts and
Wruck, 1988; Conyon and Florou, 2002; Wells,
2002; Lau, Sinnadurai and Wright, 2005) we
include two accounting measures of
performance as control variables:

9 We define UA as follows: UA
t
= (NP
t
– NP
t-1
) – (CFO
t
–
CFO
t-1
)
Therefore, we expect to find a strong negative
association between CFO
t
and UA
t
(Dechow, 1994). See
the results section for further discussion.
Contemporaneous profitability (CONTEMPNP)
10
:
is the net profit after tax reported in the current
period divided by the beginning of period
total assets. Thus this is a measure of
contemporaneous return on total assets.
Lagged profitability (LAGGEDNP)
11
: is the
net profit after tax in the previous reporting
period divided by the beginning of previous
period total assets. This is a measure of lagged
return on total assets.
3.2.4 Other variables
As noted in the theoretical development, CEO
incentives may differ depending on whether the
CEO change is planned (retirement) or forced
(resignation). Therefore, we classify the data
based on the circumstances of the CEO change
(resign/retire). Contemporary newspaper
reports, company announcements to the ASX
and the company’s annual reports are used to
identify the circumstances of the CEO change
and classify changes as resignations or
retirements. Two dummy variables are created:
RESIGN, RETIRE which are used to partition
the data as shown in equation (4).
The model in equation (4) allows estimation
of the model in equation (1) separately for the
RESIGN group of CEO departures and for the
RETIRE group.
3.3 Data and sample
We extended the sample used by Godfrey et al.
(2003) who identified CEO changes by using a
keyword search of the Australian Financial
Review database for the period 1992-1998. For
the purposes of this study, the search was
extended to 1999 and we also used the
Securities Industry Research Centre of
Asia-Pacific (SIRCA) Signal G database of
company announcements to the Australian
Stock Exchange (ASX), searching each
announcement to identify those relating to CEO
changes that may not have been picked up in
the earlier search. We eliminate all multiple

10 We define UA as follows: UA
t
= (NP
t
– NP
t-1
) – (CFO
t
–
CFO
t-1
)
Therefore, we expect to find a strong positive association
between NP
t
and UA
t
. See the results section for further
discussion.
11 We define UA as follows: UA
t
= (NP
t
– NP
t-1
) – (CFO
t
–
CFO
t-1
)
Therefore, we expect to find a strong negative
association between NP
t-1
and UA
t
. See the results
section for further discussion.
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
0
:
5
9

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
The Effects of Board Characteristics on Earnings Management around Australian CEO Changes

85

RETIRE EXECSHARE DUAL PROPIND BOARDSIZE
LAGGEDNP CONTEMPNP SIZE LEV CFO RETIRE
RESIGN EXECSHARE DUAL PROPIND BOARDSIZE
LAGGEDNP CONTEMNP SIZE LEV CFO RESIGN UA
t t t t
t t t t
t t t t
t t t t t
* )
(
* )
(
1 18 1 17 1 16 1 15
14 1 13 1 12 1 1 10
1 9 1 8 1 7 1 6
5 1 4 1 3 1 2 1 1 1
+ + + +
+ + + +
+ + + +
+ + + + +
+ + + +
+ + + + + +
+ + + +
+ + + + + =
? ? ? ?
? ? ? ? ?
? ? ? ?
? ? ? ? ?
(4)

cites for the same company, all instances where
there is no CEO change, overlapping changes,
all unlisted entities such as foreign firms,
government agencies and other not-for-profit
organisations.
Consistent with Pourciau, (1993) and
Godfrey et al. (2003), we also deleted all cases
where the departing CEO continues to serve the
firm in some capacity, such as non-executive
director or consultant. This requirement
increases the importance of board monitoring in
controlling the incoming CEO’s increased
scope for oppurtunistic earnings management.
The final sample consists of 87 firms.
12
We
classify 27 of the sample as retirements and 60
as resignations.
13
The relatively high incidence
of resignations in the final sample is partly the
result of the requirement to delete all cases
where the departing CEO continues to serve the
firm in some capacity.
14

4. Results
Table 1 summarises key variables for the
sample of 87 firms that changed CEOs. The
SIZE variable shows that the sample comprises
both large and small firms. It also shows
considerable skewness due to the inclusion of
several large banks in the sample.
15
The mean
and median size of the RESIGN group is

