The real effects of proxy advisors on the firm

• Date: 01 September 2019
• Author: Paul Calluzzo,Evan Dudley
• Published date: 01 September 2019
• DOI: http://doi.org/10.1111/fima.12251

DOI: 10.1111/fima.12251

ORIGINAL ARTICLE

The real effects of proxy advisors on the firm

Paul Calluzzo Evan Dudley

SmithSchool of Business, Queen's University,

Kingston,Ontario, Canada

Correspondence

EvanDudley, Smith School of Business, Queen's

University,Kingston, ON, Canada.

Email:evan.dudley@queensu.ca

Abstract

This paper examines the influence of proxy advisors (PA) on firm

voting outcomes, policies, and value. We measure PAinfluence with

shareholders'historical propensity to follow PA recommendations.

PA influence increases the impact of PArecommendations on proxy

voting outcomes and firm policies. When shareholders have private

incentivesto engage in costly research in the absence of a proxy advi-

sor,PA influence neither harms nor benefits shareholder value. How-

ever,at firms with dispersed shareholders PA influence can increase

value. Our findings are consistent with theories of voting in which

proxy advisors compete with private information acquisition efforts

by large shareholders.

1INTRODUCTION

Proxy advisors (PAs),such as Institutional Shareholder Services (ISS) and Glass Lewis, make voting recommendations

on issues ranging from director elections to compensation and mergers and acquisitions. These advisors wield consid-

erable influence through their proxy votingrecommendations (Ertimur, Ferri, & Oesch, 2013; Malenko & Shen, 2016),

and shareholders’ reliance on them suggests that PAs’ recommendations are valued bythe firm's investors. However,

despite increasing competition in the market for PAservices (Li, 2018), PAs have been criticized by regulators in the

United States and Europe for potential conflicts of interest, the transparency of their methodologies, and the accuracy

of their recommendations (ESMA, 2013; SEC, 2010).

Empirical evidence regarding the negative effect of PAsis sparse. Larcker, McCall, and Ormazabal (2013) find neg-

ative announcement returns around changes in compensation practices designed to conform to ISS guidelines, but

Ertimur et al. (2013) find that PAs deviatefrom their g uidelinesin some situations, contradicting the one-size-fits-all

approach. Moreover, Alexander, Chen, Seppi, and Spatt (2010) find that PA recommendations haveinformation con-

tent. Specifically,they confirm that when PAs recommend for a dissident shareholder proposal, there is a positive stock

price reaction. Given these findings, the question remains whether PAsare good or bad for the firm. Our paper investi-

gates whether third-party PAsimprove or harm firm value through their influence on the shareholder voting process.

This paper was previouslycirculated under the title “ Who Coordinates Shareholder Votes?The Role of Proxy Advisors in Corporate Governance.” We would

like to thank an anonymousreviewer and Rajkamal Iyer (Editor) for comments that greatly improved the paper. We are also grateful to Anup Agrawal, Sean

Cleary,Fabrizio Ferri, Kristina Minnick, Shastri Sandy, Mike Welker,and seminar participants at Queen's University, the 2015 FMA Orlando Conference, the

2015 Conference on Empirical Legal Studies held at WashingtonUniversity St. Louis, the 2016 Northern Finance Association Meeting, and the 2017 FMA

European Conference for valuable comments. Weacknowledge financial support from the D. I. McLeod Term Research Assistantship and the Social Sciences

andHumanities Research Council (SSHRC).

c

2018 Financial Management Association International

Financial Management. 2019;48:917–943. wileyonlinelibrary.com/journal/fima 917

918 CALLUZZO ANDDUDLEY

Shareholder voting is an important dimension of the corporate governanceprocess. Voting provides aggregation of

private information held by shareholders (Edelman, Thomas, & Thompson, 2014), yet shareholder voting is expensive

and time consuming (Iliev & Lowry, 2015) and institutional voting suffers from collective action problems (Maug &

