The Party Structure of Mutual Funds∗

Ryan Bubb New York University

Emiliano Catan New York University

ABSTRACT. We show that a parsimonious spatial model with two dimensions can explain the bulkof mutual fund voting. We estimate the model using a comprehensive dataset of the 5,332,353 votescast on 33,262 proposals from 3,844 portfolio companies from 2010 to 2015 by 3,617 mutual fundsthat in total hold almost 80% of mutual fund industry assets. The two dimensions of funds’ cor-porate governance preferences reflect the role of the two leading proxy advisors. Mutual funds areclustered into three ‘parties’—the Managerialist Party, the Shareholder Intervention Party, and theShareholder Veto Party—that follow distinctive philosophies of corporate governance and share-holders’ role. We use our methodology in turn to construct measures of the individual distributionsof preferences of public companies’ shareholder bases. Our preference measures for mutual fundsand for public companies’ shareholder bases generate a range of insights about the broader systemof corporate governance.

∗We are grateful to Andrew Eggers, Marcel Kahan, Ed Rock, and seminar participants at Oxford University andNew York University for helpful comments and discussions. We thank Nicolas Duque-Franco and Stephanie Thomasfor very able research assistance. Bubb: ryan.bubb@nyu.edu. Catan: emiliano@nyu.edu. First draft: Feb. 14, 2018.This draft: February 16, 2018.

1. INTRODUCTION

To understand corporate governance in the United States, one must understand the voting behavior

of mutual funds. Mutual funds have grown to hold about one-third of publicly traded stock and

are subject to legal duties to vote that stock in the interest of the funds’ investors. Recent years

have also seen a shift in the asset management industry away from actively managed funds to

index funds, as well as increasing concentration of assets in the largest mutual fund families. The

resulting changes in the ownership of public companies have raised concerns about its effects on

corporate governance. Azar, Schmalz, and Tecu (2017), for example, provides evidence that the

growth in common ownership by mutual funds of multiple competitors has had anticompetitive

effects.

In tandem with the growth of mutual funds as corporate shareholders, corporate law and prac-

tice have evolved to elevate the role of the shareholder franchise. Shareholder votes today play an

important role in setting issuer-level corporate governance policies, including through the use of

shareholder proposals to spur governance reforms, and have become an important tool used by in-

stitutional investors to discipline corporate management. Recent legal changes that have enhanced

the role of voting include the Dodd-Frank Act’s requirement that public companies hold regular

advisory votes on executive compensation and the widespread move to majority voting rules for

director elections. Finally, the rise of activist investor campaigns that culminate in a shareholder

vote have made votes by mutual funds increasingly pivotal.

But despite these trends that have made mutual funds central players in corporate governance,

we know relatively little about their behavior as company owners. In this paper we develop the

first systematic account of the structure of mutual fund preferences over corporate governance as

expressed by the proxy votes they cast in their portfolio companies. We do so by analyzing a large

dataset of 5,332,353 votes cast by 3,617 mutual funds from 311 fund families on 33,262 proposals

from 3,844 portfolio companies over 2010 - 2015. The dataset includes the votes by mutual funds

that in aggregate account for almost 80% of assets held by mutual funds. That is, the sample we

1

analyze is close to constituting the population of votes cast by mutual funds over the period.

Our first finding is that mutual fund preferences over corporate governance can be well rep-

resented as positions along two latent dimensions. Shareholders vote on a wide range of issues:

Should Sandra Peterson be reelected to the board of Microsoft in 2014? Should Ford declassify

its board in 2010? Should Ron Burkle’s director nominees be placed on Barnes & Noble’s board

instead of reelecting the incumbent directors? Should director elections at Procter & Gamble be

subject to a majority vote rule or a plurality rule? In this sense, the overarching “issue space”

of these preferences is highly multidimensional. But we hypothesize that these preferences and

attitudes can be organized or represented as positions along a small number of latent dimensions.

Echoing studies of the preferences of legislators expressed through votes in Congress (Poole and

Rosenthal, 2007), we show that a parsimonious spatial model with just two dimensions can explain

the bulk of mutual fund voting. Those two main dimensions, in turn, reflect the role of the two

leading proxy advisors. The relatively low dimensionality we estimate reflects linkages between

issues in the high-dimension issue space as part of a “belief system” (Converse, 1964) of mutual

funds and proxy advisors about corporate governance. As in politics, relatively few fundamental

dimensions underlie mutual fund corporate governance preferences.

We show furthermore that mutual fund preferences are not distributed unimodally across this

preference space but rather are clustered into three groups of funds with similar preferences, much

like political parties in similar studies of Congressional voting. In particular, there are three mutual

fund “parties,” which we label (1) the Managerialist Party, (2) the Shareholder Intervention Party,

and (3) the Shareholder Veto Party. Members of the Managerialist Party—which include the “Big

Three” passive managers BlackRock, Vanguard, and State Street—follow corporate management’s

recommendations at much higher rates than members of the other two parties. We refer to the latter

two parties as the “shareholder rights” parties; each follows a distinctive philosophy of corporate

governance and the role of shareholders. The Shareholder Intervention Party supports shareholder

proposals and proxy contests much more often than the Shareholder Veto Party. These votes are

forms of proactive shareholder engagement and entail shareholders attempting to intervene in the

2

company’s corporate affairs—hence the name “Shareholder Intervention Party.” In contrast, the

Shareholder Veto Party opposes management proposals at a substantially higher rate than mem-

bers of the Shareholder Intervention Party. These proposals entail corporate management asking

shareholders to ratify some management decision—hence the name “Shareholder Veto Party.” One

might expect shareholders’ willingness to intervene proactively and willingness to veto manage-

ment proposals to be positively related, but we show that across the two shareholder rights parties,

they are negatively related. The two parties thus represent distinctive visions of shareholders’ role

in corporate governance. To our knowledge we are the first to recognize and document these two

competing philosophies that drive institutional shareholders’ voting behavior.

We then investigate the factors that shape mutual funds’ preferences by regressing our prefer-

ence measures on fund characteristics that are likely to influence their incentives. Index funds, and

families that specialize in indexing, are more managerialist than actively managed funds. Even

more interestingly, we also find that funds that follow a growth stock strategy are more manageri-

alist than value funds. This is consistent with the idea that the investment thesis for most growth

stocks includes belief in the company’s current management. We also show that larger funds and

fund families are more managerialist.

Finally, we combine our estimates of the preferences of the mutual funds together with data

about mutual funds’ portfolios to construct, for each portfolio company and every year between

2002 and 2016, a share-weighted measure of the distribution of the preferences of the mutual fund

shareholders of the company as of the relevant year. We provide summary descriptions of those

distributions and characterize their cross-sectional and time-series features. We find that both the

dispersion across companies in the mean preferences of their shareholder bases and the degree

of shareholder rights orientation of companies’ shareholder bases are decreasing in companies’

market capitalization. We also find evidence that growth stock companies (i.e., companies with

lower book-to-market ratios) have systematically more managerialist mutual funds as their share-

holders. Turning to time trends, we show that public companies’ shareholder bases have become

systematically more managerialist over time (especially after 2007) along both of our dimensions

3

of shareholder preferences. In work in progress, Bubb and Catan (2018), we extend the method-

ology developed in this paper to characterize the preferences of companies’ broader shareholder

base—including non-mutual fund shareholders—and study more generally the determinants and

consequences of those preferences.

