Proxy Advisory Firms: The Economics of Selling Information to Voters

• Author: ANDREY MALENKO,NADYA MALENKO
• DOI: http://doi.org/10.1111/jofi.12779
• Published date: 01 October 2019
• Date: 01 October 2019

THE JOURNAL OF FINANCE •VOL. LXXIV, NO. 5 •OCTOBER 2019

Proxy Advisory Firms: The Economics of Selling

Information to Voters

ANDREY MALENKO and NADYA MALENKO∗

ABSTRACT

We analyze how proxy advisors, which sell voting recommendations to shareholders,

affect corporate decision-making. If the quality of the advisor’s information is low,

there is overreliance on its recommendations and insufficient private information

production. In contrast, if the advisor’s information is precise, it may be underused

because the advisor rations its recommendations to maximize profits. Overall, the ad-

visor’s presence leads to more informative voting only if its information is sufficiently

precise. We evaluate several proposals on regulating proxy advisors and show that

some suggested policies, such as reducing proxy advisors’ market power or decreasing

litigation pressure, can have negative effects.

PROXY ADVISORY FIRMS PROVIDE SHAREHOLDERS with research and recommenda-

tions on how to cast their votes at shareholder meetings. For diversified institu-

tional investors, the costs of performing independent research on each proposal

in each of their portfolio companies are substantial. The institution may thus

prefer to get information from a proxy advisory firm instead in exchange for

a fee. In recent years, demand for proxy advisory services has increased sub-

stantially due to several factors—a rise in institutional ownership, the 2003

SEC rule requiring mutual funds to vote in their clients’ best interests, and

growth in the volume and complexity of issues voted upon. The largest proxy

advisor, Institutional Shareholder Services (ISS), has over 1,900 institutional

clients and covers about 40,000 meetings around the world. Extant research

provides strong empirical evidence that proxy advisors’ recommendations have

a large impact on voting outcomes.1This influence has attracted the atten-

tion of policy makers and led to a number of proposals to regulate the proxy

advisory industry.

∗Andrey Malenko is at Carroll School of Management, Boston College. Nadya Malenko is at

Carroll School of Management, Boston College. We are grateful to Philip Bond (the Editor), the

Associate Editor, two anonymous referees, Rui Albuquerque, Alessandro Bonatti, Guojun Chen,

Jonathan Cohn, Douglas Diamond, Daniel Ferreira, Thierry Foucault, Itay Goldstein, Leonard

Kostovetsky, Doron Levit, Michelle Lowry, Richmond Mathews, Jonathan Reuter, Suresh Sun-

daresan, Bilge Yilmaz, and participants at multiple seminars and conferences for helpful com-

ments. The authors do not have any conflicts of interest as identified in the Journal of Finance’s

disclosure policy.

1See Alexander et al. (2010), Ertimur, Ferri, and Oesch (2013), Iliev and Lowry (2015), Larcker,

McCall, and Ormazabal (2015), Malenko and Shen (2016), and McCahery, Sautner, and Starks

(2016), among others.

DOI: 10.1111/jofi.12779

2441

2442 The Journal of Finance R

Market participants concerned about the influence of proxy advisors em-

phasize potential deficiencies in their recommendations—a one-size-fits-all ap-

proach, inaccuracies in their data, and conflicts of interest.2Other observers

counter that even if the quality of their recommendations is low, market forces

will ensure efficient information production and aggregation in voting because

“institutional investors are sophisticated market participants that are free to

choose whether and how to employ proxy advisory firms” (Government Ac-

countability Office (GAO) (2016)). According to Nell Minow, a well-known gov-

ernance expert, “what we have is the most sophisticated institutional investors

in the world...making a free market decision to pay for outside, objective

analysis....There couldnot bea better example of market efficiency.”3

In this paper,we emphasize t hat the market efficiency view does not take into

account the collective action problem among shareholders. We show that be-

cause shareholders do not internalize the effect of their actions on other share-

holders, there may be excessive overreliance on proxy advisors’ recommenda-

tions and, as a result, excessive conformity in shareholders’ votes. Moreover,

this problem might not be resolved by improving the quality of proxy advi-

sors’ recommendations.