12 In comparison to the sample used in Godfrey et al
(2003), the current sample represents a 38% increase in
sample size. Pourciau (1993) has a sample of 73 firms,
thus our sample is somewhat larger than samples used in
prior research.
13 Other possible approaches to classifying the
circumstances of CEO change include classifying the
appointee using an insider/outsider dichotomy. Given
that incentives for earnings management are stronger for
CEO resignations, we have retained the resign/retire
distinction used in Godfrey et al. (2003).
14 Wells 2002 notes that over 50% of departing CEOs
remain on the Board of the company. Our sample of
‘retirements’ is therefore atypical of standard Australian
experience.
15 Our sample contains nine CEO changes in listed
Australian banks. We have re-run all tests excluding
these banks and find the results consistent with those
discussed in this section of the paper.
considerably smaller than that of the RETIRE
group. UA (unexpected accruals) are negative in
period t and positive in period t+1 for both the
RESIGN and RETIRE group, although larger in
each direction for the RESIGN group. A t-test
shows that UA for the RESIGN group are
reliably negative in period t and weakly positive
in period t+1. A matched pairs t-test shows that
for the RESIGN group, UA in period t is
significantly different to UA in period t+1.
There are no significant differences for the
RETIRE group. Mean CONTEMPNP in period
t is negative for the RESIGN group and is
positive for both groups in period t+1.
Consistent with prior empirical research,
correlations between control variables CFO,
LAGGEDNP and CONTEMPNP are high (>.5)
and there is a strong (.6) positive correlation
between LEV and SIZE. Thus multicollinearity
may be an issue in interpreting the regressions.
However, among the continuous board
characteristics variables, the only correlation
greater than .3 is a negative correlation of .35
between PROPIND and EXECSHAR.
4.1 Earnings management and board
characteristics in period t
Panel A of Table 2 presents the results of
testing, on the sample of 84 firms for which
data was available, H1 that board characteristics
associated with strong corporate governance
will mitigate opportunistic negative earnings
management behaviour by incoming CEOs in
the period of CEO appointment. The model is
strongly significant (F statistic significant at 1%
level) and has high explanatory power (R
2
=
0.372). The intercept term (?) captures the
unexpected accruals where there is no CEO
duality. In this case, unexpected accruals are not
significantly different from zero. This is not
consistent with prior Australian research (Wells,
2002, Godfrey et al., 2003). Relative to this base
case, none of the board characteristic variables
are significant. Given our sample consists
entirely of companies that have changed CEO,
it is arguable that all boards included in the
sample have performed well in carrying out
UA
t+1

D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
0
:
5
9

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
ACCOUNTING RESEARCH JOURNAL  VOLUME 19 NO 2 (2006) 

86

Table 1
Descriptive Statistics for Eighty-seven Australian Firms that Changed CEO
in the Period 1992-1999
Year of CEO change (t)
RESIGN SIZE ($m) UA CFO CONTEMPNP LAGGEDNP
Mean 6513m -0.0388
*
0.0685 -0.0071 0.0187
Median 801m -0.0204 0.0619 0.0244 0.0461
RETIRE
Mean 18183m -0.0062
**
0.0692 0.0257 0.0585
Median 1350m -0.0021 0.0652 0.0306 0.0472
Year after CEO change (t+1)
RESIGN SIZE ($m) UA CFO CONTEMPNP LAGGEDNP
Mean 7268m 0.0240
*
0.0579 0.0034 -0.0071
Median 833m 0.0063 0.0637 0.0393 0.0244
RETIRE
Mean 18507m 0.0029
**
0.0869 0.0494 0.0257
Median 1527m 0.0066 0.0723 0.0372 0.0306
where SIZE is the amount of total assets as at the end of the reporting period, UA is unexpected accruals
(scaled over total assets at beginning), CFO is cash flow from operations in the current year divided by the
beginning of period total assets, CONTEMPNP is the net profit after tax reported in the current period
divided by the beginning of period total assets, LAGGEDNP is the net profit after tax in the previous
reporting period divided by the beginning of previous period total assets,
* Using a t-test, mean UA for CEO resignations in period t is significantly different from zero at the 1%
level. In period t+1 mean UA is significantly different from zero at the 10% level. Using a matched pairs t-
test, the difference between UA in period t and t+1 for CEO resignations is significant at the 1% level.
** Using a t-test, UA for CEO retirements in periods t and t+1 is not significantly different from zero.
Using a matched pairs t-test, the difference between UA in period t and t+1 for CEO retirements is not
significant.

Table 2
Earnings Management and Corporate Governance in Period t
Panel A of this table reports the results of running the following regression model for the period of CEO
change:
t t
t t t t t t t
EXECSHARE DUAL
PROPIND BOARDSIZE LAGGEDNP CONTEMPNP SIZE LEV CFO UA
9 8
7 6 5 4 3 2 1
? ?
? ? ? ? ? ? ? ?
+ +
+ + + + + + + =

where UA is unexpected accruals divided by the beginning of period total assets, CFO is cash flow from
operations in the current year divided by the beginning of period total assets, LEV is the amount of interest-
bearing debt divided by year-end total assets, SIZE is the amount of total assets as at the end of the
reporting period, CONTEMPNP is the net profit after tax reported in the current period divided by the
beginning of period total assets, LAGGEDNP is the net profit after tax in the previous reporting period
divided by the beginning of previous period total assets, BOARDSIZE is the number of members on the
company’s board of directors, as disclosed in the company’s annual report, PROPIND is the number of
directors classified as independent based on the Investment and Financial Services Association definition,
divided by board size, DUAL is a dummy variable set to 1 if there was CEO duality, zero otherwise,
EXECSHAR is the number of shares held by executive and affiliated directors, divided by the total number
of shares outstanding.
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
0
:
5
9

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
The Effects of Board Characteristics on Earnings Management around Australian CEO Changes

87

Panel A: Full Sample
?
1
?
2
?

3
?

4
?
5
?
6
?

7
?

8
?
9
?

