The current issue and full text archive of this journal is available at

www.emeraldinsight.com/0268-6902.htm

Shareholder
Shareholder rights, corporate rights
governance and earnings quality
The influence of institutional investors
767
Wei Jiang
California State University, Fullerton, California, USA, and Received 11 March 2009
Asokan Anandarajan Revised 23 April 2009
Accepted 24 May 2009
New Jersey Institute of Technology,
Newark, New Jersey, USA

Abstract
Purpose – The purpose of this paper is to examine the effect of shareholder rights on the quality of
reported earnings using a proxy for strength of shareholder rights. In the analysis, the influence of
institutional investors on shareholder rights is incorporated and their joint impact on earnings quality is
studied.
Design/methodology/approach – Alternative regression models with the level of discretionary
accrual (DA) as the dependent variables are estimated. To measure DA, a model developed by
DeChow et al. is used. Higher levels of DA imply lower quality of earnings. The independent variables
of interest are shareholder rights (measured by a modified Gomper’s index) and institutional
ownership (measured in three different ways discussed in the paper). A number of control variables,
which prior research indicates, that can influence earnings quality is also included.
Findings – It is found that stronger shareholder rights are associated with higher earnings quality.
However, when firms’ stocks are held predominantly by institutions with short investment horizons
(transient institutions), the role of shareholder rights in constraining aggressive and opportunistic
management of earnings is significantly diminished or rendered essentially ineffective.
Originality/value – This research adds to the understanding of how levels of institutional
ownership moderate the association between shareholder rights and earnings quality.
Keywords Earnings, Shareholders, Corporate governance
Paper type Research paper

1. Introduction
Shareholder rights reflect the ability of voting stockholders to exercise control over firm
assets, remove ineffective or opportunistic management, or effect ownership changes to
increase shareholder value. Traditional theory is of the view that lower shareholder
rights (weak external governance) generate information asymmetry between
shareholders and managers that leads to greater managerial incentives to reduce
transparency and manage earnings to increase their bonuses. Greater shareholder
rights, on the other hand, enables shareholders to implement corporate governance
mechanisms to monitor managers more meticulously. Much research has been done on
the implications of greater shareholder rights to companies. The main finding is that Managerial Auditing Journal
Vol. 24 No. 8, 2009
greater shareholder rights are associated with reduced agency risks (Diamond and pp. 767-791
Verrecchia, 1991; Shleifer and Vishny, 1997a) and improved firm performance (Gompers q Emerald Group Publishing Limited
0268-6902
et al., 2003). However, these studies do not directly address the association between DOI 10.1108/02686900910986402
MAJ shareholder rights regimes and the quality of reported earnings. As the first objective of
24,8 this study, we build on the existing literature on shareholder rights to investigate
whether vigilant and effective shareholder oversight reduces the incentives and
constrains the ability of managers to engage in opportunistic management of earnings.
A closely related issue to shareholder rights is the role of institutional ownership and
its influence on earnings quality as the participation of institutional investors become
768 increasingly important in the US financial markets. Sias and Starks (1998) noted that
large institutional ownership in equities increased from 24 per cent in 1980 to nearly 50
per cent by the end of 1994. As their influence grew, institutional investors abandoned
their traditional passive shareholder roles and became more active participants in the
governance of their corporate holdings. Although, it is widely recognized that
institutional participation has evolved into an integral part of shareholder rights
mechanisms, the role of institutional investors in influencing corporate performance and
earnings quality is still unclear. While there are arguments suggesting that the presence
of substantial shareholdings is associated with superior corporate performance and less
opportunistic self-serving behaviour (Shleifer and Vishny, 1997b; McConnell and
Servaes, 1990; Gadhoum, 2000), there are also studies suggesting institutional investors
are fixated on short-term performance even to the detriment of the long-term prosperity
of the firm (Porter, 1992; Demirag, 1998; Lang and McNichols, 1999). Therefore, as the
second objective of this study, we seek to examine the influence of different types of
institutional investors on the effectiveness of shareholder rights in constraining
earnings management. We categorize institutional investors as “transient” (short-term
investment horizon) and “nontransient” (long-term investment horizon) and examine
whether their varying investment horizon moderates the association between
shareholder rights and earnings quality.
Using the corporate governance index developed by Gompers et al. (2003) as a proxy for
the strength of shareholder rights, we find that strong shareholder rights improves the
reliability of financial information by reducing the ability of management to intentionally
manipulate accruals. However, we also find that when firms’ stocks are held
predominantly by institutions with short investment horizons, the role of shareholder
rights mechanisms in constraining aggressive and opportunistic management of earnings
is actually diminished or rendered essentially ineffective. Our results are consistent with
the view that institutional investors’ focusing on short-term earnings performance could
pressure management into boosting reported earnings through aggressive accounting.
Our study is motivated by two factors. First, while much research has been done on
the association between increased shareholder rights, agency costs and firm valuation,
the association between shareholder rights and the quality of earnings and the
moderating effect of institutional ownership on such a relationship have not been fully
tested empirically. Overall, research on the relation between earnings management and
shareholder rights is sparse and inconclusive[1]. Also, institutional investors are treated
as a homogeneous group in prior work on the relation between earnings quality and
institutional ownership. Thus, the results from this study contribute to our
understanding of the relation between shareholders rights, different groups of
institutional investors, and how they interact in affecting the quality of corporate
earnings. Another motivation for this study is the current effort by the Securities and
Exchange Commission (SEC) to increase shareholder rights. The SEC is of the view that
increased shareholder rights should increase corporate governance and enhance
transparency (Solomon and Lublin, 2004). The results of this study provide evidence Shareholder
to the SEC that the type of ownership should also be factored into the discussion when rights
considering steps to increase shareholder rights.
The paper proceeds as follows: Section 2 briefly reviews the literature on
shareholder rights and institutional ownership and develops our hypotheses. Section 3
discusses our research design and methodology. Section 4 provides empirical results,
and Section 5 contains a summary and conclusions. 769

2. Prior research and hypothesis development

2.1 Studies that examine the implications of shareholder rights on firms
Shareholder rights reflect the balance of power between shareholders and managers.
Voting shareholders have a motivation to make changes in management to increase firm
value. Managers, on the other hand, desire to “protect their turf” by restricting
shareholder rights. Managers use various measures to restrict shareholder rights,
including setting up defenses against takeover, requiring staggered terms for directors,
and requiring “super majority” voting requirements for approval of mergers and
ownership changes. These defensive measures along with other statutory and corporate
charter provisions are designed to limit shareholders’ participation in the governance
process and hence their ability to intervene and take disciplinary actions against them,
which in turn increases managers’ ability to protect and conceal their private control
benefits from outsiders through earnings management (Shleifer and Vishny, 1997a)[2].
Prior studies have examined the implications of shareholder rights for various
aspects of firm performance. Gompers et al. (2003) showed that firms with greater
shareholder empowerment are valued more highly by the market. More specifically,
higher levels of shareholder rights are associated with positive abnormal returns, higher
firm values, higher profits, higher sales growth and fewer acquisitions. Using a sample
of firms from 31 wealthy economies, La Porta et al. (2002) find evidence of higher
valuation for firms in countries with better protection of minority shareholders. These
empirical findings suggest that there are positive shareholder wealth implications for
firms with stronger shareholder rights. Prior research also suggests that weak corporate
governance mechanisms will lead to increased perceived risks of the firm, and thus
higher cost of capital (Jensen and Meckling, 1976; Diamond and Verrecchia, 1991).
Ashbaugh et al. (2004) provided empirical evidence that lower levels of corporate
governance as measured by the Gomper’s index were associated with higher cost of
capital. Cheng et al. (2006) also found that firms with stronger shareholder rights and
higher levels of disclosure had significantly lower levels of cost of capital relative to
firms with weaker shareholder rights and lower levels of disclosure.
We add to the shareholder rights literature by examining the association between
shareholder activism and the quality of reported earnings. Leuz et al. (2003) examined
this issue in an international setting and found that earnings management decreases in
investor protection. Baber et al. (2007) presented evidence that strong external
governance (fewer restrictions on shareholder rights) is associated with relatively
low probabilities of accounting restatement. We conjecture that vigilant and effective
shareholder oversight increases the investors’ ability to monitor and discipline managerial
actions, and thus reduced incentives for managers to engage in opportunistic financial
reporting. We test this hypothesis by performing a direct test of the relation between
MAJ the level of earnings management and strength of shareholder rights. Our hypothesis is
24,8 stated as follows:
H1. Higher levels of shareholder rights are associated with less accounting
discretion (higher earnings quality) (alternative).