Our main contribution is to systematically measure and characterize the corporate governance

preferences of mutual funds and of the shareholder bases of issuers and to use those measures to

generate new insights about the system of corporate governance more broadly. Importantly, we

start from a statistical model that is devoid of any institutional or other assumptions about the

structure of this preference space—from the perspective of our preference estimation methodol-

ogy, there is nothing unique about ISS, Glass Lewis and management as voters—and we recover

the result that the first and second most significant latent dimensions along which mutual funds’

preferences differ track the corporate governance philosophies of the two main proxy advisors.

Our methodology enables us to construct a parsimonious measure of fund-level preferences and,

in turn, an issuer-level measure of the distributions of preferences of the mutual fund shareholder

bases across public companies.

We hope the “spatial map” of mutual fund corporate governance preferences we provide will

serve as a useful field guide to scholars and practitioners of corporate governance, helping to

reveal important patterns and trends. In addition to descriptive insights, the shareholder preference

measures we introduce to the literature might also enable new quantitative tests of theories and

hypotheses in corporate governance. We consider our results on the determinants of mutual fund

preferences and on the cross-sectional patterns in the preferences of public companies’ shareholder

bases to be an initial proof of concept in that regard.

Our overall findings are broadly consistent with a burgeoning literature on the voting behav-

ior of institutional investors (Matvos and Ostrovsky, 2010; Choi, Fisch, and Kahan, 2013; Appel,

Gormley, and Keim, 2016; Brav, Jiang, and Li, 2017) and the influential role of proxy advisors

(Choi, Fisch, and Kahan, 2010; Ertimur, Ferri, and Oesch, 2013; Iliev and Lowry, 2014; Malenko

and Shen, 2016). In contemporaneous and independent work, Bolton, Li, Ravina, and Rosenthal

4

(2018) also estimate a spatial model of voting by institutional investors. Their preference estima-

tion methodology and data are different from ours. In particular, they estimate a 1-dimensional

model using W-NOMINATE, a tool developed in political science to study Congressional voting.

They take the mutual fund family as the “voter” and include “votes” of 211 mutual fund families

in their analysis sample. In contrast, we use 3,617 individual mutual funds (from 311 fund fam-

ilies) as the unit of analysis because, legally, funds and not families vote shares, and in several

major fund families (e.g., Fidelity) there are substantial differences in voting across funds within

the family. They include votes from a single fiscal year, and exclude director election proposals,

yielding only 3,318 proposals in their analysis sample, an order of magnitude fewer than in our

sample. They also do not consider the amount of assets owned by each family when characterizing

the distribution of preferences.

As a combined result of these differences they reach quite different conclusions from ours. In

contrast to our finding of a party structure in mutual fund voting, Bolton, Li, Ravina, and Rosenthal

(2018) emphasize that their distribution of ideal points is close to unimodal, distinguishing it from

the bimodal distribution of preferences in Congress that follows political parties. According to

their 1-dimensional map, Glass-Lewis sits near the mode of the distribution, next to Vanguard and

BlackRock. This is probably due in part to their exclusion of director election proposals, resulting

in their model not picking up some of the characteristic behavior of the Shareholder Veto Party

that distinguishes it from the Managerialist Party.

More fundamentally, Bolton, Li, Ravina, and Rosenthal (2018) interpret their estimated mu-

tual fund preference space in terms of investors’ degree of social orientation versus profit-seeking,

writing: “The left represents relatively socially-oriented investors, while the right represents Man-

agement recommendations and exclusively profit-oriented investors” (p. 16). Our approach results

in a very different understanding of the preference space. We show that two latent dimensions un-

derlie heterogeneity in preferences among the profit-seeking investors that hold the vast majority

of mutual fund assets, and that those two dimensions measure the extent to which funds adopt two

competing modes of shareholder engagement vis-a-vis management.

5

Finally, they do not relate their preference measure to fund characteristics or construct measures

of the preferences of the shareholder bases across public companies, as we do. An important

strength of Bolton, Li, Ravina, and Rosenthal (2018) relative to this paper, however, is that they

include data on the votes of 40 public pension funds and place them in the same 1-dimensional

preference space as mutual funds.

The plan of the paper is as follows. In Section 2 we lay out our formal spatial model of discrete

choice and our empirical methodology. In Section 3 we describe our data before turning in Section

4 to describing our estimates of the corporate governance preferences of mutual funds. Section 5

examines the behavior of the three distinct mutual fund parties we identify. In Section 6 we use

our preference measures to test various theories about the determinants of mutal funds’ corporate

governance preferences. In Section 7 we use our mutual fund preference measures to construct

measures of the preferences of public companies’ shareholder bases and examine patterns and

trends in those preferences. Section 8 concludes.

2. A SPATIAL MODEL OF MUTUAL FUND PREFERENCES

Corporate shareholders such as mutual funds vote on a range of issues, including in the election

of directors and on various corporate governance policy issues. We suppose that mutual funds’

preferences over such votes can be represented using a standard spatial model of discrete choice

in which the deterministic component of a mutual fund’s utility from each proposal outcome is a

decreasing function of the distance between the mutual fund’s ideal point in some preference space

and the location of that proposal outcome in that space.

2.1. Setup. Suppose in particular that there are n mutual funds voting on m shareholder propos-

als. Each proposal j = 1, ...,m represents a choice between a “Yes” outcome and a “No” outcome,

Ojy, Ojn ∈ Rd, where d denotes the number of dimensions of the preference space. Each mutual

fund i = 1, ..., n has an “ideal point,” xi ∈ Rd, in that same space. Utility for mutual fund i given

the outcome of proposal j is:

6

Ui(Ojy) = −h(‖xi −Ojy‖) + εijy, (1)

and,

Ui(Ojn) = −h(‖xi −Ojn‖) + εijn, (2)

where h(·) is a strictly increasing loss function, and εijy and εijn are random shocks. Fund i’s vote

on proposal j is then given by,

yij = 1⇔ Ui(Ojy) > Ui(Ojn). (3)

2.2. Empirical framework. There are many different empirical models for estimating spatial

models of this form from discrete choice data. The most influential in political science is the

NOMINATE family of parametric models (Poole and Rosenthal, 1991, 2007). Other models in-

clude Poole (2000)’s nonparametric optimal classification method and Bayesian approaches (Clin-

ton, Jackman, and Rivers, 2004). Finally, Heckman and Snyder (1997) develop a linear probability

model approach to estimating a spatial model of preferences over discrete choices in settings in

which the analyst does not observe the attributes of the choices that drive decisions.