Adj. R
2
F Stat
n=84
(t stats)
-0.040
(-0.823)
-0.386
(-1.885)
*
-0.019
(-0.407)
-0.000
(-1.022)
0.776
(7.322)
***
-0.401
(-1.917)
*
0.004
(1.405)
0.043
(1.333)
0.013
(0.395)
0.018
(0.477)
0.372 6.472
***

Panel B of this table reports the results of running the following regression that splits the regression model
in Panel A into RETIRE and RESIGN sub-samples:
RETIRE EXECSHARE DUAL PROPIND BOARDSIZE
LAGGEDNP CONTEMPNP SIZE LEV CFO RETIRE RESIGN EXECSHARE DUAL
PROPIND BOARDSIZE LAGGEDNP CONTEMNP SIZE LEV CFO RESIGN UA
t t t t
t t t t t t
t t t t t t t
* )
( * )
(
18 17 16 15
14 13 12 11 10 9 8
7 6 5 4 3 2 1
? ? ? ?
? ? ? ? ? ? ?
? ? ? ? ? ? ?
+ + + +
+ + + + + + + +
+ + + + + + + =

where UA is unexpected accruals divided by the beginning of period total assets, RESIGN is a dummy
variable set to 1 where the CEO change is not planned or orderly, RETIRE is (1-RESIGN), CFO is cash
flow from operations in the current year divided by the beginning of period total assets, LEV is the amount
of interest-bearing debt divided by year-end total assets, SIZE is the amount of total assets as at the end of
the reporting period, CONTEMPNP is the net profit after tax reported in the current period divided by the
beginning of period total assets, LAGGEDNP is the net profit after tax in the previous reporting period
divided by the beginning of previous period total assets, BOARDSIZE is the number of members on the
company’s board of directors, as disclosed in the company’s annual report, PROPIND is the number of
directors classified as independent based on the Investment and Financial Services Association definition,
divided by board size, DUAL is a dummy variable set to 1 if there was CEO duality, zero otherwise,
EXECSHAR is the number of shares held by executive and affiliated directors, divided by the total number
of shares outstanding.

Panel B: Resign/Retire
RESIGN
1
?

2
?

3
?

4
?

5
?

6
?

7
?

8
?

9
?

Adj. R
2
F Stat
n=84
(t stat)
-0.113
(-2.634)
**

-0.145
(-0.556)
-0.050
(-0.829)
0.000
(0.165)
0.621
(3.880)
***

-0.216
(-0.614)
0.008
(2.547)
**
0.090
(2.244)
**
-0.013
(-0.258)
0.032
(0.501)
RETIRE
10
?

11
?

12
?

13
?

14
?

15
?

16
?

17
?

18
?

0.524
136.9
(0.0000)
0.119
(4.616)
***

-0.757
(-5.585)
***
-0.070
(-3.390)
***
-0.000
(-0.762)
0.875
(3.274)
***

-0.631
(-14.68)
***
-0.004
(-1.312)
0.000
(0.015)
0.006
(0.268)
0.012
(0.517)

Newey-West HAC standard errors and covariance
* significant at 10% level
** significant at 5% level
*** significant at 1% level

their function of monitoring the CEO and,
where necessary, removing him/her from office.
This may also suggest they are effective in
identifying and curbing earnings management.
Of the control variables, CFO is negative but
significant only at the 10% level (Dechow et al.
1995, Guay et al. 1996), the leverage variable
(LEV) and size (SIZE) are not significant. As
expected, the two performance variables,
exhibit opposite relationships with UA.
Contemporaneous profitability (CONTEMPNP)
is strongly positively related to UA. That is,
when current profits are high, current
unexpected accruals are also high (and vice
versa). For lagged profitability (LAGGEDNP)
the opposite relationship holds but only at the
10% level. That is, high prior period profit is
associated with low unexpected accruals in the
current period. This may reflect the reversal of
current accruals in periods following high
profitability.
16

We then use equation (4) to test whether the
CEO resignation and CEO retirement categories
show any significant differences in period t.

16 For Panel A of Table 2 all variance inflation factors
(VIF) were well below 10 indicating that multicollinearity
was not a major problem.
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
0
:
5
9

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
ACCOUNTING RESEARCH JOURNAL  VOLUME 19 NO 2 (2006) 

88

Results are reported in panel B of Table 2. The
model is significant (0.000 level) and the R
2

increases from 0.372 to 0.524. The RESIGN
term shows that where the CEO is not chairman
and all continuous variables are zero, in the
period of CEO change, unexpected accruals are
reliably negative (5% level). This is consistent
with univariate the tests reported in Table 1,
with prior empirical results and with the theory
developed earlier suggesting that, following a
CEO resignation, incoming CEOs have
incentives to engage in negative earnings
management. Relative to the base case reflected
in the intercept (significant negative unexpected
accruals), several board characteristics are
positive and significant. Board size shows a
significant (5% level) positive association with
unexpected accruals
17
. That is, adding one
additional member to the Board reduces the
negative unexpected accruals by .008. This
result is consistent with Xie et al. (2002) and
Chtourou et al. (2001) in other contexts. Where
the CEO resigns, larger boards seem to
constrain opportunistic earnings management
behaviour by incoming CEOs, providing
support for H1 in relation to this board
characteristic. Also positive and significant at
the 5% level is the PROPIND variable. That is,
a higher proportion of independent directors on
the board is associated with less negative
unexpected accruals in the period of a CEO
resignation. Again, a higher proportion of
independent directors on the board appear to
constrain earnings management by incoming
CEOs. This result is consistent with Xie et al.
(2002) and Peasnell et al. (2005) in other
contexts and provides further support for H1 in
relation to this board characteristic. Of the
control variables, only CONTEMPNP is
significant (1% level). The insignificant result
for CFO and LAGGEDNP indicates that firms
experiencing a CEO resignation have a different
relationship between UA and these variables
than do the overall sample.

