Transient Institutional Ownership and the Contracting Use of Returns and Earnings*
Shane S. Dikolli
University of Texas at Austin
Susan L. Kulp
Harvard Business School
Karen L. Sedatole
University of Texas at Austin
Data Availability: Data used for this study are derived from publicly available sources
* We thank Brian Bushee for providing access to the institutional investor classifications used throughout this paper,
Frank Moers for assistance on the estimation techniques used, and Jane Zhang for valuable research assistance. We
appreciate the helpful comments of Ramji Balakrishnan, Rebecca Hann, Bruce Johnson, D.J. Nanda, Tatiana
Sandino, Jae Yong Shin, Wim Van der Stede and seminar participants at the Management Accounting Section Mid-
Year Meeting, the University of Arizona, the University of Iowa, the University of Southern California, and the
University of Utah. Kulp gratefully acknowledges the financial support of the Harvard Business School.
Transient Institutional Ownership and the Contracting Use of Returns and Earnings
Prior work suggests that, on average, institutional ownership is associated with increased pay-
for-performance sensitivity of CEO compensation. However, other studies find that transient
institutional ownership is positively associated with myopic managerial actions (Bushee 1998),
stock price volatility, and stock mispricing (Bushee 2001). Consequently, the type of institutional
ownership may affect the performance measure properties and undermine their usefulness for
CEO contracting purposes. In this paper, we posit that transient institutional ownership is
negatively associated with the congruence and noise properties of both returns and earnings and,
hence, the use of these measures for incentive compensation. Consistent with this conjecture, we
find that transient institutional ownership is negatively associated with the sensitivity of CEO
cash compensation to both of these measures. Our findings indicate that given a 10% increase in
return on equity, the cash compensation of a CEO of a firm with low transient ownership will
increase by approximately 3%, while that of a firm with high transient ownership will only
increase by approximately 1%.
Institutional investors provide an important role in the governance of firms via
monitoring activities (e.g., voting on proposals), stock trading (e.g., by "voting with their feet,"
Parrino et al. 2003) (Gillan and Starks 2003; O’Brien and Bhushan 1990; Walther 1997), and
their influence on compensation contracts (Almazan et al. 2004; Hartzell and Starks 2003).
However, institutional investors are not all alike. Among other things (e.g., the ability and
incentive to monitor, Brickley et al. 1988), they differ in their investment horizon; while
“dedicated” institutional investors have a long-term investment horizon, “transient” investors
have a short-term horizon (Bushee 1998).
Moreover, there has been increasing concern regarding the influence of transient
institutional investors, not only on market prices, but also on management decision-making
(Porter 1992). For example, Barbara Franklin, Chairman of National Association of Corporate
Directors Blue Ribbon Commission on Director Compensation and former U.S. Commerce
Secretary (under President George H.W. Bush), states:
There is confusion because of the apparent assumption that all shareholders are
alike and have the same objectives. We all know that is not so. Relationship
investors - who buy for the long haul and are not overly sensitive to short-term
ups and downs in the stock price - are one group and are ideal from the
perspective of a director. But many investors are not of this type. There are many,
termed "financial intermediaries" by Donald Perkins in a recent speech, who buy
and sell the stock for short-term profit only and who have no interest in the health
of the enterprise in the longer term. (Franklin 1996, p. 37)
The investment horizon of transient institutional investors – the “financial intermediaries”
to which Former Secretary Franklin refers – is of particular interest because of the potential
effect that a short-term investment horizon has on the key performance measure properties (i.e.,
sensitivity, precision, and congruence) of returns and earnings. In theory, these properties help
determine the usefulness of performance measures for incentive contracting purposes. Thus,
incentive contracts should vary depending on the composition of a firm’s institutional investor
In this paper, we investigate whether the concentration of transient institutional
ownership is associated with the use of returns and earnings in incentive contracting. Prior
research documents that overall pay-for-performance sensitivity in executive contracts increases,
on average, with institutional investor concentration (Hartzell and Starks 2003). However, this
result appears to be driven by long-term institutional investors with the ability and incentive to
engage in monitoring activities (Almazan et al. 2004).
By contrast, transient institutional investors have a short-term (i.e., myopic) horizon as
indicated by their high portfolio turnover, extensive use of momentum trading strategies, and
relatively high diversification ( Bushee 1998). These types of institutional investors have little
incentive and likely little opportunity to engage in monitoring. They may, however, affect
contracting decisions, not necessarily because of their influence on firms’ contracting choices,
but because of their effect on the performance measure properties of returns and earnings, two
metrics frequently used as a basis of incentive compensation.
We assert that transient institutional ownership is associated with decreases in both the
congruity of stock returns and earnings with firm value and with the precision of stock returns.
Thus, we expect the concentration of transient institutional investors to be negatively related to
the use of stock returns and earnings in cash-based incentive compensation. We focus on cash-
based incentive compensation because stock and stock option grants are often not tied explicitly
to current measures of returns and earnings (Hall 1999). That is, stock and stock option
compensation are designed to induce managers to take actions that affect the firm’s long-term
performance, rather than as a reward for current levels of performance. Therefore, these types of
compensation are less dependent on the properties (e.g., precision and congruence) of current
returns and earnings.
Prior research shows (i) that the trading behaviors of transient investors are associated
with stock price volatility (Bushee and Noe 2000), and (ii) that transient investors myopically
price securities by over-weighting current earnings (Bushee 2001). Agency theory suggests that
the contracting usefulness of a performance measure decreases with the measure’s precision
(Banker and Datar 1989) and with the degree of congruity between the measure and firm value
(Datar et al. 2001; Feltham and Xie 1994). Taken together, we argue that the trading and pricing
behaviors of transient investors diminishes the congruence and the precision of returns. We
accordingly predict and find that the implicit contracting weight on returns is negatively
associated with the concentration of transient institutional investors. Furthermore, we test the
incremental effect of transient ownership on the implicit weight on returns after controlling for
stock return volatility. The results support the argument that lower congruence, in addition to
increased noise, is associated with a decreased weight on returns in cash compensation. These
results hold after controlling for the effects of CEO age, the use of stock and option pay, and
other economic determinants of pay-for-performance sensitivity, suggesting an incremental
effect of investor composition over variables used in prior research on pay-for-performance
Additionally, we test whether the concentration of transient investors is also related to
diminished usefulness of earnings as a basis of incentive compensation. Bushee (1998) shows
that managers of firms with high concentrations of transient investors are likely to make myopic
investment decisions (e.g., cut R&D) to increase current earnings and support stock prices.
Additionally, Matsumoto (2002) finds that firms with higher transient institutional ownership are
more likely to meet or exceed expectations at the earnings announcement. However, myopic
actions to increase current earnings may not be congruent with actions to maximize long-term
firm value. If so, current earnings are less useful for contracting. Consistent with this conjecture,
we predict and find that the relationship between CEO cash compensation and earnings is
decreasing in the transient institutional ownership. Again, this result holds after controlling for
the effect of earnings volatility and for the effects of other constructs previously shown to
influence pay-for-performance sensitivity.