We adopt the linear framework of Heckman and Snyder (1997) and estimate their model using

principal components analysis (PCA). Let Y be the n x m matrix of votes of n funds on m pro-

posals. Using the singular value decomposition (SVD) of Y we can write Y = ULA′, where the

“weight matrix” A contains the eigenvectors of Y ′Y , L contains the squareroots of the eigenvalues

of Y ′Y , and U is a scaled version of the principal component “scores” of the funds Z = UL. To

estimate a d-dimensional spatial model, we take the first d columns of Z as the coordinates of the

ideal points xi of the funds.

Let ak denote the k-th column of the weight matrix A. By definition, each principal component

zk = a′kY is calculated by choosing the coefficient vector ak in order to maximize the sample

variance of zk, subject to the normalization constraint a′1a1 = 1 and the requirement that zk be un-

7

correlated with the prior principal components. This is one sense in which PCA can be understood

as a data reduction technique. As we discuss in more detail below, the analysis sample covers

votes on 33,262 proposals. Each of those proposals is an outcome variable, the votes on which

contain information about the corporate governance preferences of mutual funds. But in their full

unreduced glory, the sheer number of variables swamps attempts to characterize systematic differ-

ences in preferences across mutual funds. PCA finds the linear transformation of the original data

matrix Y of m variables (votes on proposals) into d < m variables (principal components) that are

uncorrelated and retain the maximal amount of the variation in the original data as is possible.

PCA can also be understood from a prediction standpoint. Starting from the SVD, Y = ULA′,

this can be written element-by-element as, yij =∑r

k=1 uikl1/2k ajk, where r is the rank of Y . Define:

ydij =d∑

k=1

uikl1/2k ajk, (4)

which is an approximation of yij using only the first d principal components. It can be shown that

ydij minimizes,

n∑i=1

m∑j=1

(ydij − yij)2, (5)

where ydij is any rank d approximation of yij .

2.3. Missing data. A challenge to performing PCA posed by our data is that the vote matrix is

sparse. Because most mutual funds own only a small fraction of public companies, each mutual

fund votes on only a small fraction of proposals in the sample. In particular, 95.6% of the cells

in the vote matrix are empty. Let oij = 1 if yij is observed and = 0 otherwise. PCA can be

generalized to this setting by redefining our cost function in (5) as:

n∑i=1

m∑j=1

oij(ydij − yij)2, (6)

that is, the squared prediction errors over the observed votes. There is no analytical solution to the

8

problem of choosing the rank d approximation that minimizes this cost function; instead, iterative

numerical learning techniques must be used. We follow the iterative impuation method analyzed

in Ilin and Raiko (2010). It proceeds by first imputing the missing values in Y using the row-wise

means of Y to calculate a complete data matrix Yc. The principal components of this complete

data matrix are then calculated and used to reimpute the missing votes as ydij =∑d

k=1 uikl1/2k ajk.

The principal components of this new complete data matrix using those imputed values are then

calculated, and the algorithm iterates between imputation and principal components calculation

until convergence.

3. DATA

Our mutual fund voting data is from ISS Voting Analytics, which is drawn from public filings

by mutal funds on Form N-PX. Our sample period is from 2010 - 2015. For a proposal to be

included in the analysis sample, we require that at least 20 mutual funds vote on it and that 8% of

mutual funds that vote on it be in the minority. The reason is that highly lopsided votes contain

little information about the relative preferences of mutual funds. To see the intuition, consider the

extreme case of a unanimous vote—unanimous votes contain no information about mutual funds’

relative preferences. Thus we require that there be at least a minimal amount of controversy among

mutual funds about a proposal for the proposal to be included. This excludes many uncontested

director election votes for which the nominee received overwhelming votes in favor, as well as

other routine votes, like the bulk of votes to ratify the company’s auditors. Finally, for a fund to be

included in the sample, it must have voted on at least 200 sample proposals.

The resulting analysis sample covers votes on 33,262 proposals from 3,844 portfolio compa-

nies. Table 1 provides counts of proposal types in each year. The prefixes “MP” and “SP” in

the proposal categories refer to management proposals and shareholder proposals, respectively.

Proposals to elect directors nominated by management are by far the most common type of pro-

posal. The second most frequent proposal category is management proposals related to executive

compensation, the bulk of which are say-on-pay proposals or proposals to approve or amend the

9

company’s stock compensation plan. Shareholder proposals are less numerous and mostly focused

on corporate governance issues rather than corporate social responsibility. Table 2 further breaks

down the shareholder proposals on corporate governance into more specific corporate governance

issue categories.

The analysis sample includes 3,617 mutual funds from 311 fund families. We also include as

“voters” in the data matrix rows for management, ISS, and Glass Lewis based on their respective

recommendations. This enables us to place these actors in the same preference space as the mutual

funds, which aids in interpretation of the model. Including these three actors as voters in the data

matrix has a negligible effect on our estimates; all results are robust to excluding them.

For fund characteristics, we merge in data from CRSP’s mutual fund database, using ticker,

fund name, and family name as well as data from EDGAR to link the two datasets. Table 3

compares the CRSP population of mutual funds from 2010 - 2015 holding U.S. common stock to

the CRSP funds that we were able to match to a fund in the ISS Voting Analytics data that was

included in our analysis sample. The merged sample includes about half of the CRSP population

in each year, and almost 80% of the value of U.S. common stock held by mutual funds in CRSP.

Tables 4 and 5 provide summary statistics for the CRSP population and for our merged sample,

respectively. They show that the funds in the merged sample are larger than funds in the CRSP

population of funds and moreover that index funds are disproportionately represented.

With 3,620 voters (3,617 funds plus management, ISS, and Glass Lewis) and 33,262 proposals,

there are a total of 120,408,440 potential votes in the analysis sample and therefore cells in our

data matrix defined by one row for each voter and one column for each proposal. The median

fund, however, owns a total of only 552 unique portfolio companies over the sample period, and

as a consequence there are only 5,332,353 votes in the analysis sample. In other words, 95.6% of

the cells in the data matrix are empty.

10

4. THE DIMENSIONS OF MUTUAL FUND PREFERENCE

4.1. The number of dimensions. An initial question is how many dimensions of mutual fund

preference are needed to provide a good model of mutual fund preferences. The eigenvalues of

each principal component provide one perspective on the issue. The eigenvalue of the k-th princi-

pal component measures the variance in the voting data along that dimension. Figure 1 plots the

eigenvalues of the first thirty principal components. Note that starting with the third component,

the plot becomes linear. A widely used rule-of-thumb is to include the principal components up to

the first component in the linear portion of the plot (Joliffe, 2002, pp. 116-117). Here, that is the

third component. But while the first two components have fairly straightforward interpretations,

as discussed below, the third principal component has no obvious interpretation. In what follows,

we thus focus on the first two dimensions as a parsimonious model of mutual fund preference.