17 The board size variable is the number of board members.
We also constructed a relative board size measure by
dividing the number of board members in each sample
company in each year by the overall median board size
for that year. Results using this relative board size
measure were essentially the same as those reported in
the paper.
For the retire category in period t, the
RETIRE term shows that in a CEO retirement,
where the CEO is not chairman and all
continuous variables are zero, unexpected
accruals in period t are reliably positive (1%
level). This result is contrary to the univariate
tests reported in Table 1 and to incentives
normally attributed to incoming CEOs to ‘clear
the decks’. We attribute this result to the
planned nature of the CEO retirement which
reduces incentives for the outgoing CEO to
boost earnings in period t-1, and reduces
incentives for the incoming CEO to engage in
write offs.
18
All of the board characteristics
variables are insignificant. Accordingly, there is
no evidence that board characteristics associated
with strong corporate governance have any
effect on the positive unexpected accruals found
in the period of a CEO retirement.
19

The results for the control variables are also
of interest. CONTEMPNP is again positive and
significant, but CFO and LAGGEDNP, that
were both insignificant for CEO resignations,
are now negative and significant at the 1%
level. This suggests that underlying cash flows
(CFO) and accrual reversals (indicated by
LAGGEDNP) are major drivers of unexpected
accruals in the period of a CEO retirement. A
further interesting result is the significant
negative coefficient on LEV. This result is
consistent with the view that the greater
monitoring associated with higher debt levels
reduce unexpected accruals in the period of
CEO retirements.
Overall, there is evidence that in the case of
CEO resignations, there is negative earnings
management in period t and that larger boards
and a higher proportion of independent directors
are more effective in limiting such ‘big bath’
behaviour.

18 The considerable differences between the results for
resignations and retirements provide support for treating
the retirement category as a ‘control’ group where the
incentives normally attributed to incoming CEOs are not
apparent. These results also provide some support for the
accuracy of our classification of CEO changes into
resignations and retirements.
19 The VIF for BOARDSIZE*RETIRE in equation (4) in
period t was above 10. The equation was re-run dropping
BOARDSIZE*RETIRE , with no substantive effect on the
results.
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
0
:
5
9

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
The Effects of Board Characteristics on Earnings Management around Australian CEO Changes

89

Table 3
Earnings Management and Corporate Governance in Period t+1
Panel A of this table reports the results of running the following regression model for the period after CEO
change:
1 9 1 8
1 7 1 6 5 1 4 1 3 1 2 1 1 1

+ +
+ + + + + + +
+ +
+ + + + + + + =
t t
t t t t t t t
EXECSHARE DUAL
PROPIND BOARDSIZE LAGGEDNP CONTEMPNP SIZE LEV CFO UA
? ?
? ? ? ? ? ? ? ?
where UA is unexpected accruals divided by the beginning of period total assets, CFO is cash flow from
operations in the current year divided by the beginning of period total assets, LEV is the amount of interest-
bearing debt divided by year-end total assets, SIZE is the amount of total assets as at the end of the
reporting period, CONTEMPNP is the net profit after tax reported in the current period divided by the
beginning of period total assets, LAGGEDNP is the net profit after tax in the previous reporting period
divided by the beginning of previous period total assets, BOARDSIZE is the number of members on the
company’s board of directors, as disclosed in the company’s annual report, PROPIND is the number of
directors classified as independent based on the Investment and Financial Services Association definition,
divided by board size, DUAL is a dummy variable set to 1 if there was CEO duality, zero otherwise,
EXECSHAR is the number of shares held by executive and affiliated directors, divided by the total number
of shares outstanding.

Panel A: Full Sample
?
1
?

2
?
3
?

4
?
5
?

6
?

7
?

8
?

9
?
Adj. R
2
F Stat
n=86 0.083
(1.613)
-0.525
(-4.131)
***
-0.021
(-0.429)
-0.000
(-0.773)
0.813
(5.512)
***

-0.631
(-3.753)
***
-0.002
(-0.497)
-0.012
(-0.263)
0.030
(0.687)
-0.116
(-2.952)
***

0.406 7.454
***

Panel B of this table reports the results of running the following regression that splits the regression model
in Panel A into RETIRE and RESIGN sub-samples:
RETIRE EXECSHARE DUAL PROPIND BOARDSIZE
LAGGEDNP CONTEMPNP SIZE LEV CFO RETIRE RESIGN EXECSHARE DUAL
PROPIND BOARDSIZE LAGGEDNP CONTEMNP SIZE LEV CFO RESIGN UA
t t t t
t t t t t t
t t t t t t t
* )
( * )
(
1 18 1 17 1 16 1 15
14 1 13 1 12 1 1 10 1 9 1 8
1 7 1 6 5 1 4 1 3 1 2 1 1 1
+ + + +
+ + + + + +
+ + + + + + +
+ + + +
+ + + + + + + +
+ + + + + + + =
? ? ? ?
? ? ? ? ? ? ?
? ? ? ? ? ? ?
where UA is unexpected accruals divided by the beginning of period total assets, RESIGN is a dummy
variable set to 1 where the CEO change is not planned or orderly, RETIRE is (1-RESIGN), CFO is cash
flow from operations in the current year divided by the beginning of period total assets, LEV is the amount
of interest-bearing debt divided by year-end total assets, SIZE is the amount of total assets as at the end of
the reporting period, CONTEMPNP is the net profit after tax reported in the current period divided by the
beginning of period total assets, LAGGEDNP is the net profit after tax in the previous reporting period
divided by the beginning of previous period total assets, BOARDSIZE is the number of members on the
company’s board of directors, as disclosed in the company’s annual report, PROPIND is the number of
directors classified as independent based on the Investment and Financial Services Association definition,
divided by board size, DUAL is a dummy variable set to 1 if there was CEO duality, zero otherwise,
EXECSHAR is the number of shares held by executive and affiliated directors, divided by the total number
of shares outstanding.

