Do Small Institutional Shareholders Use Low-Cost Monitoring Opportunities?
Evidence from the Say on Pay Vote
Miriam Schwartz-Ziv and Russ Wermers*
March 27, 2016
Abstract
Theories of free-riding predict that only large shareholders will monitor management. We document
that, when shareholders are given a low-cost opportunity to monitor and discipline management,
small institutional shareholders are particularly likely to do so. We focus on the “Say-On-Pay” (SOP)
vote, because it represents the best low-cost opportunity shareholders have. By examining three
levels of votes—aggregate, institutional, and fund level—we document that institutions with a
smaller ownership stake in a company are particularly likely to vote against management. Further,
firms are particularly likely to demonstrate responsiveness to SOP when non-insider blockholders
are present.
*
Miriam Schwartz-Ziv is from the Eli Broad College of Business at Michigan State University, Russ Wermers is from
the Smith School of Business at the University of Maryland. We thank Tim Adam, Jackie Cook, Charlie Hadlock, Peter
Iliev, Zoran Ivkovich, Slava Fos, Fabrizio Ferri, Feng (Jack) Jiang, Naveen Khanna, Jerchern Lin, Michelle Lowry,
Nadya Malenko, Ric Marshal, James McRitchie, Michael Ostrovski, Otto Randl, Andrei Simonov, David Stolin, Tilan
Tang, Yuehua Tang, Jun Yang, and participants in the joint Humboldt University and ESMT Conference on Recent
Advances in Mutual Fund and Hedge Fund Research, 2014 FMA, 2015 DePaul University Conference in Corporate
Social Responsibility, as well as seminar participants at the University of Baltimore, SUNY Buffalo, Chulalongkorn
University, University of Cincinnati, the Hebrew University of Jerusalem, Institutional Shareholder Services, Michigan
State University, and the Securities Exchange Commission, for helpful comments. We thank Bingkuan (Bryan) Cao,
Corrine Carr, Sam Floyd and particularly Jinming Xue for research assistance. Finally, Miriam Schwartz-Ziv thanks the
Edmond J. Safra Center for Ethics at Harvard University for hosting her as a fellow during the initial stages of the
preparation of this paper.
Electronic copy available at: http://ssrn.com/abstract=2510442
1. Introduction
Who monitors and disciplines management—is it the small shareholders of a company, or
the large stakeholders? And, will an institutional investor monitor its smaller investments differently,
and perhaps less intensively, compared to its larger portfolio investments? The traditional view is
that shareholders holding a large stake in a company are more likely to take costly actions, such as
initiating a proxy fight or confronting management, while small shareholders will enjoy a free-ride
(Grossman and Hart, 1980; Shleifer and Vishny, 1986; and Hart, 1995). In this paper, we examine
whether the most common type of shareholders—small institutional shareholders—engage in
monitoring the companies in which they have holdings, if a low-cost opportunity is made available
to them.
Starting in 2011, such an opportunity was introduced: the “Say-On-Pay” (SOP) vote. The
formal goal of the SOP vote is to allow shareholders of U.S. listed companies to state whether they
approve or disapprove of the compensation awarded to the named executive officers of a firm
during the previous fiscal year. Other than SOP, the only issues that are raised routinely at
shareholder meetings are the election of the slate of directors proposed by management, and the
ratification of the auditors, both of which offer meager choices to shareholders. Hence, of the issues
bought up routinely for vote, SOP is the best opportunity shareholders are given to provide
meaningful feedback to management. Indeed, a Spencer Stuart (2014) survey documents that, in
2013, the most frequent issue for which management proactively reached out to their large
shareholders was the SOP vote. This further supports the argument that SOP is the main
monitoring opportunity that shareholders are given.
In this paper, we test the hypothesis that institutional investors—who have a fiduciary duty
to vote on issues bought up at a shareholder meeting (Bew and Fields, 2010)—with a smallmagnitude holding (e.g., those holding a small fraction of the company’s shares) use the SOP vote
1
Electronic copy available at: http://ssrn.com/abstract=2510442
more aggressively than large-scale shareholders to govern the companies they hold. This hypothesis
is grounded in the following reasoning. First, large-scale shareholders are less likely to challenge
management through voting, because they have access to management, and are, thus, able to
confront corporate managers directly and in private.1 Such interactions do indeed occur: Ng and
Troianovski (2015) claim that managers of listed companies hold, each year, thousands of meetings
with large shareholders; Jeffrey Ubben from “Value-Act” stated he prefers to work with companies
behind the scenes (Sidel and Hoffman, 2015); and, BlackRock and Vanguard engage in private
discussions with corporate executives (Burr, 2012). Moreover, the mere threat of intervention by a
large shareholder can provide the large shareholder increased access to management, and serve as a
disciplinary force (Fos and Kahn, 2015). Second, blockholders, who have more “skin in the game”
than ordinary shareholders, may prefer to avoid the negative returns which may follow a SOP vote
with low support rates.
Indeed, we find evidence documenting the pattern of small-shareholders-opposing-SOP by
examining SOP votes cast on three levels. The first voting level we examine is at the aggregate level.
We find that, in companies with a large percentage of shares held by “blockholders” (i.e., any type of
shareholder holding at least 5% of the outstanding shares), shareholders are more likely to vote in
support of management on SOP. By contrast, when shareholder composition is dispersed,
shareholders are more likely to oppose management on the SOP vote. We estimate, for example,
that, if the “fraction of shares held by blockholders” decreases from the 25th percentile to the 1st
percentile, an additional 0.75% of the votes cast are expected to oppose SOP, which is equivalent to
a 7.3% increase in the SOP opposition rate.
On the aggregate level, our above-noted evidence on SOP voting outcomes does not reveal
We note that large positions by mutual funds, while not necessarily qualifying them as 5% blockholders, often allow
them direct access to management.
1
2
Electronic copy available at: http://ssrn.com/abstract=2510442
whether small shareholders, or alternatively, large-scale shareholders, are those most likely to vote
against SOP when large blockholders are present. To understand how the magnitude of a holding is
related to the vote cast by a shareholder, we examine a second voting level: the mutual fund level.
Mutual-funds cast an estimated 35.6% of all SOP voted shares. Accordingly, mutual fund’s votes
comprise a significant portion of all voted shares.
On the mutual fund voting level, we find further evidence that, the smaller the holding, the
more likely a fund is to oppose management on SOP: the smaller the portfolio weight represented
by a holding in a particular stock (i.e., the total value of the shares of a specific company held by the
mutual fund divided by the total value of the fund’s portfolio), the more likely the fund is to vote
against SOP. For example, if a fund’s portfolio weight were to decrease from the 99th percentile to
the 90th percentile (the corresponding portfolio weights are 0.03879 and 0.0137, respectively) a
fund’s SOP opposition rate is expected to increase by 10.34%. Similarly, we document that a shift
from the 99th percentile to the 90th percentile of the fraction of company’s shares held by a fund is
expected to increase the SOP opposition rate by 9.29%. Put differently, the smaller the magnitude of
a fund’s holding, the more likely the fund is to confront management publicly via the SOP vote.
Because we observe that a substantial fraction of funds vote consistently on the institutional
level (i.e., the investment advisor level),2 we examine this third voting level as well. We estimate that
institutional investors cast 87.8% of all voted shares, and, therefore, the institutional level clearly has
a dominating impact on the ultimate vote outcome. Since only mutual funds are required to report
their votes, we estimate how a given institution voted using the average fraction of mutual funds,
advised by that institution, that voted in support of SOP for a given corporation.
Consistent with the above findings, we find that, the smaller the portfolio weight (in
aggregate across all funds within an institution), and the smaller the fraction of a company held (in
2
In this paper, we use the term “institution” to refer to an institutional investment advisor.
3
aggregate), the more likely the institution’s funds are to oppose SOP. Even after controlling for the
magnitude of holdings on the fund level, we estimate that, if a given company’s portfolio weight in
an institution’s portfolio were to decrease from the 99th to the 90th percentile, the opposition rate of
funds within this institution is expected to increase by 27.34%. Similarly, if the fraction of a
company’s shares held by an institution decreases from the 99th percentile to the 90th percentile, the
opposition rate of funds within that institution is expected to increase by 9.91%. Notably, we find
that the magnitude and significance of the coefficients of the holding variables at the institutional
level are somewhat larger and more robust, compared to the parallel coefficients on the mutual fund
level. This result reinforces the importance of examining the voting patterns as they relate to the
magnitude of holdings at the institutional level in addition to the mutual fund level, as we do in our
paper.
We find further evidence supporting the small-shareholders-opposing-SOP pattern: On the
aggregate level of the votes, we find that the larger the fraction of shares held by blockholders, the
larger the SOP support rates. In contrast, on the fund and the institution level, we find that, the
larger the fraction of shares held by blockholders, the more likely funds and institutions (who are, in
the vast majority of our observations, not blockholders) are to oppose SOP. This contradiction
implies, once again, that the blockholders (both institutional and non-institutional) are likely those
voting in support of SOP, and drive the results documented on the aggregate level, while the typical
fund and institution with a small holding is more likely to oppose SOP, particularly when the
presence of blockholders is large.
As predicted, we document that, indeed, SOP votes which receive low support rates are
followed by negative CARs. We estimate that a decrease in the support rate from the 1st percentile to
the 25th percentile leads to an (expected) 0.83% CAR decrease. As noted, such a decrease could
motivate shareholders to refrain from opposing SOP for their large portfolio-weight investments.
4
We also examine how funds and institutions identify companies for which they will vote
against SOP for their small-scale holdings. We find that they are particularly likely to vote against
SOP when Institutional Shareholder Services (ISS) recommends to vote against SOP, compared to
when ISS recommends to vote for SOP. This indicates that funds and institutions with smaller-scale
shareholdings rely on ISS to flag “problematic” companies. This finding is consistent with those of
Iliev and Lowry (2015), who comprehensively examine funds’ votes in the pre-SOP period, and
show that, the larger a fund’s holding, the more likely it is to determine, independently of ISS, how
to vote.
Finally, we examine whether management is likely to demonstrate responsiveness to the
SOP vote, given the ownership structure. We find that, for companies that have a non-insider
blockholder, the lower the SOP support rate, the more likely the company is to: (1) experience CEO
turnover within 12 months of the SOP vote; (2) pick more modest peer-companies for determining
executive compensation; and (3) decrease the growth rate of the residual executive compensation.
These findings indicate that, while large shareholders are less likely to vote against management on
SOP, when a non-insider blockholder is present, companies are particularly likely to demonstrate
responsiveness to shareholder dissatisfaction, as reflected in the SOP vote. Thus, large blockholders
may serve as a “reluctant watchdog.”
Prior studies that examine the votes cast by mutual funds, of which we are aware, assume
that a mutual fund’s decision on how to vote is made either on the institutional level (e.g., Davis and
Kim, 2007), or, more commonly, on the fund level (e.g., Matvos and Ostrovsky, 2010; Morgan,
Poulsen, Wolf, and Yang, 2011; Iliev and Lowry, 2015; Aggarwal, Erel and Starks, 2015; and
Dimmock, Gerken, Ivkovich, and Weisbenner, 2015). By examining, within a single framework,
voting at the fund level, institution level, and aggregate (all voters) level, our study allows unique
insights not possible by these past studies. Moreover, we believe we are the first study to examine
5
how variations in the magnitudes of holdings within a given fund or an institution, relate to the votes
cast. Our multi-level analysis shows that the SOP vote has moved shareholder governance toward a
system where both small-scale institutional shareholders and blockholders are valuable—within the
same firm’s capital structure—in imposing discipline on corporate managers.3
Overall, the main contribution of our study is documenting that offering low-cost
monitoring opportunities increases the extent to which small shareholders monitor and govern the
companies they hold. Our findings suggest that low-cost monitoring opportunities offer a
coordination mechanism that allows many small institutional shareholders to voice a unified
message.
2.
Background and Motivation
The mandatory non-binding Say on Pay (SOP) vote, which took effect starting January 21, 2011,
offered shareholders an unprecedented, relatively “low-cost” opportunity to provide feedback to
management on a regular basis. In the 2011-2013 period examined, the vote applied to companies
with a public equity free float value exceeding $75 million.4 Although the SOP vote is formally about
the compensation awarded to management, as described above, it is the best opportunity
shareholders have to express their broader views on management performance.5
The seminal papers of Grossman and Hart (1980), Shleifer and Vishny (1986), and Hart
(1995) all emphasize the free-rider problem. These papers predict that large (or potentially large)
In addition, to the best of our knowledge, we are the first study that examines how holding magnitudes on the
institutional level relate to the votes cast. Finally, we believe we are the first to contrast holding variables on the
aggregate level to those on the mutual fund/institutional level.
4 In 2011, each company had a frequency SOP vote, in which shareholders determined whether they wished to hold the
SOP vote every one, two, or three years. Kronlund and Sandy (2015) find that 89.7% of the companies voted in favor of
an annual SOP vote.
5 The only other issues that are raised routinely (i.e., annually), at shareholder meetings are whether to approve the slate
dof directors recommended by management, and whether to ratify the auditors. Since, in more than 99.9% of cases,
shareholders are presented with only one slate of directors and a single option for the auditor, shareholders do not have
obvious alternative options on which to coordinate their votes on these issues. However, for the SOP vote, shareholders
a have clear and simple option–they can vote either for or against the compensation awarded.
3
6
shareholders may take costly actions, such as engaging in a proxy fight, making a tender offer, or
promoting a takeover, if their private benefits of such actions are sufficient; small shareholders will
“free-ride” and benefit from the costly actions taken by the large shareholders. McCahery, Sautner,
and Starks (2014), who survey large investors—which are “most likely to have the resources for and
interest in shareholder engagement”—highlight that such shareholders can also engage in
continuous dialogue and monitoring of management. All these papers focus on costly actions large
shareholders may take.
We hypothesize that small institutional shareholders are especially likely to use the low-cost
SOP vote to oppose management when they are dissatisfied with management performance. Our
reasoning is as follows. First, as mentioned, large shareholders have alternatives to the SOP vote—
they likely have direct access and can let their voice be heard by management. Small shareholders do
not have direct access to management. Second, small institutional shareholders are not likely to
engage in costly actions such as proxy fights, because they are too costly for such a small scale of
shareholding. Third, publicly disclosing discontent with management has a short-term price. As we
will show, receiving low support rates for the SOP vote leads to reduced stock return. Naturally,
such negative returns will pose a concern for shareholders particularly with respect to their large
portfolio weight investments.
Finally, we expect to find that the presence of a large non-insider blockholder is required for
management to respond to the SOP vote, because such a large blockholder can pressure
management to respond promptly to the negative SOP feedback (Levit and Malenko, 2011).
3.
Data
Starting January 21, 2011, the SOP vote applied to all companies listed in the United States with a
public free float exceeding $75 million. Approximately 2,200 companies fall under this definition in
7
the average year. Since we wish to avoid a selection bias (e.g., examining only the S&P 1500
companies) we collect data from data sources that cover the universe of the companies that were
required to hold a SOP vote in the period examined.
Data on company performance is obtained from CRSP and Compustat. Data on executives
and their compensation is obtained from Institutional Shareholder Services (ISS). Data on mutual
fund holdings is obtained from the CRSP mutual fund database, and from the Thomson s-12 mutual
fund holding files. Data on institutional shareholdings at the advisor level (13(f)) is obtained from
the Thomson s-34 files. In Appendix A, we describe the multiple procedures we follow to match the
Thomson s-12, Thomson s-34, and CRSP mutual fund databases to the ISS voting analytics dataset.
Data on shareholder composition, including blockholders, is obtained from GMI ratings. Data on
peer-companies selected to determine the executive’s compensation is obtained from Institutional
Shareholder Services (ISS). These data are extracted, by ISS, from the DEF 14-A filings of the
corporations.