Overall, the goal of our paper is to provide a simple framework for analyzing

the economics of the proxy advisory industry. We are particularly interested

in understanding how proxy advisors affect the quality of corporate decision-

making and in analyzing the suggested policy proposals. Toward this end, we

build a model of strategic voting in the presence of a proxy advisor.Shareholders

vote on a proposal that can increase or decrease firm value. Each shareholder

can acquire information about the proposal by conducting his own independent

research or by obtaining information from the proxy advisor.4More specifically,

a monopolistic proxy advisor has an informative signal about the proposal. The

advisor sets a fee that maximizes its profits and offers to sell its signal to share-

holders for this fee. Each shareholder then independently decides whether to

buy the advisor’s signal, incur a cost to acquire his own signal, acquire both

signals, or remain uninformed. After observing those signals he acquires, each

shareholder decides how to vote. The proposal is implemented if it is approved

by the majority of shareholders.

2See, for example, Gallagher (2014) and the 2010 SEC Concept Release on the U.S. Proxy System

(available at https://www.sec.gov/rules/concept/2010/34-62495.pdf).

3Harvard Law School Forum on Corporate Governance, March 2, 2018 (avail-

able at https://corpgov.law.harvard.edu/2018/03/02/regulating-proxy-advisors-is-anticompetitive-

counterproductive-and-possibly-unconstitutional/).

4In practice, some institutions have their own proxy research departments, while others rely

strongly on proxy advisors’ recommendations. For example, Iliev and Lowry (2015) show that there

is substantial heterogeneity among mutual funds in the extent to which they rely on ISS, and Iliev,

Kalodimos, and Lowry (2018) show that institutions that conduct more independent governance

research (as measured by their downloads of proxy-related SEC filings) are significantly less likely

to vote with ISS.

Proxy Advisory Firms 2443

Under this framework, the proxy advisor provides a valuable service: an

option to buy and follow an informative signal.5The presence of this option,

however, comes at a cost: it reduces shareholders’ incentives to invest in their

own independent research. If the firm were owned by a single shareholder, the

shareholder would perfectly internalize the effect of his decisions on firm value

and hence would choose between the two sources of information efficiently.

However, the firm is owned by multiple shareholders, leading to a collective ac-

tion problem and inefficiencies in information acquisition. Specifically, a share-

holder who acquires information (privately or from the proxy advisor) imposes

a positive externality on other shareholders by making the vote more informed.

When some other shareholders already follow the proxy advisor, this external-

ity is higher if a shareholder acquires information privately than if he acquires

information from the advisor. This is because when shareholders follow their

private signals, they make independent (or, more generally, imperfectly cor-

related) mistakes. In contrast, when shareholders follow the same signal (the

advisor’s recommendation), their mistakes are perfectly correlated, which in-

creases the probability that an incorrect decision will be made. The collective

action problem may therefore lead to excessive overreliance on the advisor’s

recommendations and crowd out too much private information production.

This trade-off between, on the one hand, providing a new informative signal

and, on the other hand, crowding out independent research and generating

correlated mistakes in votes leads to our main result: the presence of the proxy

advisor increases firm value (the probability of a correct decision being made)

only if the precision of its recommendations is sufficiently high. This result

holds for any fee charged by the proxy advisor, even if this fee is very small.6

The fact that the proxy advisor sets its fee strategically, aiming to max-

imize its own profits rather than the informativeness of voting, creates an

inefficiency of another sort. In our model, when the advisor’s information is

imprecise, firm value would be maximized if the advisor’s recommendations

were made prohibitively costly, so as to maximize shareholders’ incentives to

perform independent research. In contrast, when the advisor’s information is

sufficiently precise, firm value would be maximized if the price of the advi-

sor’s recommendations were made as low as possible. Clearly, neither of these

policies is optimal from the monopolistic advisor’s perspective. When the ad-

visor’s information is imprecise, it charges low fees to induce shareholders to

buy its recommendations. This crowds out independent research and leads to

overreliance on the advisor’s recommendations. In contrast, when the advi-

sor’s information is very precise, this information is underused: to maximize

5For example, Alexander et al. (2010) find that ISS recommendations in proxy contests convey

substantive information about the contribution of dissidents to firm value. Overall, according to

the survey of institutional investors by McCahery,Sautner, and Starks (2016), 55% of respondents

believe that proxy advisors help them make more informed voting decisions.

6Anecdotal evidence suggests that a large fraction of institutions subscribe to at least one

proxy advisor, implying that, in practice, proxy advisory fees are not very high. In addition, proxy

advisors’ recommendations are sometimes made public in high-profile cases. See also the discussion

in Section VI.