Panel B: Resign/Retire
RESIGN
1
?

2
?

3
?

4
?

5
?

6
?

7
?

8
?

9
?

Adj.R
2
F Stat
n=86
(t stat)
0.103
(1.355)
-0.551
(-2.913)
***
-0.019
(-0.268)
-0.000
(-0.504)
0.817
(5.693)
***
-0.704
(-4.727)
***
-0.004
(-0.739)
-0.006
(-0.089)
-0.004
(-0.083)
-0.123
(-2.030)
**

RETIRE
10
?

11
?

12
?

13
?

14
?

15
?

16
?

17
?

18
?

0.364
10.935
(0.0000)
0.108
(2.128)
**

-0.626
(-2.752)
***
-0.053
(-0.799)
0.000
(0.694)
0.382
(0.984)
0.069
(0.164)
0.002
(0.515)
-0.111
(-1.174)
0.125
(1.301)
-0.130
(-2.146)
**

Newey-West HAC standard errors and covariance
*significant at 10% level
** significant at 5% level
*** significant at 1% level
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
0
:
5
9

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
ACCOUNTING RESEARCH JOURNAL  VOLUME 19 NO 2 (2006) 

90

4.2 Earnings management and board
characteristics in period t+1
Table 3 presents the results of testing equation (1)
in period t+1 on the sample of 86 firms for which
data was available. Panel A presents results for
the full sample and shows that the model is
significant at the 1% level. The positive, but
insignificant intercept term indicates that where
there is no CEO duality, unexpected accruals in
period t+1 are not reliably different from zero.
Compared to this base case, the board
characteristic variable EXECSHAR is negative
and significant (1% level). Higher shareholdings
by executive and affiliated directors are associated
with lower unexpected accruals in period t+1. As
unexpected accruals are not reliably positive in
period t+1, it is not possible to interpret this as
support for H2. Of the control variables,
CONTEMPNP is positive and significant (.1%
level) and CFO and LAGGEDNP are negative
and significant (1% level). The significance of
CFO and LAGGEDNP is greater than that
reported in respect of period t.
20

Again, we use equation (4) to test whether
there are any significant differences between the
RESIGN and RETIRE categories in period t+1.
Results are shown in Panel B of Table 3. The
model overall is significant (1% level). The
RESIGN term, reflecting unexpected accruals
when there is no CEO duality and all continuous
variables are set to zero, is positive but not
significant. Of the board characteristics variables,
EXECSHAR is negative and significant at the 5%
level. Higher shareholdings by executive and
affiliated directors have a negative effect on
unexpected accruals, however, as unexpected
accruals are not reliably positive, this cannot be
interpreted as support for H2. Of the control
variables, CONTEMNP is positive and strongly
significant and CFO and LAGGEDNP are both
negative and strongly significant. The latter
results contrast with those found for CEO
resignations in period t, where neither variable
was significant. This suggests that accrual
reversals might be affecting the period t+1 results.
A further reason for controlling for accrual
reversals is that in the context of CEO change, it
is argued that the direction of earnings
management reverses between period t and t+1.

20 The VIFs for all variables in equation (1) for period t+1
were well below 10, indicating that multicollinearity was
not a major problem
Thus controlling for lagged unexpected accruals
is necessary to ensure that short-term reversal of
current accruals is not driving the results.
Therefore, we added a variable for lagged
unexpected accruals (LAGGEDUA) to equation
(4) and re-ran the period t+1 regression. Results
are shown in Table 4 for the sample of 86 firms
with data available.
In Table 4, the model is significant at the 1%
level and the R
2
has increased slightly over that
reported in Table 3. For the RESIGN category,
the lagged unexpected accruals variable
(LAGGEDUA) is not significant indicating that
unexpected accrual reversals are not driving the
resign sub-sample results. The other results
remain largely as before, except that the
coefficient on EXECSHAR is no longer
significant. After incorporating lagged unexpected
accruals, there is no evidence of positive earnings
management in the period following a CEO
resignation and no indication that board
characteristics associated with strong corporate
governance have any effect on unexpected
accruals.
Returning to Panel B of Table 3, for the retire
category in period t+1, the RETIRE term,
reflecting unexpected accruals when the CEO
retires, there is no CEO duality and all other
continuous variables are zero, is positive and
significant at the 5% level, showing evidence of
positive unexpected accruals in the period
following a CEO retirement. Of the board
characteristics variables, EXECSHAR is negative
and significant at the 5% level, indicating that
higher shareholding by executives and affiliated
directors may act to reduce positive earnings
management in periods following a CEO
retirement. The results for the control variables
within the RETIRE category are also interesting.
In all prior regressions CONTEMPNP has been
positive and significant at the 1% level or better.
In this regression CONTEMPNP is insignificant,
as is LAGGEDNP. Only CFO retains its
significance.
21