Voting outcomes are obtained from the ISS Voting Analytics database. This dataset
documents the aggregate vote outcomes for each proposal that came up for a vote at a shareholder
meeting. These outcomes are generally reported in an 8-K filing, and occasionally in a 10-Q filing. In
addition, the ISS Voting Analytics database includes data on the votes cast by mutual funds. The
latter votes are reported via the N-PX form that mutual funds submit annually to the SEC. For each
issue discussed at a shareholder meeting, the ISS dataset also includes management’s
recommendation on how shareholders should vote. With respect to the SOP votes examined in this
paper, the ISS voting analytics database includes the votes cast by 8,307 mutual funds that are
operated by the 357 largest investment advisors.
4.
Descriptive Statistics
8
We start by highlighting the large impact that institutions, in general, and mutual funds, in particular,
have on the outcome of votes. We estimate the percentage of voted shares cast by institutions using
data reported in ProxyPulse (2014), published by Broadridge—the only company through which
shareholders can submit their votes electronically (which is how the vast majority of shareholders
vote). ProxyPulse (2014) reports that, for S&P 1500 companies, 90% of all institutional
shareholdings are voted, while only 29% of all retail shareholdings are voted. ProxyPulse (2014) also
reports that institutions own, on average, 70% of the outstanding shares of these companies, while
the remaining 30% are held by retail investors. Hence, 87.8% (=(0.7*0.9)/(0.7*0.9+0.3*0.29)) of all
votes cast are cast by institutions. This figure emphasizes that vote outcomes are typically
determined by institutional investors.
In addition, Table 1 reports that, on average (median), in a given corporation-year, mutual
funds own 28.5% (29.2%) of the outstanding shares of the companies that held a SOP vote during
the 2011-2013 period. This figure is calculated by dividing the aggregate number of shares held by all
mutual funds in a given stock and a given year (in the quarter preceding the vote), by the total
number of shares outstanding (both figures are obtained from the Thompson s12 database), and
then calculating the average/median across all stock-years. Using the abovementioned ProxyPulse
(2014) figures and estimations, we estimate that on average, 35.7% [(0.9*0.285)/(0.7*0.9+0.3*0.29)]
of the voted shares are voted by mutual funds. The corresponding median is 36.6%
[(0.9*0.292)/(0.7*0.9+0.3*0.29)]. These figures highlight that mutual funds have a large impact on
the aggregate level of the votes.
Table 2 reports summary statistics for the variables included in this study. Variables are
defined in the Glossary of Variables. We highlight here several of these variables: “fraction of shares
held by blockholders” is a variable that captures the aggregate fraction of shares held by institutional
and non-institutional shareholders that each hold at least 5% of the company’s outstanding shares.
9
This variable is obtained from the GMI rating, who collect this data from the proxy filings. As Table
2 documents, on average (median), the “fraction of shares held by blockholders” is 0.27 (0.25),
indicating that the majority of the shares in the average company are owned by small-scale
shareholders, yet a substantial proportion are held by blockholders, each holding at least 5% of the
outstanding shares.
Table 2 also documents that, in general, support for management on SOP is strong: among
shareholders who vote, 89.8% vote in favor of SOP (“fraction voted for SOP”), as opposed to
voting against SOP (or in a small percentage of cases withholding or abstaining from the vote).
Table 3 focuses on the votes cast on the individual institution (advisor) level. Column 3
reports, for the 20 institutions which have participated in the largest number of SOP votes, the
frequency they voted in the opposite direction from the recommendation of Institutional
Shareholder Services (ISS), the leading proxy advisory company. Column 3 documents that some
investment advisors never vote against ISS’s recommendation, while other investment advisors do
so quite frequently.
Bew and Fields (2010, p. 22) report that some institutions determine, on the institutional
level (e.g., BlackRock), how their funds should vote, while other institutions delegate this decision to
their fund managers (e.g., BlackRock Large Cap Core Fund). Indeed, Column 4 of Table 3 indicates
that, within some institutions, funds vote unanimously (e.g., Vanguard with a 0 “S.D. of votes within
institution”), while other institutions vote on the fund level (e.g., Jackson National Management with
a 20.38% S.D.). The median standard deviation which is equal to 0.07% indicates that the median
institution almost always votes unanimously, but the average institution seems to delegate some
amount of discretion on the voting decision to individual funds, as indicated by the average standard
deviation equaling 3.05%. Following this variation, we shall examine how both the magnitude of a
fund’s holdings, and that of the institution, relate to the votes cast.
10
5.
Results
5.1.
Aggregate level votes
In this section, we start by examining the basic factors (e.g., compensation and prior firm
performance) which might affect SOP votes cast on the aggregate level. Next, we focus on the
primary variables of interest in this study—the variables capturing shareholder composition. All
analyses pertain to the 2011-2013 period. Regression results presented use OLS.
Table 4 examines which factors are related to the fraction of votes cast for SOP, that is, the
fraction of shareholders supporting the prior-year compensation awarded to the CEO and the other
top four named executives. 6 Not surprisingly, we find that, the larger the compensation awarded to
the CEO (“total compensation of CEO t-1,” Regression 1), and to the other four named executives
(“average compensation of 4 executives t-1,” Regression 2), the larger the fraction of votes cast
against the compensation awarded.
We also examine how awarding, relative to other companies, compensation with a large
residual (referred in other studies as “excessive compensation”), relates to the SOP votes cast. If
shareholders govern their managers effectively, we should expect that voting against SOP should be
positively correlated with awarding a large residual compensation, and not simply with total
compensation. We follow Core, Guay, and Larcker (2008), and define residual compensation as the
residual from a fitted compensation model for company i in year t.7 Our model’s “residual
compensation” is a proxy for what investors may perceive as excessive compensation, since it is not
explained by the observed variables.
6
The named executives include the CEO, the CFO, and the three additional executive officers who receive the highest
level of compensation within a firm.
7 Our compensation model is estimated by regressing the total compensation awarded to the CEO/four other named
executives on the lagged: ROA, abnormal returns, market capitalization, age of CEO, tenure of CEO, as well as on fixed
year and industry effects.
11
In Table 4, we find that awarding a large “residual of CEO compensation t-1” (Regression
3), and a large “residual of 4 executive’s compensation t-1” (Regression 4), significantly increases the
likelihood that shareholders oppose the SOP vote, indicating that shareholders vote against SOP
when the residual compensation is large. These results are consistent with Ertimur, Ferri, and Muslu
(2011), who examine the 1997-2007 period, during which SOP was held voluntarily by a small
number of companies, and find that awarding a large residual compensation, increases the likelihood
of shareholders voting against SOP. Table 4 results pertaining to compensation remain intact when
compensation is measured using logged total compensation.
Shareholders are also more likely to vote against management when expected compensation
is high—that is, when our model indicates that the compensation is in line with other firms with
similar levels of profitability. Thus, even highly-paid executives who are no better paid than peers in
the same situation may be perceived as being overpaid.
We also find that companies with weak financial performance (i.e., low values for “ROA of
company t-1”) and with low values of “abnormal returns” are likely to experience high opposition
rates for SOP—even after controlling for the effect of these variables on the separation of expected
vs. residual executive compensation. This finding suggests that, indeed, shareholders view the SOP
vote as an opportunity to express their general satisfaction (or lack thereof) with the company’s
performance (Iliev and Vitanova, 2015; Cuñat, Gine, and Guadalupe, 2015). Consistent with
previous studies, (Larcker, McCall, and Ormazabal, 2012; Ertimur, Ferri, and Oesch, 2012; Thomas,
Palmiter, and Cotter, 2012; and, particularly, Malenko and Shen, 2015), Regressions 1-4 all show that
a recommendation issued by ISS to vote against SOP is associated with approximately a 28%
increase in the likelihood that shareholders vote against SOP.
We now focus on our primary variables of interest—those capturing shareholder ownership
structure. Regressions 1-4 of Table 4 indicate that, when ownership is more concentrated, i.e., the
12
“fraction of shares held by blockholders” is large, the SOP vote is likely to garner larger support
rates (results are significant at the 1% level). Using the coefficient from Regression 1 of Table 4,
compared to a company in the 25th percentile of this variable, (0.144 of the outstanding shares are
held by blockholders), a company in the 1st percentile (0.0) is 0.75% ((0-0.144)*0.0519) more likely
to oppose SOP. Since during 2011-2013, on average, only 10.19% (1-0.8981, see Table 2) of the
votes cast opposed SOP, the latter change is equivalent to an increase in the likelihood that
shareholders oppose SOP by approximately 7.3% (0.75%/10.19%).
For example, in 2012, 49.86% of the voted shares of “Yahoo!” supported SOP. In that year,
17.5% of Yahoo’s shares were held by blockholders. According to our estimation in Regression 1 of
Table 4, if an additional 5% of Yahoo’s shares were held by blockholders (which is less than one
third of a standard deviation in the “fraction of shares held by blockholders”), 50.12%
(0.05*0.0519+0.4986) of the votes, i.e., the majority of the votes, would have supported SOP.8
Our findings indicate that large shareholders support management on SOP votes. In
subsequent sections, we will further investigate this finding by examining the votes cast by mutual
funds and institutions for which we can observe the magnitude of their holding. Our findings, thus
far, suggest that, since blockholders do not appear to be particularly active in terms of voting against
management, when ownership is widely dispersed, the SOP vote can serve as a coordinating
mechanism for a large number of small shareholders. Such coordination can be challenging when
ownership is dispersed (Fluck, 1999). Thus, even though theory predicts that small-scale investors
will free-ride, they apparently are more likely to use the SOP vote to voice discontent.
Finally, we examine how the magnitude of ownership by executives (who are permitted to
cast votes, just like any other shareholder) affects the aggregate SOP votes cast. The results
This estimation highlights that the shares held by the executives made the difference between the SOP vote failing or
passing. The 50% pass-fail threshold can make a difference in how the company responds according to Cuñat, Gine, and
Guadalupe (2012). There are, in fact, tens of cases in the data we examine, in which the SOP failed by a small margin.
8
13
document that the larger the fraction of shares held by executives, the more likely the SOP vote is to
pass.9 Thus, while prior research (e.g., Jensen and Murphy, 1990) indicates that management
shareownership helps to align the incentives of management with shareholders, shareholders
apparently perceive heavy ownership by management as being problematic. This finding also
highlights that executives’ shares dilute the ability of non-insider shareholders to govern
management.
5.2.
Votes Cast on the Mutual Fund Level
In Section 5.1, we documented that companies with a large percentage of shares held by
blockholders are particularly likely to receive extensive support for the SOP vote. We argued that
blockholders are likely those voting in support of SOP. However, it is also possible that small
institutional shareholders tend to vote in support of SOP when large blockholders are present. To
test whether this is the case, as well as to further understand how the magnitude of an institutional
investor’s holding motivates that investor’s voting, we next examine the votes cast by mutual funds.
We can observe the votes cast by a mutual fund, since mutual funds are (the only class of
shareholder) required to report, in an SEC filing, the votes they cast.
In Table 5, the analysis is performed on the fund-company-year level. Each observation in
Table 5 pertains to a specific fund’s SOP vote, for a given company, in a given year. Errors are
clustered at the fund level. Regressions are OLS (in order to allow the reader an easier interpretation
of coefficients), unless noted otherwise. We use the fund’s holding data at the end of the quarter
preceding the quarter during which a SOP vote is held, as a proxy for the holdings of funds just
Since, as we discussed earlier, management always recommends an affirmative vote on SOP, we would expect a
coefficient of unity on “Fraction of shares held by executives.” Since the coefficient is much smaller, about 0.16, we
infer that non-management shareholders vote much more frequently against management when management holds a
greater level of ownership. Non-executive shareholders appear to “push back” by voting more frequently against
management when management owns more of the company.
9
14
before the vote is held.
Unless noted otherwise, our specifications include fund fixed effects, because we wish to
observe how the SOP votes cast by a specific fund differ, depending on the magnitude of each
holding, i.e., how a fund votes differently for its small holdings versus its large ones. Fund fixed
effects also control for an unobserved tendency for a given fund to vote in a particular manner
across stocks and over time. We note that the approach of including fund fixed effects differs from
that of Iliev and Lowry (2015), who are primarily interested in examining the cross-sectional relation
between the magnitude of a fund’s holdings and the votes cast by that fund, while our primary focus
is on within-fund variation. Analyzing the time-series variation of a given fund’s voting patterns
reveals, from the voter’s perspective, how the magnitude of a holding relates to the vote cast by the
shareholder.
The dependent variable in Table 5 equals 1 if the fund voted “for” SOP (indicating they
supported the compensation awarded), and zero, otherwise. The results in Table 5 show that, similar
to the aggregate level results (Table 4), mutual funds are likely to vote against SOP when
compensation is large, and when performance (proxied by lagged ROA and firm abnormal return) is
weak.
To understand whether small institutional shareholders vote differently from large ones, we
examine how the magnitude of the holding of a mutual fund is related to the vote it casts. We use
the CRSP mutual fund database to estimate the following two “holding variables” that capture the
magnitude of the holding: [1] “Fund’s portfolio weight (in fraction)”—following Fich, Harford, and
Tran (2014), we examine the stock’s portfolio weight in the fund’s portfolio. The average value for
this variable is equal to 0.5%, see Table 2. [2] “Fraction of company’s shares held by fund”—the
fraction of the company’s shares held by the mutual fund. The average value of this variable is equal
15
to 0.21%, see Table 2.10
The specifications in Table 5 document that, the larger a stock’s weight in a mutual fund’s
portfolio, the more likely the fund is to vote in support of SOP. For example, assume a “fund’s
portfolio weight (in fraction)” for a given company were to decrease from the 99th percentile to the
90th percentile (compared to the distribution across all funds and all time-periods; the corresponding
portfolio weights are 0.0387 and 0.0137). According to Regression 3 of Table 5, this fund would
now be 1.07% ((0.0137-0.0387)*0.4284) more likely to oppose SOP. Adjusting this by the average
SOP opposition rate of mutual funds (100%-89.66%=10.34%, see Table 2, which documents that
the average SOP support rate of mutual funds is 89.66%), this is equivalent to a relative increase of
10.34% (1.07%/10.34%) in a fund’s SOP opposition rate, which is economically significant.
Similarly, Regressions 2 and 3 of Table 5 document that, the larger the “fraction of
company’s shares held by a fund”, the more likely a fund is to vote against SOP. According to
Regression 3 of Table 5, compared to a fund voting for a company in the 99th percentile of “fraction
of company’s shares held by a fund” (0.0266), if a fund votes for a company in the 90th percentile of
this variable (0.0056), it is 9.29% ((0.0056-0.0266)*0.4572/10.34%) more likely to oppose SOP.
These findings indicate that mutual funds exhibit voting behavior consistent with that of
blockholders—the larger the holding, the less likely shareholders are to publicly oppose management
via the SOP vote.
Regression 4 of Table 5, which is a logit version of Regression 3, confirms these results. The
results imply that, assuming all control variables are equal to their mean, a shift the 99th percentile to
the 90th percentile of a fund’s portfolio weight is expected to increase the opposition rate by 7.33%.
Similarly, a shift the 99th percentile to the 90th percentile of the fraction of company held by the fund
We use the CRSP mutual fund database as our primary source for computing the holding variables (as opposed to the
Thompson s-12 files), because Schwarz and Potter (2015) estimate that, starting from the 4 th quarter of 2007, the CRSP
mutual fund dataset is the most thorough individual dataset available.
10
16
is expected in increase the opposition rate by 10.26%.
In Regression 5, we report a cross-sectional analysis that does not include fund fixed effects,
and the results are similar, indicating that not only do our results apply with respect to the variation
of the magnitude of holdings within funds, they also apply to such variation between funds. In sum,
in this section, we find that mutual funds with small holdings are particularly likely to vote against
SOP.
5.3.