This research improves our understanding of transient institutional investors’ impact on
managerial control and the structure of incentive compensation. Prior research tends to treat all
institutional investors alike. On average, institutional investors appear to strengthen corporate
control through, for example, monitoring and increased pay-for-performance sensitivities. Our
research illustrates that the presence of transient institutional owners differs from this average
effect and, in fact, may diminish the cash-based incentive compensation usefulness of both stock
returns and earnings.
This study has three important implications. First, the findings of this study suggest that
firms with high levels of transient institutional investors may invest in increased managerial
monitoring in order to preserve the congruence of earnings with firm value. Doing so would
offset a diminution in the contracting usefulness of earnings. Second, firms may identify
alternative metrics to use as the basis of cash-based incentive compensation. That is, they may
invest in the identification and collection of other performance metrics less likely to be adversely
affected by the presence of transient institutional investors. Finally, firms with high levels of
transient ownership may utilize non-cash-based forms of compensation (e.g., stocks and options)
that are not dependent on the performance measure properties of returns and earnings. All of
these actions represent rational responses to the decline in the contracting usefulness of returns
and earnings documented in this paper.
This paper is organized as follows. The following section develops the hypotheses, and
Section 3 details the sample selection, empirical variables, and descriptive statistics. Section 4
presents the empirical methodology and analysis of results. Section 5 concludes the paper.
2. Hypothesis Development
Prior research suggests that institutional owners play an important role in managerial
control in at least two ways. First, institutional investors directly influence firms via their
monitoring activities (Gillan and Starks 2000, 2003; Schipper 1989; Shleifer and Vishny 1997).
That is, because these investors have superior information gathering and processing abilities
(O’Brien and Bhushan 1990; Walther 1997), increased incentives to monitor management (i.e.,
large shareholdings), and the ability to affect change (i.e., large voting blocks) (Shleifer and
Vishny 1997), institutional investors influence management and the board of director with their
proposals and votes (Brickley et al. 1988; Gillan and Starks 1998). Second, institutional owners
exhibit influence indirectly by selling their shares (Parrino et al. 2003). Because large turnover
affects price, management may try to minimize this turnover by catering to large investors.
Importantly, prior research also documents that not all institutional investors are alike.
Institutional investors vary in their direct monitoring efforts; some institutional investors actively
monitor a firm’s activities while others play a more passive role (Almazan et al. 2004).1
Additionally, Bushee (1998) empirically classifies institutional investors into transient and long-
1 Institutional investors play a more passive monitoring role when they have incentives to retain currently held
business with the firm (e.g., debt and insurance contracts) and a more active monitoring when they have no such
conflicts of interest. Indeed, Almazan et al. (2004) find that pay-for-performance sensitivity is higher and level of
pay is lower for firms with high levels of active institutional ownership (defined as investment advisory firms such
as mutual funds), relative to passive institutional ownership (defined as banks and insurance companies).
term based on their past trading behaviors. The long-term investors (including both dedicated
and quasi-indexers as defined in Bushee 1998; hereafter, long-term investors) make highly
concentrated investments, have low turnover, and exhibit little trading sensitivity to current
earnings. Prior research that documents a positive association between total institutional
ownership, governance, and incentive compensation (e.g., Hartzell and Starks 2003) likely
reflects this dominant group of long-term institutional investors.2
Transient institutional investors (Bushee 1998), on the other hand, exhibit high portfolio
turnover, extensive use of momentum trading strategies, and relatively high diversification.
Importantly, due to their trading behavior and their influence on management short-term actions,
the presence of transient investors is likely associated with altered performance measure
properties of returns and earnings and consequently, the weight of these measures in incentive
contracts. It is therefore unclear whether pay-for-performance sensitivity to these metrics will be
increasing in the concentration of transient institutional investors.
The following hypotheses detail the relationship between the type of institutional investor
base, specifically the concentration of transient institutional ownership, and the use of
performance measures (i.e., stock returns and earnings) in CEO cash compensation. We focus on
cash compensation to increase the power of our empirical tests. Often, firms design employee
stock and stock option plans as either a fixed number or a fixed dollar value of grants at each
grant date (Hall 1999). Thus, these types of compensation are less dependent on the properties
(e.g., precision and congruence) of current returns and earnings. As we are interested in studying
the interaction between the characteristics of the performance measures and the firm’s use of the
performance measures given the investor base, we concentrate on the effect of transient
2 In Bushee (1998) 74 percent of the sample was classified as long-term (i.e., dedicated or quasi-indexer) and 26
percent was classified as transient.
institutional ownership on the use of returns and earnings in cash compensation (an approach
similar to Lambert and Larcker 1987). In the following sections we describe our hypotheses.
2.1 CASH COMPENSATION SENSITIVITY TO STOCK RETURNS
We predict a negative association between the concentration of transient institutional
investors and the implicit contracting weight placed on stock returns for two reasons. First, prior
research suggests that the optimal contracting weight on a performance measure decreases with
the noise (i.e., increases with the precision) of the measure (Aggarwal and Samwick 1999;
Banker and Datar 1989; Lambert and Larcker 1987). Transient institutional investors engage in
frequent trading and extensively use momentum trading strategies, which increase the noise in
stock prices relative to firms without such trading activity (Bushee and Noe 2000). This, in turn,
makes stock returns less effective and/or more costly for use in incentive contracting. It follows
that firms will reduce the implicit contracting weight placed on returns in the presence of high
concentrations of transient institutional investors.
Second, in addition to changes in volatility, the presence of transient institutional
investors also impacts the congruence between stock returns and long-term firm value. Congruity
(or congruence) refers to the correlation between a performance measure, or a group of weighted
performance measures, and firm value. Feltham and Xie (1994) find that the optimal contracting
weight of a performance measure increases as the congruence between that performance measure
and firm value (i.e., the firm’s objective function) increases. As the “match” between the
performance measure and firm objectives increases, the performance measure better motivates
the manager to take actions consistent with the firm’s goals and is thus weighted more heavily.
Bushee (2001) documents that the concentration of transient investors is positively
(negatively) associated with the amount of expected short-term (long-term) earnings impounded
in stock price, consistent with myopic pricing by transient institutional investors. Furthermore,
Bushee (2001) finds that transient investors over-weight current expected earnings in their
pricing of securities; a trading strategy based on this pricing myopia generates significant
abnormal returns. These results imply that stock returns are a weaker reflection of long-term
value creation in the presence of high concentrations of transient institutional investors. Since the
optimal contracting weight of a performance measure decreases (ceteris paribus) as the degree of
congruity between the measure and firm value declines, returns become less useful valuable for
contracting purposes with increased pricing myopia induced by transient investors. We,
H1: The concentration of transient institutional ownership is negatively associated with
the sensitivity of CEO cash compensation to stock returns.