21 The VIFs for BOARDSIZE*RETIRE and
PROPIND*RETIRE in Panel B of Table 3 were above 10
indicating a problem with multicollinearity. We dropped
BOARDSIZE*RETIRE with no substantive effect on the
results. When we dropped PROPIND*RETIRE the
coefficient on DUAL*RETIRE became positive and
significant at the 1% level. There were no other substantive
changes to the results.
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
0
:
5
9

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
The Effects of Board Characteristics on Earnings Management around Australian CEO Changes

91

Table 4
Effect of Lagged Unexpected Accruals on Results in Period t+1
This table reports the results of running the following regression model:
RETIRE LAGGEDUA
EXECSHARE DUAL PROPIND BOARDSIZE LAGGEDNP CONTEMPNP
SIZE LEV CFO RETIRE RESIGN LAGGEDUA EXECSHARE DUAL
PROPIND BOARDSIZE LAGGEDNP CONTEMNP SIZE LEV CFO RESIGN UA
t t t t t
t t t t t
t t t t t t t
* )
( * )
(
20
1 19 1 18 1 17 1 16 15 1 14
1 13 1 12 1 11 10 1 9 1 8
1 7 1 6 5 1 4 1 3 1 2 1 1 1
?
? ? ? ? ? ?
? ? ? ? ? ?
? ? ? ? ? ? ?
+
+ + + + + +
+ + + + + + +
+ + + + + + + =
+ + + + +
+ + + + +
+ + + + + + +
where UA is unexpected accruals divided by the beginning of period total assets, RESIGN is a dummy
variable set to 1 where the CEO change is not planned or orderly, RETIRE is (1-RESIGN), CFO is cash
flow from operations in the current year divided by the beginning of period total assets, LEV is the amount
of interest-bearing debt divided by year-end total assets, SIZE is the amount of total assets as at the end of
the reporting period, CONTEMPNP is the net profit after tax reported in the current period divided by the
beginning of period total assets, LAGGEDNP is the net profit after tax in the previous reporting period
divided by the beginning of previous period total assets, BOARDSIZE is the number of members on the
company’s board of directors, as disclosed in the company’s annual report, PROPIND is the number of
directors classified as independent based on the Investment and Financial Services Association definition,
divided by board size, DUAL is a dummy variable set to 1 if there was CEO duality, zero otherwise,
EXECSHAR is the number of shares held by executive and affiliated directors, divided by the total number
of shares outstanding, LAGGEDUA is the unexpected accruals from the period of CEO change.

Panel A: Resign/Retire
RESIGN
1
?

2
?

3
?

4
?

5
?

6
?

7
?

8
?

9
?

10
?

Adj. R
2
F Stat
n=85
(t stat)
0.076
(0.981)
-0.514
(-2.369)
**
-0.031
(-0.470)
-0.000
(-0.326)
0.767
(6.134)
***
-0.567
(-3.284)
***
-0.002
(-0.462)
0.005
(0.079)
-0.008
(-0.157)
-0.109
(-1.636)
-0.261
-(1.134)
RETIRE
11
?

12
?

13
?

14
?

15
?

16
?

17
?

18
?

19
?

20
?

0.425
10.201
(0.0000)
0.105
(2.598)
**

-0.701
(-3.796)
***
-0.083
(-1.883)
*
0.000
(1.765)
*
0.260
(0.662)
-0.195
(-0.920)
0.003
(0.702)
-0.093
(-1.393)
0.193
(2.610)
**
-0.068
(-1.851)
*

-0.361
(-3.226)
***
Newey-West HAC standard errors and covariance
*significant at 10% level
** significant at 5% level
*** significant at 1% leve

To further investigate these results, we added
LAGGEDUA to equation (4) and re-ran the
results for the retirement category. Results in
Table 4 show that LAGGEDUA is negative and
significant (1% level) for the retire category.
The negative coefficient on LAGGEDUA
indicates that there is significant reversal of
unexpected accruals from period t to period t+1
for CEO retirements. When LAGGEDUA is
included, evidence of positive unexpected
accruals for CEO retirements in period t+1
remains significant. The RETIRE term,
reflecting unexpected accruals when the CEO
retires, there is no CEO duality and all
continuous variables are zero, remains positive
and significant at the 5% level. Of the board
characteristics variables, EXECSHAR is now
significant only at the 10% level, but DUAL is
now positive and significant at the 5% level.
That is, when the CEO is also Chairman of the
Board, unexpected accruals following a CEO
retirement (already significantly positive) are
subject to a further significant increase.
22

5. Conclusions
The objective of this paper is to empirically test
whether certain board characteristics associated
with strong corporate governance are effective

22 The VIFs for BOARDSIZE*RETIRE and
PROPIND*RETIRE in Table 4 were above 10 indicating
a problem with multicollinearity. We dropped
BOARDSIZE*RETIRE and PROPIND*RETIRE with no
substantive effect on the results.
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
0
:
5
9

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
ACCOUNTING RESEARCH JOURNAL  VOLUME 19 NO 2 (2006) 

