Center for Financial Markets and Policy
The Role of Institutional Investors in Voting: Evidence from the Securities Lending Market
Reena Aggarwal
McDonough School of Business, Georgetown University
aggarwal@georgetown.edu
Pedro A. C. Saffi
Judge School of Business, University of Cambridge
psaffi@jbs.cam.ac.uk
Jason Sturgess
Driehaus College of Business, DePaul University
jds224@georgetown.edu
http://finpolicy.georgetown.edu
The Role of Institutional Investors in Voting: Evidence from the Securities Lending Market
Reena Aggarwal
McDonough School of Business, Georgetown University
aggarwal@georgetown.edu
Pedro A. C. Saffi
Judge School of Business, University of Cambridge
p.saffi@jbs.cam.ac.uk
Jason Sturgess
Driehaus College of Business, DePaul University
jds224@georgetown.edu
Abstract
Using the unique setting of the securities lending market, we find that institutional investors restrict lending
supply and/or call back loaned shares prior to the record date in order to exercise their voting right. Loan
demand and the price of borrowing also increase around the proxy voting record date. We estimate the value of
voting rights by institutional investors in a simultaneous equation framework and show that lenders of shares
value their shares more than borrowers. Institutions place a greater value on voting rights for firms with weak
corporate governance, poor performance, and higher institutional ownership. The value of the vote is also
higher when contentious proposals such as non-routine and those related to compensation, anti-takeover, and
corporate control are on the ballot. Examining the subsequent vote outcome, we find higher recall to be
associated with less support for management and more support for shareholder proposals. Our results indicate
that institutional investors value their vote and use the proxy process as an important channel for affecting
corporate governance.
JEL: G32; G34; G38
Keywords: Proxy Voting, Securities Lending, Institutional Investors
Corresponding author: Reena Aggarwal, McDonough School of Business, Georgetown University, Washington, D.C.
20057. Tel. (202) 687-3784, aggarwal@georgetown.edu. An earlier version of the paper was titled, “Does Proxy Voting
Affect the Supply and/or Demand for Securities Lending?” We thank Alon Brav, Susan Christoffersen, Richard Evans,
Mireia Gine, Slava Fos, Stuart Gillan, Steve Kaplan, Lee Pinkowitz, Gregor Matvos, David Musto, Adam Reed, David
Ross, Laura Starks and David Yermack; seminar participants at the Federal Reserve Board, U.S. Securities and Exchange
Commission, 3rd Annual RMA - UNC Academic Forum for Securities Lending Research, American Finance Association
2013, European Finance Association 2011, Western Finance Association 2011, FMA Asia 2011, Drexel Conference on
Corporate Governance 2011, DePaul University, Georgetown University, IESE, Università Cattolica del Sacro Cuore,
Comisión Nacional del Mercado de Valores, London School of Economics, Temple University, University of Cambridge,
Queen Mary, University of Maryland, University of Texas at Austin, and Imperial College for helpful comments.
Conversations with several industry participants, particularly, Les Nelson of Goldman Sachs and Judith Polzer of J.P.
Morgan helped us to understand the workings of the securities lending market. Doria Xu and Jiayang Yu provided
excellent research assistance. We gratefully acknowledge a grant from the Q Group. Saffi acknowledges support from the
Spanish Ministry of Science and Innovation under ECO2008-05155 at the Public-Private Sector Research Center at IESE.
Aggarwal acknowledges support from the Robert E. McDonough endowment at Georgetown University’s McDonough
School of Business.
The Role of Institutional Investors in Voting: Evidence from the Securities Lending Market
I.
Introduction
Understanding the preferences of institutional investors regarding governance is
important for firms trying to attract new investors as well as policy makers considering the
regulation of different governance mechanisms. However, the mechanisms used by institutional
investors to impact corporate governance tend to be private and difficult to study. We use the
unique setting of the securities lending market to study the conditions that prompt institutional
investors to engage in influencing firm-level governance and the extent to which investors use
the proxy process to exercise their opinions.
Most large institutions have a securities lending program and consider it to be an
important source of revenue, with estimates of $800 million in annual revenue for pension funds
alone (Grene (2010)). However, investors cannot vote shares that are on loan on the voting
record date. Hence, institutional investors must decide whether to restrict lending and even recall
shares already on loan prior to an upcoming vote.1
We use a comprehensive proprietary data set comprised of shares available to lend,
shares that have actually been borrowed and are on loan, and the associated borrowing fee for the
period 2007-2009. We find a marked reduction in the lending supply prior to the proxy record
date, as institutions restrict and/or recall their loaned shares so that they can exercise their voting
rights. We also find a statistically significant increase in borrowing demand and fees around the
record date. Lending supply returns to normal levels immediately after the record date. Our
results indicate that institutions consider their vote to be valuable, and they make the effort to
determine when it is important to recall shares in order to exercise voting rights.
1
We use the terms recall and restrict interchangeably, capturing both recall of shares actually on loan, and
restriction on shares available to lend that that have not been borrowed.
We also examine the value of the vote to understand the importance that institutions place
on voting rights. In the existing literature, some controversy exists regarding how changes in the
supply and demand for borrowing shares around the record-date affect the borrowing fee. For
example, Christoffersen, Géczy, Musto, and Reed (2007) find a marginally significant increase
in borrowing fee around the record date but determine that the increase is not economically
significant. They conclude that fees are unresponsive to increases in borrowing demand on the
record date and hence that the value of the vote is negligible. However, Kalay, Karakas, and Pant
(2013) use the options market to determine the value of a vote, and place a significant value on
voting rights.
We find a statistically significant but small increase in fee around the record date.
However, we argue that borrowing fee in isolation is not sufficient to measure the value of the
vote for at least three reasons. First, Kolasinski, Reed, and Thorncock (2013) show that in the
equity lending market the loan supply schedule is essentially flat and Prado, Saffi and Sturgess
(2013) document that on average the market clears with high levels of slack lending supply.
Therefore, borrowing fee might be insensitive to both demand and supply shocks most of the
time. If so, then shares might trade for a price significantly below the value of voting rights.
Second, as with any market one needs to consider endogeneity in quantity and prices. Third, an
estimate of the value of the vote should reflect the monetary amount that an investor is willing to
exchange for the voting right. Therefore, the most appropriate measure should be the borrowing
fee paid by/to an investor who would no longer demand the voting right if the fee were
marginally higher. However, borrowing fee is measured at the security-level as the average fee
across all loans for a firm. These concerns all imply that examining changes in the borrowing fee
4
alone may underestimate the value of the vote, and even result in concluding that the value of the
vote is zero.
We address these concerns by estimating the value of a vote using a reduced-form
instrumental-variables methodology to model the equity lending supply and demand schedules
and further estimate lending supply and demand shifts on the record date. To compute the value
of the vote we first estimate the change in supply and demand and the change in the price
elasticity of supply and demand observed on the record date. Next, we combine these estimates
to compute the implied value of the vote. The implied value of the vote is defined as the change
in fee that would make the observed change in supply (and demand) on the record date to be zero.
It is important to estimate the change in the price elasticity because lenders and borrowers are
likely more sensitive to fees on the record date given the impact of voting rights on the decision
to lend and borrow. As an illustration, consider a mutual fund that places a value of 100 basis
points on voting rights for a particular firm. So long as the borrowing fee is below 100 basis
points, say 50 basis points, the mutual fund will choose to restrict lending and exercise the vote
because the value of the vote exceeds the lending income. However, as the borrowing fee
increases above 100 basis points the fund will no longer restrict lending because lending income
is greater than the value of the vote. We estimate the implied value of the vote cross-sectionally
by comparing the record-date shift in quantity and in the price elasticity for both the supply and
demand side and then estimate the change in fee that would leave supply and demand
unchanged.
We find that lenders place a higher value on their vote than borrowers. For lenders, we
estimate the value of a vote to be 226 bps, or 2.26% of market capitalization, in annualized
terms, while for borrowers it equals 122 bps. The heterogeneity in the value of the vote across
5
lenders and borrowers might arise because, unlike borrowers, lenders continue to hold economic
interest for the longer term and also might implement a policy to recall irrespective of borrowing
fee, which would inflate the value of the vote. Next, we examine heterogeneity in investors’
behavior and the value of the vote based on underlying firm characteristics and types of proposal
on the ballot. Firms with poor performance, lower institutional ownership, weaker governance
and smaller firms exhibit higher estimated values of a vote. Further, these values are higher for
record dates associated with meetings with important proposals on the ballot related to nonroutine items, compensation, anti-takeover, and corporate control.
We also investigate the relation between recall in lending supply at the record date and
the subsequent votes cast on the meeting date. In general, we find that when the proxy advisory
firm ISS recommends voting against management there are more votes against the proposal. In
addition, we find a higher recall to be associated with more FOR votes for shareholder proposals,
and fewer FOR votes for management when ISS recommends voting against management. An
increase in borrowing demand is also associated with fewer FOR votes if ISS opposes
management’s position. Further, we show that the recall in lending supply results in less support
for management proposals with a higher estimated value of the vote, such as compensation and
corporate control-related proposals.
In addition, we examine voting by mutual funds to rule out that the relation between
recall and voting outcome is driven by the alternative explanation that institutions recall shares to
vote with management. Mutual funds provide an opportunity to better observe how voting
behavior is influenced by the recall of lending shares by examining voting behavior only for
those investors who provide lending supply. We find that mutual funds are significantly less
likely to vote in favor of contentious proposals where recall in lending supply is greater and ISS
6
recommends voting against. This result alleviates the concern that mutual funds are recalling
shares to vote with management when other shareholders are following ISS’s advice and voting
against management.
In extensions to the main findings, we examine the period of the financial crisis and also
check for robustness of our results around dividend record dates. During the financial crisis of
2008, the general pattern of reduced supply and increased borrowing fees around the proxy
voting date continued to hold. In contrast to the activity around voting record dates, we find that
around the time of the ex-dividend record date, there is a statistically and economically
significant increase in borrowing demand, with little change in the lending supply.
The issues we examine are particularly relevant for a period that has seen increased
emphasis on both shareholder activism and proxy voting. Voting provides an important
mechanism for shareholders to affect firm-level corporate governance and policies. Since equity
lending transfers voting rights, it has important ramifications for corporate governance. The
increased interest in proxy voting and securities lending has resulted in fund boards now paying
attention not only to the fee received from a securities lending program but also to whether the
securities are being loaned to “responsible” borrowers. Funds are screening companies'
upcoming shareholder meetings where a vote may be important. According to a survey of
institutional investors by ISS, 37.9% of the respondents stated that a formal policy on securities
lending is part of their proxy voting policy.2
Prior research has attempted to examine the preferences of institutional investors based
on inferences of corporate governance attributes deemed important to institutional investors.
Gillan and Starks (2007) survey the evolution of institutional shareholder activism in the U.S.
2
See http://www.riskmetrics.com/press/articles/040307boardiq.html
7
from the value effect of shareholder proposals to the influence on corporate events. 3 Other
studies find that institutional investors affect CEO turnover (Parrino, Sias, and Starks (2003) and
Helwege, Intintoli, and Zhang (2012)), anti-takeover amendments (Brickley, Lease, and Smith
(1988)), executive compensation (Hartzell and Starks (2003)), and mergers (Gaspar, Massa, and
Matos (2005) and Chen, Harford, and Li (2007)). In an analysis of 23 countries, Aggarwal, Erel,
Ferreira, and Matos (2011) find that changes in institutional ownership over time are positively
associated with subsequent changes in firm-level governance, but the opposite is not true. Cuñat,
Gine and Guadalupe (2012) show that passing a governance provision is associated with an
increase in shareholder value, and more so for institution sponsored proposals. Chung and Zhang
(2011) find that the fraction of a firm’s shares held by institutions increases with the quality of
governance. Bushee, Carter, and Gerakos (2010) find evidence that ownership by governancesensitive institutions in the U.S. is associated with future improvements in shareholder rights.
However, Matvos and Ostrovosky (2010) study director elections and find heterogeneity in
institutional investors voting preferences. Overall, institutional investors’ preferences related to
governance tend to be private and are often conducted behind the scenes and hence are difficult to
study. Therefore, there is limited empirical work examining the channels used by institutional
investors to affect governance.
In a survey of institutional investors, McCahery, Sautner, and Starks (2011) find that
corporate governance is important to institutional investors, and many institutions are willing to
engage in shareholder activism. Recent papers such as Brav, Jiang, Partnoy, and Thomas (2008);
Clifford (2008); and Klein and Zur (2009) study activism by individual funds, such as pension
funds or hedge funds. Fos (2011) shows that proxy contests play a role in disciplining managers.
3
See for example, Gillan and Starks (2000, 2007), Hartzell and Starks (2003), Gaspar, Massa, and Matos (2005),
Chen, Harford, and Li (2007), and Bushee, Carter, and Gerakos (2010).
8
Gantchev (2013) finds that that the average activist campaign is estimated to cost $10.5 million,
and half of the costs come from proxy fights. Less than 5% of all campaigns reach a proxy fight;
proxy fights having a 67% success rate. Cai, Garner, and Walkling (2009) find shareholder votes
to be related to firm performance, governance, and director performance; however they conclude
that the differences are economically trivial.
Christoffersen, Géczy, Musto, and Reed (2007) use 1998-1999 data from a large lending
agent to examine borrowing demand and fees aspects of the securities lending market around a
proxy vote. They find a marginally significant increase in borrowing fee around the proxy record
date. The authors conclude that the price of a vote is zero because investors are not selling their
votes but letting them go and speculate that this is due to information asymmetry. Kalay,
Karakas, and Pant (2013) use the options market to determine the value of a vote, finding that it
is higher around shareholder meetings and to the conclusion that votes have value. While we find
the change in average fee alone leads to a value of the vote of around 2 basis points per annum
(approximately 3 times as large as found by Christoffersen, Géczy, Musto, and Reed), our new
approach estimates the value of the vote to be around 122-226 basis points per annum. These
estimates are similar in magnitude to those found by Kalay, Karakas, and Pant (2013), who find
the value of voting rights for the average firm to be 16 bps of the stock price with an average
option maturity of 38 days. By examining both lending supply and borrowing demand in a
simultaneous framework, our analysis provides a more complete picture of pricing than these
previous papers and more accurate estimates of the value of a vote.
Our paper also contributes to the literature on equity lending. Studies such as Jones and
Lamont (2002); D’Avolio (2002); Geczy, Musto, and Reed (2002); Ofek and Richardson (2002);
Cohen, Diether, and Malloy (2007); and Edwards and Hanley (2010) examine the cost of
9
borrowing. Saffi and Sigurdsson (2011) describe international equity lending markets and how
lending supply and borrowing fees are related to market efficiency and the distribution of stock
returns. Evans, Ferreira and Prado (2013) find mutual funds that lend shares are not able to act
on the short-selling signal in stocks with high borrowing demand, resulting in future
underperformance. Kaplan, Moskowitz, and Sensoy (2013) conduct an experiment in which they
introduce an exogenous supply shock to the loan supply of a single money manager. They find
no adverse impact on stock prices. Asquith, Au, Covert, and Pathak (2013) describe borrowing in
the bond market by analyzing data from one large lender for the period 2004-2007.