2.2 The role of institutional investors

770 An important element of shareholder rights mechanisms is the role of institutional
investors in influencing managerial decisions. Shareholder activism by institutional
investors increased during the late 1980s and early 1990s (Gillan and Starks, 2000).
As institutions’ ownership increased, they became more actively involved in the
governance of their corporate holdings. The primary emphasis of activist shareholders
evolved into focusing on the poorly performing firms and pressuring the management of
such firms to improve performance to increase shareholder values. Gillan and Starks
(2000) document that proxy proposals sponsored by institutional investors receive
significantly more votes, suggesting that activism by institutional investors is more
effective relative to individual investors. In 1992, the SEC passed new rules allowing
shareholders to directly communicate with each other. With this change in
communication rules, institutional shareholders no longer needed to rely on more
expensive proxy proposals to communicate with other shareholders and began having
more direct negotiation with company management (Anand, 1993). This further
increased the strength and effectiveness of shareholder rights.
Prior research suggests that the ability of managers to opportunistically manage
reported earnings is constrained by the effectiveness of external monitoring by
institutional investors. Gillan and Starks (2000, p. 279) observe “institutions that hold
large equity positions in a company have been motivated to actively participate in the
company’s strategic direction”. Monks and Minow (1995) note that the investor
with the larger stake has stronger incentives to undertake monitoring activities,
as it is more likely that the large shareholder’s increased return from monitoring
is sufficient to cover the associated monitoring costs. Consistent with the view of
institutional investors monitoring and constraining the self-serving behaviour of
corporate managers, Chung et al. (2002) found that the presence of large institutional
shareholdings inhibited managers from managing earnings towards their targets.
Ajinkya et al. (2005) provided indirect but related evidence that firms with greater
institutional ownership are more likely to issue a forecast and these forecasts tend
to be more specific, accurate and less optimistically biased. Both Bushee (1998)
and Wahal and McConnell (1997) documented a positive relation between research
and development (R&D) expenditures (which increases long-term firm value) and
institutional ownership. Jiambalvo et al. (2002) presented evidence that stock prices
reflected more current-period information that is predictive of future period earnings as
percentage of institutional ownership increases. All these studies concluded that
institutional owners have a positive association with firm value by constraining
self-serving behaviour by managers.
However, there are also contrary arguments suggesting that institutional investors
can pressure corporate managers into making accounting decisions that boost
short-term earnings at the expense of long-term value (Jones, 1991; Porter, 1992; Laverty,
1996). According to Porter (1992), some institutional investors are “transient owners”
who are overly focused on current earnings. Intense competition among institutional
investors for client funds to invest and frequent performance evaluations engenders a Shareholder
quest for short-term performance (Graves and Waddock, 1990). Lang and McNichols rights
(1997) found evidence consistent with institutional investors trading based on earnings
surprises. Bushee (1998) finds that short-term institutional holdings create incentives for
managers to reduce investment in R&D to meet short-term earnings goals, while
long-term institutional holdings (non-transient owners) serve a monitoring role in
reducing pressures for myopic behaviour. In a related study, Bushee (2001) presents 771
evidence that transient institutions invested more heavily in firms with greater expected
near-term earnings and less heavily in firms with anticipated higher long-term earnings.
Prior research has provided little direct evidence on the effect of institutional
ownership on earnings management, and the results are far from conclusive. Also, the
results from prior studies (Chung et al., 2002; Ajinkya et al., 2005) must be interpreted
cautiously since they treat all institutional investors as a homogenous group and do
not distinguish between different types of investment institutions. Our study fills this
void. More importantly, we investigate the role of institutional ownership in the
broader setting of shareholder rights regimes. Specifically, we examine how
institutional ownership and shareholder rights interact in influencing corporate
managers’ earnings management behaviour. We conjecture that institutional investors
with short-term (long-term) horizons will diminish (strengthen) the effect of
shareholder rights in curbing earnings management. We present our H2 as follows:
H2. Institutional investor type moderates the association of shareholder rights
and quality of reported earnings, with transient (nontransient) owners
diminishing (strengthening) this association (alternative).

3. Research design
3.1 Discretionary accrual model
Discretionary accruals (DAs) reflect subjective accounting choices made by managers.
They are the most common metrics used to assess earnings management. In this study,
we estimate DAs based on the forward-looking model developed by DeChow et al.
(2003), where total accruals (defined as the difference between the firm’s earnings
before extraordinary items and operating cash flow – CFO) is modeled as a function of
various components of non DAs shown as below:

TAit ¼ a þ b1 ðð1 þ kÞDREVt 2 DARit Þ þ b2 PPEit þ b3 TAit21 þ b4 GR_Salesitþ1 þ 1it

where:
Ait2 1 ¼ total assets (Compustat data no. 6) for firm i in year t 2 1;
DREVit ¼ change in net revenues (Compustat data no. 12) from year t 2 1 to t;
DARit ¼ change in accounts receivable (Compustat data no. 2) from t 2 1 to t;
PPEit ¼ gross property plant and equipment (Compustat data no. 7) in year t;
k ¼ the slope coefficient from a regression of DARit on DREVit;
TAit2 1 ¼ firm i’s total accruals from year t 2 1, scaled by year t 2 2 total
assets;
MAJ GR_salesitþ 1 ¼ the change in sales from year t to t þ 1, divided by year t sales; and
24,8 1it ¼ error term for firm I in year t.
Unless otherwise specified, all variables are scaled by year t 2 1 assets. We estimate
the cross-sectional model for each two-digit Standard Industrial Classification (SIC)
industry with at least ten firms each year. Firms with accruals measure and
772 independent variables in the extreme one percent of their respective distributions were
deleted from the sample. This approach enables us to control for industry-wide
changes (DeFond and Jiambalvo, 1994). The abnormal accruals were computed as
the difference between the total accruals and the fitted normal accruals.
The forward-looking model makes three adjustments to the modified Jones model
(DeChow et al., 1995). First, rather than assuming all credit sales are discretionary, the
model estimates the “expected” portion of the increase in credit sales, as represented by
the slope coefficient k from the regression of DARit on DREVit. Hence, the models
subtract the full amount of the change and adds back the expected change (which is k
multiplied by the change in sales). Second, some proportion of total accruals is
assumed to be predictable based on last year’s total accruals. Thus, the lagged value of
total accruals (TAit2 1) is included to capture the predictable component. Third, the
modified Jones model classifies increases in inventory in anticipation of higher sales as
earnings management. DeChow et al. (2003) argued that an increase in inventory
balance is rational and included a measure of future sales growth, (GR_Sales), to
correct for misclassifications. In essence, the forward-looking model uses future period
data to estimate current period accruals. DeChow et al. (2003) provided empirical
evidence to prove that their model had higher explanatory power. We label this
measure as forward looking discretionary accruals (FLDA)[3].

3.2 Independent variables

3.2.1 Modified corporate governance index. Our measure of the level of shareholder
rights is the corporate governance index (G-INDEX) developed by Gompers et al.
(2003). They identify a total of 24 distinct corporate governance factors that either
restricts shareholder rights or increase managerial power[4]. The Gomper’s index
(G-INDEX) can be used to proxy for how easy or difficult it is for shareholders to make
changes in management ownership that potentially affect shareholder value. By
original construction, higher G-INDEX implies greater restriction on shareholder
rights and fewer restrictions on managerial power, and thus is associated with weaker
shareholder rights or corporate governance. To aid in the exposition of our analysis, we
modify the G-INDEX so that higher (lower) values of G-INDEX reflect higher (lower)
shareholder rights regimes. Specifically, we subtract the original G-INDEX from 24, the
total number of provisions. The modified G-INDEX (MG) has a possible range from
0 to 24, with higher index indicating stronger shareholder rights mechanisms.
3.2.2 Institutional ownership. We employ three alternative measures of institutional
ownership to examine the hypothesised moderating effect of different types of
institutional ownership on the effectiveness of shareholder rights. These will be
considered individually.
Measure 1: TIO and NTIO. Following Bushee (2001), we classify institutional
investors into two distinct groups by their investment behaviour: transient and
nontransient institutions. The grouping is obtained by taking into account two main
aspects of an institution’s investment behaviour: the level of portfolio diversification Shareholder
and the degree of portfolio turnover. Transient institutions are characterized as having rights
high-portfolio turnover and highly diversified portfolio holdings. These traits reflect
the fact that transient institutions tend to be short-term focused and are mainly
interested in pursuing short-term trading profits. TIO and NTIO are computed as the
percentage of total shares owned by transient and nontransient institutions,
respectively. A potential limitation with the percentage measures is they tend to 773
move in lockstep with the total ownership, thus making it difficult to disentangle the
effect of transient ownership from that of the nontransient ownership[5]. We address
this limitation by developing indicator variables that also allow us to examine the
incremental effect of institutional ownership.
Measure 2: TIOQ5 and NTIOQ5. Based on the distribution of percentage ownership
for each group of institutions, we rank firms into quintiles and create indicator
variables to measure the predominance of the two different types of institutional
ownership. In particular, TIOQ5 (NTIOQ5) is an indicator variable that equals one if
the per cent of shares held by transient (nontransient) institutions is in the top quintile
for all firms during the year and 0 otherwise. These measures are used to evaluate the
predominance of each type of institutional ownership for a firm relative to all other
firms in the same year[6].
Measure 3: PTIOQ5 and PNTIOQ5. One limitation with the use of TIOQ5 and
NTIOQ5 is that, by construction, they do not measure the relative predominance of one
type of ownership over the other type of ownership within the same firm. The third set
of measures is constructed in such a way that they allow us to compare the magnitudes
of the two types of institutional ownership both across firms and within the same firm.
We first calculate the shareholdings for each group of institutions as a proportion of
the aggregate institutional ownership, and then rank each group into quintiles based
on the distribution of proportional ownership for that group. Specifically, the indicator
variable PTIOQ5 (PNTIOQ5) is set to one if the firm is in the top quintile of
proportional ownership by transient (nontransient) institutions and in the bottom
quintile of proportional ownership by nontransient (transient) institutions, and zero
otherwise. By construction, a firm is considered to be dominated by:
.
transient institutions if PTIOQ5 equals one; and
.
nontransient institutions if PNTIOQ5 equals one[7].