Votes Cast on the Institutional Level
As shown in Section 4, a substantial fraction of mutual fund’s votes are cast on the institutional
level, meaning that all, or virtually all, of the funds advised by a given institution vote consistently
with each other (see Section 4). Accordingly, we repeat our analysis, but use the “holding variables”
computed on the institutional level, as opposed to the fund level. We estimate, here, how the
portfolio weight of a company in an institution’s portfolio (which, on average, is equal to 0.0018, see
Table 2) and the fraction of outstanding shares held by the institution (which, on average, is equal to
0.0138, see Table 2), are each related to the votes cast at that level.
Similar to Davis and Kim (2012), who analyze votes cast on the institutional level (but do
not address the effect of the magnitude of a holding), we define the dependent variable as the
fraction of funds, within an institution, that vote in support of SOP. We include only one
observation for each institution-company-year vote. Year, industry, and institution fixed effects are
included as indicated near the bottom of Table 6, and errors are clustered at the institution level. We
emphasize that including an institution fixed effect allows observing how the same institution votes
differently, depending on the variation in the magnitude of its holding of a given stock.
Regressions 1-4 of Table 6 document that the holding variables, measured at the institutional
level, also have a large and significant effect on SOP voting. For example, Regression 3 of Table 6
17
estimates that a shift from the 99th percentile of “portfolio weight of institution’s portfolio” (0.0263
is the 99th percentile across institutions and over time) to the 90th percentile (0.0043) is expected to
increase the institution’s opposition rate to SOP by 2.53% [(0.0043-0.0263)*1.1515]. Since the mean
institutional opposition rate is 12.81% (1-0.8719, based on 0.8719 being the institutional SOP
support rate, see Table 2) this is equivalent to a 19.75% (2.53%/12.81%) relative increase in the
opposition rate.
Similarly, Regression 3 of Table 6 estimates that a shift of the “fraction of company’s shares
held by institution” from the 99th percentile (0.1076) to the 90th percentile (0.0458), is expected to
increase, by 27.05% [(0.0458-0.1076)*0.5608)/0.1281] the SOP opposition rate relative to the mean.
Taken together, these results point out that the smaller a holding on the institutional level, the more
likely the institution is to vote against SOP, and that the magnitude of this effect is economically
significant.
Regression 4 replicates Regression 3, but does not include institution fixed effects. In this
regression, both institutional holding variables are insignificant. This emphasizes that SOP voting is
largely decided at the institution level, so the relation between the size of a stockholding and the
tendency to vote with management on SOP is much more evident when we examine withininstitution variation, as opposed to cross-institution variation.
We next wish to investigate the relative importance of the magnitude of the holding on the
institutional level versus the mutual fund level. Accordingly, in Regressions 5-7 of Table 6, we
include both holding variables (portfolio weight and fraction of company) on both levels (fund and
institution). The dependent variable in these regressions is equal to one if the fund voted for SOP,
and zero, otherwise. Regression 5 includes fund fixed effects, while Regression 6 includes institution
fixed effects. We report both of these versions, since some institutions vote on the institution level
(suggesting an institution fixed effect be included), while others vote on the fund level (suggesting a
18
fund fixed effect be included). Regressions 5-7 include (but do not report) all fund control variables
included in the models reported in Table 5.
Regressions 5 and 6 of Table 6 document that the effect of the institution’s portfolio weight
is particularly large compared to that of the fund: Regression 5 of Table 6 documents that, if the
portfolio weight of a company in an institution’s portfolio decreases from the 99th percentile
(0.0263) to the 90th percentile (0.0043), each fund within this institution is expected to increase its
opposition rate by 27.34% [(1.5921*((0.0043-0.0263)/0.1281)]. In comparison, if a fund shifts an
investment from the 99th percentile (0.0387) to the 90th percentile of its portfolio weight (0.0137), the
fund is 6.68% [((0.0137-0.0387)*0.2764)/0.1034] more likely to oppose SOP. These two magnitudes
demonstrate that the portfolio weight effect is substantially larger on the institutional level,
compared to the fund level—which reinforces the conclusion that institutions frequently decide on
the institution level how their advised funds should vote.
Similar to the results above, Regression 5 of Table 6 documents that if an institution
decreases the fraction of the outstanding shares of a company held from the 99th percentile (0.1076)
to the 90th percentile (0.0458) it is 9.91% [((0.0458-0.1076))* 0.2056/12.81%] more likely to oppose
SOP. In this specification, the fraction of company held by the fund is insignificant. In Regression 6,
which includes institution fixed effects, the fraction of company shares held by an institution is
insignificant, while that held by the fund is marginally significant (at the 10% level). Hence, across
Regressions 5 and 6, of the two holding variables, the portfolio weight holding variable is more
stable in explaining the SOP vote. This indicates that when funds and institutions vote, their
decision is particularly affected by the magnitude of the holding from their own perspective (as
captured by the portfolio weight), as opposed to the magnitude of the size from the company’s
perspective (as captured by fraction of company held).
In specification 7, we do not include fund or institution fixed effects. Three of the four
19
holding variables are insignificant, demonstrating once again, that our findings are particularly driven
by within institution and within fund variation, as opposed to variation across funds in different
complexes.
We find further evidence for the small-shareholders-opposing-SOP pattern at the fund and
at the institution level in Tables 5 and 6. Importantly, the larger the “fraction of shares held by
blockholders” the more likely it is that funds and institutions oppose SOP.11 Put differently, the
median institution, which on average only holds 0.29% of a company, is less likely to support SOP
when a large percentage of the shares is held by blockholders (each holding at least 5% of the
shares).
This mutual fund voting pattern contrasts with the pattern documented at the aggregate level
of voting—the larger the fraction of shares held by blockholders, the larger the support rates for
SOP. This contrast implies, once again, that the result documented on the aggregate level is likely
driven by the voting of large blockholders in support of SOP. Our result on the fund level may also
indicate that funds are more likely to exercise their vote against management when blockholders are
present, because funds view blockholders as having the ability to pressure management to respond
to low SOP support rates (we discuss this in more detail in Section 6).
Finally, we point out that (particularly on the fund level and, to some extent, on the
institution level; see Tables 5 and 6, respectively) the larger the “fraction of shares held by
executives,” the more likely a mutual fund or institution is to vote against SOP. This result also
contrasts the result reported in Table 4 (the aggregate level)—the larger the fraction of shares held
by executives, the more likely shareholders are to vote in support of SOP. This contrast is consistent
with executives voting in support of their own compensation, but, perhaps, mutual funds and
For example, according to Regression 5 of Table 6, compared to an institution voting on a company in the 1st
percentile of “fraction of shares held by blockholders” (0.0), an institution voting on a company in the 25th percentile of
this variable (0.144) is 2.01% ((0.144-0)*-.0179/12.81%) more likely to oppose SOP.
11
20
institutions viewing high executive ownership as a potential sign of management entrenchment.
In sum, this section documents that also on the institutional level, the smaller the holding of
a given institution, the more likely the institution’s funds are to oppose SOP.
5.4. CARs Around SOP Votes
Thus far, we have documented that small institutional shareholders are likely to oppose SOP, while
the large institutional shareholders are likely to support SOP. This raises the question—why do large
shareholders refrain from opposing SOP? One potential reason is that there may be negative
consequences, in the short-term, to opposing management. For instance, if a SOP vote has lower
support for management than expected, the market may interpret this as a signal from shareholders
that management is performing below expectations at a particular firm. Under this scenario, we
might expect a negative market abnormal return following a SOP votes with low support rates. In
turn, this may motivate funds and institutions to refrain from voting against SOP for their large
portfolio weight investments, since these have a large and negative impact on the overall portfolio
return. In related work, Iliev and Vitanova (2015) and Cuñat, Gine, and Guadalupe (2015)
document that the decision to hold a SOP vote leads to positive abnormal returns. In our analysis,
we examine how the outcome of the SOP vote affects abnormal returns.
In Table 7, following, e.g., Cuñat, Gine, and Guadalupe (2012), we define the “event date” as
the meeting date. Because companies can observe the votes cast electronically as soon as they are
cast, i.e., before the meeting date, information on the votes cast may leak before the meeting date.
Companies are required by the SEC to disclose the results of the SOP vote within four days after
the meeting date. Given these issues, in Table 7, we examine the cumulative abnormal returns during
an event window of different sizes that surrounds the meeting date.
Abnormal returns are estimated by subtracting, from a company’s buy-and-hold return
21
around the meeting date, the buy-and-hold CRSP value-weighted market return for the
corresponding period [Panel A], or the buy-and-hold value-weighted size-decile portfolio return
(following, e.g., Lakonishok, Shleifer, and Vishny, 1992) [Panel B]. In the latter method, the valueweighted size decile portfolio includes all companies that are in the same size decile (using NYSE
size breakpoints) as the company of interest, as of the end of the most recent calendar year. The
“WRDS CRSP stock-portfolio assignments, capitalization deciles” is used to assign stocks to size
deciles.
The universe of events included in Table 7 are the days in which a given company held a
shareholders’ meeting which included a SOP proposal. Because the latter is already an event (that is,
the existence of the SOP proposal itself), in our CAR analysis, we control for this event with an
intercept. The variable of interest is the “fraction voted for SOP.” As the coefficient estimates
indicate, holding a meeting in which shareholders vote on SOP, but all shareholders (i.e., “fraction
voted for SOP”=1.0) vote in favor of management leads to almost no price reaction for the stock
(i.e., the estimated constant plus the estimated coefficient on “fraction voted for” multiplied by 1
sum to roughly zero). However, as opposition for SOP grows, the CAR becomes increasingly
negative. For example, Regression 3 in Panel A of Table 7 indicates that, using a nine-day window (4,+4), a decrease in the SOP support rate, as indicated by a shift from the 25th percentile of “fraction
voted for SOP” (0.8758) to the 1st percentile (0.3885), is expected to lead to a 0.83% CAR decrease
((0.3885-0.8758)*0.017). Panel B indicates a similar magnitude.12
We note that Gillan and Starks (2000) do not find a relation between the fraction of votes
cast in support of proposals submitted by shareholders and abnormal returns in the 1987-1994
period. Hence, the market has not always responded to the outcome of shareholder votes. However,
We note that, in unreported specifications, we do not find a discontinuity around the 50% opposition rate and its
effect on CAR. We generally find a linear relation–the larger the SOP opposition rate, the more negative the CAR.
12
22
as we show, the market responds to the SOP vote.
It is possible that the negative CARs following a SOP vote with low support rates may be
driven by a subset of weak, or “problematic” companies. To understand if this is indeed the case, in
Panel C of Table 7, we split the sample depending on whether ISS recommended to vote “against”
SOP (Regressions 1-4, capturing the potentially “problematic” companies), versus “for” SOP
(Regressions 5-8, capturing the “non-problematic” companies). The results of these two sets of
regressions are very similar, as is evident from the insignificant chi-squared test reported in Panel D,
which compares, formally, each pair of corresponding regression coefficients from Panel C (e.g., the
estimated coefficients from Regressions 1 versus 5 of Panel C are compared in column 1 of Panel
D).
In unreported specifications, we also split the sample by whether Glass Lewis (the chief
competitor to ISS) recommended to vote for/against, top/bottom quintile of residual
compensation, and top/bottom quintile of abnormal returns, and, again, do not find significantly
different results in one subsample versus the other corresponding subsample. Thus, our results do
not seem to be driven by a particular subset; higher opposition to SOP leads to a lower (or more
negative) abnormal return, in expectation.
5.5. When do Small Institutional Shareholders Vote Against SOP?
We now return to the voting data and ask how do small shareholders identify companies for which
they will vote against SOP? Do they follow a certain rule, e.g., companies with weak performance (as
is the case in Fos, Li, and Tsoutsoura, 2015, and Fos 2015)? Alternatively, perhaps particularly some
investors rely on proxy advisory services to flag problematic firms. For as Iliev and Lowry (2015)
show, mutual funds with large holdings are more likely to decide, independently from ISS, how to
vote. In this section, we address this issue by contrasting different subsets. In Panel A of Table 8, we
23
examine votes cast on the fund level, while, in Panel B of Table 8, we repeat the analysis on the
institutional level.
In Regression 1 of Panel A, we include only the cases in which ISS recommended to vote
against SOP, while, in Regression 2, we include only the cases in which ISS recommended to vote
for SOP. Panel A documents that the coefficient magnitudes of both holding variables (“fund's
portfolio weight” and “fraction of company's shares held by fund”) are larger in the subset in which
ISS recommends to vote against SOP, compared to the subset in which ISS recommended to vote
for SOP. This result is consistent with Iliev and Lowry (2015) who document that funds are
particularly likely to vote independently of ISS, when ISS recommends to vote against management.
Nevertheless, as indicated in the bottom section of Panel A, a chi-squared test shows that the
differences in magnitude are significantly different only for the coefficients on “fund's portfolio
weight” but not for “fraction of company’s shares held by fund.” Results are similar when we split
the sample by whether Glass Lewis (the second largest proxy advisory company) recommended to
vote against SOP (Regression 3) versus for SOP (Regression 4).
When we contrast the observations in which compensation is within the top quartile of
residual compensation (Regression 5) with those in the bottom quartile (Regression 6), the
coefficients of the two holding variables are not significantly different from each other, as indicated
by the chi-squared tests reported. Similar results are obtained when we contrast the bottom quartile
of abnormal returns (Regression 7) with the top quartile (Regression 8). Taken together, these
results imply that funds with small holdings tend to identify companies for which they are likely to
vote against SOP based on the recommendations issued by proxy advisory companies, rather than
compensation or performance measures.
In Panel B of Table 8, we repeat the abovementioned analysis at the institution level. The
chi-squared test reported documents that both holding variables (“portfolio weight of institution's
24
portfolio” and “fraction of company's shares held by institution”) are larger in the subset in which
ISS recommends to vote against SOP (Regression 1) as compared to when ISS recommended to
vote for SOP (Regression 2). Results are significant at the 1% and 10% level, respectively. None of
the other pairs of subsets examined, as specified above with respect to the mutual funds
specifications, are significantly different from each other in Panel B of Table 8. Taken together,
these results demonstrate that funds and institutions with a small magnitude of holding particularly
rely on ISS to identify the “problematic” companies for which they are likely to oppose SOP. Funds
and institutions with large holdings more frequently vote with management than small shareholders,
and often disregard a negative recommendation by the proxy advisory services.
Following Admati and Pfleiderer (2009), we address the question of whether funds that have
the option to sell their shares rather than “demonstrating” against management by voting against
SOP, vote differently from funds that do not have the option to walk the “Wall Street walk.” The
natural approach for examining this question is by analyzing the votes casts by index funds, as
opposed to those cast by actively managed funds. Index funds may monitor less intensively than
actively managed funds (and thus tend to vote for SOP), since their explicit goal is to simply track
an index. On the other hand, index-funds are “stuck” with their companies for the long run, and,
therefore, may prefer to monitor their companies more aggressively and vote against SOP.
Accordingly, we examine whether our results apply to a subset of index funds (Regression 9
of Panel A, Table 8) and to one of actively managed funds (Regression 10).13 We find that our prior
result, that mutual funds with a small ownership stake (i.e., a small portfolio weight, and a small
fraction of company’s shares held by the fund), are more likely to vote against SOP, apply only to
We categorize funds as index funds if CRSP flags the fund as an index fund, or the fund’s name suggests this is an
index find (e.g., fund name contains the words “index” or “idx” or “S&P 500” or “Russell 1000”).
13
25
the subset of actively managed funds, but not to that of the index fund.14 Hence, the SOP is
particularly used when funds (own a small stake and) actively manage their investments, giving them
an alternative to “demonstrating” against management. Moreover, we note that Appendix B
documents that, when a fund votes against SOP and the SOP vote fails, if the fund holds a large
fraction of outstanding shares, it is also likely to sell shares. Hence, there seems to be some positive
relation between the vote that is cast by a fund, and the likelihood that the fund subsequently sells its
holding of a given stock.