Rydqvist, 2009). This context suggests that there are circumstances under which third-party recommendations can

either crowd out or complement incentives for the acquisition of private information (Malenko & Malenko,2019). We

use the framework proposed by Malenkoand Malenko (2019) to formulate our hypothesis that relates the magnitude

of PAinfluence to firm value. The intuition for the hypothesis is that the effect of third-party voting recommendations

on the firm depends upon how voters would have voted in the absence of a PA.When ex ante incentives to acquire

private information are present, we argue that introducing a PAcrowds out efforts to acquire a private signal making

voting less informative and harming firm value. However,when ex ante incentives to acquire private information are

low,as they would be in firms whose shareholders hold very small fractions of the shares outstanding, introducing a PA

can be beneficial as long as the PA'sprivate signal is sufficiently informative.

To test our hypothesis, we developa measure of proxy advisor influence (ISSPOWER) defined as the proportion

of each firm's shares that are owned by mutual funds that have historically voted with the recommendations of ISS.

Our empirical strategy proceeds in three steps. First, we examine the effect of PAinfluence on firm voting outcomes

from 2004 to 2016 in management sponsored proposals that receive a negative ISS recommendation. We find that a

one standard deviation increase in ISSPOWER decreases the proportion of votes that align with management by 2.0%

and the pass rate of conflicted proposals by 1.8%. These results demonstrate that PAs havea meaningful impact on

voting outcomes and that ISS voting recommendations are more likely to be followed by investors in firms with high

ISSPOWER ensuring the validity of our measure.

In addition, we present evidence that PA influence is associated with changes in firm policy.We find that high lev-

els of PAinfluence increase the likelihood that management sponsored proposals conform to ISS guidelines. However,

as most proxy votes are nonbinding, it is unclear whether the influence PA's exert on vote outcomes leads to mean-

ingful change at the firm. There is very limited evidence on this question. Ertimur et al. (2013) find that firms engage

with investors and make changes to their compensation plan after negativePA recommendations in say on pay votes,

and Ertimur,Ferri, and Oesch (2018) find that firms respond to negative PA recommendations in director elections by

explicitlyaddressing the underlying concern. Controlling for the proportion of shares held by blockholders, we find that

PAinfluence is associated with increased director responsiveness to PA recommendations, higher executive turnover

following negative recommendations in director elections, and less classified boards following votes to declassify the

board of directors. Together, these findings confirm that PAsinfluence director behavior and executive retention prac-

tices, as well as changes in corporate governancepractices.

Moreover, we investigate whether these changes in voting outcomes and firm policies lead to improvements in

firm value measured with Tobin'sQ.We find evidence that PA influence negatively affects firm value by crowding out

the incentive for private information acquisition in firms with large institutional shareholders, but the strength of this

result varies with the empirical specification. Consistent with our primary hypothesis,PA influence has a positive effect

on firm value in firms with the least concentrated shareholders (i.e., whose fraction of blockholders lies in the bottom

quintile of the sampling distribution). Specifically, a one standard deviation increase in ISSPOWER increases Tobin's

Qby 7.4% in these firms. This finding suggests that when there exists little ex ante incentivesfor private information

acquisition in the absence of a PA,PA recommendations are sufficiently informative that following the PA'srecommen-

dation dominates uninformative voting.

We also consider whether the influence of PAschanges over time. The influence of PAs may evolve with increases

in regulatory scrutiny and as institutional investors increase in sophistication and gain a greater understanding of the

process by which PAsformulate their recommendations. Another reason to expect a change over time is the entrance

ofcompetition i n the marketfor proxy services. Different recommendations on the same proposal by different advisors

may reduce the influence of anysingle PA on the firm. We find that the fraction of fund votes that align with ISS recom-

mendations on conflicted management sponsored proposals increased between 2004 and 2010 and declined between

2011 and 2016 suggesting that greater awareness of voting recommendation practices by institutional investorsdur-

ing the second half of the sample affected their use of third-party voting recommendations. However,we also find that