Almazan et al. (2004) show that pay-for-performance sensitivity is positively associated
with the concentration of active institutional ownership (defined as investment advisory firms
such as mutual funds), but unrelated to the concentration of passive institutional ownership
(defined as banks and insurance companies). The classification of transient ownership used in
this paper is separate and distinct from the passive institutional ownership used in Almazan et al.
(2004). Additionally, Hypothesis H1 predicts not just a weaker positive (or no) relation between
the concentration of transient investors and pay-for-performance sensitivity, but a negative
2.2 CASH COMPENSATION SENSITIVITY TO ACCOUNTING EARNINGS
Managers have an incentive to increase earnings for many reasons including their cash
compensation being tied to earnings. For example, reputation and career concerns provide
incentives to maintain earnings persistence to support the firm’s stock price. The myopic pricing
pressures exerted by transient investors described above may induce managers to engage in
myopic decision-making at the expense of long-term firm value (Bushee 2001). Indeed, Bushee
(1998) documents that managers are more likely to make income-increasing cuts in R&D when
there is a high concentration of transient investors. Additionally, Matsumoto (2002) finds that
firms with high transient institutional ownership are more likely to meet or beat earnings
The evidence suggests that transient institutional investors can provide incentives for
managers to make myopic decisions that positively affect current earnings but adversely affect
long-term firm value. Managerial myopia arising from transient investor influence thus causes
earnings to less congruent with long-term value maximization and thus, less useful for
contracting. Accordingly, we hypothesize:
H2: The concentration of transient institutional ownership is negatively associated with
the sensitivity of CEO cash compensation to accounting earnings (ROE).
Note that in the above hypotheses, we predict declines in the implicit contracting weights
placed on both stock returns and earnings. We make no predictions about changes in the relative
weights placed on these measures in CEO cash compensation. As earlier noted, an important
implication of this study is that firms may look for alternative sources of pay-for-performance
sensitivity (e.g., stock or options) since results supporting our hypotheses would suggest an
overall decline in the pay-for-performance sensitivity that can be achieved with cash
compensation based only on returns and earnings.
Research Setting and Data Description
Our initial sample consists of the universe of ExecuComp firms during the years 1992 to
2000. From this database, we collect components of CEO cash compensation including salary,
bonus, and other annual payments as well as information regarding the CEO’s total wealth in the
company. From the Center for Research in Security Prices (CRSP), Compustat, and
CDA/Spectrum, we collect stock returns, accounting information, and institutional investor
holdings (i.e., from SEC Form 13f filings), respectively. Finally, we hand-collect the CEO’s age
and information regarding board characteristics (e.g., size and percent of outside directors) from
the firm’s proxy statements.3
We exclude firm-year observations with missing data. In addition, because mergers,
acquisitions, and bankruptcies affect the composition of the investor base and the firm’s
incentive plan, we drop from the sample all years in which one of these events occurs. Finally,
because we analyze the change in compensation for the pay-for-performance tests, firms must
have the pertinent information for at least two consecutive years to be retained for analysis. The
final sample consists of 1,897 firm-year observations. Table 1, Panel A describes the reasons for
data loss. Panel B classifies the sample by industry.
/Insert Table 1/
3.2.1 Dependent Variable
Change in Cash Compensation. The dependent variable used to test the hypotheses is
the change in the natural log of the CEO’s cash compensation defined as salary, bonus, and other
annual compensation (∆LnSBOA). Similar to prior studies (e.g., Lambert and Larcker 1987;
Sloan 1993), we measure the contracting use of accounting (return on common equity, ROE) and
non-accounting (returns, RET) performance measures as the implicit incentive weights (i.e.,
coefficients) in a regression of changes in cash compensation on those performance measures.
3 For a random-sample of observations, we verify the compensation information in ExecuComp with the proxy data.
3.2.2 Independent Variables
Performance Measures. We investigate the sensitivity of CEO cash compensation to two
performance measures, annual stock returns, RET, and accounting earnings. Earnings is measured
as the change in return-on-equity, ∆ROE, where ROE defined as net income before extraordinary
items and discontinued operations divided by average book value of common equity.
Institutional Investor Classifications. Bushee (1998; 2001) classifies institutional
investors based on the average size of the institution’s stake in the firm, the degree of portfolio
turnover, and the trading sensitivity to current earnings news (i.e., momentum trading). Cluster
analysis by Bushee (1998) indicates three groups of institutional investors defined as “transient,”
“dedicated,” and “quasi-indexers.” Transient institutions are characterized by the highest
turnover, the highest use of momentum trading strategies, and relatively high diversification.
Dedicated institutions have highly concentrated investments, low turnover, and almost no trading
sensitivity to current earnings. The quasi-indexers are highly diversified and typically follow a
Bushee (1998, 2001) suggests that transient institutional investors have systematic
preferences for short-term managerial actions not evident for dedicated and quasi-indexers. Thus,
we use Bushee’s (1998) classification methodology to compute the percentage of transient
institutional investors (%TRA). We combine the Bushee (1998) categories of dedicated and
quasi-indexers into one subsample of “long-term” institutional investors (i.e., %LT). The
variable, %LT, is used as a control in the tests of H1 and H2.
All institutional investor concentrations are measured as of the beginning of the fiscal
4 The classification of transient ownership by Bushee (1998) and used in this paper is separate and distinct from the
passive/active institutional ownership classifications used in Almazan et al. (2004). Importantly, the correlation
between the concentration of passive and transient ownership is only 0.15 (p < .01) in our sample. For more
evidence on differences between these classifications see Bushee (2001).
year. Note that, although institutional investors frequently fall out of the transient classification
(Bushee 1998), the total concentration of transient ownership for a given firm is relatively stable
through time as indicated by a high correlation (0.89) between %TRA and one-quarter-lags in
Control variables. We control for other variables that may affect the level of CEO cash
compensation CEO cash compensation (e.g., firm size, SIZE, measured as the log of the market
value of equity) and/or the sensitivity of CEO cash compensation to performance (e.g., logged
level of annual restricted stock grants, LnSTOCK, and of option grants, LnOPTION). In
particular, we control for the level CEO ownership which is calculated as the sum of the
executive’s stock option value, restricted stock value, and common stock holdings. We scale this
sum by the firm’s total market value and take the log (LnOWNERSHIP). We include CEO age
(CEOAGE) to control for CEO horizon effects on contracting as documented in prior research
(Dikolli 2001; Dikolli et al. 2004). We also include controls for the board size (BOARDSIZE)
and an indicator variable for when the CEO is the chairman of the Board of Directors
(CEOCHAIR). Prior work demonstrates that firms with the CEO as chairman or with larger
boards pay their CEO a higher compensation (Core et al. 1999; Yermack 1996). Finally, year
indicators and 2-digit industry indicators are included in all regressions to control for pay
similarities within industries and trends over time.