5.6. Robustness Examinations
One may wonder if the small-shareholders-opposing-management pattern is a phenomena that is
prevalent beyond the SOP vote. In Appendix C, we examine this possibility by repeating our threelevel analysis (aggregate, mutual fund, and institutional) for other types of corporate governance
proposals bought up for vote at shareholders meetings. We do not find consistent evidence
documenting that this pattern prevails for other types of proposals, which further emphasizes that
particularly the SOP vote is a low cost monitoring opportunity that small-scale shareholders appear
to view as being effective.
In addition, in an unreported specification, we examine whether a non-linear relation exists
between the holding variables we examine (both on the fund level and the institutional level) and the
SOP vote cast. We do not find convincing evidence for such a relation. We also examine if voting
patterns are different just below versus just above the 30% and 50% opposition rate thresholds. We
do not find consistently different patterns just above versus just below these thresholds.
Finally, as an additional robustness check for the fund level analysis, we re-estimate our
As the chi-squared tests indicate, the differences are statistically weak, probably because the set of index funds is both
much smaller than the set of actively managed funds, and because index funds in our sample track many different
indexes, meaning that the overlap in stockholdings among index funds is reduced.
14
26
results using Thompson s-12 mutual fund holdings data, rather than the CRSP holding dataset. We
do this analysis because each of these two datasets includes mutual funds that are not included in the
other dataset (Schwarz and Potter, 2015). We find that the holding variables that we compute and
match in both datasets are highly correlated. In unreported specifications in which we use the
Thompson data, the results are very similar to the results reported throughout this paper.
6.
Shareholder Composition and a Company’s Response to SOP
Thus far, we have documented that shareholder composition is related to the votes shareholders
cast. In this section, we address the question of whether shareholder composition is related to
whether and how a company responds to the SOP vote.15 Ertimur, Ferri, and Stubben, (2010), and
Bach and Metzger (2015) highlight that a company’s response in practice (“implementation”) to a
vote, is the measure that captures whether a non-binding vote is effective.
To measure the implementation of the nonbinding SOP vote, we focus on immediate
actions a management/board of a company can take to demonstrate to shareholders that the
company is responding to shareholder criticism, as reflected in a SOP vote that receives low support
rates. Accordingly, we examine the relation between SOP and subsequent: (1) CEO turnover within
12 months of the voting date; (2) cherry-picking of peer-companies selected for determining
executives compensation; and (3) change in the growth rate of the residual compensation awarded to
the CEO.16
Previous studies have found that non-binding SOP proposals are generally perceived as value enhancing. For example,
Cuñat, Gine and Guadalupe (2013) find that voluntary adoption of SOP increases the market value and profitability of a
company. Ferri and Maber (2013) and Iliev and Vitanova (2015) document that the UK and American markets,
respectively, reacted positively to the requirement to comply with SOP. Correa and Lel (2015) document that companies
in countries that have adopted a SOP vote have experienced a slower increase in CEO compensation, and a higher pay
for performance sensitivity compared to companies in countries that did not adopt a SOP vote.
16 Some studies report that a 30% opposition rate is sufficient to nudge a company to respond to a SOP vote (Ertimur,
Ferri, and Oesch, 2013). Other studies argue that a 50% threshold is the point at which response rates jump (Cuñat,
Gine, and Guadalupe, 2012). In unreported specifications and graphs, we examine both of these thresholds, but find that
generally, a linear relation exists between the SOP support rates and the three outcome variables we examine.
15
27
6.1.
CEO Turnover
We examine if low support rates for the SOP vote are associated with CEO turnover. We
emphasize that this relation may be endogenous, since other factors, such as performance, can affect
both the SOP vote and CEO turnover. Nevertheless, the SOP vote reflects the extent to which
shareholders are unhappy. Examining if low support rates for SOP are followed by CEO turnover
demonstrates whether the extent to which shareholders are dissatisfied, as reflected in the SOP
feedback vote, is associated with a change occurring in company leadership, or whether shareholder
opinions are ignored by the board.
Indeed, in Regression 1 of Panel A of Table 9 we find that companies that received low
support rates on SOP, are significantly more likely to experience CEO turnover within 12 months of
the voting day (the dependent variable in Table 9 equals one if the latter is the case, and zero
otherwise). To understand how responsiveness to SOP differs, given the shareholder structure, we
break the data into subsets partitioned by shareholder structure. We hypothesize that, when no
blockholder exists, companies will not respond to the SOP vote. Indeed, we do not find a relation
between the SOP vote outcome and CEO turnover in companies without a blockholder (Table 9,
Panel A, Regression 2).
We follow previous studies (Morck, Shleifer, and Vishny, 1988; McConnell and Servaes
1990; Hermalin and Weisbach 1998; Holderness, Kroszner, and Sheehan, 1999; Himmelberg,
Hubbard, and Palia, 1999), which have distinguished between companies in which the executives
hold at least 5% of the shares, versus less than that threshold. The notion of these studies is that, in
companies in which executives hold at least 5% of the shares, the executives may have substantial
power. Therefore, we do not expect to observe a response to SOP in these companies. Accordingly,
in Regression 3, we include only companies in which executives aggregately hold a block equal to, or
28
exceeding 5%.17 Indeed, we find that such companies are less likely to experience CEO turnover
following a SOP vote with low support rates.
However, as predicted, we do find that, in companies that have a non-insider block (and no
block held by insiders), CEO turnover is significantly more likely to occur following a SOP vote that
received low support rates (Regression 4). Hence, the results imply that companies are most likely to
demonstrate responsiveness to shareholder satisfaction, as reflected in the SOP vote, when a noninsider blockholder is present.
To understand if the SOP vote is unique with respect to its association with CEO turnover,
we analyze the relation between shareholder votes on director appointments and CEO turnover.
Before the SOP era, as argued by Cai, Garner, and Walking (2009), the “best” opportunity
shareholders had to express their discontent from management was by voting against the directors
nominated to the board. Accordingly, we examine, in Panel B of Table 9 for the pre-SOP period
(2008-2010), and in Panel C of Table 9 for the SOP period (2011-2013), the relation between the
votes cast against directors and CEO turnover. All control variables reported in Panel A of Table 9
are included in Panels B and C of Table 9 (but for brevity, are not necessarily reported).
The results in Panel B document for the pre-SOP era (2008-2010), when shareholders voted
against appointing directors, CEO turnover was not significantly more likely to follow. In contrast,
Panel C of Table 9 documents that during the SOP era (2011-2013), low support rates for director
appointments increased the likelihood of CEO turnover for companies with a non-insider
blockholder (Regression 4). These results highlight that voting against management/board was only
associated with CEO turnover in the SOP era, during which shareholders had an increased
opportunity to provide direct feedback on management performance.
17
As reported in the holdings of “all current executive officers and directors as a group” item in the proxy statement.
29
6.2.
Peer-companies
We start by examining whether, in the year following the SOP vote, changes are observed in the
peer companies selected for benchmarking and determining the compensation awarded to the
named executives. Since 2006, the SEC requires companies to disclose which peer-companies they
use to benchmark and determine the compensation of their named executives. Faulkender and Yang
(2010), Bizjak, Lemmon, and Ngujen (2011), and Faulkender and Yang (2012) find that companies
generally choose peer-companies that pay relatively large compensation to the CEO. Perhaps the
SOP vote offers an opportunity to govern such cherry-picking of peer-companies.
We first examine whether companies add or exclude peer companies in the year following
the SOP vote, given the SOP outcomes. On average, companies choose 19 peer-companies in a
given year. We find that following low SOP support rates, companies are significantly more likely to
add new peer-companies (Table 10, Regression 1), but not significantly likely to exclude existing
ones (Table 10, Regression 2). Accordingly, in subsequent regressions in Table 10 we focus on the
new peer-companies added in the year following the vote.
Companies can cherry-pick unreasonable peer-companies that allow inflating compensation
by picking, for example, larger peer-companies from better paying industries. To estimate the extent
an inflated peer-company is selected, we use the difference between the predicted compensation of
the peer company minus the predicted compensation of the “origin” company (i.e., the company for
which a SOP vote is held). The predicted compensation is calculated using the Core, Guay, and
Larcker (2008) methodology, as specified in Section 5.1. Because the predicted compensations takes
into account the factors that should affect the compensation awarded, the larger the difference
between these two predicted compensations, the larger the extent the origin company is picking
inflated peer-companies that are not similar to the origin company.
In Regressions 3-6 of Table 10, our dependent variable is “new peer inflation below that of
30
prior year.” This variable is equal to one if the predicted compensation of the new peer is smaller
than the average predicted compensation of the recurring peer companies (i.e., those picked both in
the SOP vote and in the year following the SOP vote). Indeed, Regression 3 of Table 10, which
includes all observations, documents that companies with low SOP support rates are more likely to
pick more modest peer companies. This finding is in line with Ertimur, Ferri, and Oesh (2013), who
document that 55% of the companies who received a negative ISS recommendation on SOP, were
likely to report in the proxy of the year following the SOP vote, that they restricted their
compensation.
In Regression 4, which is restricted to observations of companies without a blockholder, we
do not find that low SOP support rates decrease inflated peer picking. Similarly, in Regression 5,
which includes only companies in which executives aggregately hold a block equal to, or exceeding
5%,18 we do not find that more modest peer-companies are chosen following a SOP vote which
garnered low support rates. In Regression 6, we include only observations pertaining to companies
that have a non-insider blockholder. Similar to the results in the previous section, we observe that
following a SOP vote that yielded low support rates, companies with a non-insider blockholder are
likely to pick more reasonable peer-companies. Put differently, when a non-insider block is present,
companies seem to respond to a low-support-SOP vote by picking better matching peer-companies.
6.3.
Compensation
In unreported specifications, we find that companies that receive low SOP support rates still exhibit,
in the year following the SOP vote, significantly larger total compensation and also significantly
larger residual compensation (as defined in Section 5.1). This finding applies also to the companies
with a non-insider blockholder. This raises the question of whether the SOP vote is able to restrain,
18
As reported in the holdings of “all current executive officers and directors as a group” item in the proxy statement.
31
at least to some extent, the compensation awarded.
To address this question, in Table 11, we report a diff-in-diff regression in which the
dependent variable is the “percentage of change in residual compensation.” This variable essentially
captures the change in the growth rate of the residual compensation (as defined in Section 5.1)
following the SOP vote.19 Regression 1 of Table 11 reports that, indeed, companies with low SOP
support rates experience a decrease in the “percentage of change in residual compensation” in the
year following the SOP vote.
Similarly to the results above, following a SOP vote with low support rates, we do not
observe a significant decline in the “percentage of change in residual compensation” for companies
that do not have a blockholder (Regression 2), or have an insider blockholder (Regression 3).
However, Regression 4 documents that companies with a non-insider blockholder are significantly
more likely to experience a decrease in the “percentage of change in residual compensation”
following low SOP support rates. Hence, although SOP does not catalyze companies to set lower
compensation and smaller residual compensation compared to other companies, SOP does seem to
restrain the growth rate of the residual compensation when a non-insider blockholder is present.
The finding that companies tame the residual compensation following a SOP vote with low
support rates is consistent with Ferri and Maber (2013), who document that in the UK, companies
that received low support rates for the SOP vote reduced their severance pay, and removed
provisions allowing to reevaluate compensation when original targets were not met.
* * *
To summarize, in Section 5, we have documented that small shareholders are likely to
oppose SOP. In Section 6, consistent with Kandel, Massa, and Simonov (2011), we find that the
We compute the “change in residual compensation” variable by first calculating the percentage of change in the
residual compensation awarded between the year following the SOP vote and the SOP vote year, and subtracting from
this figure the percentage of change in the residual compensation awarded between the SOP vote year and the year prior
to the vote.
19
32
actions taken by many small shareholders (in our case, voting against SOP) can have a disciplinary
force on management. They pressure the company to respond to the criticism expressed via the
SOP vote. However, in the spirit of Levit and Malenko (2011), we find that management is more
likely to respond to a non-binding vote if a blockholder who can discipline management is present.
Our findings in Section 6 mirror this notion, because we demonstrate that management is
particularly likely to demonstrate responsiveness to a SOP vote if a non-insider blockholder is
present. Such a blockholder probably has at least some ability to discipline management because he
holds a large stake in the company.
7.
Conclusion
Our study documents that offering low cost monitoring opportunities increases the extent small
institutional shareholders are involved in monitoring the companies they hold. However, we also
find that companies are likely to demonstrate responsiveness to the SOP vote particularly when a
non-insider blockholder is present. This suggests that both small and large shareholders may use the
SOP vote, in different ways, to pressure companies to respond to shareholder’s criticism.
33
Glossary of Variables
Variable name
Definition
Dataset use
CRSP
Average compensation of 4 executives t-1 (in
millions)
CEO age (years)
Firm abnormal return above the value weighted
market portfolio, for the 12 months preceding
the vote
Average compensation of top 4 executives,
computed by authors
Age of CEO
CEO tenure (years)
Tenure of CEO
ISS dataset on executives
Fraction of shares held by blockholders
Fraction of outstanding shares held by
blockholders that each hold at least 5% of the
outstanding shares
Aggregate fraction of shares held by executives
GMI, based on proxy data
Fraction of shares held by institutions
Total number of shares held by institutionals/
number of shares outstanding
Thompson s-34 and CRSP,
correspondingly
Fraction voted for SOP
Fraction of votes cast for SOP/ all SOP votes
cast
ISS voting analytics dataset
ISS recommended to vote for SOP
Equals one if ISS recommended to vote for
SOP, and zero otherwise
shrout*prc/1,000
ISS voting analytics dataset
ROA of company t-1
ebitda/( the one year lagged “at”, i.e., total
assets)
Compustat
Total compensation of CEO t-1 (in millions)
Total compensation of CEO
ISS compensation data, based on proxy
data
Annual netflow of fund
We estimate first the monthly inflows (after
taking in account the monthly return), and then
estimate the total netflows during the 12 months
preceding the vote.
CRSP mutual fund
Expense ratio (weighted average of shareclasses)
Weighted average (by class) of fund's expense
ratio - "fexp_ratio.”
CRSP mutual fund
Fraction of company's shares held by fund
nbr_shares/( shrout2 *1000)
Fund twelve-months characteristic selectivity
return
Calculated by the authors using the Daniel,
Grinblatt, Titman & Wermers (1997) approach.
CRSP mutual fund and CRSP,
respectively
Thompson Reuters s12
Fund voted for SOP
A binary variable that equals one if the fund
voted for SOP, and zero otherwise.
ISS voting analytics dataset
Fund's portfolio weight (in fraction)
percent_tna/100, where percent_tna is the
"security’s percentage of the total net assets in
the portfolio"
Number of funds voting on proposal included in
the ISS voting analytics dataset
CRSP mutual fund
Total net assets managed by fund (in thousand $)
mtna/1000, where mtna is defined as “assets
minus total liabilities as of month-end.”
CRSP mutual fund
Turnover ratio (weighted average)
Weighted average (by class) of fund's turnover
ratio – “fturn_ratio.”
CRSP mutual fund
Company level variables
Abnormal return of company
Fraction of shares held by executives
Market capitalization in $Millions
ISS compensation data, based on proxy
data
ISS dataset on executives
GMI, based on proxy data
CRSP
Mutual fund level variables
Number of funds voting on proposal
34
ISS voting analytics dataset
Variable name
definition
Dataset use
Institution level variables
Fraction of company's shares held by institution
shares/( shrout2*1000)
Thompson s-34 and CRSP,
correspondingly
ISS voting analytics dataset
Fraction of funds voted for SOP
Fraction of funds within institution that voted
for SOP
Number of institutions voting on proposal
Number of institutions voting on proposal
included in the ISS voting analytics dataset
ISS voting analytics dataset
Institution’s portfolio weight (in fraction)
prc*shares/ total assets managed by institution.
Thompson s-34
Total assets managed by institution in trillions
The sum of the value (prc*shares) all holdings of
an institution in a given quarter/ one trillion
Thompson s-34 and CRSP,
correspondingly
35
Bibliography
Admati, A. R., and P. Pfleiderer (2009). The “Wall Street Walk” and Shareholder Activism: Exit as a Form of Voice. Review of
Financial Studies 22, 2645-2685.