3.2.3 Control variables. We include several control variables that could potentially
influence earnings quality based on our study of the earnings management literature.
Larger companies would have less incentive to manage earnings because they face
greater political costs and are more subject to scrutiny from financial analysts and
investors (Watts and Zimmerman, 1990). We measure firm size as the natural log of total
assets (LOGASSET). Summers and Sweeney (1998) note that unethical managers may
be induced to misstate financial statements when growth slows or reverses, in order to
maintain the appearance of consistent growth. We thus include market to book ratio
(MB) as the proxy for future profitability growth. We also control for CFO since firms
with strong CFO performance are less likely to manage accruals to boost earnings (Lobo
and Zhou, 2006; Becker et al., 1998). Financial leverage (LEV) may also be associated
with DAs as managers of highly leveraged firms have incentives to manipulate accruals
to avoid debt covenant violation (DeFond and Jiambalvo, 1994; Becker et al., 1998). Prior
MAJ studies demonstrate a lower association between returns and earnings for loss firms
24,8 (Hayn, 1995; Brown, 2001). Thus, we include in our model an indicator variable (LOSS)
that equals one if the firm reports negative income in any of the previous two years, and
zero otherwise. Palmrose (1999) reports that firms with high-litigation risk (LIT) are
more likely to be concerned about missing earnings benchmarks and, hence, may have
an incentive to manage earnings to achieve forecasted earnings targets. Hence, our
774 regression model includes an indicator variable for LIT. In addition, we control for
return on assets (ROA) because tests of earnings management may be misspecified
when DAs are correlated with firm performance (DeChow et al., 1995).
In addition to the firm-specific economic determinants, we include corporate
governance attributes documented to be linked to earnings quality (Beasley, 1996;
Klein, 2002). The board independence and audit committee independence are measured
by the percentage of independent directors on the board (BINDEP) and the audit
committee (ACINDEP), respectively. We also consider the size of board (BSIZE) and
size of audit committee (ACSIZE). A dummy variable indicating whether the chief
executive officer (CEO) is the board chair (CEOCHAIR) is included to control for the
CEO’s influence on the board. The literature research also suggests that Big 5 auditors
are less likely to allow earnings management than non-Big 5 auditors (Becker et al.,
1998; Francis et al., 1999). Thus, our analysis includes a Big 5 indicator variable (BIG5).
Finally, we include two-digit industry dummies and year dummies to control for
fixed industry and time effects. The definitions of the variables used in this study are
contained in Table I.

3.3 Model specifications

We test our hypotheses by estimating alternative ordinary least squares regression
models with the level of DAs as the dependent variable. The first set of analyses is
based on the MG and per cent institutional shareholdings (TIO and NTIO):
Model 1 : FLDA

¼ a1 þ b1 MG þ b2 TIO þ b3 NTIO þ b4 ðMG £ TIOÞ þ b5 ðMG £ NTIOÞ

X
þ bi ðControl variablesi Þ þ Year and industry dummies þ 1
i

where i ¼ 6; 5 . . . 17: All variables are as defined in Table I. According to H1, we

conjecture that stronger shareholder rights are associated with higher earnings
quality. We expect to observe a negative sign on b1. To test H2, we examine the
interaction terms MG £ TIO and MG £ NTIO. Note that in order to test the moderating
effect, it is the interaction terms that we are interested in, not the individual terms of
institutional ownership[8]. Simply put, we investigate whether the positive association
between shareholder rights and earnings quality is strengthened or diminished when
firms stocks are held predominantly by institutional investors of different types. We
expect to observe a positive sign on b4 and a negative sign on b5.
To quantify the incremental effect of different types of institutional ownership, we
perform a second set of analyses in which we partition the percent institutional
ownership into quintiles (TIOQ5 versus NTIOQ5). The resulting empirical
specification is Model 2 shown below:
 Shareholder
Variable Definition
rights
FLDA The DA measure, computed based on DeChow et al. (1996)
MG MG computed based on Gompers et al. (2003)
TIO Percentage of shares held by transient institutions
NTIO Percentage of shares held by nontransient institutions
TIOQ5 1 if percentage of shares held by transient institutions (as defined in Bushee, 2001) is 775
in the top quartile for all firms during the year and 0 otherwise
NTIOQ5 1 if percentage of shares held by nontransient (quasi-indexers and dedicated)
institutions (as defined in Bushee, 2001) is in the top quartile for all firms during the
year and 0 otherwise
PTIOQ5 1 if the firm is in the top quartile of proportional ownership by transient institutions
and in the bottom quartile of proportional ownership by nontransient institutions,
and 0 otherwise
PNTIOQ5 1 if the firm is in the top quartile of proportional ownership by nontransient
institutions and in the bottom quartile of proportional ownership by transient
institutions, and 0 otherwise
LOGASSET The natural log of total assets (Compustat data no. 6)
MB Market value of equity divided by stockholders’ equity of common shareholders
(annual Compustat data no. 60)
CFO Cash flow from operations (Compustat data no. 308) scaled by beginning of year total
assets (Compustat data no. 6)
LEV A firm’s total assets less its book value (Compustat data item 60) divided by its total
assets
LOSS 1 if the firm report a net loss in the previous year, and 0 otherwise
LIT 1 if the firm operates in one of the following high-litigation industries:
pharmaceutical/biotechnology (SIC codes 2833-2836, 8731-8734), computer (3570-
3577, 7370-7374), electronics (3600-3674), or retail (5200-5961), and 0 otherwise
ROA Earnings before extraordinary items (Compustat data no. 123) scaled by average
assets
BIG5 1 if the auditor is a Big 5 firm and 0 otherwise
BINDEP Proportion of independent outside members on the board
ACINDEP Proportion of independent directors on the audit committee
BSIZE Log of the number of directors on the board
ACSIZE Log of the number of members of the audit committee Table I.
CEOCHAIR 1 if the CEO is also the board chair, and zero otherwise Variable definitions

Model 2 : FLDA

¼ a1 þ b1 MGþ b2 TIOQ5þ b3 NTIOQ5þ b4 ðMG£TIOQ5Þþ b5 ðMG

X
£NTIOQ5Þþ bi ðControl variablesi ÞþYear and industry dummiesþ1
i

where i ¼ 6;7: : : 17: We expect to observe negative signs on b1 (H1), b5 (H2) and a
positive sign on b4 (H2). To further examine the moderating effects for different types
of ownership, we test the significance of the summed coefficients b1 þ b4 and b1 þ b5.
If high levels of transient (nontransient) ownership are associated with lower (higher)
earnings quality, the test should yield the following results:
MAJ .
b1 þ b4 is either insignificantly different from zero or positively and
significantly different from zero; and
24,8
.
b1 þ b5 is negatively and significantly different from zero.
As discussed in Section 3.2.2, TIOQ5 and NTIOQ5 by design cannot be used to
measure the relative predominance of one type of ownership over the other type of
776 ownership within the same firm. To overcome this deficiency, we construct alternative
measures PTIOQ5 and PNTIOQ5 which allow us to compare the relative significance
of the two types of institutional ownership both across firms and within the firm.
Specifically, we estimate the following model:
Model 3 : FLDA

¼ a1 þ b1 MGþ b2 PTIOQ5þ b3 PNTIOQ5þ b4 ðMG£PTIOQ5Þþ b5 ðMG

X
£PNTIOQ5Þþ bi ðControl variablesi ÞþYear and industry dummiesþ1
i
where i ¼ 6;7: : : 17: As specified, the model allows us to gain more insight into the
effect of shareholder rights mechanisms on earnings management in environments
characterized by extremely high proportion of a certain type of institutional ownership.
We predict negative signs on b1 (H1), b5 (H2) and a positive sign on b4 (H2). Likewise,
we test the significance of the summed coefficients b1 þ b4 and b1 þ b5 with
predications similar to those for Model 2.

3.4 Sample selection

Our sample is constrained in that we are limited to firms for which both the
shareholder rights index data and the institutional ownership data are available. The
data we use in this study cover the period 1998-2002. This period is of particular
interest because it is characterized by increasing accrual-based earnings management
(Cohen et al., 2008)[9]. Since the G-index is constructed on a biennial basis, we use the
G-index that is published closest to the annual meeting of the fiscal year to align
the data. Governance data (G-index and other internal governance measures) are from
the Investor Responsibility Research Center, and the institutional ownership data are
from the CDA/Spectrum Institutional 13(f) filings from the same sample period.
We exclude financial institutions (SIC codes 6000-6099). Financial institutions cannot
be considered homogenous in that they may have different incentives for earnings
management and financial accounting rules relative to the other firms in our sample.
We also exclude firms where there is insufficient information in the Compustat
database to calculate the DAs measures. This procedure yielded a sample comprising
5,658 firm-year observations.