Descriptive statistics for the sample of 1,897 firm-year observations are presented in
Table 2, Panel A (variable definitions in Panel C). On average (median), institutional investors
make up 50% (51%) of the investor base. The average (median) level of transient institutional
ownership is 11% (8%); the average (median) level of long-term institutional ownership is 38%
(38%). Table 2, Panel B presents the correlation table.
/Insert Table 2/
We test hypotheses H1 and H2 using the following empirical model.
∆LnSBOAit = β0 + β1RETit + β2RETit*%TRAit + β3RETit*%LTit
+ β4∆ROE it + β5∆ROEit*%TRAit + β6∆ROEit*%LTit
+ controls + εit
where ∆LnSBOA is the change in the log of annual cash compensation and %TRA
(%LT) is the percentage of outstanding shares held by transient (long-term) institutional
investors as defined by Bushee (1998). We use a changes specification to control for unidentified
correlated omitted variables. As in prior studies (e.g., Lambert and Larcker 1987; Sloan 1993),
we interpret the weights on the financial performance variables, returns (RET) and return on
equity (∆ROE), as the sensitivity of the CEO’s cash pay to firm performance. We control for the
effect of CEO horizon on pay-for-performance sensitivity to returns and earnings (Cheng 2004;
Dikolli et al. 2004) by including as controls terms interacting RET and ∆ROE with CEO_AGE.
We further examine how pay-for-performance sensitivity varies with the concentrations of
transient and long-term institutional investors by interacting the performance variables with
%TRA and %LT. We mean-center the independent variables (other than indicator variables) to
simplify the interpretation of the intercept and interaction and to reduce multi-collinearity arising
from the introduction of interactions among continuous variables (Aiken and West 1991).5
We allow for the possibility that the concentrations of transient and long-term
5 Indeed, there is no indication of multicollinearity in any of the empirical results reported (i.e., variance inflation
factors are less than ten, Kennedy 1997). We also run the analyses using uncentered variables. Although the results
and implications are qualitatively the same, the variance inflation factors are high, indicating multicollinearity.
institutional investors are endogenous to the contracting choices firms make. In particular,
investors may be attracted to firms with certain contracting methods or to other firm
characteristics correlated with those contracting choices. If so, an ordinary least squares (OLS)
estimation of the relation between cash compensation and %TRA will yield inconsistent
coefficient estimates (Greene 2002). We, therefore, present all analyses with both the OLS
estimates as well as estimates produced using two-stage least squares (2SLS) as recommended
by Larcker and Rusticus (2004).
4.1 INSTRUMENTAL VARIABLES ESTIMATION
In this section we describe the instrumental variables (IV) estimations of %TRA and
%LT. Bushee (2001) examines the preferences of transient and long-term investors. We draw
from his findings to identify suitable instrumental variables for %TRA and %LT for the first-
stage of our 2SLS estimation procedure. Suitable variables are those that are (preferably highly)
correlated with the endogenous variables but uncorrelated with the error term in the structural
We identify twelve instrumental variables based on Bushee (2001) for use in the first
stage estimation of %TRA and %LT. First, we use an indicator for whether or not the firm is in
the S&P 500 common stock rating (labeled, S&P500) and the S&P common stock rating
(RATE). These variables are expected to be associated with institutional holdings because
institutional investors often index a portion of their holdings. Bushee (2001) further documents
that institutional investors are attracted to firms with high R&D intensity (RDINTENSITY) and
high liquidity (LIQUIDITY). Long-term and transient institutional investors are also expected to
have preferences with respect to dividend yield (YIELD).
Three variables capturing firm risk are also expected to be associated with the level of
institutional holdings: (i) systematic risk as measured by the firm’s market-model beta (BETA),
(ii) unsystematic risk as measured by the standard deviation of market-model residuals (IRISK),
and (iii) the ratio of debt to assets (LEVERAGE) because increased leverage increases investor
losses in the event of bankruptcy. Finally, institutional holdings are expected to be positively
associated with recent performance. We therefore include an indicator equal to one if prior year
earnings is positive (DPOS), a measure of prior three-year sales growth (SGR), the prior year
market-adjusted return (MAR), and the one-year lag of total returns (Lag RET). These variables
are justified as exogenous (to compensation) IVs because they are lagged relative to the
dependent variable of interest, ∆LnSBOA (Kennedy 1997). MAR is further justified because
prior research suggests firms do not engage in relative performance evaluation in setting CEO
compensation (e.g., Aggarwal and Samwick 1999). In sum, the first-stage models estimated are
<DV>it = β0 + β1S&P500it + β2RATEit + β3RDINTENSITYit + β4LIQUIDITYit
+ β5LEVERAGE it + β6YIELDit + β7BETAit + β8IRISKit
+ β9DPOSit + β10SGRRit + β11MARit + β12Lag RETit
+ <exogenous> + εit
where <DV> is either %TRA or %LT and <exogenous> includes all the exogenous
variables in equation (1), as recommended (e.g., Gujarati 2003; Kennedy 1997). To the extent
these twelve variables explain the concentration of transient and long-term investors and are
uncorrelated with the residuals from the structural model (1), they are suitable as instrumental
variables. We test this presumption with an overidentifying restrictions test described below.
The results of the estimation of (2) are presented in Table 3.6 To assess the degree of
6 Note that in some instances the signs of several of the coefficients differ from those found in Bushee (2001). This
is likely due to differences between our study and Bushee (2001) in both the samples and in the models estimated
endogeneity present, we conduct a Hausman (1978) specification test and find that there is, in
fact, endogeneity indicated between ∆LnSBOA and the institutional investor concentration
variables, %TRA and %LT (Fdf=2,730 = 8.46, p<.01).7
/Insert Table 3/
We next assess the relevance and exogeneity of our chosen IVs. First, the Adjusted-R2 of
the full model (including all exogenous variables) is 40.92% and 24.16% for %TRA and %LT,
respectively. This is in line with the mean (median) value of similar accounting studies of 31%
(26%) reported by Larcker and Rusticus (2004). The partial Adjusted-R2 (including only the
twelve IVs), however, is more relevant to determining the efficacy of our IV estimation. The
%TRA and %LT models have a partial Adjusted-R2 of 22.51% and 7.43%, respectively. These
are again in line with what is reported in Larcker and Rusticus (2004) and indicate an acceptable
level of relevance of our IVs in predicting the endogenous variables.8 Second, the
overidentifying restrictions test (Larcker and Rusticus 2004) fails to reject the null of exogenous
instruments (χ2df=9 = 4.55, p=.87).9 Thus, our first-stage IV estimation appears to provide
reasonable support for the proposition that 2SLS will be an improvement over OLS in the
estimation of equation (1).