Aggarwal, R., Erel, I., and Starks, L. T. (2015). Influence of Public Opinion on Investor Voting and Proxy Advisors.
Georgetown McDonough School of Business Research Paper, (2447012).
Appel, I., Gormley, T. A., and Keim, D. B. (2015). Passive Investors, Not Passive Owners. Working Paper.
Bach, L., and Metzger, D. (2014). Why Do Shareholder Votes Matter?. Available at SSRN: http://ssrn.com/abstract=2247084.
Bew, R., and Fields, R. (2012). Voting Decisions at US Mutual Funds: How Investors Really Use Proxy Advisers. Available at
SSRN 2084231.
Bizjak, J., Lemmon, M., and Nguyen, T. (2011). Are all CEOs Above Average? An Empirical Analysis of Compensation Peer
Groups and Pay Design. Journal of Financial Economics, 100(3), 538-555.
Burr, Barry B. (2012). Money Managers Increasing Activism on Governance — But Quietly, Pension and Investments.
http://www.pionline.com/article/20120319/PRINT/303199980/money-managers-increasing-activism-on-governance-8212but-quietly
Cai, J., Garner, J. L., and Walkling, R. A. (2009). Electing Directors. The Journal of Finance, 64(5), 2389-2421.
Carhart, M. M. (1997). On persistence in mutual fund performance. The Journal of finance 52, 57-82.
Correa, R., and Lel, U. (2015). Say on Pay Laws, Executive Compensation, CEO Pay Slice, and Firm Value Around the World.
Journal of Financial Economics, forthcoming.
Crane, A. D., Michenaud, S., and Weston, J. P. (2012). The Effect of Institutional Ownership on Payout Policy: A Regression
Discontinuity Design Approach. Rice University Working Paper.
Cuñat, V., Gine, M., and Guadalupe, M. (2012). The Vote is Vast: the Effect of Corporate Governance on Shareholder Value.
The journal of finance, 67(5), 1943-1977.
Cuñat, V., Gine, M., and Guadalupe, M. (2015). Say Pays! Shareholder Voice and Firm Performance. Review of Finance,
forthcoming.
Daniel, K., Grinblatt, M., Titman, S., and Wermers, R. (1997). Measuring mutual fund performance with characteristic-based
benchmarks. Journal of Finance 52, 1035-1058.
Davis, G. F., and Kim, E. H. (2007). Business Ties and Proxy Voting by Mutual Funds. Journal of Financial Economics, 85(2), 552570.
Dimmock, S. G., Gerken, W. C., Ivkovic, Z., and Weisbenner, S. J. (2014). Locked-in to Govern: How the Capital Gain of a
Stock Holding Affects a Mutual Fund’s Voting Decision, Working Paper
Ertimur, Y., Ferri, F., and Muslu, V. (2010). Shareholder Activism and CEO Pay. Review of Financial Studies, 17 (2), 527-563.
Ertimur, Y., Ferri, F., and Oesch, D. (2013). Shareholder Votes and Proxy Advisors: Evidence From Say on Pay. Journal of
Accounting Research, 51(5), 951-996.
36
Ertimur, Y., Ferri, F., and Stubben, S. R. (2010). Board of Directors' Responsiveness to Shareholders: Evidence from
Shareholder Proposals. Journal of Corporate Finance, 16(1), 53-72.
Faulkender, M., and Yang, J. (2010). Inside the Black Box: The Role and Composition of Compensation Peer Groups. Journal of
Financial Economics, 96(2), 257-270.
Faulkender, M., and Yang, J. (2012). Is Disclosure an Effective Cleansing Mechanism? The Dynamics of Compensation Peer
Benchmarking. Review of Financial Studies, 26(3), 806-839.
Ferri, F., and Maber, D. A. (2013). Say on Pay Votes and CEO Compensation: Evidence From the UK. Review of Finance, 17(2),
527-563.
Ferri, F., and Oesch, D. (2014). Management Influence on Investors: Evidence from Shareholder Votes on the Frequency of
Say on Pay. Columbia Business School Research paper, (13-17).
Fich, E. M., Harford, J., and Tran, A. L. (2014). Motivated Monitors: The Importance of Institutional Investors’ Portfolio
Weights. Available at SSRN 2463718.
Fluck, Z. (1999). The Dynamics of the Management-Shareholder Conflict. Review of Financial Studies 12(2), 379-404.
Fos, V. (2015). The Disciplinary Effects of Proxy Contests Management Science, Forthcoming.
Fos, V., and Kahn, C. M. (2015). Governance through Threats of Intervention and Exit. Available at SSRN:
http://ssrn.com/abstract=2527710.
Fos, V., Li, K. and Tsoutsoura,
http://ssrn.com/abstract=2609815.
M.
(2015).
Do
Director
Elections
Matter?
Available
at
SSRN:
Gillan, S. L., and Starks, L. T. (2000). Corporate governance proposals and shareholder activism: The role of institutional
investors. Journal of financial Economics, 57(2), 275-305.
Grossman, S. J., and Hart, O. D. (1986). The Costs and Benefits of Ownership: A Theory of Vertical and Lateral Integration.
The Journal of Political Economy, 691-719.
Hart, O. (1995). Corporate Governance: Some Theory and Implications. The economic journal, 678-689.
Hermalin, B. E., and Weisbach, M. S. (1988). The Determinants of Board Composition. The RAND Journal of Economics, 589606.
Himmelberg, C. P., Hubbard, R. G., and Palia, D. (1999). Understanding the Determinants of Managerial Ownership and the
Link between Ownership and Performance. Journal of financial economics, 53(3), 353-384.
Holderness, C. G., Kroszner, R. S., and Sheehan, D. P. (1999). Were the Good old Days that Good? Changes in Managerial
Stock Ownership Since the Great Depression. The Journal of Finance, 54(2), 435-469.
Iliev, P., and Lowry, M. (2015). Are Mutual Funds Active Voters?. Review of Financial Studies, 28(2), 446-485.
Iliev, P., and Vitanova, S. (2015). The Effect of the Say-on-Pay in the US. Available at SSRN 2235064.
Jensen, M. C., and K. J. Murphy (1990). Performance Pay and Top-Management Incentives. Journal of Political Economy 98, 225264.
37
Kandel, E., Massa, M., and Simonov, A. (2011). Do Small Shareholders Count?. Journal of Financial Economics 101(3), 641-665.
Kronlund, M., and Sandy, S. (2015). Does Shareholder Scrutiny Affect Executive Compensation? Available at SSRN:
http://ssrn.com/abstract=2358696.
Lakonishok, J., Shleifer, A., and Vishny, R. W. (1992). The impact of institutional trading on stock prices. Journal of financial
economics, 32(1), 23-43.
Larcker, D. F., McCall, A. L., and Ormazabal, G. (2014). Outsourcing Shareholder Voting to Proxy Advisory Firms. Available
at SSRN: http://ssrn.com/abstract=2101453.
Malenko, N., and Shen, Y. (2015). The Role of Proxy Advisory Firms: Evidence from a Regression-Discontinuity Design.
Available at SSRN: http://ssrn.com/abstract=2526799
Matvos, G., and Ostrovsky, M. (2008). Cross-ownership, returns, and voting in mergers. Journal of Financial Economics, 89(3),
391-403.
Matvos, G., and Ostrovsky, M. (2010). Heterogeneity and Peer Effects in Mutual Fund Proxy Voting. Journal of Financial
Economics, 98(1), 90-112.
McCahery, J. A., Sautner, Z., and Starks, L. T. (2011). Behind the Scenes: The Corporate Governance Preferences of
Institutional Investors. In AFA 2011 Denver Meetings Paper.
McConnell, J. J., and Servaes, H. (1990). Additional Evidence on Equity Ownership and Corporate Value. Journal of Financial
economics, 27(2), 595-612.
Morck, R., Shleifer, A., and Vishny, R. W. (1988). Management Ownership and Market Valuation: An Empirical
Analysis. Journal of financial economics, 20, 293-315.
Morgan, A., Poulsen, A., Wolf, J., and Yang, T. (2011). Mutual Funds as Monitors: Evidence From Mutual Fund Voting.
Journal of Corporate Finance 17, 914-928.
Ng, S., and A. Troianovski (2015). Investors Prize Face Time With Bosses, The Wall Street Journal, p. A1.
ProxyPulse
(2014).
“How
Is
The
2014
Proxy
Season
Shaping
http://media.broadridge.com/documents/Broadridge-PwC-ProxyPulse-2nd-Edition-June2014.pdf
Up
Thus
Far?”
Schwarz, C., and Potter, M. (2015). Revisiting Mutual Fund Portfolio Disclosure. Available at SSRN 2093688.
Shleifer, A., and Vishny, R. W. (1986). Large Shareholders and Corporate Control. The Journal of Political Economy, 461-488.
Sidel and Hoffman (2015). Next Activist Target: AmEx. The Wall Street Journal, p. B1.
Spencer Stuart (2014). Spencer Stuart Board Index. Posted online –
https://www.nyse.com/publicdocs/nyse/listing/Spencer_Stuart_Board_Index_2014.pdf
Thomas, R. S., Palmiter, A. R., and Cotter, J. F. (2012). Dodd-Frank's Say on Pay: Will It Lead to a Greater Role for
Shareholders in Corporate Governance?. Cornell Law Review, 97, 11-49.
38
Table 1: Shares held and votes cast by mutual funds and institutions
This table estimates, for the 2011-2013 period, the shares held and votes cast by mutual funds and institutions, as
described in the paper.
Percentage of shares held by institutional investors
Estimation of percentage of SOP votes cast by institutional investors
Percentage of shares held by mutual funds
Estimation of percentage of SOP votes cast by mutual funds
39
Average
70.0%
87.8%
28.5%
35.7%
Median
29.2%
36.6%
Table 2: Summary statistics
This table reports summary statistics for the SOP observations during the 2011-2013 period included in this study.
Definitions of variables are specified in the Glossary.
n
Mean
0.25
percentile
median
0.75
percentile
S.D.
Abnormal return of company
Average compensation of 4 executives t-1 (in
$Millions)
6,472
0.0262
-0.1759
-0.0006
0.1745
0.3736
6,714
1.9013
0.7276
1.2442
2.2642
2.3838
CEO age
5,493
55.8889
51
56
61
7.2622
CEO tenure
6,306
8.3407
2.6301
5.8288
10.6027
10.9238
Fraction of shares held by blockholders
6,272
0.2679
0.144
0.249
0.364
0.1669
Fraction of shares held by executives
6,270
0.1051
0.02
0.045
0.1115
0.1568
Fraction of shares held by institutionals
5,259
0.6963
0.5853
0.7542
0.8565
0.2047
Fraction voted for SOP
6,735
0.8981
0.8759
0.9479
0.9752
0.127
ISS recommended to vote for SOP
6,770
0.8703
1
1
1
0.336
Market capitalization in $Millions
6,579
6627.54
437.11
1279.33
4067.23
22504.82
Number of institutional shareholders
6,045
219.3828
93
145
255
221.2978
ROA of company t-1
6,079
0.1109
0.0499
0.1187
0.1842
0.4942
Total compensation of CEO t-1 (in $Millions)
6,725
5.1781
1.5163
3.1844
6.5596
7.624
Annual netflow of fund
567,190
-0.0049
-0.1322
-0.0392
0.0766
0.2225
Expense ratio (weighted average of shareclasses)
582,947
0.0075
0.0025
0.007
0.0113
0.0051
Fraction of company's shares held by fund
538,135
0.0021
0
0.0002
0.0014
0.0127
Variable name
Company level variables
Mutual fund level variables
Fund twelve-month characteristic selectivity return
Fund voted for SOP
Fund's portfolio weight (in fraction)
Number of funds voting on proposal
421,395
0.0051
-0.0051
0.0024
0.0159
0.0357
1,282,574
0.8966
1
1
1
0.3045
546,144
0.005
0.0004
0.0016
0.006
0.0087
304.1387
150
244
405
209.0957
Total net assets managed by fund (in thousand $)
1,282,574
630,039
2.7008
0.099
0.3988
1.9398
8.7968
Turnover ratio (weighted average)
572,595
0.5531
0.12
0.33
0.78
0.6416
Institutional level variables
Fraction of funds voted for SOP
346,027
0.87194
1
1
1
0.32398
Fraction of company's shares held by institution
84,263
0.01383
0.00057
0.00292
0.01633
0.0231
Institution’s portfolio weight (in fraction)
84,263
0.00181
0.00006
0.00025
0.00114
0.00591
Total assets managed by institution in $Trillions
84,263
0.15485
0.01139
0.03571
0.20244
0.22758
346,027
62.97221
43
59
80
26.73242
Number of institutions voting on proposal
40
Table 3: SOP votes of Investment Advisors
This table documents for the 20 investment advisors with the largest number of votes cast, the average frequency
institutions cast SOP votes in the opposite direction of ISS recommendation (Column 3), and the standard deviation of
mutual fund’s SOP votes within the institution (Column 4).
Number of
votes cast
(2)
154,756
124,903
111,756
68,585
61,106
60,822
48,975
42,088
40,938
39,294
38,787
33,414
30,081
29,659
% votes
opposite ISS
recomm.
(3)
34%
7%
4%
0%
0%
9%
21%
8%
8%
11%
3%
0%
10%
15%
S.D. of votes
within
institution
(4)
1.02%
0.00%
3.67%
0.01%
0.00%
0.00%
0.04%
0.50%
1.96%
10.87%
1.89%
0.07%
4.07%
0.04%
29,375
13%
0.00%
16
American Century Investment Management,
Inc.
ING Funds
23,390
5%
0.46%
17
18
19
20
John Hancock Funds, LLC
Northern Trust Global Investments
Jackson National Asset Management, LLC
USAA Investment Management Company
22,518
22,048
22,008
21,880
3%
22%
26%
11%
5.37%
3.40%
20.38%
1.78%
5,016
535
11%
9%
3.05%
0.07%
Name of institution
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
(1)
BlackRock Advisors, Inc.
Vanguard Group, Inc.
Fidelity Management & Research
Dimensional Fund Advisors, Inc.
ProShare Advisors LLC
TIAA-CREF Asset Management LLC
Rydex Investments
T. Rowe Price Associates, Inc. (MD)
State Street Global Advisors
EQ ADVISORS TRUST
JPMorgan Asset Management, Inc. (US)
SEI Investments Management Corporation
Putnam Investment Management, Inc.
Charles Schwab Investment Management, Inc.
Average for all 357 fund families in study
Median for all 357 fund families in study
41
Table 4: Aggregate SOP vote outcomes
This table reports OLS regressions on the company-year level for the 2011-2013 period. The dependent variable equals
the fraction of votes cast for SOP. The regressions include year and Fama-French 48 industry fixed effects. Errors are
clustered on the company level. P-values are reported in parenthesis. * indicates p<.05, ** p<.01, and *** p<.001.