4. Empirical results
4.1 Descriptive statistics
Table II presents descriptive statistics for the DAs measure, MG, institutional
ownership and other variables for our sample firm-years. The discretionary accruals
measure (FLDA) has a mean (median) value of 0.0123 (0.0103), indicating that, on
average, FLDA is positive and income-increasing. The mean (median) of the MG is
14.998 (15) with the inter-quartile difference ranging from 13 to 17. A firm typically has
 Shareholder
1st 3rd
Variable Mean SD Quartile Median Quartile rights
Dependent variable
FLDA 0.0123 0.0593 0.0013 0.0103 0.0360
Major variables
MG 14.998 2.6121 13 15 17 777
TIO 0.1853 0.1332 0.0839 0.1572 0.2564
NTIO 0.4272 0.1717 0.3006 0.4259 0.5461
All indicator variables for 0.2500 0.4329 0 0 0
ownership
Control variables
LOGASSET (millions) 7,672.65 21,993 619.91 1,637.48 5,098.95
MB 3.1062 2.3416 1.6478 2.1729 3.3175
CFO 0.1103 0.1132 0.1633 0.1037 0.0550
LEV 0.53250 0.2401 0.3668 0.5392 0.6791
LOSS 0.2774 0.4477 0 0 1
ROA 0.0461 0.0866 0.0167 0.0475 0.0888
LIT 0.3230 0.4677 0 0 1
BIG5 0.9758 0.1538 1 1 1
BINDEP 0.6615 0.1748 0.5556 0.6923 0.8000
ACINDEP 0.8973 0.1859 0.8 1 1
BSIZE 9.2508 2.4596 7 9 11
ACSIZE 3.6858 1.1014 3 3 4
CEOCHAIR 0.6501 0.4769 0 1 1
Notes: All variables are defined in Table I; all continuous variables are winsorized at top 1 per cent Table II.
and bottom 99 per cent to mitigate outliers Descriptive statistics

more nontransient ownership than transient ownership (42.72 versus 18.53 per cent).
The mean percentage of the aggregate institutional ownership is 61.26 per cent (sum of
TIO and NTIO), which is some what higher than those reported in earlier studies
(e.g. 36 per cent in Bushee (2001), 39 per cent in Jiambalvo and Subramanyam (1998)
and 45 per cent in Matsumoto (2002)). The difference is likely due to an increase in
institutional ownership over time, as well as the sample bias toward Standard & Poor’s
(S&P) 1500 firms. By construction, all indicator variables for the institutional
ownership (TIOQ5, NTIOQ5, PTIOQ5 and PNTIOQ5) have a mean value of 20 per cent.
With respect to the control variables, the average firm in our sample has total assets
(LOGASSET) of 7,672.65 million[10] and trade at a little over three times the book
value. In total, 28 per cent of the firm-years in the sample report losses (LOSS) in either
of the previous two years. The mean of LEV is 53 per cent, and 32 per cent of the firms
reside in industries characterized by high-LIT. The cash flows from operations (CFO)
and ROA are 11.03 and 4.6 per cent of the total assets for the average firm, respectively,
indicating that the average firm in our sample is profitable and liquid.
Turning to the internal corporate governance variables, independent directors make
up 66 per cent of the board and 89 per cent audit committee in our sample. The median
board size is nine members, while the median audit committee is three, the latter being
consistent with the minimum audit committee size recommended by the Blue Ribbon
Committee Findings (1999). Duties of the CEO and the board chair are segregated for
MAJ 65 per cent of the sample. Nearly, 98 per cent of the sample firms are audited by Big 5
24,8 firms, which is not surprising given that our sample selection is based on S&P 1500.
The correlation matrix of some of the key variables is presented in Table III, where
Pearson (Spearman) correlations appear below (above) the diagonal. The FLDA is
negatively correlated with MG, indicating that earnings quality is higher for firms with
stronger shareholder rights. This provides initial support to H1. The TIO variable is
778 positively associated with FLDA, consistent with the view that institutions that are
overly focused on short-term earnings pressure managers to deliver consistently
higher earnings via earnings management. In contrast, NTIO is negatively correlated
with FLDA, which seems to indicate that nontransient ownership gives rise to higher
earnings quality. However, it is marginally significant with a p-value of 0.16 for
Pearson correlation. In addition, the correlations between the indicator variables of
institutional ownership (TIOQ5, PTIOQ5, NTIOQ5 and PNTIOQ5) and the DA
measure all have expected signs and are mostly significant, thus supporting H2. Note
that the correlations between TIOQ5 and PTIOQ5, NTIOQ5 and PNTIOQ5 are 0.414
and 0.187, respectively, which are both significantly less than one. This indicates that
these indicator variables to a large extent are separate and distinct constructs of the
institutional ownership. On the whole, the univariate results provide some preliminary
evidence in support of our hypotheses. However, these results do not represent direct
evidence on the moderating effect of institutional ownership which can only be tested
through the interaction terms. There are also correlations (untabulated) among our
variables of primary interest and various control variables included in our regression
models. Therefore, we rely on the multivariate analyses to formally test our
hypotheses. When we examined the variance inflation factors (VIFs) for each of our
multivariate regressions, none of the calculated VIFs exceed three, suggesting that
multicollinearity is unlikely to be a problem.
For the sake of brevity, Table III omits the correlations for control variables.
Untabulated results show that market to book ratio (MB) and ROA are positively
associated with FLDA, while firm size (LOGASSET), CFOs, LEV, prior losses (LOSS)
and LIT are all negatively related to FLDA. Among the governance variables,
the proxies for board and audit committee independence (BINDEP and ACINDEP) and
the auditor characteristic (BIG5) all exhibit negative association with FLDA, while
variables both size measures (BSIZE and ACSIZE) and the incidence of the CEO
serving as the chair of the board (CEOCHAIR) are positively correlated with FLDA.
Since many of the control variables are significantly correlated with our variables of
primary interest, and with one another, we next conduct multivariate analyses to
further investigate the relationship between earnings quality, shareholder rights and
institutional ownership.

4.2 Regression results

Table IV presents the results from multivariate analysis. All results are estimated
using White (1980) heteroscedasticity-corrected covariance matrix. Column one
contains the estimated results for Model one. MG is negatively and significantly
associated with FLDA ( p ¼ 0.00), supporting H1 that strong shareholder rights
regimes deter managers from using DAs to manipulate earnings. This result is
consistent with Baber et al.’s (2007) finding that stronger external corporate
governance is associated with lower probabilities of accounting restatement.
 FLDA MG TIO NTIO TIOQ5 NTIOQ5 PTIOQ5 PNTIOQ5

FLDA 2 0.016 (0.01) 0.039 (0.02) 20.062 (0.09) 0.018 (0.04) 20.023 (0.07) 0.015 (0.00) 2 0.017 (0.05)
MG 2 0.026 (0.00) 20.048 (0.00) 0.036 (0.00) 20.103 (0.00) 0.06 (0.00) 20.143 (0.00) 0.023 (0.08)
TIO 0.040 (0.04) 2 0.069 (0.00) 0.728 (0.00) 0.465 (0.00) 20.157 (0.20) 0.291 (0.00) 2 0.163 (0.00)
NTIO 2 0.076 (0.16) 0.032 (0.01) 0.669 (0.00) 20.096 (0.23) 0.490 (0.00) 20.098 (0.11) 0.397 (0.00)
TIOQ5 0.016 (0.03) 2 0.100 (0.00) 0.447 (0.00) 20.116 (0.19) 20.049 (0.00) 0.601 (0.00) 2 0.333 (0.00)
NTIOQ5 2 0.036 (0.12) 0.06 (0.00) 20.184 (0.25) 0.520 (0.00) 20.052 (0.00) 20.231 (0.00) 0.217 (0.00)
PTIOQ5 0.023 (0.00) 2 0.139 (0.00) 0.182 (0.00) 20.091 (0.21) 0.414 (0.00) 20.261 (0.00) 2 0.313 (0.00)
PNTIOQ5 2 0.021 (0.09) 0.020 (0.06) 20.161 (0.00) 0.487 (0.00) 20.333 (0.00) 0.187 (0.00) 20.333 (0.00)
Notes: Person (Spearman) correlation coefficients are in the lower (upper) triangle, with p-values shown in parentheses; all variables are defined in Table I
Shareholder