The paper proceeds as follows. Section 2 provides background on the proxy voting
process and the securities lending market. Section 3 describes the data on proxy voting, securities
lending, and other firm-level corporate attributes. In Section 4, we show the changes in lending
supply around proxy voting record date. Section 5 shows the relation between changes in
securities lending activity and proposal type. Section 6 presents results of our empirical findings
on voting outcomes and the role of lending supply. Section 7 provides additional analysis around
dividend record dates, and during the financial crisis. Section 8 concludes.
2.
Background on Proxy Voting and Securities Lending
2.1 Proxy Voting
In the United States, state laws control the holding of annual meetings to elect directors
and matters of corporate governance, as discussed by Karmel (2010). However, federal securities
laws control the solicitation of proxies. In light of changes in shareholder demographics, the
structure of share holdings, technology, and the potential economic significance of each proxy
vote, the SEC reviewed the proxy infrastructure and issued a “proxy plumbing” concept release
in July 2010. The concept release identified several issues that might require a regulatory
10
response, including proxy voting and securities lending; “empty voting,” under which economic
ownership is decoupled from voting rights; over-voting and under-voting, both of which can
result from a mismatch between the number of shares held compared to the number of shares
credited to a broker-dealer; and the need for investors to know proxy items before the record date
so that they can decide whether to lend their shares or not.4 The SEC also raised the issues of
whether funds should report the number of shares cast and how the funds voted.
There are many rules and regulations that apply to the proxy process. To give
shareholders sufficient time to make an informed voting decision, registrants must follow a
timeline. SEC proxy Rule 14a-13 requires that a “Broker Search” be distributed to banks,
brokers, and nominees who then compile a list of beneficial owners. This broker search must
take place 20 business days prior to the record date for an annual meeting and ten days for a
special meeting. Most states (for example, California and Delaware) require that the record date
be set at a maximum of 60 days and a minimum of ten days prior to the meeting; New York sets
the maximum at 50 days. The record date determines the ownership date for voting purposes. As
long as shares are not lent out on the voting record date, the owner can vote them. Preliminary
proxy material must be filed with the SEC via EDGAR ten days before distributing definitive
copies to shareholders. Proxy material must be mailed out 40 days before the meeting date.
Mutual funds typically have an oversight process, with board involvement, to monitor the
funds’ proxy voting process. The SEC’s Rule 206(4)-6 requires funds to adopt and implement
proxy voting policies and procedures and to make voting record available to clients. According
to the SEC, “This disclosure enables fund shareholders to monitor their funds’ involvement in
4
Empty voting refers to the situation in which an investor has voting rights but no economic interest. See Hu and
Black (2006, 2007) for a discussion of how investors might use the securities lending market or derivatives for
empty voting.
11
the governance activities of portfolio companies.” In 2003, the SEC started requiring mutual
funds to disclose proxy voting records by filing Form N-PX.
2.2
Securities Lending
Securities lending is generally defined as a transaction in which the beneficial owner of the
securities, normally a large institutional investor such as a pension fund or mutual fund, agrees to
lend its securities to a borrower, such as a hedge fund, in exchange for collateral consisting of
cash and/or other securities. 5 Although lenders refer to these shares as being “on loan”, the
lender actually transfers ownership and voting rights. Shares may be borrowed for a variety of
reasons, including short selling, covering a short position, or for trading strategies such as
convertible bond arbitrage, dividend tax-arbitrage strategies (see Christoffersen, Géczy, Musto,
and Reed (2005) and Thornock (2013)), and merger arbitrage, and possibly for empty voting.
The lender earns a spread by investing the collateral in low-risk short-term securities. In a typical
U.S. loan, the collateral is 102% on domestic securities and 105% for international securities.
Risk in collateral can arise if the counterparty defaults and/or daily mark-to-market does not
occur, resulting in the value of collateral securities to drop below the value of lent securities.
Many of these problems related to collateral and counterparty risk were highlighted in a study
conducted by the U.S. Government Accountability Office in 2011.6
The securities lending market has grown tremendously in the last decade. By 2007, the total
value of securities on loan was estimated at $5 trillion (Lambert 2009), with associated annual
borrowing fees of $8-10 billion. 7 Most large institutional investors have a securities lending
program and consider securities lending as a key source of revenue. Institutional investors
5
The securities lending process is shown graphically in Appendix 1, and an example of cash flows and fees on a
securities loan with cash collateral is provided in Appendix 2.
6
“401 (K) Plans: Issues Involving Securities Lending in Plan Investments,” U.S. General Accountability Office,
2011.
7
http://www.forbes.com/2007/09/25/retail-investors-securities-biz-cx_lm_0925brokerage.html
12
suffered large losses in 2008 that led to lawsuits against big custodial banks. The allegation was
that the custodians did not invest the collateral in safe, plain-vanilla securities, resulting in losses
for their clients.
As is evident from the SEC’s concept release of July 2010, there are questions about
whether securities lending has contributed to proxy abuse. The concern is that market
participants, such as activist investors, can obtain voting rights to exert influence or gain control
of a company and do so without corresponding economic ownership in the company (see Hu and
Black, 2006 and 2007). Most securities lending involves shares borrowed from pension funds,
mutual funds, and other large institutional investors. Institutions have started to include policies
on securities lending in their proxy guidelines, but they vary considerably in scope and detail.
Some funds require a total recall of shares, while others weigh the lost revenue against the
benefits of voting on a case-by-case basis. Below, we provide some examples from funds’ proxy
voting guidelines.
Putnam Funds
“The funds’ have requested that their securities lending agent recall each domestic issuer’s
voting securities that are on loan, in advance of the record date for the issuer’s shareholder
meetings, so that the funds may vote at the meetings.”8
TIAA-CREF
“Even after we lend the securities of a portfolio company, we continue to monitor whether
income from lending fees is of greater value than the voting rights that have passed to the
borrower. Using the factors set forth in our policy, we conduct an analysis of the relative value of
lending fees versus voting rights in any given situation. We will recall shares when we believe
the exercise of voting rights may be necessary to maximize the long-term value of our
investments despite the loss of lending fee revenue.”9
State Board of Administration of Florida (SBA)
8
9
See https://content.putnam.com/shared/pdf/proxy_voting_guidelines.pdf
See http://www.tiaa-cref.org/ucm/groups/content/@ap_ucm_p_tcp/documents/document/tiaa01007871.pdf
13
“Circumstances that lead the SBA to recall shares include, but are not limited to, occasions when
there are significant voting items on the ballot such as mergers or proxy contests or instances
when the SBA has actively pursued coordinated efforts to reform the company’s governance
practices, such as submission of shareholder proposals or conducting a detailed engagement. In
each case, the direct monetary impact of recalled shares will be considered and weighed against
the discernable benefits of recalling shares to exercise voting rights. The SBA recognizes that it
may not be possible to determine, prior to a record date, whether or not shares warrant recall.”10
Fund groups such as Vanguard and Fidelity do not have specific discussion of policies on
recalling shares in their public proxy guidelines. California Public Employees’ Retirement
System (CalPERS) has a two-step list. About 30 securities on the “Focus” list are completely
restricted from lending because CalPERS takes an active interest in these securities and always
wants the shares available to vote. For the second list of 300 securities, which represents the
largest market value of CalPERS position, CalPERS wants to ensure that the securities are
returned prior to a proxy vote.11 The SEC requires funds to recall shares for “material” events but
has not defined materiality. In a survey by ISS, 92.3% of the respondents indicated that mergers
and acquisitions were the most important reason to recall shares.12
One of the challenges to recalling shares is that shareholders typically do not receive the
proxy material until after the record date. However, in order to vote, institutions must recall the
shares by the record date. Hedge funds have argued that they do not borrow shares simply for
voting purposes because they do not even know about the items on the proxy ballot as of the
record date. Listed companies on the New York Stock Exchange are required to provide the
NYSE a notice of record and shareholder meeting dates at least ten days prior to the record date.
The SEC is considering whether this information should be disseminated to the general public.
3.
3.1
Data
Securities Lending Descriptive Statistics
10
See http://www.sbafla.com/fsb/LinkClick.aspx?fileticket=mt0icmFCYMk%3d&tabid=378
See http://www.securitiestechnologymonitor.com/issues/19_31/21468-1.html?zkPrintable=true
12
See http://www.riskmetrics.com/press/articles/040307boardiq.html
11
14
For the most part, understanding the securities lending market has been limited partly
because of the lack of transparency in this fragmented market. We obtain a proprietary equity
lending data set from Data Explorers for the period January 2007 to December 2009. Data
Explorers collects this information daily from 125 large custodians and 32 prime brokers in the
securities lending industry and provides comprehensive coverage of equity lending activity
available to market participants that includes lending supply, shares actually borrowed, and the
corresponding fees at the security level. Our data covers more than 85% of the securities lending
market. There are 4,333 firms in the equity lending sample, however the proxy voting data limits
the analysis to Russell 3000 firms. As of December 2009, there was $1.55 trillion available to
lend, out of which $113 billion was actually lent out and would be considered as being on loan.
Saffi and Sigurdsson (2011) provide a detailed description of the data.
The main dependent variables in our study are equity lending supply, borrowing demand,
utilization rate, and annualized borrowing fees. We define these variables as follows: lending
supply (SUPPLY) is the dollar value of supply relative to a firm’s market capitalization; loan
quantity (ONLOAN) is the dollar value of shares on loan on a given day relative to market
capitalization; utilization rate (UTILIZATION) is ONLOAN divided by SUPPLY; and borrowing
fee (FEE) is the difference between the risk-free interest rate and the rebate rate expressed in
basis points (bps) per annum.13 The rebate rate is the portion of the interest rate on the collateral
that is returned to the borrower. The lender needs to reinvest the collateral at a rate higher than
the rebate in order to earn a positive return. Expected returns can be increased by investing
collateral in securities with more credit risk or a longer maturity relative to the loan, however this
can result in a loss if market rates rise. We use the effective Federal Funds rate as our proxy for
13
For cash-based transactions the loan fee is directly negotiated between lenders and borrowers and reported by
Data Explorers.
15
the risk-free rate. Firms that have a fee greater than 100 basis points (1%) are commonly
considered to be SPECIAL. Such firms are more closely watched by investors and are more
expensive to borrow.
In Panel A of Table 1, we present descriptive statistics for the equity lending market for
7,415 record dates from 2007 to 2009 based on the -30 to +30 days window around the recorddate event. On average, 23.78% of a firm’s market capitalization is available for lending, with
4.06% being on loan and resulting in a utilization rate of 17.78%. The minimum and maximum
values of SUPPLY (winsorized at the 1% level) are 1.65% and 48.57%, respectively. ONLOAN
varies from a high of 20.49% to a low of 0.01%. Some firms are heavily borrowed while others
are not borrowed at all. UTILIZATION is as high as 70% in our sample.
The mean annualized fee is 48.3 bps. Therefore, the daily cost of borrowing $1 billion worth
of shares on the record date is quite low. However, this cost can quickly rise for firms in high
demand reaching a maximum of 1,114 bps in our winsorized sample. The minimum fee of -50.84
bps implies that the lender pays the borrower. In fixed-contract lending, it is possible for the fee
to be negative because the rebate is set in advance. If the rebate is larger than the interest earned
on the collateral, e.g. when interest rates quickly decrease, then the fee will be negative. During
the sample period, 0.09% of the firms had a fee greater than 100 basis points and were
considered to be “on special”. The mean and median number of days for which firms are on loan
is 16 days and one day, respectively. Most loans are “open ended” and rolled over every day
without a specific maturity date. Panel B presents changes in lending activity on the record date
relative to the average during the period -30 to -20 days before the record date. On average,
SUPPLY drops by 1.93% of market capitalization. ONLOAN and FEE increase by 0.06% of
market capitalization and 2.40 bps, respectively.
16
Panel C of Table 1 shows that the lending supply of securities as a percentage of market
capitalization (SUPPLY) is relatively stable over the 2007-2009 period. However, average
demand for borrowing shares (ONLOAN) experiences a severe drop post financial crisis,
decreasing from 4.60% in 2007 to 3.22% in 2009.14 The mean FEE varies from a high of 58.50
bps in 2008 to a low of 43.05 bps in 2009. As a result, the average annualized borrowing fee
(FEE) is lowest in 2009 at 15.75 bps.
3.2
Other Firm-Level Data
We use CRSP to obtain share price (PRICE), market capitalization (SIZE), turnover
(TURNOVER), and bid-ask spread (SPREAD). We use only common shares with price over $1,
and further merge the data to Compustat and collect data on book equity (EQUITY) to calculate
the book-to-market equity ratio (BM). We exclude closed-end funds, American Depositary
Receipts (ADRs) and real estate investment trusts (REITs). We obtain ownership data from the
Thomson Reuters CDA/Spectrum database on SEC 13F filings. The 13F filings must be reported
on a quarterly basis by all investment companies and professional money managers with assets
over $100 million under management. For each firm, we calculate total institutional ownership
as a percentage of market capitalization (INST) and institutional ownership concentration (INST
CONC), measured as the Hirschman-Herfindahl index normalized between zero and one. We use
firm-level corporate governance index GOV41 as in Aggarwal, Erel, Ferreira, and Matos (2011).
GOV41 assigns a value of one to each of the 41 governance attributes if the company meets
minimally acceptable governance guidelines on that attribute and zero otherwise.15
14
During the financial crisis, many restrictions were placed on short selling. These restrictions affected several
arbitrage strategies used by hedge funds, hence the drop in demand for borrowing shares. In the United States, a ban
on short sales was imposed on financial firms during the period September 19 to October 8, 2008. Australia, Japan
and a few other countries banned short sales in all firms. Some U.S. mutual funds temporarily halted their securities
lending programs and did not lend out shares to short-sellers.
http://www.boston.com/business/markets/articles/2008/09/23/2_mutual_fund_firms_act_to_halt_short_sales/
15
Aggarwal, Erel, Stulz, and Williamson (2009) describe the data in more detail.
17
3.3
Proxy Voting Descriptive Statistics
Proxy voting analysis examines 56,220 proposals for 7,415 record-dates obtained from
ISS. The proxy voting data cover the Russell 3000 firms and includes proposal-level
characteristics such as proposal description, sponsor, management’s recommendation, ISS’s
recommendation, threshold for the proposal to pass, votes cast, and voting result.