Fraction of shares held by
blockholders
(1)
0.0519***
(.000)
Fraction voted for SOP
(2)
(3)
0.0552***
0.0579***
(.000)
(.000)
(4)
0.1054***
(.000)
Fraction of shares held by
executives
0.1558***
(.000)
0.1621***
(.000)
0.1621***
(.000)
Total compensation of CEO t1 (in millions)
-0.0016**
(.029)
Average compensation of 4
executives t-1 (in millions)
0.1558***
(.000)
-0.0056***
(.001)
Predicted CEO compensation
t-1
-0.0241***
(.007)
Residual of CEO
compensation t-1
-0.0016**
(.029)
Predicted 4 executives
compensation t-1
-0.1844***
(.009)
Residual of 4 executives
compensation t-1
-0.0056***
(.001)
ROA of company t-1
0.0123**
(.039)
0.0117**
(.037)
0.0192***
(.003)
0.0072
(.221)
Abnormal return
0.0256***
(.000)
0.0250***
(.000)
0.0186***
(.000)
-0.0023
(.833)
Market cap in Billions of $
0.0002*
(.094)
0.0002**
(.032)
0.0021***
(.007)
0.0049***
(.009)
Fraction of shares held by
institutions
-0.0001
(.558)
-0.0001
(.419)
0.0003
(.159)
-0.0009***
(.009)
Number of institutional
shareholders
0.0000*
(.074)
0.0000***
(.001)
0.0003***
(.005)
0.0009***
(.008)
CEO tenure (years)
-0.0001
(.446)
-0.0001
(.354)
0.0005*
(.063)
0.0007**
(.046)
-0.0006***
(.010)
-0.0006**
(.010)
-0.0004
(.108)
-0.0001
(.817)
0.2831***
(.000)
0.2841***
(.000)
0.2831***
(.000)
0.2841***
(.000)
Yes
0.601
4,612
Yes
0.599
4,610
Yes
0.601
4,612
Yes
0.599
4,610
CEO age (years)
ISS recommended to vote for
SOP
Year and industry fixed effects
R-squared
N
42
Table 5: SOP votes cast by mutual funds
This table reports regressions on the fund-company-year level for the 2011-2013 period. The dependent variable equals
one if the fund voted for SOP. Errors are clustered on the fund level. The regressions include year and Fama-French 48
industry fixed effects. P-values are reported in parenthesis. * indicates p<.05, ** p<.01, and *** p<.001.
(1)
Fund's portfolio weight (in
fraction)
Fund voted for SOP
(3)
(4)
(5)
0.4284***
(.000)
7.0165***
(.007)
0.4895*
(.076)
0.5131**
(.010)
0.4572**
(.021)
13.2616**
(.018)
0.5828***
(.008)
(2)
0.4625***
(.000)
Fraction of company's shares
held by fund
Fraction of shares held by
blockholders
-0.0198***
(.000)
-0.0199***
(.000)
-0.0196***
(.000)
-0.5695***
(.000)
-0.0202***
(.000)
Fraction of shares held by
executives
-0.0308***
(.000)
-0.0316***
(.000)
-0.0313***
(.000)
-0.3280***
(.000)
-0.0313***
(.000)
Number of funds voting on
proposal
0.0065***
(.000)
0.0072***
(.000)
0.0068***
(.000)
0.1625***
(.000)
0.0065***
(.000)
Total compensation of CEO
t-1 (in Million $)
-0.0028***
(.000)
-0.0028***
(.000)
-0.0028***
(.000)
-0.0398***
(.000)
-0.0027***
(.000)
ROA of company t-1
0.0046*
(.065)
0.0045*
(.073)
0.0045*
(.074)
0.2309***
(.000)
0.001
(.719)
Firm abnormal return
0.0146***
(.000)
0.0150***
(.000)
0.0148***
(.000)
0.4528***
(.000)
0.0174***
(.000)
Market cap in Billions of $
0.0002***
(.000)
0.0002***
(.000)
0.0002***
(.000)
0.0000***
(.000)
0.0002***
(.000)
Fraction of shares held by
institutions
-0.0120***
(.008)
-0.0124***
(.008)
-0.0125***
(.007)
-0.0951
(.195)
0.0017
(.798)
CEO tenure (years)
-0.0002***
(.000)
-0.0003***
(.000)
-0.0003***
(.000)
-0.0020*
(.051)
-0.0002***
(.000)
CEO age (years)
-0.0006***
(.000)
-0.0006***
(.000)
-0.0006***
(.000)
-0.0165***
(.000)
-0.0008***
(.000)
Fund twelve-months
characteristic selectivity
return
-0.0109
(.697)
-0.013
(.646)
-0.0111
(.694)
-0.5807
(.465)
-0.1729**
(.019)
Annual inflow (fraction of
total assets)
0.0001*
(.086)
0.0001
(.113)
0.0001*
(.094)
0.0026***
(.005)
0.0001
(.160)
Expense ratio (weighted
average of shareclasses)
-1.464
(.648)
-1.4135
(.660)
-1.4304
(.656)
-49.4471
(.354)
-5.0894***
(.000)
Turnover ratio (weighted
average)
0.0003
(.917)
0.0006
(.848)
0.0006
(.851)
0.0384
(.777)
0.0142***
(.000)
Total net assets managed by
fund (in thousand $)
0.0008
(.420)
0.0008
(.439)
0.0008
(.446)
0.0119
(.603)
0
(.940)
0.4715***
(.000)
0.4714***
(.000)
0.4714***
(.000)
4.6028***
(.000)
0.4713***
(.000)
Yes
Yes
Yes
Yes
Yes
ISS recommended to vote for
SOP
Year and ind. fixed effects
Fund fixed effects
Type of regression
R-squared
N
Yes
Yes
Yes
Yes
No
OLS
0.444
OLS
0.443
OLS
0.443
Logit
OLS
0.309
259,270
256,934
256,934
241,125
256,934
43
Table 6: Votes cast on the institution level
Regressions 1-4 report OLS regressions on the institution-company-year level for the 2011-2013 period. The dependent
variable in these regressions equals the fraction of all funds within the institution that cast a vote for SOP. Regressions 57 report OLS regressions on the fund-company-year level for the 2011-2013 period. The dependent variable in the latter
regressions equals one if the fund voted for SOP. Regressions 5-7 include (but do not report) all fund control variables
included in Table 5. All regressions include year, and Fama-French 48 industry fixed effect. Institution/ fund fixed
effects are included as specified below. Errors are clustered on the institution level in Regressions 1-4 and 6-7, and on
the fund level in Regressions 5. P-values are reported in parenthesis. * indicates p<.05, ** p<.01, and *** p<.001.
Fraction of funds voted for SOP
(1)
Institution’s portfolio weight
(in fraction)
(2)
1.6098***
(.000)
Fraction of company's shares
held by institution
0.6056**
(.032)
Fund voted for SOP
(3)
(4)
(5)
(6)
(7)
1.1515***
(.006)
0.7293
(.475)
1.5921***
(.000)
1.4678**
(.024)
0.0044
(.996)
0.5608**
(.047)
0.2537
(.143)
0.2056***
(.002)
0.2686
(.183)
0.1585
(.274)
0.2764***
(.008)
0.3145***
(.005)
0.4393**
(.022)
0.2838
(.175)
0.3888*
(.092)
-0.0627
(.826)
Fund's portfolio weight (in
fraction)
Fraction of company's shares
held by fund
Fraction of shares held by
blockholders
-0.0182**
(.011)
-0.0240***
(.001)
-0.0233***
(.001)
-0.0177*
(.051)
-0.0179***
(.000)
-0.0198**
(.035)
-0.0211***
(.000)
Fraction of shares held by
executives
-0.0202**
(.041)
-0.0174*
(.086)
-0.0175*
(.081)
-0.0252*
(.073)
-0.0344***
(.000)
-0.0366***
(.000)
-0.0301***
(.001)
Total assets managed by
institution in
-0.0813
(.294)
-0.1012
(.225)
-0.0991
(.233)
0.1136***
(.004)
-0.0325
(.404)
-0.1007
(.505)
0.1080***
(.000)
Number of institutions voting
on proposal
0.0003**
(.024)
0.0003***
(.002)
0.0003***
(.004)
0.0003*
(.061)
0.0004***
(.000)
0.0004***
(.004)
0.0005***
(.000)
Total compensation of CEO
t-1 (in )
-0.0037***
(.000)
-0.0036***
(.000)
-0.0036***
(.000)
-0.0036***
(.000)
-0.0026***
(.000)
-0.0026***
(.003)
-0.0029***
(.000)
ROA of company t-1
0.015
(.106)
0.0153
(.111)
0.0151
(.112)
0.0159
(.124)
0.0044
(.138)
0.0044
(.522)
0.0043
(.558)
Firm abnormal return
0.0112***
(.004)
0.0115***
(.003)
0.0113***
(.004)
0.0156***
(.001)
0.0107***
(.000)
0.0107***
(.000)
0.0105***
(.000)
Market capitalization in
0.0002***
(.000)
0.0003***
(.000)
0.0002***
(.000)
0.0002***
(.007)
0.0002***
(.000)
0.0002***
(.003)
0.0002***
(.003)
0
(.871)
0
(.843)
0
(.847)
0.0001
(.779)
-0.0145***
(.008)
-0.0145
(.191)
0
(.882)
-0.0001
(.348)
-0.0001
(.271)
-0.0001
(.278)
-0.0001
(.659)
-0.0003***
(.000)
-0.0003**
(.033)
-0.0002
(.186)
-0.0003**
(.023)
-0.0003**
(.034)
-0.0003**
(.036)
-0.0005***
(.002)
-0.0006***
(.000)
-0.0006***
(.000)
-0.0006***
(.000)
0.5084***
(.000)
0.5088***
(.000)
0.5087***
(.000)
0.5105***
(.000)
0.4469***
(.000)
0.4468***
(.000)
0.4603***
(.000)
Yes
Institution
0.484
68,514
Yes
Institution
0.485
68,514
Yes
Institution
0.485
68,514
Yes
None
0.315
68,514
Yes
Fund
0.435
183,577
Yes
Institution
0.428
183,577
Yes
None
0.336
315,948
Fraction of shares held by
institutions
CEO tenure (years)
CEO age (years)
ISS recommended to vote for
SOP
Year and ind. fixed effects
Fund/ institution fixed effects
R-squared
N
44
Table 7: CARs around SOP votes
This table reports the cumulative abnormal returns (CARs) around the SOP votes held in the 2011-2013 period.
Abnormal returns are calculated around the meeting date, by subtracting, from a company’s compounded return, the
compounded: value weighted market return (Panel A), value weighted size-decile portfolio (Panel B), value weighted
market return conditional on ISS recommending to vote against SOP (Specifications 1-4 of Panel C), value weighted
market return conditional on ISS recommending to vote for SOP (Specifications 5-8 of Panel C). In each panel we
regress the abnormal returns estimated on the fraction of votes cast in favor of management as well as an intercept.
Panel D compares the coefficients of each pair of regressions from Panel C (e.g., Regressions 1 versus 5 of Panel C are
compared in specification 1 of Panel D). In all panels, CARs are expressed in fractions. P-values are reported in
parenthesis. * indicates p<.05, ** p<.01, and *** p<.001.
Panel A: Value weighted CAR
Window
[-1, 1]
[-3, 3]
[-4, 4]
[-5, 5]
[-10, 10]
(1)
(2)
(3)
(4)
(5)
Fraction voted for
0.0088**
(.018)
0.0144***
(.008)
0.0170***
(.006)
0.0171**
(.010)
0.0171*
(.063)
Constant
-0.0074**
(.032)
-0.0121**
(.015)
-0.0151***
(.008)
-0.0154**
(.012)
-0.0140*
(.097)
7,123
7,123
7,122
7,121
7,117
N
Panel B: Value weighted size-decile CAR
Window
[-1, 1]
[-3, 3]
[-4, 4]
[-5, 5]
[-10, 10]
(1)
(2)
(3)
(4)
(5)
Fraction voted for
0.0083**
(.024)
0.0141***
(.008)
0.0173***
(.005)
0.0171***
(.010)
0.0190**
(.036)
Constant
-0.0084**
(.013)
-0.0152***
(.002)
-0.0193***
(.001)
-0.0202***
(.001)
-0.0256***
(.002)
7,276
7,276
7,274
7,273
7,267
N
Panel C: Value weighted CAR Broken Down by ISS Recommendation
CAR ISS recommend to vote against SOP
CAR ISS recommend to vote for SOP
Window
Fraction voted for
Constant
N
[-1, 1]
[-3, 3]
[-5, 5]
[-10, 10]
[-1, 1]
[-3, 3]
[-5, 5]
[-10, 10]
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
0.0111
(0.288)
-0.009
(0.21)
675
0.0300**
(0.032)
-0.0224**
(0.02)
675
0.0371**
(0.038)
-0.0283**
(0.021)
675
0.0665**
(0.011)
-0.0498***
(0.005)
675
0.0169**
(0.036)
-0.0159**
(0.036)
4598
0.0346***
(0.003)
-0.0339***
(0.002)
4598
0.0585***
(0.000)
-0.0587***
(0.000)
4596
0.0383*
(0.054)
-0.0426**
(0.024)
4594
Panel D: Differences in CARs given ISS’ recommendation (as documented in Panel C)
Specifications compared
Fraction voted for
Constant
(1)-(5)
(1)
-0.0058
0.14
(.0704)
0.0069
0.36
(0.550)
Difference of magnitude
Chi squared
Prob.> chi^2
Difference of magnitude
Chi squared
Prob.> chi^2
45
(2)-(6)
(2)
-0.0046
0.05
(0.827)
0.0115
0.44
(.505)
(3)-(7)
(3)
-0.0214
0.74
(0.388)
0.0304
2.2
(0.138)
(4)-(8)
(4)
0.0282
0.62
(0.430)
-0.0072
0.06
(0.801)
Table 8: When Do Small Institutional Shareholders Oppose SOP?
This table reports OLS regressions for the 2011-2013 period on the fund-company-year level (Panel A) and on the
institution-company-year level (Panel B). The dependent variable in: Panel A is equal to one if the fund voted for SOP,
and in Panel B is equal to the fraction of funds within the institution that voted in support of SOP. The regressions
reported in Panel A include all control variables reported in the Table 5 specifications, while those reported in Panel B
include all control variables reported in the Table 6 specifications. For brevity, not all control variables are reported in
these panels. The bottom section of each panel reports the differences in magnitudes for each pair of regressions, and a
chi-squared test indicating whether these differences are significant. Below each chi-squared test, (in parenthesis) the
probability> chi^2 is reported. The regressions include year, Fama-French 48 industry, and fund/ institution fixed
effects as noted. Errors are clustered on the fund level in Panel A, and on the institution level in Panel B. P-values are
reported in parenthesis. * indicates p<.05, ** p<.01, and *** p<.001.
Panel A: Fund level analysis
Fund voted for SOP
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
(10)
Fund's portfolio
weight (in
fraction)
1.8035***
(.000)
0.0263
(.638)
1.2718***
(.000)
0.2382***
(.002)
0.4183*
(.099)
0.5296***
(.000)
0.185
(.435)
0.5165***
(.006)
-0.091
(.624)
0.4736***
(.006)
Fraction of
company's shares
held by fund
0.7946**
(.017)
0.4418***
(.000)
0.5124*
(.093)
0.4598***
(.000)
0.5600*
(.065)
0.5573***
(.003)
-0.0269
(.913)
0.3171
(.184)
0.5497
(.624)
0.539***
(.000)
Fraction of
shares held by
blockholders
-0.0369**
(.028)
-0.0146***
(.000)
-0.0098
(.518)
-0.0282***
(.000)
-0.0160
(.177)
-0.0364***
(.000)
0.0066
(.492)
-0.0880***
(.000)
-0.0274***
(.001)
-0.0328***
(.007)
Fraction of
shares held by
executives
-0.0066
(.717)
-0.0146***
(.000)
-0.1825***
(.000)
-0.0589***
(.000)
-0.2076***
(.000)
-0.0251***
(.000)
-0.0829***
(.000)
-0.1490***
(.000)
-0.0980***
(.000)
-0.1136***
(.000)
ISS
recommend
against
ISS
recommend
For
GL
recommend
against
GL
recommend
For
Top
quartile
residual
comp.
Bottom
quartile of
residual
comp.