Correlation matrix
rights

Table III.
779
MAJ
Dependent variable ¼ FLDA
24,8 Model 1 Model 2 Model 3

Panel A: regression with institutional ownership interactions

Independent variables
Intercept 0.017 (0.52) 0.023 (0.39) 0.013 (0.56)
780 MG 2 0.003 (0.00) 2 0.002 (0.00) 2 0.002 (0.00)
TIO 0.052 (0.12)
NTIO 20.081 (0.11)
MG £ TIO 0.003 (0.05)
MG £ NTIO 2 0.008 (0.28)
TIOQ5 0.034 (0.06)
NTIOQ5 2 0.017 (0.29)
MG £ TIOQ5 0.002 (0.03)
MG £ NTIOQ5 2 0.001 (0.14)
PTIOQ5 0.030 (0.03)
PNTIOQ5 20.011 (0.09)
MG £ PTIOQ5 0.003 (0.01)
MG £ PNONTRAN5 2 0.002 (0.06)
LOGASSET 20.004 (0.00) 2 0.004 (0.00) 20.004 (0.00)
MB 0.001 (0.02) 0.001 (0.01) 0.001 (0.01)
CFO 20.726 (0.00) 2 0.726 (0.00) 20.721 (0.00)
LEV 0.011 (0.05) 0.012 (0.05) 0.012 (0.04)
LOSS 20.039 (0.00) 2 0.037 (0.01) 20.033 (0.00)
ROA 0.722 (0.00) 0.712 (0.00) 0.714 (0.00)
LIT 0.018 (0.00) 0.019 (0.00) 0.019 (0.00)
BIG5 20.006 (0.51) 2 0.005 (0.62) 20.005 (0.53)
BINDEP 20.008 (0.05) 2 0.012 (0.01) 20.011 (0.01)
ACINDEP 0.002 (0.71) 0.002 (0.74) 0.002 (0.74)
BSIZE 0.006 (0.04) 0.006 (0.07) 0.005 (0.07)
ACSIZE 0.004 (0.13) 0.003 (0.17) 0.003 (0.16)
CEOCHAIR 0.003 (0.03) 0.003 (0.02) 0.003 (0.02)
Industry and year dummies Yes Yes Yes
Adj. R 2 0.4508 0.4757 0.4961
No. of observations 5,658 5,658 5,658
Panel B: significance test
Table IV. t-test of bMG þ bMG£ TIOQ5 0.000 (0.65)
The association between t-test of bMG þ bMG£ NTIOQ5 2 0.003 (0.12)
income-increasing DAs, t-test of bMG þ bMG£ PTIOQ5 0.001 (0.88)
corporate governance t-test of bMG þ bMG£ PNTIOQ5 20.004 (0.07)
index and alternative
measures of institutional Notes: Two-tailed p-values of the estimated parameters are reported in parentheses; all continuous
ownership variables are winsorized at top 1 per cent and bottom 99 per cent to mitigate outliers

The coefficient on the interaction involving MG and TIO is positive and significant
( p ¼ 0.04), implying that the higher the level of transient ownership, the weaker the
constraining effect on earnings management of shareholder rights. The coefficient on
the interaction involving MG and NTIO is negative as expected but insignificant. This
result does not provide support to the hypothesis that the higher the level of
nontransient ownership, the stronger the constraining effect on earnings management
of shareholder rights. However, as noted before, we need to exercise caution
in interpreting the interaction term between two continuous variables due to a number
of confounding issues and technical difficulties (Aiken and West, 1991). Shareholder
Hence, we perform formal tests of H2 using Models two and three where indicator rights
variables for institutional ownership are used.
With respect to the control variables, the results are consistent with those in
the correlation analysis. Firm size (LOGASSET) is negatively and significantly
associated with FLDA, supporting Watts and Zimmerman’s (1990) argument that
large companies face greater political costs and are more subject to scrutiny from 781
the financial media and investors, and thus less likely to manage earnings. LEV is
negatively and significantly related to FLDA, consistent with heavy indebtedness
provide incentives for managers to avoid violation of debt covenants through accruals
management (DeAngelo et al., 1994). The positive and significant coefficient on market
to book ration (MB) indicates that firms experiencing high growth aggressively
exercise accounting discretion to avoid negative earnings surprises. LIT is positively
associated with FLDA, suggesting that firms with high-LIT are more likely to be
concerned about missing earnings benchmarks and hence manage earnings to achieve
forecasted earnings targets. Other firm-specific financial control variables (ROA and
CFO) are statistically significant at conventional levels, suggesting that earnings
management is less likely for liquid and profitable firms. Turning to the governance
variables, the estimate for BINDEP is significantly negative whereas the coefficient on
BSIZE is positive, consistent with firms with more independent directors and smaller
boards exercising less accounting discretion. The incidence of the CEO serving as the
board chair (CEOCHAIR) is positively correlated with FLDA, suggesting that having a
CEO who is simultaneously the chairman erodes the effectiveness of the board and
increases the likelihood of manipulating earnings. Contrary to our expectations, the
audit committee characteristics (ACINDEP and ACSIZE) exhibit positive associations
with FLDA, although the estimates on these variables all lack statistical significance.
To evaluate the incremental and moderating effect of institutional ownership, we
estimate Model 2 by replacing the level of institutional ownership with their quintile
values. The regression results are reported in column two. MG is negatively and
significantly associated with FLDA ( p ¼ 0.00), lending further support to our first
hypothesis. To test the moderating effect of institutional ownership, we turn to the
interaction terms and compute the sum of coefficients. The coefficient on MG £ TIOQ5
is positive and significant ( p ¼ 0.03) and the sum of coefficients, bMG þ bMG£ TIOQ5, is
statistically insignificant from zero ( p ¼ 0.65), consistent with high levels of transient
ownership weakening the strength of shareholder rights, to the extent that they
diminish or even reverse the positive effect of strong shareholder rights in decreasing
earnings management. The coefficient on MG £ NTIOQ5 is negative as expected, but
insignificant at conventional levels ( p ¼ 0.14). The sum of coefficients, bMG and
bMG£ NTIOQ5, is insignificantly different from zero ( p ¼ 0.12). These results indicate
that nontransient institutions’ role in strengthening shareholder rights and reducing
earnings management is questionable. Overall, the estimated results lend some
support to our second hypothesis that the moderating effect varies between different
types of institutions. Transient institutions are more likely to pressure managers into a
short-term focus and create incentives for them to manage earnings to meet near-term
targets, while nontransient institutions’ effectiveness in monitoring corporate
managers and inhibiting earnings management is almost nonexistent.
MAJ We next estimate Model 3 which includes indicator variables that identify a firm’s
24,8 ownership structure as either predominantly transient or nontransient. The coefficient
on MG remains negative and highly significant ( p ¼ 0.00). Turning to the moderating
effect, the interaction term MG £ PTIOQ5 carries the expected positive sign and is
statistically significant ( p ¼ 0.01). The sum of bMG and bMG£ PTIOQ5 is insignificantly
different from zero ( p ¼ 0.88), suggesting that the constraining effect of shareholder
782 rights on earnings management is rendered virtually ineffective in firms with
an ownership base dominated by short-term focused institutional investors.
The interaction term MG £ PNTIOQ5, and the summed coefficient of bMG and
bMG£ PNTIOQ5 carry negative signs as expected and turn marginally significant,
suggesting that long-term, sophisticated institutional investors are generally effective
in their role of strengthening shareholder rights and deterring opportunistic earnings
management. Notice that PTIOQ5 and PNTIOQ5 produce larger parameter estimates
and stronger significance levels than the institutional ownership measures used in
Model 2, suggesting that they provide more precise estimates. Model 3 also yields the
most significant results and highest explanatory power (adjusted R 2 ¼ 0.4961), which
is possibly attributable to a more sound definition of the extent of predominance for
different types of institutional ownership. The three regression models used in our
tests overall have significant explanatory power, with the adjusted R 2 ranging from
45.08 to 49.61 per cent.
To provide evidence on whether there is any differential relation between DAs,
corporate governance index, and institutional ownership conditional on whether
DAs are income-increasing or income-decreasing, we partition our sample into two
groups based on the sign of the firms’ DAs. Panel A of Table V reports the regression
results estimated with sample firms that report FLDA grater than or equal to zero.
The results are similar to those from our main analysis reported in Table IV. MG is
negative and significant across the three alternative models. Our alternative metrics of
transient institutional ownership and their respective interaction with the corporate
governance index carry the expected positive sign and are mostly significant (except
for TIO in model one), while none of the nontransient measures is statistically
significant. Panel B of Table V reports the regression results for firms having FLDA
values less than zero. We find no significant association between FLDA and MG or any
of the alternative institutional ownership measures. These findings seem to suggest
that the earlier results reported in our main analysis are driven by income-increasing
DAs. While income-decreasing accruals can be interpreted as a form of biased financial
reporting, they also reflect a conservative application of GAAP. Typically, transient
institutional investors focus on short-term earnings performance and pressure
management into boosting reported earnings through aggressive accounting.
Therefore, they are likely more concerned with the use of income-increasing DAs.
This might explain why we did not find any statistically significant association
between income-decreasing DAs and institutional ownership measures[11], [12].
To sum up, we find evidence that firms with stronger shareholder rights regimes
are associated with higher quality of reported earnings. However, the presence of large,
predominantly transient institutions substantially diminishes the effect of shareholder
rights in reducing earnings management. The evidence is mixed and nonconsistent on
the role of nontransient institutions in curbing earnings management through
monitoring and disciplining actions.
 Shareholder
Model 1 Model 2 Model 3
rights
Panel A: regression based on income-increasing DAs
Dependent variable ¼ positive FLDA
Independent variables
Intercept 0.023 (0.30) 0.032 (0.07) 0.022 (0.16)
MG 2 0.005 (0.00) 2 0.003 (0.00) 2 0.003 (0.00) 783
TIO 0.061 (0.09)
NTIO 20.088 (0.06)
MG £ TIO 0.004 (0.03)
MG £ NTIO 2 0.012 (0.15)
TIOQ5 0.039 (0.05)
NTIOQ5 2 0.023 (0.25)
MG £ TIOQ5 0.003 (0.03)
MG £ NTIOQ5 2 0.003 (0.11)
PTIOQ5 0.039 (0.04)
PNTIOQ5 20.017 (0.07)
MG £ PTIOQ5 0.005 (0.00)
MG £ PNONTRAN5 2 0.003 (0.05)
LOGASSET 20.004 (0.00) 2 0.005 (0.00) 20.005 (0.00)
MB 0.001 (0.02) 0.002 (0.01) 0.001 (0.00)
CFO 20.769 (0.00) 2 0.753 (0.00) 20.758 (0.00)
LEV 0.015 (0.05) 0.018 (0.04) 0.019 (0.03)
LOSS 20.042 (0.00) 2 0.041 (0.01) 20.042 (0.00)
ROA 0.756 (0.00) 0.761 (0.00) 0.758 (0.00)
LIT 0.018 (0.00) 0.017 (0.00) 0.020 (0.00)
BIG5 20.009 (0.23) 2 0.008 (0.18) 20.008 (0.13)
BINDEP 20.011 (0.03) 2 0.015 (0.00) 20.014 (0.00)
ACINDEP 0.006 (0.65) 0.005 (0.81) 0.005 (0.56)
BSIZE 0.008 (0.03) 0.010 (0.05) 0.009 (0.05)
ACSIZE 0.005 (0.17) 0.005 (0.19) 0.004 (0.18)
CEOCHAIR 0.005 (0.01) 0.004 (0.00) 0.005 (0.00)
Industry and year dummies Yes Yes Yes
Adj. R 2 0.4867 0.5011 0.5153
No. of observations 3,128 3,128 3,128
Panel A. Significance test
t-test of bMG þ bMG£ TIOQ5 0.000 (0.37)
t-test of bMG þ bMG£ NTIOQ5 2 0.006 (0.28)
t-test of bMG þ bMG£ PTIOQ5 0.002 (0.56)
t-test of bMG þ bMG£ PNTIOQ5 20.006 (0.05)
Panel B: regression based on income-decreasing DAs
Dependent variable ¼ negative FLDA
Independent variables
Intercept 0.019 (0.48) 0.019 (0.29) 0.016 (0.48)
MG 2 0.002 (0.26) 2 0.002 (0.23) 2 0.001 (0.29)
TIO 0.045 (0.59) Table V.
NTIO 20.042 (0.71) The association
MG £ TIO 0.002 (0.38) between DAs, corporate
MG £ NTIO 0.001 (0.57) governance index
TIOQ5 0.027 (0.25) and alternative measures
NTIOQ5 2 0.011 (0.29) of institutional
MG £ TIOQ5 0.001 (0.27) ownership conditional
(continued) on the sign of DAs
MAJ Model 1 Model 2 Model 3
24,8
MG £ NTIOQ5 0.000 (0.89)
PTIOQ5 0.027 (0.33)
PNTIOQ5 20.007 (0.58)
MG £ PTIOQ5 0.002 (0.42)
784 MG £ PNONTRAN5 2 0.002 (0.56)
LOGASSET 20.002 (0.06) 2 0.002 (0.05) 20.002 (0.05)
MB 0.000 (0.08) 0.001 (0.07) 0.001 (0.05)
CFO 20.627 (0.10) 2 0.639 (0.07) 20.633 (0.08)
LEV 0.011 (0.10) 0.012 (0.09) 0.012 (0.09)
LOSS 20.025 (0.10) 2 0.031 (0.09) 20.029 (0.10)
ROA 0.688 (0.20) 0.678 (0.12) 0.675 (0.13)
LIT 0.013 (0.22) 0.019 (0.21) 0.018 (0.19)
BIG5 0.004 (0.68) 0.004 (0.77) 0.003 (0.89)
BINDEP 20.008 (0.28) 2 0.010 (0.25) 20.011 (0.22)
ACINDEP 0.000 (0.56) 0.000 (0.55) 0.001 (0.58)
BSIZE 0.004 (0.09) 0.004 (0.11) 0.004 (0.16)
ACSIZE 0.003 (0.17) 0.002 (0.28) 0.002 (0.29)
CEOCHAIR 0.002 (0.10) 0.003 (0.11) 0.003 (0.11)
Industry and year dummies Yes Yes Yes
Adj. R 2 0.1711 0.1987 0.1916
No. of observations 2,530 2,530 2,530
Panel B: significance test
t-test of bMG þ bMG£ TIOQ5 2 0.001 (0.78)
t-test of bMG þ bMG£ NTIOQ5 2 0.002 (0.63)
t-test of bMG þ bMG£ PTIOQ5 0.001 (0.90)
t-test of bMG þ bMG£ PNTIOQ5 20.003 (0.37)
Notes: Two-tailed p-values of the estimated parameters are reported in parentheses; all continuous
variables are winsorized at top 1 per cent and bottom 99 per cent to mitigate outliers; all variables are
Table V. defined in Table I