We present proxy voting characteristics in Panel A of Table 2. On average, 86.62% of
votes are cast on proxy proposals, with 91.86% of those votes being in favor and only 7.54%
against. This overwhelming majority in favor of proposals is reflected in the 70.16% vote margin
by which they pass. We create different categories of proposals, with the explicit aim of
exploring those that might be considered as contentious, based on disagreement between
different parties, and those that are associated with significant events. First, we classify proposals
as routine and non-routine. NYSE Rule 452 outlines non-routine proxy proposals as those in
which broker voting is not allowed. Examples include proposals relating to anti-takeover
provisions, stock capitalization and mergers. Second, we examine proposals relating specifically
to anti-takeover provisions (G-INDEX) included in the G-Index developed by Gompers, Ishi, and
Metrick (2003), compensation proposals (COMP), and those that relate to mergers/proxy
contests (CORP CONTROL).
In Panel B of Table 2, we describe the voting outcome of non-routine proposals, which
comprise 12.25% of the total sample. These proposals have almost three times more votes cast
against the proposal than the total sample. Almost 60% of non-routine proposals are related to
compensation. Shareholder-sponsored proposals are a much smaller subset (only 25.56%) and
usually fail to pass, receiving an average of 40% of FOR votes, although when ISS is in favor of
the proposal the average proportion of FOR votes increase to 46.17%. Examples of shareholder-
18
sponsored proposals include Say on Pay; requests that the firm provide cumulative voting;
reduce supermajority voting; require independent chairman of board; require a majority vote for
the election of directors; and declassify the board of directors. We also provide descriptive
statistics on non-routine proposals that are likely to attract most attention from investors.
Proposals relating to compensation, anti-takeover, and corporate control receive far more
negative votes than the average for all proposals discussed earlier.
4.
Securities Lending and Record Date
4.1
Lending Supply, Borrowing Demand and Fees around Proxy Voting Record Date
Figure 1 plots lending supply, borrowing, utilization, and borrowing fees for the period
starting 30 days before the record date and ending 30 days after the record date. We define the
record date (day 0) as the event date. For our 7,415 voting record dates, the average time
between the record date and the shareholder meeting is 53 days. The supply of shares available
to lend as a fraction of market capitalization starts to decrease about 20 days before a vote and is
at its lowest point on day 0, the record date. SUPPLY starts at 24.09% on day -30 and reduces to
22.16% by the record date. This drop in supply is consistent with institutions restricting or
recalling their shares at the time of a vote. On the first day after the record date, SUPPLY returns
to pre-event levels in line with institutions not wanting to lose revenue from lending.
The results suggest that institutions start restricting supply in advance of the proxy record
date to ensure that shares can be recalled and that they can exercise the vote. In practice,
institutions are generally advised to allow two weeks for a recall prior to a proxy vote, and
possibly longer if the firm is “special”. Consistent with industry practice, we find that the drop in
lending supply starts to occur about two weeks before the record.16 Institutions might also recall
16
We thank securities lending practitioners at J.P. Morgan and Goldman Sachs for helping us understand industry
practices for recalling and restricting lendable shares.
19
shares in advance to provide sufficient notice to borrowers, thus alleviating possible problems for
borrowers to find shares and improving an institution’s reputation as a stable and reliable
lender.17 Before the availability of electronic firm-loan monitoring systems, recalls frequently
failed. The Securities Industry and Financial Markets Association estimated that in 2002, 25% of
recalls failed.18
Examining the plot for borrowing demand (ONLOAN) shows a small increase around the
record date. On day -30, on average, 4.10% of a firm’s market capitalization is on loan, and by
the record date it grows to 4.13%, increasing by only 0.03% of a firm’s market capitalization.
Finally, UTILIZATION and FEE both increase in the 20 days prior to the record date. The
finding adds insight to Blocher, Reed and Van Wesep (2013), who argue that shifts in supply
matters only for firms on special by revealing that supply shifts become important even at
relatively low levels of utilization.
4.2
Determinants of Lending Supply, Borrowing Demand and Fees
To begin our analysis, we investigate the determinants of the equity lending market
around the record-date by estimating separate pooled regressions using daily lending supply,
borrowing, and borrowing fee as the dependent variables. For each of the 7,415 record dates, we
consider an event window of -30 days to +30 days, where t=0 is the proxy voting record date.
We include a record date dummy (RDATE) to examine whether there is abnormal equity lending
market activity on the record date compared to the 30 days before and after the record date.
We follow Prado, Saffi and Sturgess (2013) by including the following variables to
explain securities lending. To control for ownership, we use INST, institutional ownership at the
17
Hu and Black (2008) discuss the case of Fidelity and Morgan Stanley, who together held 10% shares of Telecom
Italia and led a campaign against a takeover of Pirelli. However, they were only able to vote 1% of the shares
because the remaining shares were lent out and could not be called in in time for the vote. The Pirelli bid was
approved.
18
Securities Technology Monitor, November 13, 2007.
20
end of the previous quarter measured as a percentage of market capitalization, and INST CONC,
concentration of institutional holdings using the Hirschman-Herfindahl index. We use lagged
values of log of market capitalization (SIZE), book-to-market ratio (BM), turnover (TURNOVER),
and spread (SPREAD) as explanatory variables to control for firm characteristics. We include a
dummy for firms with a share price below five dollars (PRICE<$5). Similar to Kolasinski, Reed
and Ringgenberg (2013), we also include short-term momentum (Short-Term Mom) measured as
the cumulative return over the five previous days and long-term momentum (Long-Term Mom)
as the cumulative return over the previous 252 trading days. In all estimations, we include year
fixed effects, and present results with and without firm fixed-effects. Throughout, we cluster
standard errors by firm to ensure robustness.19
Table 3 reports the results for the determinants of lending supply, borrowing demand and
fees. In columns 1 and 2, the dependent variable is lending supply, expressed as percentage of
market capitalization. The estimation in column 1 includes year fixed effects but not firm fixed
effects; column 2 includes both year and firm fixed effects. In column 1 (2), the explanatory
variable RDATE has a coefficient of -1.64 (-1.623), which is significant at the 1% level. In terms
of economic significance, the coefficient indicates that on average, lending supply is lower on
the record date by 1.64% of market capitalization, or approximately 7% of the mean over the [30,+30] sample window. Examining the within-firm results in Column 2, lending supply is
higher when institutional ownership (INST) is higher, when institutional ownership is not
concentrated (INST CONC), for larger firms (SIZE), and value firms (BM).20 The coefficient of
long-term momentum is positive and of short-term momentum is negative, both significant,
19
The results are robust to clustering the standard errors both by firm-record date to ensure robustness to
heteroskedasticity and serial correlation within a given proxy window and by firm and time to ensure robustness to
heteroskedasticity as well as serial and cross-sectional correlation.
20
However, the coefficient on size is negative when we exclude firm FE. This is because of the cross-sectional
correlation of other firm attributes, particularly INST.
21
indicating that investors are willing to lend more shares in firms with higher returns during the
previous year but less so in firms with higher returns in the previous five days. In addition to
standard control variables, we include firm-level corporate governance, GOV41. The positive
and statistically significant coefficient of 4.60 on GOV41 in column 1 indicates that firms with
better governance have a higher lending supply. Unsurprisingly we find a positive and
insignificant result when we include firm fixed-effects as governance changes rarely within firms.
This result is consistent with the argument that better governance alleviates shareholders’
concerns that share lending will be detrimental to the value of their holdings.
The determinants of borrowing demand appear in columns 3-4 of Table 3. The positive
coefficient on RDATE indicates that demand is statistically higher on the record date. In the
model shown in column 3 (4), the coefficient of RDATE is 0.082 (0.085), which amounts to an
increase of 2% compared to the mean over the [-30, +30] sample window. Examining the results
with firm fixed-effects in column 4, borrowing demand is higher if institutional ownership is
higher, and for firms that are more liquid, and demand is lower for firms priced below $5. There
is a negative and significant association between previous performance as proxied by both short
and long-term momentum and borrowing demand on the record date. Investors are likely to
borrow more shares in companies that are not performing well. Again, we include the corporate
governance index GOV41 in this analysis. We note that the coefficient on GOV41 is negative and
significant. Although better corporate governance alleviates shareholders’ concerns when
lending, it appears to deter those investors who borrow. This result is consistent with the
hypothesis that better governance deters stock borrowing and subsequent short selling because,
all else equal, it is associated with fewer opportunities for investors to profit on the downside.
22
Columns 5-6 of Table 3 report the results of similar tests using FEE as the dependent
variable. In both models, the coefficient of RDATE is positive and significant at 1%, implying
that the fee for borrowing stock increases on the record date. This corresponds to a 3.76%
increase relative to mean over the [-30, +30] sample window. However, while the increase in fee
on the record-date is statistically significant, the coefficient of 1.814 (1.572) in column 1 (2)
implies that the value of the vote is a negligible 1.814 bps (1.572 bps) per annum. As described
in the introduction, examining fee in isolation might bias downwards estimates of the value of
the vote. We address this issue in the next section.
5.
Endogeneity and Value of the Vote around Record Dates
Our paper provides four methodological contributions to the estimation of the value of
the vote. First, we apply an instrumental-variables approach to identify supply (demand) curves
through exogenous shocks to demand (supply). Using the change in fee at the record date
without jointly modeling supply and demand lead to biases in the estimation of the value of the
vote. The OLS results presented in Section 4 do not account for the cost of borrowing shares
and/or simultaneous shifts in prices and quantities in the equity lending market and are therefore
potentially limited in helping us understand the preferences of institutional investors. The
estimations assume that lending supply is fixed when inferring changes in borrowing demand,
and that borrowing demand is fixed when inferring changes in lending supply around the record
date. However, significant changes take place simultaneously on both the supply side and on the
demand side, as lenders restrict the quantity of lendable shares and borrowers increase the
demand for loans, potentially motivated by increasing their voting power at the shareholders’
meeting. Therefore, a relevant concern is that the increase in borrowing demand found around
record dates may be biased downwards due to the impact of less supply resulting in an increase
23
in fee, which in turn makes it more expensive to borrow shares. At the same time, higher
borrowing demand can increase fees, and this might result in lending supply decreasing by less
than it would otherwise had fees remained constant.21 The instrumental-variables (IV) approach
uses exogenous instruments to identify supply and demand curves and address this issue.
Second, we also identify differences in how lenders and borrowers value their vote, an
expected feature of this market given the different incentives of those that supply shares to those
that borrow shares. Third, we allow for non-linearities in the supply and demand curves. A
particular interesting one is the change that takes place on the record date. For example, if the
slope of the demand curve changes at the record date we must take this effect into account when
identifying the supply curve to prevent misspecification. The endogeneity of fee requires extra
care when estimating parameters, being addressed by Angrist, Graddy and Imbens (2000) and
applied in the finance literature by Kolasinski, Reed and Ringgenberg (2013). It is reasonable to
expect that lenders and borrowers will be more sensitive to fees on the record date given the
impact of voting rights on the decision to lend and borrow, causing the price-elasticity to be
different at the record date.
Our final contribution is to provide a new measure of the value of a vote that compares
the supply (and demand) curves identified on the record date to those identified for "normal"
periods. We jointly model the dynamics of ONLOAN and SUPPLY around the record-date using
a two-stage regression approach to model quantity (ONLOAN or SUPPLY) and price (FEE). Our
measure of the value of a vote is given by the change in fee that would make the observed
change in supply (and demand) on the record date to be zero. In Section 5.1 below we outline our
methodology for estimating the value of the vote. In Section 5.2 we discuss potential instruments
21
Note that this effect of fee on SUPPLY and ONLOAN biases our OLS results against finding a record date effect.
24
for lending supply and loan demand. In Section 5.3 we present the IV regressions and results on
the value of the vote.
We motivate the simultaneous estimation of price and quantity by examining changes in
equity lending around the record date conditioning on borrowing fee. We split the sample of
firms into those that are easy to borrow versus those that are expensive. We define a firm to be
“On Special” if it has a borrowing fee greater than 100 bps at t=-30. We also show statistics for
companies that are extremely expensive to borrow and have fees above 1,000 bps at t=-30. Table
4 shows that of the 7,415 record dates only 79 are associated to firms with borrowing fees above
1,000 bps. Panel A of Table 4 reports averages of equity lending variables at t-30. The average
lending supply as percentage of market capitalization is 14.52% for firms On Special relative to
25.02% for firms that are not. Borrowing demand is also higher for the On Special group. The
lower supply and higher demand results in a much higher annualized fee of 429 bps for the On
Special group, compared with a fee of 9.30 bps for the other group. If investors incorporate the
cost of borrowing into the decision to lend or borrow then we might expect very different record
date behavior across these two groups.
Panel B of Table 4 reports the change in each lending attribute from its average during
the (-30,-20) days before the record date. The lending supply of the On Special group changes by
less when compared with the non-special firms both in absolute terms and percentage terms. This
implies that lenders recall/restrict more shares when fee is low partly because the potential loss
of lending revenue is low. Borrowing demand increases for the non-special firms, consistent with
the results in Table 3. However, borrowing demand actually decreases on the record date for the
On Special group, potentially due to the higher fee. This suggests that, on average, the mean fee
of 429 bps exceeds the value borrowers place on voting rights and these potential borrowers
25
prefer not to borrow the stock on the record date. The borrowing fee increases by 8.37 bps for the
On Special group and 1.09 bps for the other group.22 These descriptive statistics illustrate that
not only does borrowing fee play a role in the decision to lend/borrow but also that the change in
fee around the record date is not a sufficient proxy for the value of the vote. For example,
examining the change in fee for the On Special firms would lead to the conclusion that the value
of the vote is negative for borrowers. Even for the non-special firms, the change in fee is small
simply because of the slack in supply that is typical in the market for equity lending. Combined,
these descriptive statistics and the slack in supply provide an important explanation for the low
value of the vote based on average fee (e.g., as reported by Christoffersen, Géczy, Musto, and
Reed (2007)). Additionally, the fact that fee changes around the record date in response to
changes in supply and demand means that one should not only incorporate borrowing fee in an
analysis of equity lending but also the endogenous relationship between quantity and prices. In
the next section we provide a new methodology to estimate the equity lending behavior around
the proxy record date and shed light on the value of the vote.
5.1
Methodology and Estimation of the Value of a Vote
Using a reduced-form instrumental-variables approach, we control for the simultaneity of
supply and demand and, more importantly, use estimated parameters to provide a better measure
of the value of a vote. We infer the price sensitivity of ONLOAN and SUPPLY to fees using
instruments that identify demand and supply shocks. This sensitivity varies according to firm
characteristics and types of corporate events on the ballot, supporting the idea that investors
assign different values to their vote depending on whether they are lenders or borrowers, and for
record dates that include “important” proposals.
22
The daily cost for a $1mil loan is equal to (9.3/(252 *100)=) $369.04 for firms not ON SPECIAL and $17,011 for
those that are ON SPECIAL, almost 50 times larger.