Bottom
quartile of
abnormal
return
Top
quartile of
abnormal
return
Index
funds
Non-index
funds
Subsample
Year, industry,
and fund fixed
effects
R-squared
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
0.548
0.372
0.345
0.243
0.257
0.238
0.197
0.247
0.161
0.326
N
30257
226677
47640
209218
55780
66574
69003
54867
56047
95879
Specifications
compared
Fund's port.
weight
Frac. of comp.
held by fund
(1) versus (2)
(3) versus (4)
(5) versus (6)
(7) versus (8)
difference
1.7772
chi-squared
18.18***
(0.000)
difference
1.0336
chi-squared
7.41***
(0.006)
difference
-0.1113
chi-squared
0.11
(0.734)
difference
-0.3311
0.3528
0.85
(0.357)
0.0526
0.01
(0.934)
0.0027
0.0
(0.9578)
-0.344
46
chisquared
0.62
(0.430)
1.22
(0.269)
(9) versus (10)
difference
-0.5642
chi-squared
2.20
(0.137)
0.0107
0.23
(0.633)
-Table 8 continuedPanel B: Institutional level analysis
Fraction of funds voted for SOP
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
2.0646***
(.000)
0.4995***
(.000)
0.8970***
(.001)
0.5386***
(.000)
0.9387***
(.003)
0.6612***
(.000)
0.5665***
(.001)
0.6149***
(.000)
3.3506**
(.013)
0.0514
(.852)
0.6187
(.646)
0.579
(.102)
-0.2822
(.789)
0.0057
(.992)
1.8219*
(.090)
-0.0903
(.912)
-0.0216
(.618)
-0.0130*
(.071)
-0.031
(.429)
-0.0208**
(.026)
-0.0219
(.525)
-0.0245
(.159)
-0.0215
(.393)
-0.0717***
(.000)
-0.1135***
(.008)
-0.0027
(.749)
-0.2690***
(.000)
-0.0454***
(.000)
-0.2524***
(.000)
-0.0469***
(.008)
0.1288***
(.000)
-0.1528***
(.000)
ISS
recommend
against
ISS
recommend
For
GL
recommend
against
GL
recommend
For
Top quartile
of residual
compensation
Bottom
quartile of
residual
compensation
Bottom
quartile of
abnormal
return
Top
quartile of
abnormal
return
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
R-squared
0.509
0.47
0.359
0.34
0.25
0.356
0.242
0.328
N
4525
34017
6921
31621
7376
10468
10424
8318
Fraction of company's shares
held by institution
Portfolio weight of
institution's portfolio (in
fraction)
Fraction of shares held by
blockholders
Fraction of shares held by
executives
Subsample
Year, industry, and institution
fixed effects
Specifications compared
Fund's port. weight
Frac. of comp. held by fund
(1) versus (2)
(3) versus (4)
(5) versus (6)
(7) versus (8)
difference
1.5651
chi-squared
8.74***
(0.003)
difference
0.3584
chi-squared
0.85
(0.357)
difference
0.2775
chi-squared
0.24
(0.622)
difference
-0.0484
chi-squared
.05
(.825)
3.2992
3.11*
(0.077)
0.0397
.01
(0.912)
-0.2879
0.06
(0.805)
1.9122
1.18
(0.225)
.
.
47
Table 9: The SOP vote and CEO turnover
Panel A reports OLS regressions on the company-year level for the 2011-2013 period, Panel B reports OLS regressions
on the director-year level for the 2008-2010 period, while Panel C reports OLS regressions on the director-year level for
the 2011-2013 period. The dependent variable in all three panels equals one if CEO turnover occurred within 12 months
of the vote, and zero otherwise. The primary variable of interest in Panel A is “fraction voted for SOP”, while in Panels
B and C it is “fraction voted for director.” Panels B and C include, but do not report, all the control variables reported in
Panel A. All regressions include fixed year and fixed Fama-French 48 industry effect. Errors are clustered on the
company level. P-values are reported in parenthesis. * indicates p<.05, ** p<.01, and *** p<.001.
Panel A: Fraction of votes cast against SOP
Was CEO replaced within 12 months following the vote?
(1)
(2)
(3)
(4)
-0.0929***
0.0644
-0.0463
-0.1379***
Fraction voted for SOP
(.003)
(.344)
(.277)
(.003)
Fraction of shares held
-0.0132
-2.2273
-0.0543*
0.0176
by blockholders
(.491)
(.233)
(.075)
(.500)
Fraction of shares held
0.0268
-0.0052
0.0165
-0.7215***
by executives
(.242)
(.791)
(.579)
(.009)
-0.0002
-0.0061
0.006
-0.028
ROA of company t-1
(.977)
(.797)
(.327)
(.242)
-0.0399***
0.0113
-0.0371***
-0.0441***
Abnormal return
(.000)
(.561)
(.000)
(.000)
Market capitalization in
0
0
0
0
$Millions
(.823)
(.958)
(.140)
(.912)
-0.0002
0.0016
-0.0004
0.0004
CEO tenure (years)
(.592)
(.165)
(.213)
(.504)
0.0016***
0.0029*
0.0016***
0.0017***
CEO age (years)
(.000)
(.055)
(.005)
(.009)
Vote examined
SOP
SOP
SOP
SOP
Period included
11'-13'
11'-13'
11'-13'
11'-13'
No
Insiders
Non-insider
Companies included
All
blockholders
block
block
R-squared
0.016
0.452
0.01
0.025
N
4,879
234
2,184
2,632
Panel B: Fraction of votes cast for directors before SOP era (2008-2010)
Was CEO replaced within 12 months following the vote?
(1)
(2)
(3)
(4)
0.0038
0.2448**
-0.0311
0.0318
Fraction voted for director
(.856)
(.039)
(.356)
(.251)
R-squared
0.013
0.054
0.032
0.024
N
13,213
1,212
5,897
6,783
Panel C: Fraction of votes cast against directors during SOP era (2011-2013)
Was CEO replaced within 12 months following the vote?
(1)
(2)
(3)
(4)
-0.0557**
0.3923***
0.0417
-0.1639***
Fraction voted for director
(.037)
(.008)
(.232)
(.000)
R-squared
0.04
0.213
0.031
0.074
N
10,424
692
4,304
5,915
48
Table 10: The SOP vote and subsequent changes in peer-companies
This table reports OLS regressions on the peer-company year level for the 2011-2013 period. The dependent variable: in
Regression 1 is equal to one if a peer-company was added and zero otherwise, in Regression 2 is equal to one if a peercompany was excluded and zero otherwise, and in Regressions 3-6 equals one if a new peer selected in the year following
the SOP vote is more “modest” than the peers selected in the year of the SOP vote. Specifically, this variable is equal to
one if the predicted compensation of a new peer is smaller than the average predicted compensation of the recurring
peer companies (i.e., those picked both in the SOP vote and in the year following the SOP vote). The primary variable of
interest is the “fraction voted for SOP.” All regressions include fixed year and fixed Fama-French 48 industry effect.
Errors are clustered on the company level. P-values are reported in parenthesis. * indicates p<.05, ** p<.01, and ***
p<.001.
New peer
added
Peer
excluded
(1)
(2)
(3)
(4)
(5)
(6)
Fraction voted for SOP
-0.1079**
(.032)
-0.0294
(.569)
-0.1260**
(.036)
-0.2951
(.682)
-0.0714
(.447)
-0.1615**
(.039)
Fraction of shares held
by 5% blockholders
0.0171
(.681)
0.0782*
(.078)
-0.0136
(.818)
0.0793
(.343)
-0.0397
(.663)
Fraction of shares held
by executives
0.0762
(.157)
0.0736
(.313)
0.0939
(.196)
0.2595
(.256)
0.1682*
(.057)
0.0821
(.935)
ROA of company t-1
-0.0387
(.423)
-0.0194
(.496)
0.0075
(.813)
0.4087
(.146)
0.0189
(.608)
0.061
(.417)
Abnormal return
0.0261
(.191)
0.0455**
(.023)
0.0023
(.916)
0.0265
(.888)
0.0115
(.624)
0.0381
(.341)
-0.0000***
(.000)
-0.0000*
(.067)
0
(.228)
0
(.940)
0
(.934)
0
(.181)
-0.0008
(.203)
-0.0003
(.677)
-0.0005
(.685)
-0.0009
(.935)
0.0012
(.494)
-0.0024
(.202)
-0.0020**
(.042)
0.0011
(.238)
0.0034**
(.011)
-0.009
(.233)
0.0041**
(.014)
0.0013
(.510)
All
All
All
No
blockholders
Insiders
block
Non-insider
block
Market capitalization in
$Millions
CEO tenure (years)
CEO age (years)
Residual compensation of new peer above average peer
residual compensation
Type of companies
included
Company and industry
fixed effects
R-squared
Yes
Yes
Yes
Yes
Yes
Yes
0.031
0.489
0.022
0.033
0.032
0.034
N
43,270
40,134
9,023
361
4,382
4,497
49
Table 11: The SOP vote and the percentage of change in residual compensation
This table reports OLS regressions on the company-year level for the 2011-2013 period. The dependent variable in these
regressions is the “percentage of change in residual compensation” which captures the change in the growth rate of the
residual compensation. This variable is computed by first calculating the percentages of change in the residual
compensation awarded between the year following the SOP vote and the SOP vote year, and subtracting from this figure
the percentages of change in the residual compensation awarded between the year of the vote and the year prior the
vote. The primary variable of interest is the “fraction voted for SOP.” All regressions include fixed year and fixed FamaFrench 48 industry effect. Errors are clustered on the company level. P-values are reported in parenthesis. * indicates
p<.05, ** p<.01, and *** p<.001.
Percentage of change in residual compensation
Fraction voted for SOP
(1)
(2)
(3)
(4)
11.5350**
(.045)
18.1432
(.887)
14.9802
(.214)
15.1592**
(.044)
18.4882
(.268)
-11.3710**
(.011)
Fraction of shares held by
blockholders
0.4187
(.947)
Fraction of shares held by
executives
-0.7523
(.775)
-15.1505
(.522)
1.8478
(.715)
-22.7423
(.595)
ROA of company t-1
-4.3706
(.509)
147.3772
(.529)
-5.373
(.681)
-1.8666
(.408)
Abnormal return
-0.2836
(.893)
-6.4557
(.625)
0.861
(.773)
-3.4383
(.181)
Market capitalization in
$Millions
0
(.983)
0
(.878)
0
(.536)
0
(.191)
CEO tenure (years)
0.161
(.517)
0.0924
(.799)
0.333
(.535)
0.0023
(.951)
-0.1993
(.309)
0.0808
(.873)
-0.5074
(.269)
0.0441
(.645)
All
No
blockholders
Insiders block
Non-insider
block
CEO age (years)
Type of companies
included
Company and industry
fixed effects
R-squared
Yes
Yes
Yes
Yes
0.012
0.473
0.036
0.024
N
2,016
65
782
1,216
50
Appendix A: Matching the voting analytics dataset to holding data
In this appendix we explain how we match the ISS voting analytics dataset to three datasets: CRSP
mutual fund dataset, Thomson Reuters S12 Database on mutual fund holdings, and Thomson s-34
dataset on the institutional 13-f filers.
CRSP mutual funds dataset. Unfortunately, the ISS voting analytics dataset on mutual
fund’s votes does not include conventional identifiers for mutual funds. However, ISS does provide
links to the N-PX form which include, in virtually all cases, a fund family CIK code and a mutual
fund “seriesid” identifier.20 Reporting a fund ticker in the N-PX filing is voluntary, and most mutual
funds do not do so. To increase the number of funds for which we are able to obtain a ticker, we
follow the procedure used by Matvos and Ostrovsky (2008, see footnotes 6 and 7), and Iliev and
Lowry (2014), and match the fund’s seriesid to at least one of the tickers reported in the company’s
filing section of the Edgar database.21
To further increase the number of mutual funds for which we are able to match a ticker, we
manually search in several additional databases for a ticker that is associated with the fund name and
the institution’s name, as reported in the N-PX filing. These additional databases include the CRSP
Mutual Fund Database, Thomson Reuters Database on mutual fund holdings S12, Factset, and
general searches on the internet.
Using all these approaches, we are able to match 40.2% of the SOP vote-observations
included in the Mutual Funds ISS Voting Analytics database to a fund ticker. However, for a given
company in a given year, the average aggregate holdings of mutual funds that we are able to match
to a ticker amount to 19.9% of the outstanding stocks. We estimate in Table 1 that 27.5% of the
20
The Seriesid identifier is assigned by the SEC, and uniquely identifies a mutual fund. To the best of our knowledge
the Seriesid identifier is not included in any of the mutual fund databases commonly available to academics.
21
In Edgar, https://www.sec.gov/edgar/searchedgar/companysearch.html, one may type a seriesid in the “Fast Search”
box, which leads to the hyperlink “List all Funds and Classes/Contracts for…” which details the available tickers of all
funds branching from the seriesid.
51
outstanding stocks are held, on average, by mutual funds. Hence, we are able to match voting
corresponding to the holdings of 72% (19.9%/27.5%) of the stocks held by mutual funds. Finally,
we search in the CRSP mutual funds dataset for each ticker we have found for each fund included in
the ISS voting analytics dataset, in a given quarter. If the quarter and the ticker match, we record the
corresponding crsp_portno, which is the fund identifier in the CRSP mutual funds dataset.
Thomson Reuters S12 mutual funds holdings. We match each ticker we have identified
for each of the funds included in the ISS voting analytics dataset in a given quarter to a WFICN
(using the MFLINK table available from Wharton Research Data Services), and then to the
Thomson fund identifier–“fundno.”
Thomson s-34 institutionals holdings. For each fund, we map the Thomson fund
identifier–“fundno,” to a Thomson institution identifier (“mgrno”), using the S12type5 file from
WRDS. The S12type5 file mapping is not always updated in cases in which one institution acquires
another institution. Accordingly, we manually examine, for each institution, whether the latter is the
case in the 2011-2013 period we study. In the cases a fund is held by an institution that is acquired
by another institution, we identify the correct institution by searching for the name of the fund in
Form N-SAR. This form identifies the primary adviser (i.e., institution) of each fund.
52
Appendix B: Are SOP votes translated to selling stocks?
Admati and Pfleiderer (2009) emphasize that selling stock is a one of the major possibilities
shareholders have to demonstrate their dissatisfaction with a stock they hold. Because the SOP vote
offers an additional route for shareholders to express dissatisfaction, in this appendix we examine if
the SOP vote also translates to shareholders “voting with their feet”, i.e., selling (or buying) stock.
Because the decision on how to allocate the portfolio’s assets is typically done on the fund level, we
conduct this analysis on the fund level.
In Table I we define the dependent variable “percentage change of the number of shares
held” as the percentage change in the number of shares held between the end of the quarter in
which the SOP vote was held, and the end of the quarter preceding the quarter at which the SOP
vote was held (i.e., (number of shares t+1 - number of shares t)/ number of shares t). Specifically, we
examine whether the fund is likely to sell/ buy stocks, conditional on the SOP vote it cast
(Regression 1), given that: the fund voted against SOP and the vote failed (Regression 2), the fund
voted against SOP and the vote passed (Regression 3), the fund voted for SOP and the vote failed
(Regression 4), or the fund voted for SOP and the vote passed (Regression 5).
Regression 1 does not document that the fund’s vote (“fund voted against proposed
compensation”) is related to the change in the magnitude of the holding (“percentage change of the
number of shares held”). Regressions 2-5 examine whether a certain combination of a specific vote
cast, and a certain vote outcome, is likely to catalyze changes in holdings following the SOP vote,
given the magnitude of the fund’s holding, as measured by “fund's portfolio weight (in fraction)”
and “fraction of company's shares held by fund.”
The only significant result for these latter variables in Regressions 2-5 is obtained in
Regression 2–this regression documents that when the fund voted against SOP, and the vote failed,
the larger the “fraction of company shares held by fund”, the more likely the fund is to sell their
53
shares. This indicates that in the special case in which a fund opposes SOP of a given company and
the vote fails, funds decrease their exposure to the company stock by selling stocks. Hence, in this
case, the SOP vote and the “wall street walk” are correlated. In unreported specifications we repeat
this analysis, but replace the dependent variable to “change of the value of stock held.” The results
are qualitatively similar.