4.3 Additional analyses

In order to validate our primary findings and provide more evidence on how G-index
and institutional ownership are related to DAs, we conduct several additional tests.
The weak results on the nontransient institutional ownership may be attributable
to the specification of the variable failing to capture the “monitoring role” of the
long-term institutional owners. To account for the possibility that there can be
substantive differences in trading and governance behaviour within the nontransient
type, we further classify it into two subgroups, namely, “dedicated owners” and
“quasi-indexers” (Bushee, 2001). “Dedicated owners” are institutions with large,
long-term holdings which are concentrated in only a few firms. They are the group
most likely to provide a high degree of monitoring of managerial behaviour.
This monitoring can occur either explicitly, through governance activities, or implicitly,
through information gathering. “Quasi-indexers” use indexing or buy-and-hold
strategies that are characterized by high-diversification and low-portfolio turnover. On
one hand, their passive, fragmented ownership provides little incentive to monitor
managers (Porter, 1992). One the other hand, these institutions are unable to sell
because of their indexing strategies, which may provide strong incentives to monitor
management to ensure it is acting in the long-term interests of the firm (Monks and Shareholder
Minow, 1995). Given that the role of “quasi-indexers” is not clearly defined, combining rights
them with dedicated owners to form one single type of ownership could potentially
reduce the power of the test and thus explain the lack of significance of our results. We
reran all our analyses using the new classification. The results are substantively similar
to those reported in Table IV. The estimates for “dedicated owners” and
“quasi-indexers” and their corresponding interaction terms range from insignificant 785
to marginal across three models. In addition, we did not find any significant differences
between the results for “dedicated owners” and those for “quasi-indexers”.
For dedicated institutions, the lack of significant impact probably stems from the
fact that their percentage holdings are small relative to other types of ownership.
For instance, untabulated results show that within the top quintile of proportional
ownership by dedicated institutions, the mean holdings is less than 19 per cent,
compared to mean holdings of 91 per cent (68 per cent) for quasi-indexer (transient)
institutions. For quasi-indexer ownership, the lack of results is probably due to the
ambiguous role they play in constraining earnings management.
To alleviate concerns that our results are driven by our accrual specification, we
employ an alternative accruals measure that adjusts for firm performance. Adjusting for
performance is important because prior research documents that DA estimates are
correlated with firm performance (Kasznik, 1999; Kothari et al., 2005). As a means of
mitigating the misspecification to reduce the likelihood of incorrect inferences, we
employ a portfolio match technique advanced by Kasznik (1999). We first compute
abnormal accruals which are estimated as the difference between the total accruals and
the fitted normal accruals using the modified Jones model (DeChow et al., 1995). To
adjust for firm performance, we partition firms within each two-digit SIC code into
deciles based on their prior year’s ROA. The performance matched DA is the difference
between a sample firm’s abnormal accruals and the median abnormal accrual for each
ROA portfolio. This portfolio approach controls for relative performance across random
samples. We then re-estimate all of our regressions using performance-adjusted
accruals. The results are qualitatively similar to those reported in the DA analysis.
To assess whether our results are time sensitive or are driven by any one year within
our sample period, we re-run the analyses separately for each year. Running annual
regressions also allows us to address the possible violation of the independence
assumption because the results from the pooled estimation are vulnerable to the serial
correlation problem[13]. The respective sets of analyses from each year produce results
similar to those reported for the full sample, with the statistical tests being a little weaker
but remaining mostly significant at conventional levels. These results are not surprising
because within each firm, the levels of shareholder rights and institutional ownership
remain fairly stable over time. We also calculate the mean coefficients from the annual
regressions and t-statistics based on the variation in the coefficients (Fama and MacBeth,
1973). The coefficient estimates and the t-statistics calculated using the Fama-MacBeth
method are largely consistent with our main hypotheses, although we should avoid
over-interpreting the results since only five years are used in the calculation of t-statistics.
To ensure that the results reported are not sensitive to the definition of variables, we
utilize several alternative measures for the variables in the regression models.
Specifically:
MAJ .
we replace the market-to-book ratio (MB) with the sales growth variable
24,8 measured as the percentage increase in sales from year t 2 1 to year t;
.
measure loss as the proportion of negative earnings before extraordinary items
in the prior three years; and
.
substitute the log of market value (or log of sales) for the LOGASSET to
investigate the sensitivity of our results to the alternative control measures.
786
The results are largely the same except that the coefficient on the sales growth measure
turns negative and insignificant. However, results for the variables of primary interest
are substantively similar to previously reported results.
We also perform tests by imposing additional restrictions on the sample.
Specifically:
.
we restrict the sample to firms with at least five percent institutional ownership
to ensure that institutional ownership is likely to be influential in the firm; and
.
we restrict the sample to firms audited by Big 5 auditors and Grant Thornton
and BDO Seidman.