26
We employ the (IV) estimator developed by Angrist, Graddy and Imbens (2000).23 The
estimation allows for time-varying supply and demand functions, using FEE as the endogenous
variable. We build on the approach in Angrist, Graddy and Imbens (2000) by identifying the
average price elasticity in the market for equity lending both in general and around the record
date.
The first-stage equation for FEE is given by:
The second-stage equations are given by:
̂
̂
̂
̂
{
The SUPPLY and ONLOAN quantities are linked by the endogenous price FEE. FEE is
jointly determined by the interaction between supply and demand, being the endogenous variable
in our system, and can change due to demand and supply shocks that allow us to identify each
curve as long as we have suitable instruments. In order to identify the parameters associated with
the endogenous variable FEE in the second stage, we need instruments (INSTRUD and INSTRUS)
that are exogenous to each dependent variable. For example, in the ONLOAN equation we need
variables that affect supply but not demand to obtain the estimated ̂ and
̂
to
address the endogeneity issue. Note that we also must have a first-stage equation for
RDATE*FEE, since the product of an endogenous variable (FEE) and an exogenous one
(RDATE) is still endogenous (see Wooldridge (2001)). If a restriction in supply results in a
higher borrowing fee, and higher prices result in lower demand, then it is relatively
straightforward to show that standard OLS estimates that ignore endogeneity will result in
23
Angrist, Graddy and Imbens (2000) estimate the demand for fish by identifying the demand elasticity using
weather patterns as exogenous shocks to supply. Kolasinski, Reed and Ringgenberg (2013) apply a similar
methodology to estimate the loan supply schedule and how it varies with proxies for search frictions.
27
downward biased estimates for
. Similarly, ignoring endogeneity leads to a downward bias in
the restriction in supply at the record date.
Our measure of the value of a vote, VVOTE, attempts to answer the following question:
how much would the fee have to change such that the estimated record date impact on supply
and demand would be zero? This is equivalent to setting the first derivative of quantity with
respect to the record date to zero and solving for the fee. Therefore, we solve the following
equation:
, for i = S, D. Thus, for supply we estimate how much the borrowing
fee would have to increase at the record date such that the lender prefers to lend rather than
restrict or recall their shares available to borrow. On the demand side, we estimate how much
price would have to increase such that the borrower would not demand more shares on the record
date.
The following equations summarize our measure of the value of a vote as a function of
estimated parameters:
This approach allows us to identify the value of the vote and also differences in the
preferences towards voting between borrowers and lenders and is in the spirit of self-selection by
market participants suggested by Roy (1951). Some investors care about voting while others do
not, both across and within lenders and borrowers. The value of the vote we estimate tells us of
the value of voting rights for investors that do care about the value of the vote. Alternatively,
there is a subset of lenders (and a universe of potential borrowers) who do not place a value on
voting rights and choose not to restrict lending around the record date. This approach also
complements the findings of Matvos and Ostrovosky (2010), who show that there is important
28
heterogeneity in institutional investor voting preferences. We exploit this heterogeneity to
estimate the value of the vote and also show how heterogeneity in preferences varies across
proposal type.
We present point estimates and statistical significance for the value of the vote. As
is determined by a non-linear combination of parameters, we test for its statistical significance
using the delta method to compute standard errors. We also estimate the value of votes for firms
split by firm characteristics, such as institutional ownership, size and corporate governance
quality, and also for meetings with specific types of proposal on the ballot, such as, non-routine
proposals, proxy contests, compensation and corporate control related proposals.
It is important to note that we are not inferring the value of the vote from the average
change in fee directly but rather we use the price elasticity associated with record date lending to
estimate the value. Thus, using this methodology we might find that the value of the vote is
greater than the increase in the fee on the record date. This would imply that voting rights trade
at a price below value, which is to be expected given observed slack in the supply curve.24 Note
that SUPPLY is not equal to equilibrium supply, which by definition equals equilibrium demand,
allowing us to estimate different values for voting between lenders and borrowers.
5.2
Choice of Instruments and Falsification Tests
In order to identify valid instruments, we perform falsification tests to identify suitable
candidates. A valid instrument should have statistical significance to explain one dependent
variable (e.g. SUPPLY) but not the other (e.g. ONLOAN). More importantly, these potential
variables must have a sound reasoning. On the supply side, Prado, Saffi and Sturgess (2013)
24
Slackness in supply on the record date even for stocks where the fee is small implies that there are some lenders
that place a value of the vote close to zero. For this subset of lenders the change in record date lending would be
zero and price elasticity would be undefined. Our methodology estimates the value of the vote by examining the
price elasticity for lenders that have a non-zero value of the vote.
29
show that institutional ownership concentration, INST CONC, is an important determinant of
SUPPLY even after controlling for total ownership. More concentrated holdings result in larger
shareholders having greater power to affect the supply of shares available. If short sale
constraints lead to overpricing, shareholders can try to limit supply to support prices of their own
shares. Thus, more concentrated owners may prefer to not lend stock and therefore retain control
of voting rights, which would otherwise pass to the borrower.25
On the demand side, we use a measure of hedging demand, Hedging Demand, proposed
by Hwang, Liu and Xu (2013) who argue that short selling can help correct under-pricing of
firms by facilitating the hedging of industry risk. Hedging Demand is defined as the equalweighted cumulative return in the past 252 days of related firms (excluding the firm’s own
returns) with the same four-digit GICS industry classification code. If other firms in the same
industry become under-valued, arbitrageurs would purchase the under-valued firms and short
substitute securities. Thus, we expect the demand for shorting stock i to be high when the
demand for going long shares of competitors’ j, as measured by low cumulative returns in the
previous year, is high. The results in Appendix 3 show our falsification tests using firm-fixed
effect regressions of SUPPLY and ONLOAN. Based on these results we use INST CONC as our
instrument for supply-related shocks and Hedging Demand as the instrument for demand-related
shocks.26
5.3
IV Regressions and the Value of a Vote
25
We employ firm fixed effects throughout to ensure robustness to unobserved firm heterogeneity. In Table 3
ownership concentration is negatively associated with ONLOAN, but is insignificant once we include firm fixed
effects.
26
Unlike Kolasinski, Reed and Ringgenberg (2013), we find that short term momentum has predictive power for
both variables, being an unsuitable instrument in our sample.
30
Because FEE is an endogenous variable in our system, the FEE*RDATE cross-product is
also endogenous and we instrument it with RDATE*Hedging Demand and RDATE*INST CONC,
respectively for supply and demand, implementing the IV approach as suggested in Wooldridge
(2001). Table 5 displays results for the first stage estimates of FEE as a function of our
instruments and firm controls and also include time and firm fixed effects with standard errors
clustered at the firm and year. As expected, we find that firms with higher INST CONC exhibit
higher fees because there is less supply available to borrow. Hedging Demand has a coefficient
equal to -37.62, implying that a decrease in the 252-day past returns of competing firms leads to
an increase in fees. Similar to our previous OLS results, the fee is 1.576 bps higher on the record
date.
Table 6 displays our main second stage results using the instrumented fee estimated in the
first stage to control for the endogeneity of the fee. All equations include an additional set of firm
characteristics as control variables, year and firm fixed-effects, and with standard errors clustered
by firm and year. Columns 1 and 2 show estimates for SUPPLY with and without the
RDATE*FEE coefficient. In column 1, the record date effect equals -1.613%, close to the effect
presented in Table 3. Further, we find that in general supply is insensitive to fee within firm
consistent with a flat supply curve. In column 2, we include the variable RDATE*FEE. The
coefficient on RDATE*FEE is positive and statistically significant, implying that the recall of
shares at the record date is sensitive to borrowing fee and that recall is lower if the fee received
by lenders is higher. Lenders weigh their value of the vote and the potential lending income
before restricting lending. The Kleibergen-Paap statistic tests if the instruments are sufficiently
correlated with the included endogenous regressors. We can safely reject the null that
endogenous variables are under-identified and obtain similar conclusions using the Cragg-
31
Donald Wald statistic. Columns 3 and 4 report the same results using ONLOAN as the dependent
variable. The positive and statistically significant RDATE coefficients indicate an increase in
borrowing demand on the record date. Borrowing demand is lower for firms with higher
borrowing fee as one might expect given the endogeneity between price and quantity. Further,
this price elasticity is greater on the record date. The statistically significant coefficient for
RDATE*FEE implies that for very expensive firms demand actually decreases on the record date,
in line with the descriptive statistics shown for ON SPECIAL firms in Table 4.
In terms of the value of a vote, the estimate for
is equal to 226.9 bps in
annualized terms (0.90 bps per day) with an estimated standard error equal to 85 bps, being
significant at the 1% value.
is equal to 122.1 bps in annualized terms (0.48 bps
per day) and has a standard error equal to 26 bps, also significant at the 1% level. We find that
lenders assign almost twice as much value to votes than borrowers on the record date. The fact
that investors recall their shares quite a long time before the record date affects the economic
interpretation of the value of the vote. From Figure 1, we observe that SUPPLY begins to fall
around 20 days before the record date, which combined with estimated
imply that
lenders are willing to give up ((20/252)*226.9/100=) 0.180% of the value of shares available to
borrow to vote. For the demand side, the estimated value of voting to borrowers is 0.097%. We
now investigate whether these estimates are higher for particular types of firms or proposals.
5.4
Firm and Proposal Characteristics and Value of Vote
Our methodology allows us to investigate whether the sensitivity of supply and demand
variables to FEE at the record date vary depending on firm characteristics and specific types of
proposals included on the ballot. In Table 7, we examine differences in the value of a vote based
on four firm characteristics: corporate governance, institutional ownership, stock returns during
32
the past 12-months, and size of the firm, measured by market capitalization. For each of these
characteristics, we expect that the value of the vote may vary. For example, the value of voting
rights is likely to be greater for firms with weak governance compared to firms with strong
governance, where shareholders hold more power. Similarly we might expect that firms with
large institutional holdings exhibit greater monitoring and thus voting rights become less
valuable. Even within a firm, there may be time-series variation in voting rights. For example, in
periods of low returns investors may place a higher value on implementing change through
voting. Finally, the value of the vote might vary with size, perhaps because of correlation
between size and other factors such as governance and institutional holdings, but also because
the vote may hold more influence in smaller firms where ownership is less dispersed.
For both SUPPLY and ONLOAN, we report estimated coefficients for RDATE and
RDATE*FEE after applying the simultaneous-equation approach to alternative sub-samples. We
split the sample around the firm-year median firm characteristic.27 The coefficient of RDATE in
all cases is negative and statistically significant for SUPPLY, and positive and significant for
ONLOAN estimations. These results are consistent with our previous findings presented in Table
6. The positive coefficient of RDATE*FEE on the supply side shows that if fee is higher then
lenders are less likely to recall. Similarly, the negative coefficient of RDATE*FEE on the
demand side implies that less borrowing takes place if fee is higher. We also estimate the value
of a vote in each case and find the value of a vote to be statistically different from zero in all
cases for both SUPPLY and ONLOAN.
Panel A splits firms into low and high corporate governance based on GOV41. Lenders
and borrowers of shares value a vote quite differently for the two groups. Institutional investors
prefer not to lend out shares on the record date for firms with weak governance. The annualized
27
The split is not always equal as there is some clustering around the median value.
33
value of a vote for lenders is almost twice as large, at 314 bps for firms with weak governance
relative to 117 bps for firms with strong governance, and the difference is statistically significant.
On the borrowing side, the value of the vote is much lower at 118 bps for firms with weak
governance and 126 bps for firms with strong governance with the difference not be statistically
significant.
The sample is split by institutional ownership in Panel B of Table 7. The value of a vote
is significantly higher for both lenders and borrowers in firms with low institutional ownership
relative to high institutional ownership. Panel C splits the sample by low and high monthly
returns in the preceding twelve months. The value of a vote for lenders in firms with low and
high stock returns is 244 and 174 bps, respectively. The difference in the value of a vote for the
two groups is statistically significant. Share lenders are particularly interested in exercising their
vote in firms that are not performing well. Our results suggest that both lenders and borrowers
value their vote more in firms that are performing poorly. They can use the vote to bring about
change at the firm. In Panel D, the sample is split based upon market capitalization. On the
supply side, the value of the vote is significantly higher for firms with low market capitalization
at 308 bps compared to high market capitalization firms at 49 bps. In contrast, borrowers place a
fairly comparable value on the vote for both low and high market capitalization firms at 97 and
80 bps, respectively, and the difference is not statistically significant.
We also examine changes in lending supply around record dates associated with proxy
events that are likely deemed to be more “important” to shareholders. In Table 8, we report
results for the value of a vote in four alternative subsamples based on the presence of at least one
of the following types of proposals: non-routine, compensation-related, anti-takeover, and
corporate control (proxy contests and mergers). We split the sample based on whether the record
34
date is associated with a proxy event or not. However, we omit firms from both subsamples that
do not have at least one proxy event in question. In all cases, for both the supply and demand
side, the value of a vote is positive and statistically significant.
In Panel A of Table 8, we show results for record dates with and without non-routine
proposals. Non-routine proxy proposals are outlined by NYSE Rule 452 as those in which broker
voting is not allowed and include proposals relating to anti-takeover provisions, stock
capitalization and mergers. On the supply side, we find higher value of votes when at least one
non-routine proposal is present in the ballot but find no statistical difference for ONLOAN.
Panel B splits the sample according to the presence of at least one compensation-related
proposal and yields similar results. With the increased prominence of corporate governance
concerns, managerial compensation policies have become a focus of investors’ attention and we
expect it to be reflected into a higher value of the vote. While we do not find a statistically
significant difference for compensation versus non-compensation proposals, we do observe that
the value of the vote is equal to 227 bps for record dates with compensation-related proposals
compared to 176 bps for those without on the supply side.
The third group of contentious proposals we consider are based on anti-takeover
provisions (G-INDEX) included in the G-Index developed by Gompers, Ishi, and Metrick (2003).
In Panel C, the sample is differentiated by G-INDEX and Non-G-INDEX related proposals. The
value of the vote differs greatly for lenders with G-INDEX related proposals at 304 bps and nonG-INDEX related proposals at 231, and the difference is statistically significant. Similarly, we
find that the value of the vote from the borrowers’ perspective is greater for record dates
involving a G-INDEX proposal.