Note that equation (1) includes interaction terms between the endogenous variables,
%TRA and %LT, and variables assumed to be exogenous (e.g., RET and ∆ROE). Thus, in the
(recall, our model includes all the exogenous variables from the structural model in equation (1)). We are not,
however, interested in interpreting the coefficient estimates. Rather, the purpose is to generate predicted values of
%TRA and %LT for use in the second-stage estimation of the structural model.
7 We conduct this test using a reduced form of the structural model without the interactions.
8 In order for the 2SLS estimation to be an improvement over the OLS estimates, it must be the case that the
correlation between the IVs and the error in the structural model (1) must be less than 22.51% (7.43%) of the
correlation between %TRA (%LT) and the structural model error (Larcker and Rusticus 2004). Thus, either the
degree of endogeneity between %TRA (%LT) and ∆LnSBOA must be relatively high, or the degree of exogeneity
of the IVs must be relatively high (i.e., correlation between the IVs and the error in the structural model).
9 Note that one additional variable, the time the firm has been listed, was originally included as an IV. However, this
variable was subsequently omitted from stage one because the overidentifying restrictions test rejected the null that
this variable was exogenous.
second-stage of the 2SLS estimation we compute the Heckman and Vytlacil (1998) 2SLS
estimator. This involves taking the predicted values from the first-stage and using them to
compute the interactions for the second-stage estimation. For completeness, we present the
results of both the OLS and 2SLS estimations of equation (1) used to test Hypotheses H1 and H2
over the next two subsections, respectively.
4.2 HYPOTHESIS H1
Hypothesis H1 predicts that the pay-for-performance sensitivity of CEO compensation to
returns is negatively associated with the concentration of transient institutional investors. We
argue, in part, that the trading behaviors of transient investors induce noise in stock returns
thereby decreasing the contracting usefulness of returns. Prior research documents an association
between transient investor concentration and stock price volatility (Bushee and Noe 2000). The
positive (negative) correlation between transient (long-term) institutional investor ownership and
stock return volatility is evident in the correlation table presented in Table 2, Panel B.
Hypothesis H1 predicts that the sensitivity of cash compensation to returns is decreasing
in the concentration of transient institutional ownership, %TRA; that is, β2 in equation (1) will be
negative. Moreover, prior research that assumes a long-term horizon for institutional investors
suggests that pay-for-performance sensitivity is increasing in the concentration of institutional
ownership (Hartzell and Starks 2003). Consistent with this research, we expect the sensitivity of
cash compensation to returns to be increasing in the concentration of long-term institutional
ownership, %LT. That is, we expect β3 in equation (1) to be positive.
/Insert Table 4/
The estimation of equation (1) is presented in Table 4. The results in Models 1 (OLS) and
2 (IV) indicate that CEO pay is increasing in stock performance as indicated by the positive
coefficients on RET (coefficient of 0.165 in Model 1 and 0.181 in Model 2, p < .01, one-tailed).
The results further indicate that the sensitivity of CEO cash compensation to stock return
performance is, indeed, decreasing in the concentration of transient institutional ownership.
Specifically, the coefficients on the interaction term RET*%TRA in both Model 1 (β2 = -0.307, p
< .01, one-tailed) and Model 2 (β2 = -0.348, p < .05, one-tailed) is significantly negative. Note
that the coefficient on the interaction term RET*%LT in Model 1 is significantly positive (β3 =
0.144, p < .05, one-tailed); it is positive but insignificant in Model 2. Thus, although the
sensitivity of cash compensation to stock returns may be increasing in the concentration of long-
term institutional owners consistent with prior research (Hartzell and Starks 2003), it is
decreasing in the concentration of transient institutional investors, consistent with Hypothesis
4.3 HYPOTHESIS H2
Hypothesis H2 states that the sensitivity of CEO cash compensation to accounting
earnings is also decreasing in the concentration of transient ownership. We reason that the
myopic pricing behavior (Bushee 2001) of transient investors pressures managers to make short-
term decisions to increase current earnings at the expense of long-term firm value.10 Hypothesis
H2 accordingly predicts that the contracting weight placed on earnings will be decreasing in the
concentration of transient institutional investors.
10 Consistent with this notion, in untabulated results we find that (i) the number of quarterly consensus analyst
earnings forecasts that a firm beats, (ii) the probability of beating the consensus annual forecast, and (iii) the
probability of beating each of the four quarterly consensus forecasts (consistent with Matsumoto 2002) is increasing
in the concentrations of both transient and long-term investors. Importantly, this effect is significantly stronger for
transient institutional investor concentration relative to long-term investor concentration. This result is not by itself
conclusive, but it is consistent with prior research (Bushee 1998) documenting myopic decision-making in firms
with high concentrations of transient investors. Another interpretation of this result is that managers of firms with
high concentrations of transient investors work more aggressively to manage analyst expectations downward in
order to increase the probability of meeting or beating analysts’ forecasts. This would not necessarily lead to a
decline in earnings contracting usefulness and, hence, works against our finding evidence to support Hypothesis H2.
The estimation of equation (1) presented in Table 4, Models 1 (OLS) and 2 (IV) allows
us to test H2. The results are consistent with H2. Specifically, although cash compensation is
increasing in ∆ROE (coefficients of 0.170 and 0.163 for Models 1 and 2, respectively, p < .01,
one-tailed) and this effect is increasing in the concentration of long-term institutional investors in
Model 1 (coefficients of 0.280, p<.10), the coefficient on the %TRA*∆ROE interaction term is
negative (-1.045 and -1.309 in Models 1 and 2, respectively) and statistically significant (p < .01,
one-tailed). Thus, consistent with H2 the sensitivity of cash compensation to accounting earnings
is decreasing in the concentration of transient institutional investors.
Finally, we note that transient investors may be attracted to firms that substitute stock
and stock option compensation for cash compensation. Thus, in all models of Table 4 we control
for this possibility by including four interaction variables, between each of RET and ∆ROE with
LnOPTION and LnSTOCK. The option and stock grant variables are intended to capture the
extent to which the firm uses stock and stock option compensation. If stock and stock options are
indeed substitutes for cash in compensation plans, we would expect these interaction terms to
have a negative coefficient.
Note that in both Models 1 and 2, the RET-LnOPTION, RET-LnSTOCK, and ∆ROE-
LnSTOCK interaction terms are insignificantly different from zero (or weakly significant at
p<.15). However, the coefficient on ∆ROE–LnOPTION in Model 1 is negative and significant in
Model 2. This implies that the sensitivity of cash compensation to earnings is decreasing in the
use of stock options. Importantly, our results are robust to the inclusion of control variables
capturing the substitution of stock and option compensation for cash compensation. In addition,
results are also robust to the inclusion of controls for the effects of (i) high-tech firms, (ii)
“passive” institutional investors (Almazan et al. 2004), (iii) investment opportunities (Baber et
al. 1996), and (iv) levels of CEO ownership in the firm on cash pay-for-performance sensitivities
(i.e., results hold even with the addition of terms interacting RET and DROE with variables
capturing these influences).