Table 6 provides the classification percentage (CP) and average proportional reduction in error

(APRE) for models using 1 - 10 dimensions. The CP is simply the percentage of votes that the

model predicts correctly, where a predicted value ydij > 0.5 is taken to predict a “Yes” vote, and

ydij < 0.5 is a predicted “No” vote. APRE measures the reduction in error the model achieves in

classifying votes relative to a simple benchmark model of predicting that all funds vote with the

majority on the proposal.1 A two-dimensional model performs well, correctly classifying 87% of

the votes, with an APRE of 49%.

4.2. The interpretation of the dimensions. We turn now to the interpretation of the dimensions

of mutual fund preference. Figure 2 shows the estimated preferences of mutual funds, with funds’

scores on the first dimension on the horizontal axis and their scores on the second dimension on

the vertical axis. Each dot is a mutual fund. Also depicted with triangles are the preferences at

the mutual fund family level, calculated as the average of the family’s funds’ preferences, for a

subset of the mutual fund families in the data. For the mutual fund families labeled in the figure,

1For each proposal, the proportional reduction in error (PRE) is equal to Number Minority Votes−Number Classification ErrorsNumber Minority Votes .

The APRE sums over all of the proposals:∑m

j=1 Number Minority Votesj−Number Classification Errorsj∑mj=1 Number Minority Votesj

.

11

we also color code the family average and each of the family’s mutual funds, to give some sense

of the dispersion within fund family. We also show the univariate histograms of funds’ locations

on dimensions 1 and 2 in Figure 3.

Note first the locations of management, ISS, and Glass Lewis. Management is in the lower left

of the figure, with low scores on both dimensions, ISS is in the lower right, and Glass Lewis is in

the upper left. These locations reflect our basic interpretation of the two dimensions. Dimension 1

tracks the extent to which mutual funds vote against management along the lines that ISS generally

recommends. Similarly, dimension 2 measures the extent to which mutual funds vote against

management along the lines that Glass Lewis recommends.

To further corroborate this interpretation, we turn to how different types of proposals load on

each of these two dimensions. The estimated dimension 1 coefficients for the proposals (denoted

a1) indicate each proposal’s contribution to a fund’s score on dimension 1—a “Yes” vote on a

proposal adds that proposal’s a1 coefficient to the fund’s score on dimension 1. Hence examin-

ing the pattern of coefficients across proposals can help interpret the substantive content of the

dimensions. The top panel of Figure 4 plots the histograms of the a1 coefficients among proposals

supported by management (all management proposals as well as a few shareholder proposals) and

among proposals opposed by management (most shareholder proposals). While the a1 coefficients

on proposals opposed by management are overwhelmingly positively signed, similar fractions of

proposals supported by management have positive and negative coefficients.

The magnitudes of the a1 coefficients are further broken down in the top panel of Figure 5,

which plots the fraction of proposals in each proposal category that are “Extremely Positive,”

“Non-Extreme,” and “Extremely Negative.” We define as “extreme” coefficients that have an ab-

solute value in the top quartile of the distribution of a1 coefficients’ absolute values. While most

of the extreme coefficients for management proposals are negative, there are a substantial number

that are positive, and vice-versa for shareholder proposals.

Once we further break down the proposals by ISS’s recommendation, the interpretation of

dimension 1 comes into focus. The top panel of Figure 6 gives the distribution of extreme a1

12

coefficients by proposal category among proposals that ISS recommends against, and the bottom

panel for those ISS supports. Virtually all of the extremely negative (positive) coefficients are on

proposals ISS opposes (supports). Interestingly, the one exception to this is for proxy contests, for

which there are a few extreme coefficients that go against the ISS recommendation.

This analysis establishes that the first dimension of mutual fund preference measures the degree

to which funds vote in line with ISS’s recommendations. This interpretation is also reflected in the

location of ISS on the first dimension. ISS’s position is determined just like any other voter in

the model by multiplying its “votes” (i.e., recommendations) by the coefficient matrix. As shown

in the histogram of funds’ scores on dimension 1 in Figure 3, ISS is on the extreme positive side

of dimension 1. Moreover, the distribution is bimodal: there are a substantial number of funds

aligned closely with ISS in the right mode and a substantial number of funds in the left mode on

the opposite side of dimension 1.

Turning to the second dimension, the bottom panels of Figures 4 and 5 show the distribution

of a2 coefficients by management recommendation and proposal category, respectively, and Figure

7 further breaks down the extreme coefficients by Glass Lewis recommendation. We are ham-

pered by the fact that we have data on Glass Lewis’s recommendations for only 15,328 out of

33,262 total proposals. But Glass Lewis’s recommendation generally organizes the a2 coefficients

well. Virtually all of the extremely negative coefficients are on proposals Glass Lewis opposes

and management supports. Similarly, almost all of the extremely positive coefficients are on pro-

posals Glass Lewis supports and management opposes. Furthermore, in the histogram of mutual

funds’ scores on dimension 2 in Figure 3, Glass Lewis is on the extreme positive side, opposite

management on the extreme negative side.

In sum, our spatial model shows that mutual fund preferences can be well-represented as posi-

tions along two latent dimensions, one that tracks the recommendations of ISS and one that tracks

Glass Lewis’s views. It is noteworthy that these two dimensions are orthogonal. That is, while

one might imagine that management, ISS, and Glass Lewis sit on a single dimension that ranges

from an extreme managerialist view on one end to an extreme shareholder rights view on the other,

13

with ISS and Glass Lewis ordered according to the intensity of their shareholder rights views, a

better representation of mutual fund preferences is that ISS and Glass Lewis’s views reflect two

orthogonal dimensions of shareholder preference. A fund can be extreme on dimension 1 without

being extreme on dimension 2, and vice-versa.

The fact that dimensions 1 and 2 of mutual fund preferences track the views of the two leading

proxy advisors could reflect either or both of two different phenomena: (1) the proxy advisors’

recommendations could have a causal effect on mutual fund choice; and (2) the proxy advisors’

recommendations might track existing patterns in mutual fund preference. In either case, a remain-

ing question is what the substantive content of the two dimensions are in terms of specific views on

corporate governance. Comparing the top and bottom panels of Figure 5, the patterns of loadings

for the two dimensions are broadly similar, generally loading negatively on management proposals

and positively on shareholder proposals. In this sense, both dimensions measure a vision of “share-

holder rights,” and the patterns of coefficients reveal that, for both dimensions, the higher a fund

scores, the more the fund opposes management and supports shareholder proposals. We explore

in the next section the differences between the visions of corporate governance and shareholders’

role captured by these two dimensions by examining how clusters of funds that are extreme on just

one of the dimensions vote their shares.

4.3. Stability over time. So far we have used voting data from the entire sample period 2010 -

2015 to estimate a single ideal point for each fund. A natural question is whether the structure of

preferences we estimate is stable over the period. To investigate this, we divide the sample into

three two-year cohorts and estimate the spatial model separately for each cohort. The resulting

fund ideal points for each cohort are plotted in Figure 8. The basic configuration of preferences,

including the locations of management, Glass Lewis, and ISS, is remarkably stable over time.