35
Finally, in Panel D of Table 8, we consider proposals related to corporate control by
examining proxy contests and mergers. In a proxy contest, shareholders vote to resolve a conflict
between the firm’s management and board of directors, referred to as “incumbents”, and a group
of shareholders, referred to as “dissidents”. Some examples of high profile proxy contests
include Carl Icahn’s efforts to unseat Yahoo’s board in 2008, and Hewlett Packard – Compaq
merger in 2001. Dissident shareholders can initiate the proxy contest by filing a preliminary
proxy statement PREC14A and definitive proxy statement in connection with contested
solicitations DEFC14A. Data on proxy contests is hand-collected and supplemented from data
from Sharkrepellent.net, an organization that covers proxy fights and activism. For mergers we
identify proposals for targets and acquirers. On the supply side, we find that the value of
corporate control related proposals is significantly higher than non-corporate control related
proposals at 381 and 90 bps, respectively. Not surprisingly, we find the value of a vote on the
lending side is highest for corporate control proposals. For borrowers, the difference in the value
of a vote for corporate control and non-corporate control related proposals is not statistically
significant. Overall, we find that lenders place a higher value on the vote than borrowers.
In summary, we show that it is important to control for endogeneity when modeling the
equity lending behavior around the record date. Both lenders and borrowers internalize the cost
of borrowing when deciding to lend or borrow. We exploit this price elasticity to estimate the
value of the vote and show that the value of the voting rights cannot be estimated simply by the
change in fee around the record date. Further, the value of the vote varies by investor and firm
heterogeneity.
6.
Voting Outcome
36
In this section we study whether the recall of supply by institutional investors or an
increase in borrowing demand have any impact on the vote outcome at the shareholder meeting.
We estimate regressions for the 6,887 non-routine proposals where the dependent variable is
FOR, the percentage of votes in favor of a proposal.28 For each proposal we test if the restriction
in lending supply and the increase in demand around the record date plays a role on how votes
are cast on the subsequent meeting date. Importantly, the meeting date is on average 53 days
after the record date. If institutions recall lending supply to exercise their vote, then we should
expect that voting outcome is associated with recalled supply.
The independent equity lending variables are the change in lending supply, ΔSUPPLY,
and the change in borrowing demand, ΔONLOAN. These changes are based on the average
lending supply and on loan during days (t=-30 to -20) to the record date (t=0). We include
indicator variables for management proposals that management supports and the proxy advisory
service ISS opposes (DISS), for shareholder-sponsored proposals (DSHR), and for proposals
relating to compensation (COMP), G-INDEX (G-INDEX), and corporate control (CORP
CONTROL) that we examined in Section 5. We also interact the change in supply and the change
in on loan with these characteristics to better understand when equity lending activity is
important to determine support for a proposal. Further, we include the firm-specific
characteristics and proposal fixed effects included in the earlier estimations, but omit these for
brevity in Table 9. All regressions include firm fixed effects, time dummies, and standard errors
are double clustered at the firm and year levels.
In column 1 of Table 9, we present evidence that shows the record date change in lending
supply is positively associated with more votes against the proposal. The coefficient of
28
FOR is defined as the percentage of number of FOR votes, relative to the base by which the proposal is decided.
The base depends by proposal, but may be the sum of FOR, AGAINST, and ABSTAIN votes, the sum of FOR and
AGAINST votes, or the number of shares outstanding, for example.
37
ΔSUPPLY of 0.350 implies that a recall in lending supply (i.e. a decrease in ΔSUPPLY) is
negatively associated with support for non-routine proposals, however significance is only at the
10% level. We find no relation between borrowing demand and a larger proportion of FOR votes
for proposals. Further, shareholder sponsored proposals exhibit 44% less FOR votes, on average,
than management sponsored proposals for the same firm. However, the significant coefficient of
-2.444 on ΔSUPPLY * DSHR shows that the record date recall in supply is positively associated
with more votes being cast in favor of shareholder-sponsored proposals.
Next, in column 2, we introduce proposal characteristics and advice from proxy advisory
services. Consistent with Alexander, Chen, Seppi, and Spatt (2010), we find that the
recommendations of proxy advisors play an important role in the outcome of proposals. 29 In
general proposals that ISS recommends against management are associated with significantly
lower support. Further, the significant coefficient of 1.545 on ΔSUPPLY * DISS implies that a
higher recall (negative ΔSUPPLY) at the record date leads to fewer votes being cast in favor of a
proposal if ISS opposes management.30 This is consistent with institutional investors responsibly
fulfilling a monitoring role whereby they provide prudence on behalf of shareholders. Switching
focus to proposal characteristics we find greater support for compensation, G-Index, and
corporate control proposals. However, where the recall of lending supply is higher the support
for these proposals is lower: the coefficients estimated by interacting ΔSUPPLY with these three
proposal types are all positive and significant. In Section 5 we showed that institutional investors
placed a higher value on voting rights for these types of proposals; here we show that when
institutions do recall shares they tend to vote against the proposal. This finding is consistent with
29
An alternate explanation, proposed by Choi, Fisch and Kahan (2010) is that proxy advisors aggregate information
from investors rather than provide independent advice.
30
More generally, both Cai, Garner, and Walkling (2009) and Bethel and Gillan (2002) examine director elections
and show that unfavorable recommendations by ISS lead to more votes against management
38
institutional investors providing monitoring of managerial activities via the proxy process. We
also find evidence that higher borrowing is associated with less support for proposals that ISS
recommends against but not much for compensation, G-Index, and corporate control.
To summarize, the results in columns 1 and 2 of Table 9 show indirect evidence that
lending supply recall is positively associated with votes for shareholder proposals, and against
management proposals related to compensation, governance and corporate control, especially
when ISS recommends voting against management. However, we are unable to examine directly
how the recalled shares are voted because we do not observe the identity of the institutions that
recall shares. Consequently, our results could be driven by an alternative explanation of voting
behavior by institutions with recalled shares. For example, it is possible that more shares are
recalled in contentious proposals where ISS recommends voting against management because
the institutions support management and therefore recall shares to vote with management, while
at the same time other shareholders vote against management. Rather than monitoring managers,
it could be that institutions side with management and vote against value-increasing proposals.
To rule out this alternative explanation, we examine voting only by mutual funds. Mutual funds
provide an opportunity to better observe how recalled shares are voted because mutual funds are
one of the largest lenders of shares (see for example D’Avolio (2002)). Therefore we can
examine how voting behavior of a subset of investors who are significant providers of lending
supply is influenced by the recall of lending shares.
We obtain data on mutual fund voting behavior reported on SEC Form N-PX. Form NPX identifies all proposals on which the fund has voted portfolio securities and discloses how the
fund voted on each proposal (the number of shares voted is not required to be disclosed). Our
sample includes mutual fund voting data for 6,651 individual funds that are part of 308
39
institutions (mutual fund families) for the 3,826 record dates that include non-routine proposals.
In total we have 1,524,290 fund-proposal voting behavior in our sample. In column 3 of Table 9,
we examine how voting outcome is affected by recall in supply for mutual funds by estimating if
the mutual fund voted FOR the proposal, where FOR is equal to 100 if the fund voted in favor of
the proposal, and zero otherwise.31 We repeat the estimation of FOR presented in column 2 but at
the mutual fund level and include fund family fixed effects in addition to firm fixed effects to
control for fund family-level policies on both voting and lending supply recall. The results show
that mutual fund voting is associated with support for proposals where there is a greater recall in
general. However, where ISS recommends against the proposal or the proposal is related to
compensation or anti-takeover a larger recall in supply is associated with less support from
mutual funds. This result alleviates the concern that mutual funds are recalling shares to vote
with management when other shareholders are following ISS’s advice and voting against
management.
Collectively, these results show that changes in lending of supply have a meaningful
impact on voting outcomes and that institutions act on ISS recommendations. Further, while the
fewer number of votes in favor of proposals may not result in the proposal being rejected, there
is evidence that votes recorded against proposals have spillover governance effects.. Cai, Garner,
and Walkling (2009) and Fischer et al. (2009) show that meaningful vote totals against director
election proposals, even where the proposal passes, are followed by changes in the board,
management, or corporate actions within the next year. Finally, the results on proposal
characteristics in general compare favorably with those found by Iliev, Lins, Miller, and Roth
(2011), regarding votes against management globally. Once again, our results are consistent with
31
We present OLS estimations to ensure that distributional assumptions do not unduly affect our results (Angrist
and Pischke (2009)). The results are robust to employing logit estimations.
40
lenders recalling shares ahead of the proxy record date to exercise their vote. Further, the results
suggest that the effort put in by institutional investors into determining when to recall shares
does impact voting outcome.
7.
Additional Analysis
7.1
Dividend Record Dates
There is some evidence that the equity lending market is affected by the dividend record
date due to tax-arbitrage strategies (Christoffersen et al. (2005), Saffi and Sigurdsson (2011), and
Thornock (2013)). To ensure that our results are not driven by an alternative explanation based
on dividend tax-arbitrage strategies, we examine the frequency of dividend and proxy record
dates. For the 7,415 proxy record dates in our sample, we observe 2,609 dividend record dates in
the t=-30 to t=30 days window around the proxy record date. The mean (median) number of days
between the proxy record date and the dividend record date is 11.6 (11) days and only 235 proxy
record dates coincide with a dividend record date.
In Figure 3, we plot the equity lending market activity around the dividend record date.
We find a large spike in borrowing demand and fees around dividend record dates, but little
change in lending supply. These results contrast sharply with Figure 1, which shows that the
activity around proxy voting dates is characterized by a marked reduction in lending supply and
only a small change in borrowing demand and fees.
In Panel A of Table 10, we present additional robustness results. We repeat the tests
conducted for the proxy record date, but now, we adjust for dividend record dates. We include a
dummy variable equal to one if the firm reports paying a dividend at least once in the past three
years (DIV DUMMY), and a dividend record date dummy equal to 1 if we find that the dividend
record date is within (-1, +1) days of a proxy voting record date (DIV RDATE). We first estimate
41
regressions by using only the dividend record date and then include the proxy voting record date.
When we examine the effects of dividends, we find that on average, firms that pay dividends
exhibit a higher lending supply. In tests in which we exclude the proxy voting record date, we
find a significant recall in supply of -1.358% of market capitalization on the dividend record date
and an increase in borrowing demand.
When we introduce the proxy voting record date, we see that our earlier results shown in
Table 3 of reduced lending supply, increased borrowing demand, and fees continue to hold.
However, after controlling for the proxy voting record date, we find that the results reported by
Thornock (2013) and Ringgenberg (2011), that lending supply is lower around ex-dividend dates,
no longer hold. Borrowing demand and loan fees both increase around the dividend record date.
The 0.554% increase in borrowing demand on dividend record dates is economically large and
an order of magnitude greater than the change in borrowing of 0.06% found on proxy voting
records date. The equity lending market behaves differently around proxy voting record dates
than it does around dividend record dates. There is a much larger increase in shares borrowed
around a dividend record date than around the time of a proxy vote likely related to tax-related
arbitrage strategies (Christoffersen, Géczy, Musto, and Reed (2005)).
7.2
Financial Crisis
During the financial crisis of 2008, there was considerable concern about counterparty
risk following the events surrounding Bear Stearns and Lehman Brothers. The period exhibited
high volatility in funding rates that also generated large swings in loan fees. Aitken and Singh
(2009) examine the 10-Q reports of three major custodian banks (Bank of New York, State Street,
and J.P. Morgan) before and after the bankruptcy of Lehman Brothers and find a decrease in total
securities lending from $1.48 trillion in June, 2008 to $0.82 trillion by December, 2008. Some
42
investors had concerns about the instruments used to invest the collateral and equity lenders sued
some custodial banks. The drop off in the securities lending during the crisis was due to a
number of factors including decrease in demand as borrowers decreased their leverage and
pulled to the side and very conservative cash reinvestment guidelines that got put into place.
The short-selling bans imposed by regulators in many markets also had an impact on
short selling and securities lending. Beber and Pagano (2013) find that the short-selling bans
imposed in more than 20 different countries during the financial crisis reduced liquidity, slowed
price discovery, and failed to support stock prices. Boehmer, Jones, and Zhang (2013) study the
short-selling ban in the U.S. and find a reduction in shorting activity and an increase in spreads,
price impact, and intraday volatility. Kolasinski, Reed, and Thornock (2013) find a significant
increase in loan fees following the ban.
In Panel B of Table 10, we introduce a dummy LEHMAN, which we set equal to one for
all days in 2008 on or after September 15th that characterize our “crisis” period. We use this
dummy to examine the effect of the financial crisis on the equity lending market around record
dates. Supply, demand, and fees all decreased after Lehman’s bankruptcy. Borrowing demand
decreased more than lending supply, which explains why fees decrease by about 29 bps. Even
after controlling for the financial crisis period, we find reduced supply and a small increase in
demand at the record date; thus, our results continue to hold. The interaction of RDATE with
LEHMAN does not result in any significant changes in lending supply before or after the crisis.
However, we do find evidence to support less borrowing demand and fees on record dates
following Lehman’s bankruptcy. This finding is consistent with borrowers becoming less keen to
engage in short selling due to fewer profitable opportunities.
8.
Conclusion
43
The preferences of institutional investors tend to be private and generally cannot be
studied because they are often conducted behind the scenes. We use a unique setting and a
comprehensive proprietary data set that allows us to directly observe the role of institutional
investors in corporate governance. We examine change in lending supply, borrowing demand,
and fees in the securities lending market around the proxy record date. In our study, we focus on
the role of investors in voting and the alignment of economic exposure and voting rights, with
the goal of examining the extent to which corporate governance matters to institutional investors
and to which they use the proxy process to exercise their opinions. If institutions have loaned out
their shares, then they cannot exercise their vote. Hence, institutional investors must decide
whether to restrict lending and even recall shares already on loan in the event of an upcoming
vote. Just prior to the proxy record date, we find a significant reduction in lending supply,
because institutions restrict or call back their loaned shares in order to vote.
We examine both the lending and borrowing sides and are able to control for their
simultaneous impact on determining fees and improve upon previous estimates of the value of
the vote. Average borrowing fee in isolation is not a sufficient measure of the value of a vote and
results in underestimating the value. Our measure of the value of a vote is the change in the
borrowing fee that would make the observed changes in supply (and demand) on the record date
to be zero. We find the value of the vote to be much higher for lenders than for borrowers. We
also find heterogeneity in the value of the vote based on firm and proposal characteristics. The
value is higher for firms with weak performance, weak governance, smaller firms, and firms with
low institutional ownership. The value is also higher when non-routine, compensation, antitakeover and corporate control proposals are on the ballot.
44
We show that the recall in equity lending supply is related to the subsequent vote
outcome. Higher recall is associated with fewer FOR votes for management and more FOR votes
for shareholder proposals. The influence of proxy advisory firm ISS is also evident in voting
outcome. If ISS opposes management, then we find the higher recall to be associated with less
FOR votes for the proposal. Further, we show that the recall in lending supply has an even bigger
impact on voting outcome for the proposals with a higher estimated value of the vote, like
compensation and corporate control-related. We find little relation between borrowing demand
and vote outcome.