ECONOMIC SIGNIFICANCE OF H1 AND H2 RESULTS
Figure 1 presents an economic interpretation of the returns sensitivity (H1) and earnings
sensitivity (H2) results. Holding all other variables constant at their means (recall, variables are
mean-centered), we plot the percent change in cash compensation as a function of RET for three
levels of %TRA (2%, 8% and 23%, Panel A) corresponding to the 10th, 50th, and 90th percentiles
of %TRA. We similarly plot the percent change in cash compensation as a function of RET for
three levels of %LT (30%, 39% and 47%, Panel B) corresponding to the 10th, 50th, and 90th
percentiles of %LT.
As can be seen in Figure 1, Panel A the percent change in compensation as a function of
RET is decreasing in the concentration of transient ownership as indicated by the smaller slope
of the 90th percentile %TRA line (as compared to the other two lines). In Figure 1, Panel B,
however, the reduction in the slope of the high (90 percentile) %LT line is not evident and, in
fact, the slope of this line is actually larger than the slopes of the 10th and 50th percentile %LT
/Insert Figure 1/
In Panels C and D of Figure 1 we plot the percent change in cash compensation as a
function of ∆ROE for three levels of %TRA (2%, 8% and 23%, Panel C) corresponding to the
10th, 50th, and 90th percentiles of %TRA. We similarly plot the percent change in cash
compensation as a function of ∆ROE for three levels of %LT (30%, 39% and 47%, Panel D)
corresponding to the 10th, 50th, and 90th percentiles of %LT.
Again, in Figure 1, Panel C the percent change in compensation as a function of ∆ROE is
decreasing in the concentration of transient ownership as indicated by the smaller slope of the
90th percentile %TRA line (as compared to the other two lines). In Figure 1, Panel D, however,
there is no reduction in the slope of the high (90 percentile) %LT line relative to the slopes of the
10th and 50th percentile %LT lines. As an example given a 10% increase in return on equity (i.e.,
as ∆ROE increases from 0% to 10%), the percent change in cash compensation of a CEO of a
firm with low (i.e., 2%) transient ownership will increase by approximately 3%, while that of a
firm with high (i.e., 23%) transient ownership will only increase by approximately 1%. Thus, the
percent change in cash compensation is less sensitive to changes in ∆ROE when transient
institutional ownership is high.
SUPPLEMENTARY ANALYSIS – THE EFFECT OF STOCK RETURN AND EARNINGS VOLATILITY
Recall that Hypothesis 1 predicts that the sensitivity of cash compensation to stock
returns is decreasing in the concentration of transient ownership because of the negative effects
of transient investor behavior on both the congruence and the noise properties of stock returns. In
Models 3 (OLS) and 4 (IV) of Table 4 we add a control for the direct effect of stock price
volatility on the sensitivity of cash compensation to stock returns. Specifically, we add a term
interacting RET and RET_VOLAT. This specification allows us to determine if the association
between transient institutional ownership and cash compensation sensitivity to returns
incremental to the direct effect of volatility on pay-for-performance sensitivity documented in
prior research (e.g., Ittner et al. 1997). We have no ex ante reason to predict a change in earnings
volatility as a result of transient institutional ownership. However, for completeness we also
include an analogous interaction between ∆ROE and EARN_VOLAT to determine if there is a
decline in the implicit weight on earnings incremental to any earnings volatility effect that
transient owners might have.
These interaction terms (RET-RET_VOLAT and ∆ROE-EARN_VOLAT) introduced in
Table 4, Models 3 and 4 are, as expected, negative (and at least weakly significant in all four
cases). This suggests that, consistent with prior research, the implicit weights on stock returns
and earnings are decreasing in the noise in these measures. Importantly, the negative effects of
transient institutional investors on the sensitivity of cash compensation to returns and earnings
documented in the previous section hold as indicated by the statistically significant coefficient on
both the RET-%TRA interaction (coefficients of -0.267, p < .01, and -0.286, p<.10 in Models 3
and 4 respectively) and the ∆ROE-%TRA interaction (coefficients of -1.046, p<.01 and -1.222,
p<.01 in Models 3 and 4, respectively). This is consistent with the negative association between
the concentration of transient investors and the sensitivity of cash compensation to stock returns
(H1) and earnings (H2) being driven, not only by the volatility in these measures associated with
transient investor ownership, but also by a decrease in the congruence of stock price and earnings
in the presence of high concentrations of transient institutional investors.
Prior research suggests that institutional investors act, on average, as monitors of a firm’s
actions. That is, because of increased incentives to monitor management (i.e., large
shareholdings) and the ability to effect change (i.e., large voting blocks), these investors play an
important role in corporate governance (Shleifer and Vishny 1997). Although prior empirical
studies document associations of institutional investors as a whole to the structure of
compensation contracts, few studies explicitly consider differences in institutional investors. In
this paper, we study how differences in the investment horizons of institutional investors may
affect the properties of stock returns and earnings and thus, alter their use in incentive
Transient institutional investors, by engaging in frequent (momentum) trading (Bushee
2001) and by overweighting current earnings in pricing securities (as, for example, in Bushee
1998), increase the noise in and decrease the congruence of stock price. Consequently, stock
returns are less effective as a basis for incentive compensation. We, accordingly, expect and find
evidence that the implicit weight on returns in CEO cash compensation is negatively associated
with transient ownership concentration.
Prior work documents how short-term (transient) owners create incentives for managers to
make short-term decisions in order to increase current earnings. When this occurs (as, for
example, in Bushee 1998), the earnings number becomes less congruent with firm objectives,
and, consequently, less useful for contracting. Accordingly, we document that transient
ownership concentration is negatively correlated with the contracting weight placed on earnings.
Interestingly, we find the opposite effects for long-term institutional investors; that is, the
sensitivity of cash compensation to both returns and earnings is increasing in the concentration of
long-term investors. Thus, the two types of institutional investors provide offsetting effects on
the structure of incentive contracts.
The primary limitation of this study is its focus on cash compensation. We do so because we
are interested in studying the usefulness of returns and earnings as bases of incentive
compensation, given a firm’s investor base. Since annual stock and stock option grants are not
typically tied to performance metrics, these types of compensation are less dependent on the
properties (e.g., precision and congruence) of current returns and earnings. Note that our study
does not necessarily imply that overall pay-for-performance sensitivity declines as the
concentration of transient institutional investors increases. Indeed, the results in this study lead to
important implications for firms with high levels of transient institutional ownership.
In particular, firms must be aware of the decreased usefulness of both stock returns and
earnings associated with the presence of transient institutional owners and may need to find
alternative ways to achieve desired levels of pay-for-performance sensitivity. They may
substitute other forms of compensation such as stock and stock options. Alternatively, firms may
use alternative performance measures less likely to be adversely affected by transient investor
behavior. Investment in the identification and collection of such measures, although costly,
might very well provide for more efficient contracting. Finally, and perhaps most importantly,
firm might invest in alternative control mechanisms such as strengthening the board of directors
to facilitate increased monitoring of managers. Increased monitoring may serve to prevent
myopic decision-making on the part of managers and, thus, preserve the congruence of earnings,
even in the presence of myopic incentives transient investors provide. Overall, increased demand
for alternative control mechanisms in the presence of transient ownership is an important
implication of the current study that warrants additional research.