In each of the cohorts, the first and second dimensions measure the same tendencies to vote in

accordance with the corporate governance philosophies of the two major proxy advisors. The

locations of major fund families are stable over time as well, showing that the latent dimensions

14

of corporate governance preferences that we estimate are stable, and giving us more confidence in

our results from the pooled sample.

Comparing the 2010 - 2011 cohort to the 2014 - 2015 cohorts, however, reveals modest but

potentially important shifts of two of the largest fund families. BlackRock starts out the sample

period aligned with management on dimension 1 but in the middle of funds’ distribution on di-

mension 2. But by 2014 - 2015, BlackRock had fallen to the far negative side of dimension 2 and

is located at the heart of the cluster of funds located near management. BlackRock’s corporate

governance philosophy has thus evolved from a significant willingness to oppose management in

accordance with the view of shareholder rights measured on dimension 2 to a more strictly man-

agerialist approach by the end. In contrast, State Street has steadily drifted along dimension 1

away from management, becoming more shareholder rights oriented over time. Neither of these

two shifts is dramatic, but given the sizes of the two fund families, they are noteworthy.

5. THE PARTY STRUCTURE OF MUTUAL FUNDS

One might expect mutual funds’ preferences to be distributed unimodally, and perhaps normally,

across their two-dimensional preference space. But the contour plot of the density of mutual fund

preferences (weighted by the TNA of each fund) in Figure 9 reveals that there are in fact three dis-

tinct modes, or clusters, of mutual funds, located near management, ISS, and Glass Lewis. Mutual

funds’ voting behavior thus reveals the existence of three distinct mutual fund “parties”: groups of

funds that have similar preferences about corporate governance and generally vote together. For

reasons we explain below, we label these three parties (1) the Managerialist Party, (2) the Share-

holder Intervention (SI) Party, and (3) the Shareholder Veto (SV) Party, respectively. We refer to

the latter two parties as the “shareholder rights” parties. We depict our definitions of these three

parties in Figure 10 by superimposing on the scatter plot of fund preferences circles that define

each party’s membership. We define the Managerialist Party as funds within a distance of 30 from

the center of the cluster of mutual funds in the lower-left. The SI Party and SV Party are defined

15

as funds within a distance of 30 from the locations of ISS and Glass Lewis, respectively.2

5.1. Party coherence. Table 7 provides measures of the degree to which party members vote

together. For each proposal, we calculate the outcome voted for by a majority of each party’s

members. We then report the fraction of party members’ votes in each proposal category that

were cast in the opposite direction of the party’s majority. The column labeled “All” reports the

corresponding fractions for all mutual funds considered as a single party, which serves as a useful

benchmark. Considering mutual funds as a whole, 25% of mutual fund votes in the sample are cast

in the opposite way from how a majority of mutual funds voted on the proposal. In contrast, for all

proposals, only 16% of the votes of Managerialist Party members were cast against the majority of

the party, and the corresponding figures for the SI Party and SV Party are 4% and 8%, respectively.

Hence the two shareholder rights parties exhibit substantially more coherence in their voting than

the Managerialist Party, a result consistent with those parties having a clear focal point in their

proxy advisor’s recommendations to coordinate their votes. The degree of coherence varies across

proposal types. For example, the proposal type with the highest fraction of SV Party members

voting in their party’s minority is shareholder proposals on corporate governance, at 17%. The

proposal types with the highest fraction of SI Party members voting in their party’s minority, in

contrast, are management proposals related to mergers and acquisitions and shareholder social

proposals.

5.2. Party corporate governance philosophies. Consider now the substantive visions of corpo-

rate governance that animate each of the three parties. Table 8 provides the fraction of “Yes” votes

cast by each party by proposal category. In interpreting these figures, it is important to keep in

mind that for a proposal to make it into the sample, it must pass our “lopsidedness” threshold by

having at least 8% of the mutual funds voting on the proposal vote in the minority among mu-

tual funds. Relative to the entire population of proposals, this sample selection criterion generally

drives down support for management proposals and drives up support for shareholder proposals.

2Our characterizations of party voting behavior below are all robust to using smaller radii to define these parties.

16

As befits its name, mutual funds in the Managerialist Party support management proposals, and

oppose shareholder proposals, at far higher rates than do members of the other two parties. To be

clear, Managerialist Party members do not always support management. Across all management

proposals, members of the Managerialist Party vote in favor 82% of the time, and similarly they

vote in favor of shareholder proposals (and hence generally against management’s wishes) 19% of

the time. In comparison, these rates across all mutual funds are 70% and 47%, respectively.

The two shareholder rights parties, by contrast, vote against management at much higher rates.

Each shareholder rights party represents a distinctive vision of corporate governance and share-

holders’ role. The two key comparisons that capture their basic difference is (1) between their

overall rates of support for management proposals; and (2) between their overall rates of support

for shareholder proposals. The SI Party supports shareholder proposals at a rate of 84%, com-

pared to only 50% for the SV Party. These shareholder proposals generally entail shareholders

attempting to affirmatively intervene in the corporate affairs of the company by demanding that the

company’s board of directors adopt changes to corporate policies and practices, hence the name

“Shareholder Intervention Party.” In contrast, the SV Party supports management proposals at a

rate of only 50%, compared to 62% support for the SI Party. Put differently, the SV Party opposes

management proposals at a rate of 50%, which is 32% more often than the SI Party’s rate of 38%.

Management proposals generally entail corporate management asking for shareholders to ratify

some decision by the board, such as management’s pay practices, an increase in authorized com-

mon stock, a management nominee for director, and the like. A vote against these proposals is thus

a sort of “veto” by the shareholder of a management initiated decision, hence the name “Share-

holder Veto Party” for the shareholder rights party that tries to veto management at relatively high

rates.

This basic distinction between the two shareholder rights parties is reinforced by examination

of specific subcategories of management proposals and shareholder proposals. For all but two

categories of management proposals, the SV Party supports them at lower rates than the SI Party,

and the two exceptions prove the rule. One exception is management nominees for director in

17

a proxy contest in which a shareholder dissident has nominated a competitive slate of directors.

Such proxy contests are perhaps the most interventionist shareholder actions possible, and hence

it is consistent with our conceptual distinction between the two sharehoder rights parties that the

SI Party opposes the management nominees (and supports the shareholder nominees) at greater

rates than the SV Party. Similarly, the other exception is management proposals on mergers and

acquisitions. These are among the most consequential for the corporation, since failure to receive

shareholder ratification generally means a major transformative transaction cannot go forward.

Votes against these are thus arguably more “interventionist” than other categories of management

proposals, like voting against management’s say-on-pay proposal. Similarly, for every subcategory

of shareholder proposal, the SI Party supports the proposals at greater rates than does the SV Party.