Becker et al. (1998) find that companies with non-Big 5 auditors report DAs that
significantly increase income compared to Big 5 auditors. We eliminate firms audited
by non-Big 5 auditors (except for Grant Thornton and BDO Seidman) to remove this
bias. In both cases, our overall results remain virtually identical to our reported results
using the full sample[14].

5. Summary and conclusions

This study examines whether stronger shareholder rights induce higher earnings
quality and how institutional ownership influences the effectiveness of shareholder
rights in constraining opportunistic management of accruals. Our research adds to the
extant literature by examining whether greater shareholder empowerment increases
investors’ ability to monitor and discipline managers, thus reducing earnings
management. In addition, we investigate the role of institutional ownership in
influencing corporate managers’ earnings management behaviour in the broader
setting of shareholder rights regimes. After controlling for the firm-specific factors, we
find that greater shareholder empowerment (as measured by the Gomper’s index) is
associated with higher earnings quality. However, this positive effect is attenuated or
neutralized when institutional investors are predominantly transient with short
investment horizons. Our results are consistent with the view that institutional
investors’ focus on short-term earnings performance could pressure management into
boosting reported earnings through aggressive accounting. It appears that
institutional investors may be using their increased empowerment and their ability
to affect managerial behaviour for the purpose of increasing the value of their shares in
the short-term. The results from this study enhance our understanding of the relation
between shareholders rights and different types of institutional ownership, and how
they interact in affecting the quality of corporate earnings. In view of the current effort
by the SEC to increase shareholder rights, the results of this study provide evidence to
the SEC that type of ownership should also be factored into the discussion when
considering steps to increase shareholder rights.
 The main limitation of this study is that the sample size is restricted. Our sample is Shareholder
constrained in that we are limited to firms for which both the shareholder rights index rights
data and the institutional ownership data are available. Another limitation is that,
while the G-index used in this study surrogating for corporate governance is definitely
superior to single measures of corporate governance used in this study, it may not be
all inclusive. There may be other factors of corporate governance not factored into this
index. This provides an avenue for future research. Future researchers could examine 787
the adequacy of this index and focus on developing more comprehensive indices that
better reflect and surrogate for corporate governance. If and when more exhaustive
corporate governance indices are constructed, future studies could replicate this study
to examine if the result of this study can be corroborated using more sophisticated
governance measures.

Notes
1. For example, using Gomper’s index as a proxy for overall quality of corporate governance,
Bowen et al. (2007) did not find significant association between their discretionary accrual
measure and Gomper’s index. Similarly, Larcker et al. (2007) found no statistical association
between abnormal accruals and their anti-takeover indices designed to proxy for the power
of the market for corporate control in disciplining the firm.
2. For example, insiders can use their financial reporting discretion to overstate earnings and
conceal unfavorable earnings realizations (i.e. losses) that would prompt outsider
interference.
3. In the section of additional analyses, we examine the robustness of our results using the
performance-matched discretionary accruals model and obtain similar conclusions.
4. The 24 provisions examined include anti-greenmail, blank cheque preferred stock, business
combination laws, by law and charter amendment limitations, classified board, compensation
plans with change in control provisions, director indemnification contracts, control share
cash-out laws, cumulative voting requirements, director’s duties, fair price requirements,
golden parachutes, director indemnification, limitations on director liability, pension
parachutes, poison pills, secret ballot, executive severance agreements, silver parachutes,
special meeting requirements, supermajority requirements, unequal voting rights and
limitations on action by written consent. The index is constructed for each firm by adding one
point for every factor that restricts shareholder rights or increases management power.
5. For example, the high correlation between transient and nontransient ownership can potentially
bias our estimates and statistical results (see the correlation matrix in Table III for details).
6. There are two additional reasons for our choice of the quartile values over the levels of
institutional ownership. It allows us to: (1) quantify the incremental effect of different types of
institutional ownership; and (2) examine the moderating effect of institutional ownership by
interacting it with the shareholder rights variable. There are a number of confounding issues
and technical difficulties in interpreting the interaction between two continuous variables (see
Aiken and West (1991) for more details), and thus indicator variables are preferred.
7. The results are inferentially similar when we rank each group into quartiles or terciles, but
are most significant when defined using quintiles.
8. From a technical perspective, the coefficient of the institutional ownership variable captures
the effect of an increase in the TIO or NTIO by which it is multiplied on the dependent
variable when the other variable in the interaction term (i.e. MG) is set to zero. Since, MG can
rarely be zero, the interpretation tends to be questionable. In all subsequent analyses, we will
focus on the interaction terms only.
MAJ 9. Gomper’s index was compiled for years 1990, 1993, 1995, 1998, 2000, 2002 and 2004. In 1998, the
sample size increased by about 25 per cent through additions of some smaller firms and firms
24,8 with high institutional-ownership levels. For purpose of consistency, our sample selection starts
with 1998. Our results are qualitatively similar when the full sample period is used.
10. The median value of total assets is 1,637.48 million, indicating a skewed distribution and
thus justifying the log transformation.
788 11. We also ran our tests using the absolute value of discretionary accruals. While the signs of
the coefficients on our variables of interest are in the predicted direction, none of them are
statistically significant. Given that we found no significant results for firms having
income-decreasing accruals, combining the positive and negative discretionary accruals and
transforming them into absolute accruals likely lead to a reduced likelihood of rejecting the
null hypothesis of no earnings management, hence an overall lack of significance for the
combined sample.
12. We also estimated our models with the total institutional ownership (i.e. the sum of transient
and nontransient ownership). We found no evidence of a significant association between the
total ownership measure and discretionary accruals. This result is not surprising
considering the differential roles played by the two different groups of institutional
investors. Treating them as a homogeneous group would bias tests against finding any
significant relation between earnings quality and institutional ownership. We would like to
thank the anonymous reviewer for suggesting this analysis.
13. Governance characteristics such as G-index are correlated over time. Mean change in the
G-index over the sample period is 0.31, and the index changes no more than ^1 in 88 per cent
of the cases. Therefore, including multiple observations for each firm over the sample period
potentially violates the independent and identically distributed assumption and overstates
statistical significance.
14. We do not tabulate the estimates for the additional analyses because differences between
these results and those presented in Table IV are inconsequential. All results are available
upon request.