54
Table I: Changes in fund’s ownership following the SOP vote
This table reports OLS regressions on the fund-company-year level for the 2011-2013 period. The dependent variable is
the “percentage change of the number of shares held” in end of the quarter in which the SOP vote was held, compared
to the number of shares held in the end of quarter preceding the SOP vote (i.e., (number of shares t+1 - number of
shares t)/ number of shares t). All regressions include year, and Fama-French 48 industries fixed effects. Errors are
clustered on the fund level. P-values are reported in parenthesis. * indicates p<.05, ** p<.01, and *** p<.001.
Fund voted against proposed
compensation
Fund's portfolio weight (in
fraction)
(1)
1.0342
(.926)
Percentage change of the number of shares held
(2)
(3)
(4)
(5)
-719.1434*
(.093)
-10.0008
(.788)
-557.4128
(.467)
-1.0459
(.816)
-798.6629
(.102)
52.8169
(.887)
-0.3702
(.550)
17.3555
(.534)
8.2603
(.734)
0.4225
(.961)
0.0001
(.572)
15.351
(.543)
-17.9241
(.378)
-149.4298**
(.037)
0.018
(.725)
-5.3642
(.226)
-4.0059
(.300)
0.4607
(.700)
0
(.754)
1.6163
(.684)
-0.7874
(.810)
610.7375
(.624)
-0.4141
(.621)
13.5407
(.751)
0.7216
(.988)
8.2286
(.573)
0
(.832)
-82.0065*
(.052)
18.7957
(.505)
5.9201
(.551)
0.0027
(.609)
-0.2978
(.461)
0.4561
(.254)
-0.0913
(.436)
0
(.950)
-0.0123
(.977)
-0.2371
(.542)
45.7847
(.908)
-0.456
(.541)
18.0413
(.568)
9.9105
(.711)
-0.0944
(.992)
0.0001
(.509)
24.1365
(.391)
-22.6662
(.325)
Number of institutions voting
on proposal
0.1348
(.393)
0.0154
(.348)
0.2974
(.321)
-0.0029
(.118)
0.1328
(.458)
CEO tenure (years)
0.2377
(.479)
-0.0144
(.723)
-0.2643
(.683)
0.0047
(.192)
0.2713
(.469)
CEO age (years)
-0.0095
(.984)
0.064
(.255)
-0.704
(.403)
0.0005
(.927)
0.0252
(.963)
-142.1507
(.134)
-3.1309
(.727)
-203.6319
(.295)
-4.8686***
(.000)
-152.4563
(.156)
-0.0201
(.970)
-0.0057
(.926)
0.0584
(.965)
-0.0052
(.547)
-0.0227
(.969)
Expense ratio (weighted
average of shareclasses)
1559.1729**
(.039)
7.9936
(.923)
-436.3863
(.770)
-12.6949
(.150)
1778.6622**
(.034)
Turnover ratio (weighted
average)
-4.0416
(.332)
-0.4418
(.261)
-0.7912
(.917)
-0.023
(.620)
-4.5207
(.336)
0.0002
(.633)
All
observations
All
observations
Yes
0
218,515
0
(.452)
0
(.986)
0
(.987)
0.0002
(.637)
Against
Against
For
For
Fail
Pass
Fail
Pass
Yes
-0.001
7,300
Yes
-0.001
12,209
Yes
0.012
3,777
Yes
0
194,943
Fraction of company's shares
held by fund
Total compensation of CEO t1 (in Million $)
Fraction of shares held by
executives
ROA of company t-1
Abnormal return
Market capitalization in
$Millions
Fraction of shares held by
blockholders
Fraction of shares held by
institutions
Fund twelve-months
characteristic selectivity return
Annual inflow (fraction of
total assets)
Total net assets managed by
fund (in thousand $)
How did the fund vote?
Vote outcome
Year and industry fixed effects
R-squared
N
55
Appendix C: Size of holding and votes cast on non-SOP proposals
This appendix replicates for the 2011-2013 period, the analysis conducted throughout this paper for
additional types of issues that are commonly bought up for vote at shareholders meetings. In Table
II, Panel A reports the vote results on the aggregate level, Panel B reports votes cast on the mutual
funds level, and Panel C reports votes cast on the institutional level.
In each of Panels A, B, and C of Table II, Regressions 1-3 documents issues for which
voting “against” is generally perceived as indicating that shareholders are monitoring the company
particularly actively, whereas in Regressions 6-9 the opposite is the case, i.e., voting “for” the
proposal is generally perceived as indicating increased monitoring). More specifically, in each of
Panels A, B, and C of Table II the following issues are examined: electing directors (Regression 1),
approving an omnibus stock plan (Regression 2), increasing authorized common stock (regression
3), requiring an independent board chairman (regression 4), declassifying the board of directors
(Regression 5), reducing the supermajority vote requirement (regression 6), providing the right to act
by written consent (Regression 7), requiring a majority vote for the election of directors (Regression
8), and restoring or providing for cumulative voting (Regression 9).22
As the results reported document, there does not seem to be a consistent voting pattern
within any of the voting levels, and also across the three voting levels examined. In other words,
particularly small shareholders are not consistently more likely to support proposals that are
generally perceived as those that reflect shareholder’s monitoring particularly actively/ enhanced
corporate governance practices. For example, Panel A, Regression 1, examines the percentage of
votes cast on the aggregate level, in support of directors that were up for nomination. In this
regression, “Fraction of shares held by blockholders” does not enter the regression significantly,
In all panels, a coefficient with a value “0” with no corresponding p-value reported, indicates the variable was dropped
due to collinearity.
22
56
indicating that the presence of blockholders is not significantly related to the extent of support
directors receive.
In Panel B, Regression 1, which is a regression on the fund voting level, the dependent
variable equals one if the fund voted for a director, and zero otherwise. In this regression, “fund’s
portfolio weight (in fraction)” has a negative significant coefficient, indicating that for companies
that consist a smaller portion of a fund’s portfolio, a fund is more likely to vote in support of the
directors. In contrast, “fraction of company's shares held by fund” has a significant positive
coefficient, indicating that the larger the fraction of outstanding shares held by a fund, the more
likely the fund is to vote in support of the directors of that company. Finally, on the institutional
level, “institution’s portfolio weight (in fraction)” is positive, and significant only at the 10% level,
indicating that a larger portfolio weight of a company, is associated with an increased likelihood that
the institution’s funds vote for the directors, however “fraction of company's shares held by
institution” is far from significant.
Hence, as this example illustrates, during the SOP era (2011-2013 in our study) we do not
find that for other types of issues bought up for vote at shareholder’s meetings the magnitude of the
holding is consistently related to the vote cast, both within a given voting level, and also across the
three voting levels we examine. This further suggests that during the SOP period we examine,
particularly the SOP vote is a convenient mechanism for small institutional shareholders/ mutual
funds to voice a critical feedback.
57
Table II Panel A: Votes cast on the aggregate level
This table reports OLS regressions on the company-year level for the 2011-2013 period. The dependent variable equals the fractions of votes cast “for” the proposal.
Each column examines a different type of proposal, as specified in the table. All regressions include fixed year and fixed Fama-French 48 industry effect. Errors are
clustered on the company level. P-values are reported in parenthesis. * indicates p<.05, ** p<.01, and *** p<.001
Fraction of aggregate votes cast "For"
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
Fraction of shares held
by blockholders
0.0088
(.190)
-0.0116
(.793)
0.0547
(.482)
0.1376
(.303)
-0.0598***
(.000)
-0.1285***
(.000)
0.3325***
(.000)
-0.2154
(.371)
0.4147
(.409)
Fraction of shares held
by executives
0.0714***
(.000)
0.1010*
(.099)
0.1359**
(.046)
0.1396
(.397)
0.0519***
(.000)
-0.0232**
(.020)
-0.8151***
(.000)
-1.1472
(.172)
0.3409
(.172)
Total compensation of
CEO t-1 (in Million $)
-0.0020***
(.000)
-0.0028
(.130)
-0.002
(.227)
-0.0066***
(.000)
-0.0033***
(.000)
-0.0053***
(.000)
-0.0053***
(.000)
0.0009
(.159)
-0.0026
(.764)
ROA of company t-1
0.0252***
(.004)
0.0976**
(.026)
0.0253
(.396)
-0.0301
(.826)
0.0398***
(.000)
0.0968***
(.000)
-0.0203
(.190)
0.3204
(.312)
1.0065
(.100)
Firm abnormal return
0.0562***
(.000)
0.0176
(.235)
0.016
(.173)
0.081
(.153)
0.0176***
(.000)
0.0990***
(.000)
0.1421***
(.000)
0.0827
(.418)
0.3067
(.271)
ISS recommended to
vote for
0.0279***
(.000)
0.0438
(.144)
-0.0118
(.642)
-0.1128***
(.001)
0
-0.1029***
(.000)
-0.0707***
(.000)
0
0
Management
recommended to vote
for
0.0752
(.386)
0
0
0
-0.0723***
(.000)
0.0388*
(.091)
0.0017
(.955)
0.3399
(.105)
0
Elect
Director
Approve
Omnibus
Stock Plan
Increase
Authorized
Common
Stock
Require
Independent
Board
Chairman
Declassify
the Board
of
Directors
Reduce
Supermajority
Vote
Requirement
Provide
Right to
Act by
Written
Consent
Require a
Majority
Vote for
the
Election of
Directors
Restore or
Provide for
Cumulative
Voting
Vote that indicates
more monitoring
R-squared
Against
0.085
Against
0.053
Against
-0.013
For
0.246
For
0.088
For
0.431
For
0.176
For
0.029
For
-0.093
N
24,178
325
112
111
45,091
25,626
29,675
54
27
Issue discussed
58
Table II Panel B: Votes cast on the fund level
This table reports OLS regressions on the fund-company-year level for the 2011-2013 period. The dependent variable equals one if the fund voted “for” the proposal,
and zero otherwise. Each column examines a different type of proposal, as specified in the table. The regressions include, but do not report, all control variables
included in Table 5. The regressions include year, Fama-French 48 industry, and fund fixed effects. Errors are clustered on the fund level. P-values are reported in
parenthesis. * indicates p<.05, ** p<.01, and *** p<.001.
Fund voted "For"
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
Fraction of shares held by
blockholders
-0.0133***
(.000)
0.0418*
(.061)
-0.0617*
(.093)
-0.0541
(.428)
-0.0001
(.995)
-0.0551**
(.038)
-0.1415**
(.018)
0.4910***
(.000)
1.1317***
(.000)
Fraction of shares held by
executives
-0.0363***
(.000)
-0.006
(.865)
-0.0771*
(.093)
0.7890***
(.000)
-0.0463**
(.030)
-0.3859***
(.000)
0.5654
(.453)
-0.2276
(.241)
0.8544***
(.000)
Fund's portfolio weight (in
fraction)
-0.0843*
(.074)
-0.0542
(.896)
-0.1584
(.853)
-0.6041**
(.027)
-0.2169
(.476)
-1.0385**
(.022)
-0.3501
(.374)
-1.8518***
(.001)
-0.3173
(.307)
Fraction of company's
shares held by fund
0.2050***
(.001)
0.9583***
(.000)
-0.2472
(.847)
-4.2847***
(.004)
0.0416
(.927)
0.1943
(.823)
0.3346
(.899)
3.1418
(.139)
-0.0317
(.976)
Number of funds voting on
proposal
0.0027***
(.000)
0.0130***
(.000)
0.0065
(.446)
-0.0094***
(.001)
0.0015
(.250)
-0.0248***
(.000)
0.0176***
(.009)
-0.0424***
(.000)
0.0674***
(.000)
ROA of company t-1
0.0100***
(.000)
-0.0701***
(.003)
-0.0067
(.765)
0.0417
(.675)
0.0174
(.263)
0.0434
(.108)
-0.1241*
(.082)
0.5890***
(.000)
1.1542***
(.001)
Management recommended
to vote for
0.2960***
(.000)
0
0
0
0.0703***
(.000)
0.1909***
(.000)
0.4033***
(.000)
0.0793
(.290)
0
Elect
Director
Approve
Omnibus
Stock Plan
Increase
Authorized
Common
Stock
Require
Independent
Board
Chairman
Declassify
the Board
of
Directors
Reduce
Supermajority
Vote
Requirement
Provide
Right to
Act by
Written
Consent
Require a
Majority
Vote for
the
Election of
Directors
Restore or
Provide for
Cumulative
Voting
Against
0.29
Against
0.454
Against
0.46
For
0.63
For
0.389
For
0.383
For
0.664
For
0.676
For
0.832
1,858,263
23,232
7,584
13,792
21,105
13,604
10,928
5,624
6,343
Issue discussed
Vote that indicates more
monitoring
R-squared
N
59
Table II Panel C: Votes cast on the institution level
This table reports OLS regressions on the institution-company-year level for the 2011-2013 period. The dependent variable equals the fraction of the institution's funds
that voted “for” the proposal. Each column examines a different type of proposal, as specified in the table. The regressions include, but do not report, all control
variables included in Table 6. The regressions include year, Fama-French 48 industry, and institution fixed effects. Errors are clustered on the institution level. P-values
are reported in parenthesis. * indicates p<.05, ** p<.01, and *** p<.001.
Fraction of institution's funds that voted “For”
(4)
(5)
(6)
(1)
(2)
(3)
(7)
(8)
(9)
Fraction of shares held by
blockholders
-0.0529***
(.002)
-0.3244***
(.000)
0.0095
(.957)
0.2824
(.171)
0.0249
(.706)
0.3740***
(.007)
0.3784**
(.029)
-0.1816
(.654)
0.829
(.482)
Fraction of shares held by
executives
-0.0845***
(.000)
0.1423*
(.074)
-0.0054
(.975)
0.26
(.463)
-0.9155***
(.000)
-0.3286
(.304)
-0.8898
(.662)
1.9637***
(.000)
0.8787
(.121)
Institution’s portfolio
weight (in fraction)
0.5554*
(.079)
0.8117*
(.099)
1.0508*
(.060)
0.3254
(.706)
0.3373*
(.053)
1.0670**
(.031)
3.0504***
(.005)
2.3776*
(.094)
1.9483
(.567)
Fraction of company's
shares held by institution
0.0761
(.905)
-3.8112
(.135)
-4.6479
(.599)
-0.1731
(.914)
2.9413
(.153)
1.996
(.444)
-0.1139
(.952)
7.3737***
(.006)
-0.4694
(.885)
Total assets managed by
institution in trillions
0.0883
(.459)
0.7608
(.121)
-0.0068
(.993)
0.2133
(.776)
0.1124
(.608)
-0.7412*
(.081)
-0.0905
(.889)
-0.5204
(.534)
-0.6676
(.607)
Number of funds voting
on proposal (in fractions)
0.0194***
(.000)
0.0118*
(.084)
0.0023
(.918)
0.0758***
(.000)
0.0214**
(.021)
0.0456**
(.033)
0.0718***
(.000)
0.1467***
(.000)
0.0067
(.940)
Management
recommended to vote for
SOP
-0.1273**
(.024)
0
0
0
-0.0691***
(.000)
0.0573***
(.008)
0.037
(.328)
-0.2814**
(.039)
0
ISS recommended to vote
for SOP
0.0351***
(.000)
0.0629***
(.002)
-0.0415
(.576)
-0.1818***
(.000)
0
-0.2647
(.275)
-0.1236
(.391)
0
0
Elect
Director
Approve
Omnibus
Stock Plan
Increase
Authorized
Common
Stock
Require
Independent
Board
Chairman
Declassify
the Board
of
Directors
Reduce
Supermajority
Vote
Requirement
Provide
Right to
Act by
Written
Consent
Require a
Majority
Vote for
the
Election of
Directors
Restore or
Provide for
Cumulative
Voting
Vote that indicates more
monitoring
R-squared
Against
0.28
Against
0.275
Against
0.225
For
0.335
For
0.317
For
0.35
For
0.416
For
0.334
For
0.405
N
268,569
3,380
1,161
1,709
8,916
4,346
2,705
828
772
Issue discussed
60
61