Table 9 further breaks down the shareholder proposals on corporate governance into specific types

of proposals. In 8 out of the 9 corporate governance proposal types with a substantial difference

between the two parties, the SI Party supports the proposals at a greater rate than does the SV Party.

The two shareholder rights parties thus represent distinctive philosophies of shareholders’ role

in corporate governance. The Shareholder Veto Party focuses on monitoring the management of the

corporation and critically considering management’s proposed courses of actions, voting against

them when the shareholder views them as problematic. In contrast, the Shareholder Intervention

Party focuses instead on actively intervening in corporate affairs by demanding reforms to basic

corporate governance rules and supporting proxy contests for corporate control. Importantly, while

one might expect shareholders’ willingness to affirmatively intervene and their inclination to veto

management’s proposed actions and policies to be positively related, we find that the SV Party

intervenes less than the SI Party, while the SI Party vetos less than the SV Party. So among

members of the two shareholder rights parties, these two modes of shareholder engagement are

negatively related.

These two competing visions of corporate governance are in turn reflected in the recommen-

dations of ISS and Glass Lewis. Recall that the SI Party is centered on ISS’s preference location

and the SV Party is centered on Glass Lewis’s preference location. The extent to which these two

18

competing visions of corporate governance among institutional investors predate ISS and Glass

Lewis, or instead were brought about by the recommendations of ISS and Glass Lewis, remains

an open question. In either case, one theory that might explain why ISS and Glass Lewis follow

these two distinctive approaches is based on product differentiation. ISS and Glass Lewis compete

in the market for proxy advice. Given some preexisting cleavages among institutional investors

in their corporate governance philosophies, competition between the two may have led to a prod-

uct differentiation and market segmentation that has resulted in the reinforcement or even creation

of the two shareholder rights parties and their distinctive approaches to superintending corporate

management.

Finally, note that there are a few funds in the preference plot in Figure 2 that score highly on

both dimensions of fund preference. These are funds that often vote against management when ei-

ther ISS or Glass Lewis recommends against management. We have labeled the fund families with

average scores in the upper right of the preference space—note that they are socially responsible

fund management companies, like Domini and Calvert. Our framework shows that these socially

responsible fund families are extreme in their shareholder rights orientation, as expressed through

their votes.

5.3. Characteristics and sizes of the parties’ memberships. A final set of questions about the

party structure of mutual funds concerns the characteristics of their memberships and their size in

terms of total net assets. Table 10 provides average characteristics across the three parties. The

primary contrast is between the Managerialist Party on the one hand and the two shareholder rights

parties on the other. Funds in the Managerialist Party are more heavily indexed, far larger by assets,

come from larger fund families, and charge substantially lower expense ratios. Another intriguing

difference is that funds in the Managerialist Party are more heavily weighted toward growth stock

strategies than value stock strategies in comparison to the two shareholder rights parties. We return

to this pattern in Sections 6 and 7 below. The final row in the table gives the fraction of TNA in

the mutual fund industry held by funds in each of the parties as of 2015. The Managerialist Party

19

is much larger than the others, at 68% of mutual fund industry TNA, followed by the SI Party at

11% and the SV Party at 5%

5.4. Mapping the Votes on Specific Proposals. To build further intuition about the window into

mutual fund voting preferences our framework provides and to illustrate how it works, in this

section we discuss four specific proposals using our spatial model to explain the patterns in funds’

votes. Figures 11 - 14 depict the spatial maps of votes for four different proposals. In each figure,

we plot the “Yes” and “No” votes cast by each fund at the location of the fund’s ideal point, using

different marker styles to identify “Yes” and “No” votes according to whether the model correctly

predicted the fund’s vote. Each point is labeled using the first two letters of the fund’s family name,

to aid in identifying funds (e.g., Vanguard fund votes are labeled “Va”).

5.4.1. Dissident nominee for director at DuPont. The first proposal vote plotted in Figure 11 is

for a candidate for director at DuPont nominated by the activist hedge fund Trian Partners in a

2015 proxy contest. Trian had nominated four candidates for DuPont’s twelve member board,

but despite winning the backing of most non-index institutions, its candidates narrowly lost to

management’s nominees when the “Big Three” indexers, Vanguard, BlackRock, and State Street,

all sided with management. In our data, 72% of funds (ignoring the number of shares each held)

voted in favor of the Trian nominee. Following the outcome of the vote, it was reported that had

any of the Big Three backed Trian, its nominees would have won.3

Examining the vote plot for this proposal, it is clear that the cutting line that separates the

model’s predictions of “Yes” versus “No” cuts diagonally through the two clusters in the lower left

of the diagram between the incorrectly predicted “Yes” and “No” votes. As illustrated in the figure,

Trian won the support of not only the bulk of the Shareholder Intervention Party, but also most of

the Shareholder Veto Party and a substantial fraction of the upper-right half of the Managerialist

Party. Interestingly, the model predicted that State Street funds would vote “Yes”–they are among

the cloud of incorrectly predicted “No” votes in the upper-right portion of the Managerialist Party.

3See Reuters, “DuPont wins board proxy fight against activist investor Peltz,” May 13, 2015.

20

The model suggests that this was indeed a close call, with State Street in particular the key swing

vote.

Overall the model predicts correctly 78% of the votes on this proposal and achieves a propor-

tional reduction in error (PRE) of 24%. So while the model certainly helps organize and explain

the voting patterns on the proposal, it performs less well in predicting votes in this proxy contest

than it does on average. (Recall the overall percentage of votes classified correctly across the entire

sample of proposals is 87%, with an APRE of 49%.)

5.4.2. Shareholder proposal on executive compensation at Waste Management, Inc. The share-

holder proposal at issue in the vote plotted in Figure 12 was submitted by International Brotherhood

of Teamsters General Fund at the 2015 annual shareholder meeting of Waste Management, Inc. It

requested that the board of directors adopt a policy under which it would limit the acceleration of

the vesting of equity awards for Waste Management executives, noting that under the company’s

current practices its CEO would receive accelerated vesting of stock worth $18.2 million upon a

change of control. 44% of funds in our data voted for the proposal.

The vote plot reveals that this proposal attracted the votes of the Shareholder Intervention Party,

but not those of the Shareholder Veto Party or Managerialist Party. Interestingly, State Street again

bucked our model’s predictions: the cloud of incorrectly predicted “Yes” votes on the right side

of the Managerialist Party are State Street funds. But overall, the model performs very well in

predicting funds’ votes on this proposal, correctly predicting 91% of the votes and achieving a

PRE of 78%.