Our findings are consistent with the fact institutional investors recall their shares
selectively, depending on the proposals on the ballot. Even though it might be argued that by
recalling shares for a few days, institutions are not giving up much in terms of lending fees, our
evidence suggests that institutions are putting in significant effort in the proxy voting process in
determining when to recall. The results imply that institutions do care about corporate
governance and they use the proxy voting process as a channel. Our analysis suggests policy
makers should address several issues related to proxy voting, including the need for investors to
learn about proxy items before the record date so that they can decide whether to lend their
shares or not.
45
References
Aggarwal, R., I. Erel, M. Ferreira, and P. Matos, 2011, Does Governance Travel Around the
World? Evidence from Institutional Investors, Journal of Financial Economics, 100, 154-182.
Aggarwal, R., I. Erel, R. Stulz, and R. Williamson, 2009, Differences in Governance Practices
between U.S. and Foreign Firms: Measurement, Causes, and Consequences, Review of Financial
Studies, 22, 3131-3169.
Aitken, J., and M. Singh, 2009, Deleveraging After Lehman-Evidence from Reduced
Rehypothecation, IMF Working Paper 09/42, International Monetary Fund.
Alexander C., M. Chen, D. Seppi, and C. Spatt, 2010, Interim News and the Role of Proxy
Voting Advice, Review of Financial Studies, 23, 4419-4454.
Angrist, J. D., K. Graddy, and G. W. Imbens, 2000, The Interpretation of Instrumental Variables
Estimators in Simultaneous Equations Models with an Application to the Demand for Fish,
Review of Economic Studies, 67, 499-527.
Angrist, J., and J-S. Pischke, 2009, Mostly Harmless Econometrics, Princeton Press, Princeton,
NJ.
Asquith, P., A. Au, T. Covert, and P. Pathak, 2013, The Market for Borrowing Corporate Bonds,
Journal of Financial Economics, 107, 155-182.
Beber, A., and M. Pagano, 2013, Short-Selling Bans around the World: Evidence from the 200709 Crisis, Journal of Finance, 68, 343-381.
Bethel, J. E. and S. L. Gillan, 2002, The Impact of the Institutional and Regulatory Environment
on Shareholder Voting, Financial Management, 31, 29-54.
Blocher, J., A. V. Reed, and E. D. Van Wesep, 2013, Connecting Two Markets: An Equilibrium
Framework for Shorts, Longs and Stock Loans, Journal of Financial Economics, 108, 302-322.
Boehmer, E., C. M. Jones, and X. Zhang, 2013, Shackling Short Sellers: The 2008 Shorting Ban,
Review of Financial Studies, 26, 1363-1400.
Brav, A., W. Jiang, F. Partnoy, and R. Thomas, 2008, Hedge Fund Activism, Corporate
Governance, and Firm Performance, Journal of Finance, 63, 1729-1775.
Brickley, J., R. Lease, and C. Smith, 1988, Ownership Structure and Voting on Antitakeover
amendments, Journal of Financial Economics, 20, 267-291.
Bushee, B., M. Carter, and J. Gerakos, 2010, Institutional Investor Preferences for Corporate
Governance Mechanisms, Working Paper, University of Pennsylvania.
46
Cai, J., J. L. Garner, and R. A. Walkling, 2009, Electing Directors, Journal of Finance, 64,
2389-2421.
Chen, X., J. Harford, and K. Li, 2007, Monitoring: Which Institutions Matter, Journal of
Financial Economics, 86, 279-305.
Choi, S., J. Fisch, and M. Kahan, 2010, The Power of Proxy Advisors: Myth or Reality?, Emory
Law Journal, 59, 869-918.
Chung, K. H. and H. Zhang, 2011, Corporate Governance and Institutional Ownership, Journal
of Financial and Quantitative Analysis, 46, 247-273.
Christoffersen, S. K., C. C. Géczy, D. K. Musto, and A. V. Reed, 2005, Cross Border Dividend
Taxation and the Preferences of Taxable and Non-Taxable Investors, Journal of Financial
Economics, 78, 121-144.
Christoffersen, S. K., C. C. Géczy, D. K. Musto, and A. V. Reed, 2007, Vote Trading and
Information Aggregation, Journal of Finance, 62, 2897-2929.
Clifford, C., 2008, Value Creation or Destruction: Hedge Funds as Shareholder Activists,
Journal of Corporate Finance, 14, 323-336.
Cohen, L., K. B. Diether, and C. J. Malloy, 2007, Supply and Demand Shifts in the Shorting
Market. Journal of Finance, 62, 2061-2096.
Cuñat, V., M. Gine, and M. Guadalupe, 2012, The Vote is Cast: The Effect of Corporate
Governance on Shareholder Value, Journal of Finance, 67, 1943-1977.
D’Avolio, G., 2002, The Market for Borrowing Firm, Journal of Finance Economics, 66, 271306.
Edwards, A., and K. Hanley, 2010, Short Selling in Initial Public Offerings, Journal of Financial
Economics, 98, 21-39.
Evans, R., M. Ferreira, and M. Prado, 2013, Equity Lending, Investment Restrictions and Fund
Performance, Working Paper, University of Virginia.
Gantchev, N., 2013, The Costs of Shareholder Activism: Evidence from a Sequential Decision
Model, Journal of Financial Economics, 107, 610-631.
Gaspar, J., M. Massa and P. Matos, 2005, Shareholder Investment Horizons and the Market for
Corporate Control, Journal of Financial Economics, 76, 135-165.
Geczy, C., D. Musto, and A. Reed, 2002, Firms are special too: an Analysis of the Equity
Lending Market, Journal of Financial Economics, 66, 241-269.
47
Gillan, S., and L. Starks, 2000, Corporate Governance Proposals and Shareholder Activism: The
Role of Institutional Investors, Journal of Financial Economics, 57, 275-305.
Gillan, S., and L. Starks, 2007, The Evolution of Shareholder Activism in the United States,
Journal of Applied Corporate Finance, 19, 55-73.
Gompers, P., J. Ishii, and A. Metrick, 2003, Corporate Governance and Equity Prices, Quarterly
Journal of Economics, 118(1), 107-155.
Grene, Sophia, September 12, 2010, Guide Issued on Securities Lending, FT.Com,
http://www.ft.com/cms/s/0/5da3c26c-bd04-11df-954b-00144feab49a,dwp_uuid=571fa9ea-b6a711db-8bc2-0000779e2340.html.
Fischer, P., J. Gramlich, B. Miller, and H. White, 2009, Investor Perceptions of Board
Performance: Evidence from Uncontested Director Elections, Journal of Accounting and
Economics, 48, 172-189
Fos, V., 2011, The Disciplinary Effects of Proxy Contests, Working Paper, University of Illinois
at Urbana-Champaign.
Hartzell, J., and L. Starks, 2003, Institutional Investors and Executive Compensation, Journal of
Finance, 58, 2351-2374.
Helwege, J., V. J. Intintoli, and A. Zhang, 2012, Voting with their Feet or Activism? Institutional
Investors’ Impact on CEO Turnover, forthcoming, Journal of Corporate Finance.
Holden, C. W, and A. Subrahmanyam, 1992, Long-Lived Private Information and
Imperfect Competition, Journal of Finance, 47, 247–70.
Hu, H. T.C., and B. Black, 2006, The New Vote Buying: Empty Voting and Hidden Ownership,
Southern California Law Review, 79, 811-908.
Hu, H. T.C., and B. Black, 2007, Hedge Funds, Insiders, and the Decoupling of Economic and
Voting Ownership: Empty Voting and Hidden (Morphable) Ownership, Journal of Corporate
Finance, 13, 343-367.
Hu, H. T.C., and B. Black, 2008, Equity and Debt Decoupling and Empty Voting II: Importance
and Extensions, University of Pennsylvania Law Review, 156, 625-739.
Hwang, B., B. Liu, and W. Xu, 2013, Can Short Selling Help Correct Under-Pricing? Working
Paper, Purdue University.
Iliev, P., K. V. Lins, D. P. Miller, and L. Roth, 2011, Shareholder Voting and Corporate
Governance Around the World, Working Paper, Penn State University
48
Jones, C., and O. Lamont, 2002, Short-Sale Constraints and Stock Returns, Journal of Financial
Economics, 66, 207-239.
Kalay, A., O. Karakas, and S. Pant, 2013, The Market Value of Corporate Votes: Theory and
Evidence from Option Prices, forthcoming, Journal of Finance.
Kaplan, S., T. Moskowitz, and B. Sensoy, 2013, The Effects of Stock Lending on Security Prices:
An Experiment, forthcoming, Journal of Finance.
Karmel, R. S., 2010, Voting Power Without Responsibility or Risk - How Should Proxy Reform
Address the Decoupling of Economic and Voting Rights?, Villanova Law Review, 55, 93-124.
Klein, A., and E. Zur, 2009, Entrepreneurial Shareholder Activism: Hedge Funds and Other
Private Investors, Journal of Finance, 64, 187-229.
Kolasinski, A. C., A. V. Reed, and J. R. Thornock, 2013, Can Short Restrictions Result in More
Informed Short Selling? Evidence from the 2008 Regulations, Financial Management, 42, 155181.
Kolasinski, A. C., A. V. Reed, and M. Ringgenberg, 2013, A Multiple Lender Approach to
Understanding Supply and Demand in the Equity Lending Market, Journal of Finance, 68, 559595.
Lambert, E., June 22, 2009, Securities Lending Meltdown, Forbes, 183, 13, p. 50.
Matvos, G., and M. Ostrovsky, 2010, Hetergeneity and Peer Effects in Mutual Fund Proxy
Voting, Journal of Financial Economics, 98, 90-112.
McCahery, J., Z. Sautner, and L. Starks, 2011, Behind the Scenes: the Corporate Governance
Preferences of Institutional Investors, Working Paper, University of Amsterdam.
Ofek, E., and M. Richardson, 2002, The Valuation and Market Rationality of Internet Stock
Prices, Oxford Review of Economic Policy, 18, 265-287.
Parrino, R., R. Sias, and L. Starks, 2003, Voting with Their Feet: Institutional Ownership
Changes around Forced CEO Turnover, Journal of Financial Economics, 68, 3-46.
Prado, M., P. A. C. Saffi, and J. Sturgess, 2013, Ownership Structure, Limits to Arbitrage and
Stock Returns: Evidence from Equity Lending Markets, Working Paper, Georgetown University.
Ringgenberg, M., 2011, When Short Sellers Agree to Disagree: Short sales, Volatility, and
Heterogeneous Beliefs, Working Paper, Washington University in St. Louis..
Roy, A. D., 1951, Some Thoughts on the Distribution of Earnings, Oxford Economic Papers, 3,
135-146.
49
Saffi, P., and K. Sigurdsson, 2011, Price Efficiency and Short Selling, Review of Financial
Studies, 24, 821-852.
Thornock, J., 2013, The Effects of Dividend Taxation on Short Selling, forthcoming, The
Accounting Review.
Wooldridge, J. M., 2001, Econometric Analysis of Cross Section and Panel Data, MIT Press
Books, The MIT Press, Edition 1, Volume 1, page 231.
50
Figure 1
Equity Lending Market Activity around Record Date
The figure presents a daily plot of lending supply, on loan, utilization and loan fees for the period (30,+30) for 7,415 record dates (day t=0 is the proxy voting record date) during the years 2007-2009.
SUPPLY is the percentage of market capitalization available to lend; ONLOAN is the percentage of
market capitalization actually borrowed; UTILIZATION is the ratio of ONLOAN to SUPPLY expressed in
percentage; FEE is the annualized borrowing fees expressed in basis points. In the top panel SUPPLY is
shown on the left-hand axis and UTILIZATION is shown on the right-hand axis. In the bottom panel, the
left-hand axis shows ONLOAN and the right-hand axis shows FEE.
25.0%
20.0%
24.5%
19.5%
24.0%
SUPPLY
23.0%
18.5%
22.5%
18.0%
22.0%
17.5%
UTILIZATION
19.0%
23.5%
21.5%
17.0%
21.0%
16.5%
20.0%
16.0%
-30
-28
-26
-24
-22
-20
-18
-16
-14
-12
-10
-8
-6
-4
-2
0
2
4
6
8
10
12
14
16
18
20
22
24
26
28
30
20.5%
SUPPLY
UTILIZATION
4.15%
51
50
4.10%
48
47
4.00%
46
3.95%
45
3.90%
44
ONLOAN
FEE (bps)
51
FEE (bps)
4.05%
-30
-28
-26
-24
-22
-20
-18
-16
-14
-12
-10
-8
-6
-4
-2
0
2
4
6
8
10
12
14
16
18
20
22
24
26
28
30
ONLOAN
49
Figure 2
Equity Lending Market Activity around Ex-Dividend Dates
The figure presents a daily plot of lending supply, on loan, utilization and loan fees for the (-30, +30)
days period around 14,278 dividend ex-dividend dates (day t=0 is based on settlement taking place on exdividend date) during the years 2007-2009. SUPPLY is the percentage of market capitalization available
to lend; ONLOAN is the percentage of market capitalization actually borrowed; UTILIZATION is the ratio
of ONLOAN to SUPPLY expressed in percentage; FEE is the annualized borrowing fees expressed in
basis points. In the top panel SUPPLY is shown on the left-hand axis and UTILIZATION is shown on the
right-hand axis. In the bottom panel, the left-hand axis shows ONLOAN and the right-hand axis shows
FEE.
26%
18.5%
18.0%
17.5%
17.0%
22%
16.5%
20%
16.0%
15.5%
18%
15.0%
14.5%
16%
14.0%
13.5%
-30
-28
-26
-24
-22
-20
-18
-16
-14
-12
-10
-8
-6
-4
-2
0
2
4
6
8
10
12
14
16
18
20
22
24
26
28
30
14%
4.5%
48
4.4%
46
4.3%
44
4.2%
42
4.1%
40
4.0%
38
3.9%
36
3.8%
34
3.7%
32
3.6%
30
ONLOAN
FEE (bps)
52
FEE (bps)
UTILIZATION
-30
-28
-26
-24
-22
-20
-18
-16
-14
-12
-10
-8
-6
-4
-2
0
2
4
6
8
10
12
14
16
18
20
22
24
26
28
30
ONLOAN
SUPPLY
UTILIZATION
SUPPLY
24%
Table 1
Equity Lending and Firm Characteristics
The table presents characteristics of the equity lending market around the record dates of Russell 3000 firms
from 2007 to 2009. Panel A presents average equity lending characteristics based on the -30 to +30 days
window around record dates. Panel B shows the change in lending characteristics from the average of days -30
to -20 to record date 0. In Panel C we show the yearly averages of the equity lending variables. SUPPLY is the
percentage of market capitalization available to lend; ONLOAN measures borrowing demand and is the
percentage of market capitalization actually borrowed; FEE is the annualized borrowing fee expressed in basis
points; and UTILIZATION is the ratio of ONLOAN to SUPPLY expressed in percentage. SUPPLY, ONLOAN,
and FEE are winsorized at 1%.