Panel A: Sample Construction
Number of ExecuComp observations after deleting observations (1) involved in
mergers, acquisitions, and bankruptcies, and (2) with missing Execucomp data
(e.g., CEO age).
Number of observations with missing Spectrum and IRRC data.
Number of observations with missing governance data.
Number of observations with missing CRSP and Compustat data.
Number of outlier observations trimmed.b
Total sample size
a We required five consecutive years of price and financial data to compute price and earnings volatility.
a To control for the influence of outliers, the top and bottom one percent of the sample distribution for total cash
compensation and returns were trimmed. Observations with negative MTB were also omitted. The results are robust
to alternative trimming procedures.
Panel B: Classification by Industry
NAICS Description Frequency Percent
Agriculture & Mining
Transportation and Warehousing
Insurance & Real Estate
Panel A: Variable Distributions (N = 1,897)
Instruments for %TRA and %LT
Std Dev Min
661.63 1,616.76 192,244.00
Panel B: Pearson Correlations
Bold type represents statistical significance at the .05 level.
Panel C: Variable Definitions
≡ percentage of shares outstanding held by institutional investors (all types) per
≡ percentage of shares outstanding held by transient (i.e., short-term)
institutional investors per Bushee (1998);
≡ the predicted value of %TRA (using instrumental variables);
≡ percentage of shares outstanding held by dedicated or quasi-indexer (i.e.,
long-term) institutional investors per Bushee (1998);
≡ the predicted value of %LT (using instrumental variables);
≡ LnSBOAt – LnSBOAt-1, where SBOA is the sum of CEO salary, bonus and
other annual income (as defined in ExecuComp and including, for example,
perquisites, other personal benefits, and above-market earnings on restricted
stock options or deferred compensation) (in thousands);
≡ the log of the Black-Scholes valuation of the annual stock option grant (in
thousands, as computed in ExecuComp);
≡ the log of the value of the annual restricted stock grant (in thousands, as
defined in ExecuComp);
≡ ROEt – ROEt-1, where ROE is the accounting return on equity defined as net
income before extraordinary items and discontinued operations divided by
average book value of common equity;
≡ annual stock return;
≡ 1 if the CEO is also the board chairman, 0 otherwise;
≡ number of directors on the board;
≡ the log of BOARDSIZE;
≡ log of market value of equity;
≡ assets (in millions);
≡ age of the CEO;
≡ a measure of CEO’s overall ownership in the firm as a percent of total market
value; this measure is calculated as log (total holdings / total market value),
where total holdings includes stock, restricted stock and options;
≡ the stock price volatility from the Execucomp database, computed as the
standard deviation of monthly price changes over 60 months;
≡ earnings volatility, computed as the standard deviation of annual earnings
(scaled by average assets) over five years;
Instruments for %TRA and %LT
≡ 1 if the firm is in the S&P 500, 0 otherwise.
≡ S&P common stock rating (9 = A+ … 0 =not rated);
≡ R&D divided by sales for the prior year;
≡ log of average monthly volume divided by shares outstanding over prior year;
≡ debt divided by total assets for prior year;
≡ dividend yield over the prior year;
≡ market model beta estimated from up to 36 prior monthly returns;
≡ unsystematic risk (standard deviation of daily market-model residuals over
≡ 1 if firm’s earnings are greater than zero in the prior year, zero otherwise;
≡ average sales growth over prior three years;
≡ market-adjusted returns over prior year;
≡ annual stock return over prior year.
Instrumental Variables Prediction of Transient and Long-Term
Institutional Investor Concentration
2nd Stage Exogenous Variables
F-Statistic (full model including all exogenous variables)
F-Statistic for joint test of instrumental variables
Adjusted-R2 (full model including all exogenous variables)
Partial Adjusted-R2 (model with only instrumental variables)
*, **, *** represents statistical significance at the .10, .05, .01 levels, respectively (two-tailed). Froot’s (1989) robust standard
errors are used to compute p-values. Year indicators and 2-digit NAICS industry indicators (suppressed) are included in “2nd stage
a To control for the influence of outliers, the top and bottom one percent of the sample distribution for SBOA and RET were
trimmed. Observations with negative MTB were also omitted. The results are robust to alternative trimming procedures. Variable
definitions in Table 2, Panel C. Expected signs given for %TRA (%LT) are for those variables that were found to be significant for
%TRA (both % quasi-indexers and % dedicated) in Bushee (2001).
The Association Between Transient Institutional Investor Concentration
and the Sensitivity of CEO Cash Compensation to Returns and Earnings (H1 & H2)a
RET * %TRA
RET * %LT
RET * CEOAGE
RET * LnOPTION
RET * LnSTOCK
RET * RET_VOLAT
∆ROE * %TRA
∆ROE * %LT
∆ROE * CEOAGE
∆ROE * LnOPTION
∆ROE * LnSTOCK
∆ROE * EARN_VOLAT
F-Statistic for test of:
RET*%TRA = RET*%LT
∆ROE*%TRA = ∆ROE*%LT
~, *, **, *** represents statistical significance at the .15, .10, .05, .01 levels, respectively (one-tailed for coefficients with predicted
signs, two-tailed otherwise). Froot’s (1989) robust standard errors are used to compute p-values. Year indicators and 2-digit NAICS
industry indicators (suppressed) are included as controls.
a To control for the influence of outliers, the top and bottom one percent of the sample distribution for SBOA and RET were
trimmed. Observations with negative MTB were also omitted. The results are robust to alternative trimming procedures. To reduce
multi-collinearity arising from the introduction of interaction terms (Aiken and West 1991) and for ease of intercept interpretation,
all independent variables (except indicator variables) are mean-centered. Variable definitions in Table 2, Panel C. Models 2 and 4
are estimated using the predicted values of %TRA and %LT from the model estimated in Table 3.