5.4.3. Management say-on-pay proposal at PolyOne Corporation. In Figure 13 is plotted a say-

on-pay proposal at PolyOne Corporation from its 2015 annual shareholder meeting. 83% of funds

in our data voted for the proposal. As predicted by the model, the proposal attracted the support

of most of both the Managerialist Party and the Shareholder Intervention Party, and was opposed

by most of the Shareholder Veto Party. Overall the model correctly predicts 91% of the votes and

achieves a PRE of 47%

21

5.4.4. Uncontested director election at Avnet, Inc. Finally, in Figure 14 is plotted an uncontested

director election at Avnet Inc. from 2011. 85% of funds in our data voted for the proposal. Again,

the model organizes funds’ votes well, with the SV Party opposed but the Managerialist Party and

the SI Party in favor. The cutting line separates the incorrectly predicted “Yes” and “No” votes

toward the upper left of the preference space, and the model mispredicts only the funds that are

close to that cutting line. Overall the model correctly predicts 94% of the votes and achieves a PRE

of 61%.

6. THE DETERMINANTS OF MUTUAL FUND PREFERENCE

What factors shape mutual funds’ preferences and underlie the heterogeneity we have documented?

6.1. Theory. As a theoretical matter, mutual fund managers’ incentives to become informed and

exercise the funds’ voting rights in a meaningful way are relatively weak. One reason is the classic

collective action problem that besets shareholder voting generally. Because (1) shareholders bear

all of the costs of becoming informed but share the benefits of informed voting with other share-

holders; and (2) the probability that any public company shareholder’s votes are pivotal is so small,

shareholders have little incentive to invest in voting.

By aggregating share capital, mutual funds might seem to mitigate these problems. But these

intermediaries are subject to their own agency problem vis-a-vis the ultimate beneficial owners who

invest in them (Rock, 1991; Black, 1992). In particular, any improvement in portfolio company

value produced by a mutual fund exerting effort to vote in an informed manner is shared with all

of the mutual fund competitors who hold the company’s stock. For index funds, this means that

costly voting effort increases their costs without improving their returns relative to competitors.

One caveat to this conclusion is that, since index funds (unlike actively managed funds) cannot

sell when they are displeased with management, their inability to exit might lead them to resort

more to “voice.” But even so, the basic logic above that effective voice is costly without producing

competitive benefits would seem, as a theoretical matter, overpowering.

22

In contrast, the managers of actively managed funds face somewhat stronger voting incentives

than those of index funds (Kahan and Rock, 2007; Shapiro-Lund, 2017). An actively managed

fund manager can improve the fund’s return relative to competitors by taking actions that increase

the value of a stock in which the fund is overweight relative to the market. Moreover, the process

of investment selection might generate information that is useful for voting.

Relatedly, we hypothesize that different approaches to security selection might result in dif-

ferent views about corporate governance. Active managers who follow growth stock strategies

attempt to identify companies whose rich market valuation is more than justified by their growth

prospects. That suggests that the fund manager will believe in the management of the company,

leading the fund manager toward a view of corporate governance generally in line with the Man-

agerialist Party’s view (i.e., to score low on each dimension). In contrast, a value stock strategy

might include investing in stocks whose undervaluation is driven in part by mistakes of the current

management, and that investing philosophy might lead naturally to a view of corporate governance

in line with the views of one of the shareholder rights parties.

The size of a mutual fund and of a fund family might also shape their corporate governance

preferences. In particular, larger funds and fund families benefit from economies of scale in voting

and moreover have a higher probability of being pivotal voters. The largest fund families often

own more than 5% stakes in large public companies, making them potentially crucial swing vot-

ers (Fichtner, Heemskerk, and Garcia-Bernardo, 2017). How this translates into their corporate

governance preferences is not obvious. We offer two competing hypotheses. First, it may be that

the path of least resistance (i.e., the lowest cost approach to voting) is to simply generally follow

management’s recommendations. Under this account, economies of scale in voting might make it

cost effective to invest in information that could lead the fund to vote against management, at least

for the actively managed funds that might receive a competitive benefit from votes that increase

firm value. In this case, we might expect large funds and families to have higher scores (i.e., to

be less managerialist) than smaller funds. But note that this hypothesis is at odds with the basic

comparisons of fund characteristics across the three parties in Table 10.

23

Alternatively, the path of least resistance for small funds might be to simply follow one of the

two proxy advisors. Economies of scale might thus result in large funds and fund families moving

away from simply outsourcing voting decisions to proxy advisors. Indeed, Iliev and Lowry (2014)

show that, among actively managed funds, larger funds and larger fund families are less likely to

vote in line with ISS’s recommendations. Under this view, larger funds and larger fund families

will score lower on dimensions 1 and 2.

Finally, consider the relationship between funds’ expense ratios and their preferences. One

natural perspective is that because it is costly to become informed about the issues being voted on,

the funds that do so will have higher expense ratios. But theoretically, this leads to the same set of

ambiguous predictions for their preferences as does fund size—it turns on what the “path of least

resistance” for fund voting is.

6.2. Empirical approach. Ultimately whether variation in incentives faced by fund managers

shapes their corporate governance preferences as reflected in how they vote portfolio company

shares is an empirical question. We view one contribution of this paper as the creation of parsi-

monious measures of shareholder preference that allow for testing various theories and hypotheses

in corporate governance, like the ones considered above. In this section we illustrate the potential

value of our measures of shareholder preference by using them to investigate whether institutional

investors’ incentives matter for their voting behavior.

The univariate comparisons between characteristics of members of the three mutual fund par-

ties in Table 10 discussed above suggest that fund and family size and indexing are associated with

mutual fund corporate governance preferences. But because these fund characteristics covary—for

example, index funds are on average much larger than actively managed funds—multiple regres-

sion can help tease out partial correlations that are more informative about the underlying process

of preference formation.

We report regressions of mutual funds’ scores on dimensions 1 and 2 on fund-level and family-

level characteristics in Tables 11 and 12, respectively, with standard errors clustered at the family

24

level. Because ultimately we are trying to characterize mutual funds’ votes, and those votes are a

direct linear function of assets (since dollars buy shares which equal votes), we use weighted OLS

regressions, weighting each fund observation by its TNA.

Column (1) of Tables 11 and 12 reports a regression of funds’ score on dimension 1 on mea-

sures of the funds’ / families’ management style (index, growth, or value) and quintiles of TNA.

Column (2) adds a measure of the fund’s / family’s expense ratio to the regression, which is miss-

ing for about half of sample funds, and columns (3) and (4) repeat the two regressions using funds’

score on dimension 2 as the dependent variable.

6.3. Results.

6.3.1. Fund management style. There is strong evidence that indexing leads to lower scores on

the two dimensions. Focusing on the full sample regressions in columns (1) and (3) of Table 11, the

coefficient on Index fund is negative for both dimensions but statistically significant only for Score

2. The models using family-level covariates in Table 12 show that indexing has a large and statis-

tically significant negative relationship with both Score 1 and Score 2. Indeed, the absolute value

of the coefficients on fraction of family assets indexed are much larger than for the coefficients on

Index fund in the fund-level covariate regressions. This suggests that the role of indexing in shap-

ing mutual fund preferences is largely a family-level, as opposed to fund-level, phenomenon. This

makes economic sense: for families that specialize in index funds, like Vanguard, BlackRock, and

State Street, there would be little reason to delegate voting decisions down to the fund manager

level. The activities of index fund managers would generate little useful information for voting

the funds’ shares. Similarly, for families that specialize in active management, while it may make

sense to allow some individual managers for active funds to direct their funds’ votes separately

from the fund family as a whole, the voting of shares held by the family’s index funds is likely to

be centralized and therefore follow the general approach of the fund family.