Panel A: Equity Lending Characteristics
Obs.
SUPPLY
ONLOAN
FEE
UTILIZATION
SPECIAL
SUPPLY
ONLOAN
FEE
UTILIZATION
SUPPLY
ONLOAN
FEE
UTILIZATION
Mean
23.78%
4.06%
48.28
17.78%
0.09%
Median
24.23%
2.63%
9.90
12.59%
0.00%
Std Dev
10.71%
4.22%
158.25
16.25%
0.29%
7,415
7,415
7,415
7,415
7,415
Panel B: Changes in Equity Lending on Proxy Record Date
7,415
7,415
7,415
7,415
2007
21.78%
4.60%
48.34
22.52%
-1.93%
-1.31%
2.71%
0.06%
0.03%
1.11%
2.40
0.48
40.37
1.81%
1.10%
5.69%
Panel C: Average Equity Lending Over Time
2008
23.42%
4.72%
58.50
20.99%
2009
21.26%
3.22%
43.05
15.75%
53
Min
1.65%
0.01%
-50.84
0.23%
0.00%
-19.85%
-7.85%
-553.41
-49.98%
Max
48.57%
20.49%
1113.81
68.90%
1.00%
34.42%
12.90%
1080.85
94.38%
Table 2
Descriptive Statistics – Voting Proposals
The table presents descriptive statistics for 56,220 proxy proposals of Russell 3000 firms in the 2007-2009
period. Panel A shows data for all proposals while Panel B shows voting outcome statistics for different types of
non-routine proposals. VOTES CAST is the percentage of the total votes cast relative to shares outstanding. FOR,
AGAINST, and ABSTAIN are the total number of votes for, against, and abstained for the proposal, respectively,
relative to the BASE by which the proposal outcome is measured (expressed as a percentage). VOTE MARGIN is
defined as FOR minus the minimum threshold required for the proposal to pass. Voting outcome variables are
winsorized at the 1%-level. In Panel B, Obs. refers to the number of proposal observations and RD Obs. refers to
the number of record date proposals (there may be multiple proposals on each record date). NON ROUTINE
proposals are proposals not relating to operational or uncontested directorships. MGT are managementsponsored proposals. SHDR are shareholder-sponsored proposals. G-INDEX, COMP and CORP CONTROL are,
respectively, dummies for anti-takeover, compensation and merger/proxy contest related proposals.
Panel A: Voting Outcome for All Proposals
Obs.
Mean
Median
Std Dev
Min
56,220
86.62%
88.74%
9.49%
37.42%
VOTES CAST
56,220
91.86%
97.37%
14.15%
18.94%
FOR
56,220
7.54%
2.48%
13.25%
0.00%
AGAINST
56,220
0.41%
0.00%
1.56%
0.00%
ABSTAIN
56,220
70.16%
87.10%
30.79%
-31.37%
VOTE MARGIN
Panel B: Voting Outcome by Proposal Type for Non-Routine Proposals
VOTES
Obs. RD Obs.
FOR
AGAINST ABSTAIN
CAST
6,887 3,719
77.65% 73.24%
23.02%
1.86%
NON ROUTINE
5,127 3,717
78.82% 84.61%
11.73%
0.97%
- MGT
1,760
824
74.27% 39.99%
55.91%
4.44%
- SHDR
4,024 2,854
77.76% 80.67%
18.02%
1.34%
- COMP
1,190 1,034
79.54% 65.23%
28.19%
1.10%
- G-INDEX
588
371
72.31% 80.60%
14.64%
3.53%
- CORP CONTROL
54
Max
100%
100%
75%
11.9%
100%
VOTE
MARGIN
23.43%
35.01%
-10.50%
30.48%
9.96%
51.35%
Table 3
Abnormal Lending Supply, Borrowing Demand and Fees around Proxy Voting Record Dates
The table presents results from an event study on the effect of proxy voting on equity lending supply demand in
the (-30, +30) days period around 7,415 voting record dates (record date is t=0). SUPPLY is the percentage of
market capitalization available to lend. ONLOAN is the percentage of market capitalization actually borrowed
and FEE is the annualized borrowing fees expressed in basis points. RDATE is a dummy equal to one on the
record dates. Control variables comprise governance index (GOV41), institutional ownership (INST),
concentration of institutional ownership as measured by the Herfindahl index (INST CONC), the natural log of
market capitalization (SIZE), book to market (BM), stock turnover (TURNOVER), bid-ask spread (SPREAD), a
small firm dummy (PRICE<$5), and Short-Term Mom and Long-Term Mom are defined as the cumulative
returns in the previous 5 and 252 days, respectively. All regressions include year and firm fixed-effects and
robust standard errors clustered at the firm-level, presented in parentheses. *** (**,*) indicates significance at
the 1% (5%, 10%) level.
Dependent Variable
SUPPLY
RDATE
INST
INST CONC
SIZE
BM
TURNOVER
SPREAD
PRICE<$5
Short-Term Mom
Long-Term Mom
GOV41
Constant
Firm FE
Year FE
Adj. R-squared
# of Firms
ONLOAN
FEE
(1)
(2)
(3)
(4)
(5)
(6)
-1.640***
(0.037)
28.004***
(0.437)
-51.468***
(2.257)
-0.670***
(0.067)
1.326***
(0.151)
0.002
(0.062)
-0.169
(0.124)
-2.046***
(0.260)
-2.096***
(0.251)
1.140***
(0.207)
4.603***
(1.165)
7.921***
(0.800)
No
Yes
0.672
-1.623***
(0.036)
21.204***
(0.877)
-23.484***
(2.587)
1.146***
(0.213)
0.299**
(0.152)
0.076**
(0.030)
-0.129***
(0.039)
0.480**
(0.202)
-2.058***
(0.150)
0.305*
(0.183)
1.647
(1.882)
0.082***
(0.013)
5.115***
(0.245)
-4.227***
(0.694)
-0.661***
(0.036)
-0.110
(0.080)
1.068***
(0.042)
-0.333***
(0.045)
-0.755***
(0.125)
-0.155
(0.139)
-0.707***
(0.110)
-2.393***
(0.679)
6.570***
(0.432)
No
Yes
0.290
0.085***
(0.010)
11.829***
(0.608)
-1.363
(1.524)
0.314**
(0.137)
0.094
(0.104)
0.367***
(0.023)
-0.042*
(0.024)
-0.518***
(0.137)
-0.367***
(0.094)
-0.345***
(0.120)
-8.771***
(1.184)
1.814***
(0.395)
-139.656***
(15.796)
275.384***
(54.872)
-8.354***
(1.340)
-11.857**
(5.492)
22.590***
(2.094)
-14.984***
(3.522)
48.682***
(10.909)
3.706
(6.975)
-19.199***
(6.013)
6.593
(29.631)
170.760***
(25.210)
No
Yes
0.104
1.572***
(0.332)
45.744*
(23.674)
206.886**
(82.101)
-8.709
(5.931)
13.865**
(6.302)
2.762***
(0.949)
-1.278
(1.592)
3.421
(6.788)
7.972*
(4.454)
-4.320
(5.011)
63.828
(45.069)
Yes
Yes
0.905
3,053
55
Yes
Yes
0.779
3,053
Yes
Yes
0.749
3,053
Table 4
Lending Supply, Borrowing Demand and Fee for Firms “On Special”
Panel A of the table reports the averages of equity lending variables at t=-30 and Panel B reports the change in
each lending attribute between the (-30, -20) days period average to the record date. On Special includes firms
with a borrowing fee in excess of 100 bps, measured on t=-30. SUPPLY is the percentage of market
capitalization available to lend. ONLOAN is the percentage of market capitalization actually borrowed; FEE is
the annualized borrowing fees expressed in basis points; and UTILIZATION is the ratio of ONLOAN to SUPPLY
expressed in percentage.
Panel A: Lending Supply, Borrowing Demand and Fee at t=-30
On Special
No
Yes
FEE>1000 bps
#(Record Dates)
SUPPLY
ONLOAN
FEE
UTILIZATION
6,756
659
25.02%
14.52%
3.83%
6.85%
9.30
428.68
15.20%
42.96%
79
12.40%
6.44%
1108.37
50.81%
Panel B: Change from the Average Level During t=-30 to -20 and Record Date
On Special
No
Yes
FEE>1000 bps
#(Record Dates)
SUPPLY
ONLOAN
FEE
UTILIZATION
6,756
659
-2.00%
-0.92%
0.10%
-0.26%
1.09
8.37
1.83%
1.24%
79
-0.73%
-0.30%
9.37
0.47%
56
Table 5
First-stage Estimation of Fee
The table presents results for the first stage estimates of FEE as a function of our instruments and firm controls.
FEE is the annualized borrowing fee expressed in basis points. Hedging Demand is defined as the equalweighted cumulative return in the past 252 days of related firms with the same four-digit-GICS industry
classification. INST CONC is the institutional ownership concentration measured as the Hirschman-Herfindahl
index normalized between zero and one. Control variables include a record date dummy (RDATE),
institutional ownership (INST), the natural logarithm of market capitalization (SIZE), book to market (BM),
stock turnover (TURNOVER), bid-ask spread (SPREAD), a small firm dummy (PRICE<$5), cumulative returns
in the previous 5 days (Short-Term Mom), cumulative returns in the previous 252 days (Long-Term Mom), and
the internal governance measure (GOV41). All regressions include year and firm fixed-effects and robust
standard errors double clustered at the firm and year level, presented in parentheses. *** (**,*) indicates
significance at the 1% (5%, 10%) level.
Dependent Variable: FEE
Instrument in Supply Estimation
Hedging Demand
-37.627***
(11.876)
Instrument in On Loan Estimation
INST CONC
208.848**
(82.300)
Controls
RDATE
1.576***
(0.333)
50.496**
(23.662)
-8.794
(5.897)
13.969**
(6.266)
2.751***
(0.949)
-1.349
(1.588)
3.478
(6.767)
0.072
(0.045)
-0.042
(0.049)
58.750
(44.766)
Yes
Yes
0.025
INST
SIZE
BM
TURNOVER
SPREAD
PRICE<$5
Short-term Mom
Long-term Mom
GOV41
Firm FE
Year FE
Adj. R-squared
57
Table 6
Second Stage Estimations
The table presents our main second stage results using the instrumented fee estimated in the first stage to control
for the endogeneity of the fee. SUPPLY is the percentage of market capitalization available to lend. ONLOAN is
the percentage of market capitalization actually borrowed. FEE is the annualized borrowing fee expressed in
basis points, RDATE is a variable equal to one at record date, zero otherwise. Control variables include
institutional ownership (INST), the institutional ownership concentration (INST CONC), the natural logarithm of
market capitalization (SIZE), book to market (BM), stock turnover (TURNOVER), bid-ask spread (SPREAD), a
small firm dummy (PRICE<$5), cumulative returns in the previous 5 days (Short-Term Mom), cumulative
returns in the previous 252 days (Long-Term Mom), and the internal governance measure (GOV41). Hedging
Demand is defined as the equal-weighted cumulative return in the past 252 days of related firms with the same
four-digit-GICS industry classification. VVOTE is defined as the ratio of between the RDATE and FEE*RDATE
coefficients as shown in Section 5.1. The significance of VVOTE is computed using the delta method. All
regressions include year and firm fixed-effects and robust standard errors clustered at the firm and year level,
presented in parentheses. *** (**,*) indicates significance at the 1% (5%, 10%) level.
Dependent Variable
SUPPLY
FEE
(1)
-0.006
(0.007)
FEE*RDATE
RDATE
INST
INST CONC
SIZE
BM
TURNOVER
SPREAD
PRICE<$5
Short-term Mom
Long-term Mom
GOV41
-1.613***
(0.038)
21.497***
(0.931)
-22.150***
(3.079)
1.090***
(0.218)
0.388**
(0.196)
0.094***
(0.035)
-0.138***
(0.042)
0.502**
(0.208)
-0.020***
(0.002)
0.003
(0.002)
2.058
(1.964)
ONLOAN
(2)
-0.007
(0.007)
0.009**
(0.004)
-2.068***
(0.222)
21.498***
(0.931)
-22.163***
(3.078)
1.088***
(0.218)
0.387**
(0.196)
0.094***
(0.035)
-0.136***
(0.042)
0.504**
(0.208)
-0.020***
(0.002)
0.0028
(0.002)
2.060
(1.964)
HEDGING
Firm FE
Year FE
Kleibergen-Paap rk LM
P-Value
Cragg-Donald Wald F
VVOTE (bps p.a.)
Yes
Yes
8.967
0.003
2189.63
Yes
Yes
22.757
0.000
713.55
226.9***
58
(3)
-0.015*
(0.008)
0.098***
(0.021)
12.670***
(0.841)
-0.388*
(0.210)
0.418*
(0.218)
0.427***
(0.041)
-0.111***
(0.040)
-0.545***
(0.195)
0.002
(0.002)
-0.005***
(0.002)
-3.267**
(1.608)
-0.031**
(0.016)
Yes
Yes
6.057
0.014
1216.02
(4)
-0.015*
(0.008)
-0.001***
(0.000)
0.167***
(0.030)
12.669***
(0.840)
-0.388*
(0.210)
0.418*
(0.218)
0.427***
(0.041)
-0.111***
(0.040)
-0.545***
(0.195)
0.002
(0.002)
-0.005***
(0.002)
-3.268**
(1.608)
-0.031**
(0.016)
Yes
Yes
6.058
0.014
607.91
122.1***
Table 7
Firm Characteristics and Value of Vote
The table examines differences in the value of a vote based on splitting our sample of 7,415 firm-record dates in
below and above-median values of four firm characteristics: corporate governance, institutional ownership, stock
returns, and market capitalization. Panel A is based on the GOV41 measure of internal corporate governance. Panel B
uses institutional ownership taken from 13f files, Panel C uses cumulative returns in the previous twelve months, and
Panel D is based on stock market capitalization. SUPPLY is the percentage of market capitalization available to lend
and ONLOAN is the percentage of market capitalization actually borrowed. FEE is the annualized borrowing fee
expressed in basis points, RDATE is a variable equal to one at record date, zero otherwise. The value of the vote,
VVOTE, is defined as the ratio of between the RDATE and FEE*RDATE coefficients as shown in Section 5.1. The
significance of VVOTE is computed using the delta method. All regressions include year and firm fixed-effects and
robust standard errors clustered at the firm and year level, presented in parentheses. *** (**,*) indicates significance
at the 1% (5%, 10%) level.
Firm
Lending Variable
SUPPLY
Coefficients
Characteristic
RDATE
Panel A: Corporate Governance
Low Governance
-2.083***
(0.229)
-2.095***
(0.496)
High Governance
FEE*RDATE
(bps p.a.)