References
Aiken, L.S. and West, S.G. (1991), Multiple Regression: Testing and Interpreting Interactions,
Sage, London.
Ajinkya, B., Bhojraj, S. and Partha, S. (2005), “The association between outside directors,
institutional investors and the properties of management earnings forecasts”, Journal
of Accounting Research, Vol. 43, pp. 343-76.
Anand, V. (1993), “Proxy season quiet about elective pensions”, Pensions and Investment Age,
Vol. 21, p. 3.
Ashbaugh, H., Collins, D. and LaFond, R. (2004), “Corporate governance and the cost of equity
capital”, working paper, University of Wisconsin, Madison, WI.
Baber, W.R., Kang, S.H. and Liang, L. (2007), “Shareholder rights, corporate governance and
accounting restatement”, available at: http://ssrn.com/abstract¼760324
Beasley, M.S. (1996), “An empirical analysis of the relation between the board of director
composition and financial statement fraud”, The Accounting Review, Vol. 71, pp. 443-65.
Becker, C., DeFond, M., Jiambalvo, J. and Subramanyan, K.R. (1998), “The effect of audit quality
on earnings management”, Contemporary Accounting Research, Vol. 15, pp. 1-24.
Blue Ribbon Committee Findings (1999), Blue Ribbon Committee Findings, American Institute
of Certified Public Accountants, New York, NY.
Bowen, R.M., Rajgopal, S. and Venkatachalam, M. (2007), “Accounting choice, corporate Shareholder
governance and firm performance”, Contemporary Accounting Research.
rights
Brown, L.D. (2001), “A temporal analysis of earnings surprises: profits versus losses”, Journal
of Accounting Research, Vol. 39 No. 2, pp. 221-41.
Bushee, B. (1998), “The influence of institutional investors on myopic R&D investment
behavior”, Accounting Review, Vol. 73, pp. 305-33.
Bushee, B. (2001), “Do institutional investors prefer near-term earnings over long-run value?”, 789
Contemporary Accounting Research, Vol. 18, pp. 207-46.
Cheng, C.S.A., Collins, D. and He Huang, H. (2006), “Shareholder rights, financial disclosure, and
the cost of equity capital”, Review of Quantitative Finance and Accounting, Vol. 27,
pp. 175-204.
Chung, R., Michael, F. and Kim, J. (2002), “Institutional monitoring and opportunistic earnings
management”, Journal of Corporate Finance, Vol. 8, pp. 29-48.
Cohen, D.A., Aiyesha, D. and Thomas, L.Z. (2008), “Real and accrual-based earnings
management in the pre- and post-Sarbanes Oxley periods”, The Accounting Review, Vol. 83,
pp. 757-87.
DeAngelo, H., DeAngelo, L. and Skinner, J.D. (1994), “Accounting choice in troubled companies”,
Journal of Accounting and Economics, Vol. 17 Nos 1/2, pp. 113-43.
DeChow, P.M., Richardson, S. and Tuna, I. (2003), “Why are earnings kinky? An examination of
the earnings management explanation”, Review of Accounting Studies, Vol. 8, pp. 255-384.
DeChow, P.M., Sloan, R.G. and Sweeney, A.P. (1995), “Detecting earnings management”,
The Accounting Review, Vol. 70, pp. 193-225.
DeChow, P.M., Sloan, R.G. and Sweeney, A.P. (1996), “Causes and consequences to earnings
manipulation: an analysis of firms subject to enforcement action by the SEC”,
Contemporary Accounting Research, Vol. 13, pp. 1-36.
DeFond, M. and Jiambalvo, J. (1994), “Debt covenants effects and the manipulation of accruals”,
Journal of Accounting and Economics, Vol. 17, pp. 145-76.
Demirag, I. (1998), “Corporate governance, accountability and pressures to perform”, Studies in
Managerial and Financial Accounting, Vol. 8, pp. 21-35.
Diamond, D. and Verrecchia, R. (1991), “Disclosure, liquidity, and the cost of capital”, The Journal
of Finance, Vol. 66, pp. 1325-55.
Fama, E.F. and MacBeth, J. (1973), “Risk, return and equilibrium: empirical tests”, Journal of
Political Economy, Vol. 81, pp. 607-36.
Francis, J., Maydew, E. and Sparks, H. (1999), “The role of big six auditors in the credible
reporting of accruals”, Auditing: A Journal of Practice & Theory, Vol. 18 No. 2, pp. 17-34.
Gadhoum, Y. (2000), “The consequences of ownership structure and corporate governance on
competitiveness”, working paper, Universite de Quebec, Montréal.
Gillan, S.L. and Starks, L.T. (2000), “Corporate governance proposals and shareholder activism:
the role of institutional investors”, Journal of Financial Economics, Vol. 57, pp. 275-305.
Gompers, P., Ishii, J. and Metrick, A. (2003), “Corporate governance and equity prices”, Quarterly
Journal of Economics, Vol. 118, pp. 107-55.
Graves, S.B. and Waddock, S. (1990), “Institutional ownership and control: implications for
long-term corporate strategy”, Academy of Management Executive, Vol. 4, pp. 75-83.
Hayn, C. (1995), “The information content of losses”, Journal of Accounting and Economics,
Vol. 20 No. 2, pp. 125-53.
MAJ Jensen, M. and Meckling, W. (1976), “Theory of the firm: managerial behavior, agency costs and
ownership structure”, Journal of Financial Economics, Vol. 3, pp. 305-60.
24,8 Jiambalvo, J. and Subramanyam, K.R. (1998), “The effect of audit quality on earnings
management”, Contemporary Accounting Research.
Jiambalvo, J., Rajgopal, S. and Venkatachalam, M. (2002), “Institutional ownership and the extent
to which stock prices reflect future earnings”, Contemporary Accounting Research, Vol. 19,
790 pp. 117-45.
Jones, J. (1991), “Earnings management during import relief investigations”, Journal of
Accounting Research, Vol. 29, pp. 193-228.
Kasznik, R. (1999), “On the association between voluntary disclosure and earnings
management”, Journal of Accounting Research, Vol. 33, pp. 353-67.
Klein, A. (2002), “Audit committee, board of director characteristics, and earnings management”,
Journal of Accounting and Economics, Vol. 33, pp. 375-400.
Kothari, S., Leone, A. and Wasley, C. (2005), “Performance matched discretionary accrual
measures”, Journal of Accounting and Economics, Vol. 39 No. 1, pp. 163-97.
Lang, M.H. and McNichols, M.F. (1997), “Institutional trading and corporate earnings and
returns”, working paper, Stanford University, Stanford, CA.
Lang, M.H. and McNichols, M.F. (1999), “Institutional trading and corporate performance”,
working paper, Stanford University, Stanford, CA.
La Porta, R., Lopez-de-Silanes, F., Schleifer, A. and Vishny, R. (2002), “Investor protection and
corporate valuation”, Journal of Finance, Vol. 57, pp. 1147-70.
Larcker, D.F., Richardson, S.A. and Tuna, I. (2007), “Corporate governance, accounting outcomes,
and organizational performance”, The Accounting Review, Vol. 82 No. 4, pp. 963-1008.
Laverty, K. (1996), “Economic ‘short-termism’: the debate, the unresolved issues, and the
implications for management practice and research”, Academy of Management Review,
Vol. 21, pp. 825-60.
Leuz, C., Dhananjay, N. and Wysocki, P.D. (2003), “Earnings management and investor
protection: an international comparison”, Journal of Financial Economics, Vol. 69,
pp. 505-27.
Lobo, G.J. and Zhou, J. (2006), “Did conservatism in financial reporting increase after the
Sarbanes Oxley Act? Initial evidence”, Accounting Horizons, Vol. 20, pp. 57-73.
McConnell, J. and Servaes, H. (1990), “Additional evidence on equity ownership and corporate
value”, Journal of Financial Economics, Vol. 27, pp. 595-612.
Matsumoto, D.A. (2002), “Management’s incentives to avoid negative earnings surprises”,
The Accounting Review, Vol. 77, pp. 483-514.
Monks, R. and Minow, N. (1995), Corporate Governance, Blackwell, Cambridge, MA.
Palmrose, Z.V. (1999), “Empirical research on auditor litigation”, Studies in Accounting Research
No. 33, American Accounting Association, Sarasota, FL.
Porter, M.E. (1992), “Capital disadvantage: America’s failing capital investment system”,
Harvard Business Review, Vol. 70, pp. 65-82.
Shleifer, A. and Vishny, R. (1997a), “A survey of corporate governance”, The Journal of Finance,
Vol. 52, pp. 737-83.
Shleifer, A. and Vishny, R. (1997b), “The limits of arbitrage”, Journal of Finance, Vol. 52,
pp. 35-55.
Sias, R.W. and Starks, L.T. (1998), “Institutions, individuals and return autocorrelations”,
working paper, University of Texas, Austin, TX.
Solomon, D. and Lublin, J. (2004), “SEC considering rule that would allow shareholders to Shareholder
nominate board members”, Wall Street Journal, Vol. 22.
Summers, S.L. and Sweeney, J.T. (1998), “Fraudulent misstated financial statements and insider
rights
trading: an empirical analysis”, The Accounting Review, Vol. 73, pp. 131-46.
Wahal, S. and McConnell, L. (1997), “Do institutions investors exacerbate managerial myopia?”,
unpublished working paper, Purdue University, West Lafayette, IN.
Watts, R. and Zimmerman, J. (1990), “Positive accounting theory: a ten-year perspective”, 791
The Accounting Review, Vol. 65, pp. 112-34.
White, H. (1980), “A heteroscedasticity consistent covariance matrix estimator and a direct test
for heteroscedasticity”, Econometrica, Vol. 48, pp. 817-38.

About the authors

Wei Jiang is an Associate Professor of Accounting at California State University. He earned his
PhD from Rutgers University, Newark. His research interests are in the area of financial
accounting, auditing and information systems. He has published in Advances in Accounting
among others.
Asokan Anandarajan is a Professor of Accounting at the New Jersey Institute of Technology.
He earned his PhD at Drexel University, Philadelphia. His research interests are in the area of
earnings overstatement fraud. He has published in peer-reviewed journals including, Accounting
Horizons, Behavioral Research in Accounting, Auditing: A Journal of Practice & Theory,
Advances in Accounting, among others. Asokan Anandarajan is the corresponding author and
can be contacted at: anandarajan@adm.njit.edu

To purchase reprints of this article please e-mail: reprints@emeraldinsight.com

Or visit our web site for further details: www.emeraldinsight.com/reprints