92

in controlling earnings management in the
financial reports of Australian firms that
changed CEOs.
The results show that in the year of CEO
change, negative earnings management is only
observed where the CEO resigns. In the case of
CEO retirements, unexpected accruals in period
t are positive. In the context of CEO
resignations, larger boards and a higher
proportion of independent directors significantly
reduce the negative unexpected accruals
suggesting that they provide more effective
monitoring in this context. These results are
consistent with prior research in other contexts
(Xie, et al. 2002, Peasnell et al., 2005) and
provide support for H1 in relation to these two
board characteristics. For CEO retirements,
none of the board characteristics have any
significant relationship with the positive
unexpected accruals found in the period of CEO
retirement. In the period after CEO change, we
find no evidence of positive earnings
management for CEO resignations, even after
including lagged unexpected accruals from
period t. For CEO retirements, there is evidence
of significant positive unexpected accruals in
period t+1. This relationship remains significant
when lagged unexpected accruals from period t
are included in the regressions. In this
specification, CEO duality (that is, where the
CEO is also board chairman) actually
exacerbates the positive unexpected accruals in
periods following CEO retirements.
Examining the impact of board
characteristics associated with strong corporate
governance on earnings management in this
context allows us to provide additional evidence
that might help to overcome an inherent
problem with much of the prior context-specific
earnings management research. Many
researchers have been unable to distinguish
between management opportunism and accurate
reporting of firms’ underlying economic
performance (requiring recognition of necessary
write downs) as explanations for the results
obtained (see Dechow and Sloan, 1991;
Pourciau, 1993). After controlling for current
and lagged profitability we find evidence of
negative earnings management in the period of
CEO resignation. We further find that larger
boards of directors and a higher proportion of
independent directors mitigate the negative
unexpected accruals found in CEO resignations.
This suggests that our evidence is consistent
with opportunism as the explanation, as we
would not expect that accurate recognition of
necessary write-offs would be significantly
reduced by board characteristics associated with
strong corporate governance.
There are also several limitations inherent in
the study. The first is the method used to
measure unexpected accruals. Measurement
error is possible as the model used may lack
sufficient power to detect earnings
management. Similarly, as unexpected accruals
vary with firm performance, and firm
performance varies with the probability of CEO
change, then the ensuing tests will be biased.
However, controlling for contemporaneous and
lagged profitability (and lagged unexpected
accruals) should help to overcome this
problem. The second limitation is that we
have not explicitly considered the
abnormal/extraordinary classification as a
vehicle for earnings management.
Notwithstanding these limitations, this paper
is the first to examine the effects of any
corporate governance mechanisms on earnings
management in the context of CEO change and
also adds to the very limited research in
Australia on the association between corporate
governance and earnings management. Further
it provides empirical evidence on the efficacy of
a number of the recent ASX Corporate
Governance Council (2003) recommendations.
Thus, this study should also be of interest to
regulators.
References
ASX Corporate Governance Council, (2003), Principles of
Good Corporate Governance and Best Practice
Recommendations.
Beasley, M. (1996), ‘An Empirical Analysis of the Relation
between the Board of Directors Composition and
Financial Statement Fraud’, The Accounting Review,
Vol. 71, pp. 443–465.
Brickley, J. A. Jarrell G. and Coles, J. L. (1996), ‘Checks
and Balances in the Boardroom?’, Chief Executive,
June, pp. 66–67
Bushman, R. and Smith, A. (2001),. ‘Financial Accounting
Information and Corporate Governance’, Journal of
Accounting and Economics, Vol. 32, pp237–333.
Chtourou, S., Bedard J. and Courteau, L. (2001), ‘Corporate
Governance and Earnings Management’, SSRN
Working Paper Series.
Conyon, M. and A. Florou, (2002), ‘Top executive
dismissal, ownership and corporate performance’,
Accounting and Business Research, Vol. 32,
pp. 209-226.
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
0
:
5
9

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
The Effects of Board Characteristics on Earnings Management around Australian CEO Changes