Economic Significance of the Association Between Institutional Investor Concentration
and the Sensitivity of CEO Cash Compensation to Returns and Earnings (H1 & H2)a
Panel A: %TRA and the Sensitivity of Cash Compensation to Returns (RET)
-30%-20% -10% 0% 10% 20%30% 40%50% 60% 70%
% Change in Cash Compensation
%TRA = 23% [90th percentile]
%TRA = 8% [50th percentile]
%TRA = 2% [10th percentile]
Panel B: %LT and the Sensitivity of Cash Compensation to Returns (RET)
-30% -20%-10% 0% 10%20% 30% 40% 50% 60% 70%
% Change in Cash Compensation
%LT = 47% [90th percentile]
%LT = 39% [50th percentile]
%LT = 30% [10th percentile]
Panel C: %TRA and the Sensitivity of Cash Compensation to Earnings (∆ ∆ROE)
-10% -8%-6%-4% -2% 0%2%4%6% 8% 10%
% Change in Cash Compensation
%TRA = 23% [90th percentile]
%TRA = 8% [50th percentile]
%TRA = 2% [10th percentile]
Panel D: %LT and the Sensitivity of Cash Compensation to Earnings (∆ ∆ROE)
-10% -8% -6% -4%-2%0% 2%4%6% 8%10%
% Change in Cash Compensation
%LT = 47% [90th percentile]
%LT = 39% [50th percentile]
%LT = 30% [10th percentile]
a Figure 1 presents an economic interpretation of the returns (H1) and earnings (H2) sensitivity results (Table 5,
Model 2). Holding all other variables constant at their means (recall, variables are mean-centered), we plot the
percent change in cash compensation as a function of RET (Panels A and B) and ∆ROE (Panels C and D) for three
levels of %TRA (Panels A and C) and of %LT (Panels B and D) corresponding to the 10th, 50th, and 90th percentiles
of those variables.
Aggarwal, R. K. and A. A. Samwick. 1999. The other side of the trade-off: The impact of risk on
executive compensation. Journal of Political Economy 107 (1): 65-105.
Aiken, L. S. and S. G. West. 1991. Multiple regression: Testing and interpreting interactions.
London: Sage Publications.
Almazan, A., J. C. Hartzell, and L. T. Starks. 2004. Conflicts of interest and monitoring costs of
institutional shareholders: Evidence from executive compensation. Working Paper.
Baber, W. R., S. N. Janakiraman, and K. Sok-Hyonb. 1996. Investment opportunities and the
structure of executive compensation. Journal of Accounting & Economics 21 (3): 297-318.
Banker, R. D. and S. M. Datar. 1989. Sensitivity, precision, and linear aggregation of signals for
performance evaluation. Journal of Accounting Research 27 (1): 21-39.
Brickley, J., R. Lease, and C. Smith. 1988. Ownership structure and voting on antitakeover
amendments. Journal of Financial Economics 20: 267-292.
Bushee, B. J. 1998. The influence of institutional investors on myopic R&D investment
behavior. The Accounting Review 73 (3): 305-333.
Bushee, B. J. and C. F. Noe. 2000. Corporate disclosure practices, institutional investors, and
stock return volatility. Journal of Accounting Research 38 (3): 171-202.
Bushee, B. J. 2001. Do institutional investors prefer near-term earnings over long-run value?
Contemporary Accounting Research 18 (2): 207-246.
Cheng, S. 2004. R&D expenditures and CEO compensation. The Accounting Review 79 (2): 305-
Core, J., R. Holthausen, and D. Larcker. 1999. Corporate governance, Chief Executive Officer
compensation, and firm performance. Journal of Financial Economics 51: 371-406.
Datar, S., S. C. Kulp, and R. A. Lambert. 2001. Balancing performance measures. Journal of
Accounting Research 39 (1): 75-92.
Dikolli, S. S. 2001. Agent employment horizons and the contracting demand for forward-looking
performance measures. Journal of Accounting Research 39 (3): 481-494.
Dikolli, S. S., S. L. Kulp, and K. L. Sedatole. 2004. The use of forward-looking performance
measures to align CEO and investor horizons. Working Paper.
Feltham, G. A. and J. Xie. 1994. Performance measure congruity and diversity in multi-task
principal/agent relations. The Accounting Review 69 (3): 429-453.
Franklin, B. H. 1996. Who really is worthy of alignment? Directors & Boards (Spring): 37-38.
Froot, K. A. 1989. Consistent covariance matrix estimation with cross-sectional dependence and
heteroskedasticity in financial data. Journal of Financial and Quantitative Analysis 24: 333–355.
Gillan, S. L. and L. T. Starks. 1998. A survey of shareholder activism: Motivation and empirical
evidence. Contemporary Finance Digest 2 (3): 10-34.
Gillan, S. L. and L. T. Starks. 2000. Corporate governance proposals and shareholder activism:
The role of institutional investors. Journal of Financial Economics 57 (2): 275–305.
Gillan, S. L. and L. T. Starks. 2003. Corporate governance, corporate ownership, and the role of
institutional investors: A global perspective. Working Paper.
Greene, W. H. 2002. Econometric analysis (5th ed.). New York: Prentice Hall.
Gujarati, D. N. 2003. Basic econometrics (4th ed.). New York: McGraw-Hill.
Hall, B. J. 1999. The design of multi-year stock option plans. Journal of Applied Corporate
Finance 12 (2): 97-106.
Hartzell, J. C. and L. T. Starks. 2003. Institutional investors and executive compensation.
Journal of Finance. 58 (6): 2351-2374.
Hausman, J. A. 1978. Specification tests in econometrics. Econometrica 46 (6): 1251-1271.
Heckman, J. and E. Vytlacil. 1998. Instrumental variables methods for the correlated random
coefficient model. Journal of Human Resources 33: 974-987.
Ittner, C. D., D. F. Larcker, and M. V. Rajan. 1997. The choice of performance measures in
annual bonus contracts. The Accounting Review 72 (2): 231-255.
Kennedy, P. 1997. A guide to econometrics. Cambridge, Ma.: MIT Press.
Lambert, R. A. and D. F. Larcker. 1987. An analysis of the use of accounting and market
measures of performance in executive compensation contracts. Journal of Accounting Research
25 (Supplement): 85-129.
Larcker, D. F. and T. O. Rusticus. 2004. On the use of instrumental variables in accounting
research. Working Paper.
Matsumoto, D. A. 2002. Management's incentives to avoid negative earnings surprises.
Accounting Review 77 (3): 483.
O’Brien, P. and R. Bhushan. 1990. Analyst following and institutional ownership. Journal of
Accounting Research 28 (Supplement): 55–76.
Parrino, R., R. W. Sias, and L. T. Starks. 2003. Voting with their feet: Institutional ownership
changes around forced CEO turnover. Journal of Financial Economics 68 (1): 3-46.
Porter, M. 1992. Capital choices: Changing the way America invests in industry. Boston:
Council on Competitiveness/Harvard Business School.
35 Download full-text
Schipper, K. 1989. Commentary on earnings management. Accounting Horizons 3 (December):
Shleifer, A. and R. W. Vishny. 1997. A survey of corporate governance. Journal of Finance 52:
Sloan, R. G. 1993. Accounting earnings and top executive compensation. Journal of Accounting
and Economics 16: 55-100.
Walther, B. 1997. Investor sophistication and market earnings expectations. Journal of
Accounting Research 35 (Autumn): 157–180.
Yermack, D. 1996. Higher market valuation of companies with a small board of directors.
Journal of Financial Economics 40 (2): 185-211.