0.007**
(0.003)
0.018*
(0.010)
314***
Difference in Value of Vote
P-value
ONLOAN
VVOTE
117***
Yes
0.087
Low Governance
0.217***
(0.055)
0.132***
(0.019)
High Governance
-0.002***
(0.001)
-0.001***
(0.0002)
Difference in Value of Vote
P-value
118***
126***
No
0.812
Panel B: Institutional Ownership
SUPPLY
Low Inst. Ownership
-1.976***
(0.177)
-2.189***
(0.911)
High Inst. Ownership
Difference in Value of Vote
P-value
ONLOAN
Low Inst. Ownership
0.134***
(0.044)
0.152**
(0.078)
High Inst. Ownership
Difference in Value of Vote
P-value
0.007***
(0.002)
0.016***
(0.003)
-0.0009*
(0.0005)
-0.004*
(0.002)
296***
140***
Yes
0.025
103***
46***
Yes
0.093
59
Panel C: Previous Twelve Month Returns
SUPPLY
Low Returns
-2.059***
(0.215)
-2.179***
(0.272)
High Returns
0.008**
(0.003)
0.012**
(0.006)
Difference in Value of Vote
P-value
ONLOAN
244***
174***
Yes
0.068
Low Returns
0.145**
(0.059)
0.164***
(0.046)
High Returns
-0.0011***
(0.0004)
-0.0016**
(0.0007)
Difference in Value of Vote
P-value
132***
101***
No
0.588
Panel D: Market Capitalization
SUPPLY
Low Mkt Cap
-2.907***
(0.260)
-1.988**
(0.451)
High Mkt Cap
0.009***
(0.003)
0.040**
(0.019)
Difference in Value of Vote
P-value
ONLOAN
308***
49***
Yes
0.012
Low Mkt Cap
0.148***
(0.037)
0.138***
(0.037)
High Mkt Cap
Difference in Value of Vote
P-value
-0.002***
(0.0003)
-0.002*
(0.001)
97***
80**
No
0.737
60
Table 8
Proposal Type and Value of Vote
The table examines the value of four alternative subsamples based upon proposal types. Panel A is based on nonroutine proposals defined by NYSE Rule 452 as those in which broker voting is not allowed, including proposals
related to anti-takeover provisions, stock capitalization, and mergers; Panel B is based on compensation proposals
referring to those related to managerial compensation policies; Panel C is based on the G-INDEX of anti-takeover
provisions as developed by Gompers, Ishi, and Metrick (2003); and Panel D is based on corporate control proposals
defined as those record dates with a proxy contest or merger. SUPPLY is the percentage of market capitalization
available to lend. ONLOAN is the percentage of market capitalization actually borrowed. FEE is the annualized
borrowing fee expressed in basis points, RDATE is a variable equal to one at record date, zero otherwise. The value
of the vote, VVOTE, is defined as the ratio of RDATE and FEE*RDATE coefficients as shown in Section 5.1. All
regressions include year and firm fixed-effects and robust standard errors clustered at the firm and year level,
presented in parentheses. *** (**,*) indicates significance at the 1% (5%, 10%) level.
Coefficients
Lending Variable
SUPPLY
Obs.
Proposal Type
RDATE
Panel A: Non-Routine Proposals
3,719
Non-Routine
3,696
Routine
-2.294***
(0.257)
-2.316***
(0.352)
FEE*RDATE
(bps p.a.)
0.009**
(0.004)
0.014**
(0.007)
248***
Difference in Value of Vote
P-value
ONLOAN
Difference in Value of Vote
P-value
SUPPLY
170***
Yes
0.099
3,719
Non-Routine
3,696
Routine
0.131***
(0.043)
0.285**
(0.115)
-0.001**
(0.001)
-0.003**
(0.001)
No
0.994
Panel B: Compensation Related Proposals
2,854
Compensation
2,427
Non-Compensation
-2.383***
(0.284)
-2.550***
(0.334)
110***
111***
No
0.994
0.010**
(0.005)
0.015**
(0.006)
Difference in Value of Vote
P-value
ONLOAN
VVOTE
227***
175***
No
0.127
2,854
Compensation
2,427
Non-Compensation
Difference in Value of Vote
P-value
0.091***
(0.029)
0.147***
(0.028)
-0.0008*
(0.0004)
-0.0013***
(0.0005)
119***
115***
No
0.959
61
Panel C: G-INDEX Related Proposals
SUPPLY
1,034
G-INDEX
1,202
Non-G-INDEX
-2.369***
(0.196)
-2.396***
(0.121)
0.008***
(0.001)
0.010***
(0.001)
Difference in Value of Vote
P-value
ONLOAN
Difference in Value of Vote
P-value
ONLOAN
231***
Yes
0.063
1,034
G-INDEX
1,202
Non-G-INDEX
0.108***
(0.033)
0.085**
(0.036)
-0.0005*
(0.0003)
-0.0008***
(0.0002)
Difference in Value of Vote
P-value
Panel D: Corporate Control Related Proposals (Proxy Contests and Mergers)
SUPPLY
304***
371
Corporate Control
821
Non-Corporate Control
371
Corporate Control
821
Non-Corporate Control
Difference in Value of Vote
P-value
-2.694***
(0.227)
-2.456***
(0.449)
0.243***
(0.087)
0.217***
(0.068)
0.007*
(0.004)
0.027*
(0.015)
-0.002*
(0.001)
-0.004**
(0.0023)
205***
95***
No
0.121
381***
90***
Yes
0.011
113***
49***
No
0.323
62
Table 9
Voting Outcome
The table presents results from a regressions analysis of voting outcome for non-routine proposals. The dependent variable
is VOTES FOR, the percentage of votes FOR the proposal. NON ROUTINE proposals are defined by NYSE Rule 452 as
those in which broker voting is not allowed. Columns (1) – (2) present results for all voting; Column (3) presents results
only for mutual funds voting. The independent variables are: ΔSUPPLY and ΔONLOAN, the change in lending supply and
on loan from days (t=-30 to -20) to record date (t=0). DSHR is a dummy equal to one if shareholders sponsor the proposal,
zero otherwise. DISS is a dummy equal to 1 when management is in favor and ISS is against the proposal. DSHR equals
one for shareholder-sponsored proposals. G-INDEX, COMP and CORP CONTROL are, respectively, dummies for antitakeover, compensation and merger/proxy contest related proposals All estimations include proposal fixed effects and firmlevel controls. Control variables include the internal governance measure (GOV41), institutional ownership (INST),
concentration of institutional ownership as measured by the Herfindahl index (INST CONC), the natural logarithm of
market capitalization (SIZE), book to market (BM), stock turnover (TURNOVER), bid-ask spread (SPREAD), a small firm
dummy (PRICE<$5), and prior twelve-month return (RETURN). All regressions include year and firm fixed-effects and
robust standard errors clustered at the firm level (firm-record date level in column (3)), presented in parentheses. *** (**,*)
indicates significance at the 1% (5%, 10%) level.
ΔSUPPLY
ΔONLOAN
DISS
COMP
G-INDEX
CORP CONTROL
DSHR
ΔSUPPLY * DISS
ΔSUPPLY * COMP
ΔSUPPLY * G-INDEX
ΔSUPPLY * CORP CONT
ΔSUPPLY * DSHR
ΔONLOAN * DISS
ΔONLOAN * COMP
ΔONLOAN * G-INDEX
ΔONLOAN * CORP CONT
ΔONLOAN * DSHR
Firm FE
Fund Family FE
Observations
Adjusted R-squared
Dependent Variable: % of Votes FOR proposal
All Voting
Voting by Mutual Funds
(1)
(2)
(3)
0.350*
-0.103
-0.795**
(0.218)
(0.333)
(0.322)
0.151
0.711
(0.413)
(0.726)
-19.831***
-49.218***
(1.182)
(0.798)
11.050***
0.133
(0.982)
(0.702)
11.297***
9.984***
(1.409)
(1.514)
9.333***
9.393***
(2.095)
(1.175)
-43.595***
-42.660***
-36.987***
(1.781)
(1.761)
(0.961)
1.545***
1.178***
(0.395)
(0.272)
1.859***
0.466*
(0.445)
(0.304)
1.642***
1.233**
(0.550)
(0.593)
2.131**
0.380
(0.870)
(0.406)
-2.444***
-1.997***
0.162
(0.556)
(0.569)
(0.342)
-2.269**
(0.952)
-0.989
(1.151)
-1.632
(1.429)
-0.644
(1.907)
0.280
0.320
(1.563)
(1.641)
Yes
Yes
Yes
No
No
Yes
6,887
6,887
1,524,290
0.599
0.727
0.791
63
Table 10
Equity Lending Market around Dividend Record Date and the Financial Crisis of 2008
The table presents results from an event study on the effects of proxy voting on the equity lending market in the period (-30,+30) days around
7,415 voting record dates (record date is at t=0) during the 2007-2009 period. The independent variables are equity lending supply, borrowing
demand and borrowing fee. RDATE is a dummy equal to one on the voting record date. GOV41 is the internal governance measure from Aggarwal
et al. (2011). In Panel A we investigate the robustness of results to the inclusion of dividend record dates. DIV DUMMY is a dummy variable equal
to one if the firm has paid a dividend in the past three years. DIV RDATE is a dummy variable equal to one for the 326 dividend record dates in the
window (-1,+1) around proxy voting date. In Panel B we examine the equity lending market post financial crisis. LEHMAN is a dummy equal to
one for all days in 2008 on or after 15th September, and RDATE * LEHMAN is dummy equal to one of the voting record date falls in this period.
Control variables (not shown) include institutional ownership (INST), concentration of institutional ownership (INST CONC), the natural log of
market capitalization (SIZE), book to market (BM), stock turnover (TURNOVER), bid-ask spread (SPREAD), a small firm dummy equal to one if
firm price is less than $5 (PRICE<$5), and a cumulative five day return (RETURN). Dividend record date regressions and financial crisis
regressions include time fixed effects. All regressions include robust standard errors clustered at the firm-level, presented in parentheses. *** (**,*)
indicates significance at the 1% (5%, 10%) level.
Panel A: Dividend Record Date
Panel B: Financial Crisis
Dependent Variable
Dependent Variable
SUPPLY
RDATE
ON LOAN
-1.659***
(0.039)
FEE
0.060***
(0.015)
1.340***
(0.394)
RDATE * LEHMAN
LEHMAN
DIV DUMMY
DIV RDATE
GOV41
Adj. R-squared
1.093***
(0.219)
-1.358***
(0.315)
3.566***
(1.171)
0.67
1.091***
(0.219)
-0.187
(0.325)
3.567***
(1.171)
0.67
0.179
(0.129)
0.596***
(0.212)
-2.355***
(0.692)
0.29
0.179
(0.129)
0.554**
(0.216)
-2.355***
(0.692)
0.29
64
-4.389
(4.902)
5.459
(6.084)
-43.455
(29.524)
0.29
-4.388
(4.902)
4.513
(6.195)
-43.456
(29.524)
0.06
SUPPLY
ON LOAN
FEE
-1.634***
(0.044)
-0.073
(0.048)
-0.698**
(0.347)
0.096***
(0.017)
-0.044*
(0.026)
-1.102***
(0.198)
2.170***
(0.421)
-1.763**
(0.738)
-40.493***
(9.017)
4.210***
(1.182)
0.67
-2.276***
(0.697)
0.28
-49.346*
(29.787)
0.05
Appendix 1
How Does Equity Lending Work?
US Securities $100
US Securities $100
Collateral $102
Beneficial
Owners
Lending
Agent
A fund within one
of the lending
agents (e.g.,
PIMCO Total
Return Fund)
PIMCO
Borrower
Mark-to-Market
Collateral
Dividends & other
entitlements
Dividends & other
entitlements
Cash Investment Vehicle


Borrower leaves collateral with lending agent (e.g. State Street) and pays a fee for the loan.
Lender still receives dividends but loses voting rights; borrower gets voting rights.
65
GS
Appendix 2
Cash Flows on a Securities Loan with Cash Collateral
Settlement date
Term
Security
Security price
Quantity
Loan value
Rebate rate
Collateral
Margin required
Collateral required
Reinvestment rate
Daily lending income
Daily Rebate
June 30th
Open
XYZ Limited
$10.00 per share
100,000 shares
$1,000,000.00
80 basis points
cash
2%
$1,020,000.00
130 basis points
$13.97 ($1,020,000.00 * 0.005 * (1/365))
$22.36 ($1,020,000 * 0.008 * (1/365))
Assumption: No change in value, therefore no change due to daily mark to market, and no change in
terms.
Payments to the borrower:
On July 30th $670.80 ($22.36 * 30 days)
Profit for the lender:
On July 30th $419.10 ($13.97 * 30 days)
Source: Adapted from “An Introduction to Securities Lending,” Spitalfields Advisors Limited, 2006.
66
Appendix 3
Falsification Tests
The dependent variable in column 1 is SUPPLY, percentage of market capitalization available to lend; and in
column 2 the dependent variable is ONLOAN, the percentage of market capitalization actually borrowed. The
explanatory variables are: RDATE is a dummy equal to one on the record dates. Variables include institutional
ownership (INST), concentration of institutional ownership as measured by the Herfindahl index (INST CONC),
the natural log of market capitalization (SIZE), book to market (BM), stock turnover (TURNOVER), bid-ask
spread (SPREAD), a small firm dummy (PRICE<$5), and a governance index (GOV41). Short-Term Mom and
Long-Term Mom are defined as the cumulative returns in the previous 5 and 252 days, respectively. Hedging
Demand is defined as the equal-weighted cumulative return in the past 252 days of related firms with the same
four-digit-GICS industry classification. All regressions include quarterly time-effects and robust standard errors
clustered at the firm-level, presented in parentheses. *** (**,*) indicates significance at the 1% (5%, 10%) level.
Explanatory Variable
RDATE
INST
INST CONC
SIZE
BM
TURNOVER
SPREAD
PRICE < $5
GOV41
Short-Term Mom
Long- Term Mom
Hedging Demand
(1)
(2)
SUPPLY
ONLOAN
-1.623***
(-44.705)
21.173***
(24.094)
-23.497***
(-9.094)
1.146***
(5.389)
0.298**
(1.961)
0.076**
(2.538)
-0.129***
(-3.284)
0.480**
(2.372)
1.680
(0.891)
-0.021***
(0.001)
0.003*
(0.002)
0.243
(0.247)
0.085***
(0.010)
11.885***
(0.609)
-1.438
(1.530)
0.313**
(0.137)
0.096
(0.103)
0.366***
(0.023)
-0.043*
(0.023)
-0.510***
(0.137)
-8.833***
(1.182)
-0.004***
(0.001)
-0.003***
(0.001)
-0.026***
(0.010)
Yes
Yes
Yes
Yes
3,053
0.905
3,053
0.779
Time FE
Firm FE
Number of firms
Adjusted R-squared
67