93

Dalton, D.R. Daily, C.M. Johnson J.L and Ellstrand, A.E.
(1999), ‘Number of Directors and Financial
Performance: A Meta-analysis’, Academy of
Management Journal, Vol. 42, pp. 674–687.
DeAngelo, L.E, (1988), ‘Managerial Competition,
Information Costs, and Corporate Governance: The Use
of Accounting Performance Measures in Proxy
Contents’, Journal of Accounting and Economics, Vol.
10, pp. 3–36.
Dechow, P. Sloan R. and Sweeney, A. (1995), ‘Detecting
Earnings Management’, The Accounting Review,
Vol. 70, pp. 193–225.
Dechow, P. Sloan R. and Sweeney, A (1996), ‘Causes and
Consequences of Earnings Manipulation: an Analysis of
Firms Subject to Enforcement Actions by the SEC’,
Contemporary Accounting Research, Vol. 13, pp. 1–36.
Eddey, P. and Taylor, S. (1999), ‘Directors’
Recommendations on Takeover Bids and the
Management of Earnings: Evidence from Australian
Takeovers’, Abacus, Vol. 35, 1, pp. 29–45.
Eisenberg,T. Sundgren S. and Wells, M. (1998), ‘Large
Board Size and Decreasing Firm Value in Small Firms’,
Journal of Financial Economics, Vol. 48, 1, pp. 35–54.
Elliott, J. and Shaw, W. (1988), ‘Write-offs as Accounting
Procedures to Manage Perceptions’, Journal of
Accounting Research, Vol. 26, Suppl. pp. 91–120.
Fama, E.F. (1980), ‘Agency Problems and the Theory of the
Firm’. Journal of Political Economy, Vol. 88,
pp. 288–307.
Fama, E. and Jensen, M. (1983), ‘Separation of Ownership
and Control’, Journal of Law and Economics, Vol. 26,
pp. 301–325.
Gilson, S.C. (1990), ‘Bankruptcy, Boards, Banks, and
Blockholders: Evidence on Changes in Corporate
Ownership and Control when Firms Default’, Journal
of Financial Economics, Vol. 27, 2, pp. 355–387.
Godfrey, J. Mather P. and Ramsay, A. (2003), ‘Earnings and
Impression Management in Financial Reports: the Case
of CEO Changes’, Abacus, Vol. 39, pp. 95–123.
Guay, W.R. Kothari S.P. and. Watts, R.L (1996), ‘A Market-
Based Evaluation of Discretionary Accrual Models’,
Journal of Accounting Research, Vol. 34, pp 83–106.
Hart, O. (1995), ‘Corporate Governance: Some Theory and
Implications’, The Economic Journal, Vol. 105,
pp.678–689.
Hermalin, B.E. and Weisbach, M.S. (2003), ‘Board of
Directors as an Endogenously Determined Institution: A
Survey of the Economic Literature’, Federal Reserve
Bank of New York Economic Policy Review, April,
pp. 7–26.
Jensen, M.C., (1993), ‘The Modern Industrial Revolution,
Exit and the Failure of Internal Control Systems’,
Journal of Finance, Vol. 48, pp. 831–880.
Jensen, M.C. and Meckling, W.H. (1976), ‘Theory of the
Firm: Managerial Behaviour, Agency Costs and
Ownership Structure’, Journal of Financial Economics,
Vol. 3, 305–360.
Klein,A. (2002), ‘Audit Committee, Board of Director
Characteristics and Earnings Management’, Journal of
Accounting and Economics, Vol. 33, pp. 375–400.
Lau, J, P. Sinnadurai, and S. Wright, (2005), ‘Corporate
Governance and CEO Dismissal in Australia’, Working
paper, Macquarie University.
Mace, M.L. (1971), Directors, Myths and Reality, Boston,
MA: Harvard Business School Press.
Peasnell, K. V., Pope, P. F. and Young, S., (2005) ‘Board
monitoring and earnings management: Do outside
directors influence abnormal accruals?’, Journal of
Business Finance and Accounting, 32, 1311–1346.
Pourciau, S. (1993), ‘Earnings Management and Nonroutine
Executive Changes’, Journal of Accounting and
Economics, Vol.16, pp. 317–336.
Sloan, R., (2001), ‘Financial Accounting and Corporate
Governance: a Discussion’, Journal of Accounting and
Economics, Vol. 32, pp. 335–347.
Sarbanes-Oxley Act, (2002), One Hundred Seventh
Congress of the United States of America, Second
Session.
Strong, J. and Meyer, J. (1987), ‘Asset Writedowns:
Managerial Incentives and Security Returns’, Journal of
Finance, Vol. 42, pp. 643–661.
Vancil, R.F. (1987) Passing the Baton: Managing the
Process of CEO Succession, Boston, MA: Harvard
Business School Press.
Warner, G., R. Watts and K. Wruck, (1988), ‘Stock Prices
and Top Management Changes’, Journal of Financial
Economics, Vol. 20, pp. 461–492.
Weisbach, M. (1988), “Outside directors and CEO
turnover”, Journal of Financial Economics Vol. 20, pp.
413–460.
Wells, P. (2002), ‘Earnings Management Surrounding CEO
Changes’, Accounting and Finance, Vol. 42
pp.169–193.
Xie, B. Davidson W. and DaDalt, P. (2002), ‘Earnings
Management and Corporate Governance: the Role of
the Board and the Audit Committee’ Journal of
Corporate Finance, Vol. 9, pp. 295–316.
Yermack, D. (1996), ‘Higher Market Valuation of
Companies with a Small Board of Directors’, Journal of
Financial Economics, Vol. 40, pp. 185–211.
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
0
:
5
9

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
This article has been cited by:
1. Rusmin Rusmin, Emita W. Astami, Bambang Hartadi. 2014. The impact of surplus free cash
flow and audit quality on earnings management. Asian Review of Accounting 22:3, 217-232.
[Abstract] [Full Text] [PDF]
2. Howard Chan, Robert Faff, Arifur Khan, Paul Mather. 2013. Exploring the Moderating Role
of Growth Options on the Relation between Board Characteristics and Management Earnings
Forecasts. Corporate Governance: An International Review 21:10.1111/corg.2013.21.issue-4,
314-333. [CrossRef]
3. Annelies Stockmans, Nadine Lybaert, Wim Voordeckers. 2013. The conditional nature of board
characteristics in constraining earnings management in private family firms. Journal of Family
Business Strategy 4, 84-92. [CrossRef]
4. Mark Wilson. 2011. Earnings Management in Australian Corporations. Australian Accounting
Review 21:10.1111/auar.2011.21.issue-3, 205-221. [CrossRef]
5. Marion Hutchinson. 2009. Governance issues in accounting. Accounting Research Journal 22:2,
89-92. [Abstract] [Full Text] [PDF]
6. Sebahattin Demirkan, Harlan Platt. 2009. Financial status, corporate governance quality, and
the likelihood of managers using discretionary accruals. Accounting Research Journal 22:2,
93-117. [Abstract] [Full Text] [PDF]
7. James Lau, Philip Sinnadurai, Sue Wright. 2009. Corporate governance and chief executive
officer dismissal following poor performance: Australian evidence. Accounting & Finance
49:10.1111/acfi.2009.49.issue-1, 161-182. [CrossRef]
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
0
:
5
9

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)

doc_457616521.pdf