There is also evidence for our hypothesis that growth stock investment strategies lead to more

managerialist voting behavior relative to value stock investment strategies. The coefficient on

25

Growth fund is below the coefficient on Value fund for both Score 1 and Score 2 in Table 11, but

this difference is statistically signficant only for Score 2 (F-test p-value = .007). The results using

family-level measures of growth versus value strategies in Table 12 are similar, with the difference

between the two coefficients statistically significant only for Score 2 (F-test p-value = 0.041).

6.3.2. Fund and family size. There is strong evidence that fund size is negatively related to scores

on both dimensions. The coefficients on quintiles of fund TNA in Table 11 are monotonically

decreasing in fund size for both dimensions. Importantly, note that this is controlling for being an

index fund versus an actively managed fund. Big funds are more managerialist.

The relationship between family size and preferences is more complicated. The coefficients in

column (1) of Table 12 show an initially increasing relationship between Score 1 and family TNA

up to the 3rd quintile of TNA. The coefficients then fall thereafter, with the coefficient on the 5th

quintile statistically significantly different from the coefficient on the 4th quintile. The pattern is

similar for Score 2 in column (3).

6.3.3. Expense ratios. Finally, the coefficients on expense ratio reported in columns (2) and (4)

of the two tables are consistently positive but only statistically significant in the fund-level model

for Score 2. Note again that this is controlling for being an index fund, which have lower expense

ratios than actively managed funds.

7. THE PREFERENCES OF PUBLIC COMPANIES’ MUTUAL FUND

SHAREHOLDER BASES

We have documented systematic variation in mutual funds’ preferences and developed a simple

taxonomy of mutual fund corporate governance preferences, organizing them into three “parties.”

We turn now to companies’ shareholder bases. Each public company has a particular set of mutual

fund and other shareholders and hence a particular distribution of corporate governance prefer-

ences among its shareholder base. Because mutual funds’ portfolios vary, the distribution of mu-

26

tual fund preferences varies across public companies’ shareholder bases. In this section we use

our preference measures, combined with data on mutual funds’ portfolios, to construct issuer-level

measures of the preferences of these shareholder bases. We then use measures to describe sys-

tematic patterns across issuers in their shareholders’ preferences. An important caveat is that our

estimated distributions of shareholder preferences pertain only to the mutual fund shareholders. In

work in progress (Bubb and Catan, 2018) we extend the methodology developed in this paper to

characterize the preferences of companies’ broader shareholder base—including non-mutual fund

shareholders—and study the determinants and consequences of those preferences. Our hope is that

our measures of the preferences of public companies’ shareholder bases will be a useful addition

to the corporate governance literature and enable the testing of a range of theories and hypotheses

about the system of corporate governance.

7.1. Methodology. For each year-end between 2002 and 2016, we construct the distribution of

the preferences of mutual fund shareholders of each portfolio company as follows. Starting from

the Thomson Reuters Mutual Fund Holdings database, we identify, for the last quarter of the year,

the portfolio owned by each fund (identified through the fund’s fundno) during that quarter.4 Each

portfolio contains a list of each of the securities owned by the fund, together with an indication of

the number of such securities that the fund owned. After excluding securities that do not corre-

spond to equity securities in operating companies,5 for each public company held by mutual funds

in the sample in each year we construct the distribution of preferences across shares owned by mu-

tual funds. To characterize trends and patterns across these distributions, we focus on the average

share’s preference among shares in the company owned by mutual funds for each company-year

as a simple summary measure. To be specific, we calculate the weighted average of the scores of

mutual fund shareholders of the company on each dimension, weighting by the number of shares

owned by each mutal fund. We will refer to these measures of average preference as “company

4Our algorithm for estimating mutual fund preferences identifies mutual funds through their CRSP crsp_portno.We link a Thomson Reuters’ fundno with its corresponding crsp_portno as of the relevant date using WRDS’MFLINKS table.

5For example, a mutual fund might hold an interest in an exchange traded fund.

27

shareholder scores.”6 Note that because our preference estimates are static—a single ideal point

is estimated for each mutual fund based on the votes they cast from 2010 - 2015—the variation

over time in the company-level distributions of mutual fund preference is all driven by changes in

mutual funds’ portfolios over time.

With this procedure, we are able to construct estimates of the distribution of mutual fund share-

holder preference scores of company shares owned by mutal funds for a total of 44,286 company-

year’s. Table 13 provides for each year basic information about the set of companies for which

we were able to estimate mutual fund shareholder preference distributions. At the beginning of

the sample in 2002, we can estimate preference distributions for 2,827 public companies, and the

number rises to 3,023 by the end of the sample period in 2016. The average percentage of shares

of these companies that are owned by mutual funds rises from 19% in 2002 to 26% by 2016, con-

sistent with the growth of the mutual fund industry over this period. In contrast, the mutual funds

for which we have preference estimates own only 11% of the sample companies’ shares in 2002,

rising to 23% by 2016. Our preference measures’ coverage of mutual fund shareholders improves

over time because we estimated mutual funds’ preferences using voting data from 2010 - 2015,

and turnover in the fund industry means that there are funds in the earlier years that do not appear

in our voting data. The average number of mutual fund shareholders with estimated ideal points

for sample companies increases from 59 in 2002 to 156 in 2009 before falling to 122 by 2016. The

fall may be due to increasing concentration in the fund industry over this period.

7.2. Results. As a preliminary matter, it is useful to simply compare the distribution of company

shareholder scores to the distribution of mutual funds’ scores on dimensions 1 and 2. Figures 16

and 17 plot the company shareholder scores as of 2002 and 2016, respectively, over a contour plot

of the (unweighted) density of the mutual fund scores reported in Section 4. Company shareholder

scores are relatively less disperse than their fund-level counterparts. This is intutive, as a company’s

shareholder score is simply a convex combination of the scores of the funds that owned shares in

6For the sake of brevity, the fact that (1) the measures pertain only to the preferences of the mutual fund sharehold-ers and (2) that the measure summarizes the distribution of the preferences of those shareholders through the mean ofthat distribution will be implied.

28

the company. Nonetheless, there is a substantial amount of dispersion in the company shareholder

scores. In short: not all public companies have the same distribution of preferences among their

shareholders. This is not surprising, but to our knowledge we are the first to document the degree

of dispersion across companies in the preferences of their shareholder bases.

Company shareholder scores have several interesting cross-sectional features. First, as shown

in Figure 15, mean company shareholder scores, and the dispersion of those scores across compa-

nies, are decreasing in companies’ market capitalization. Both of these features are likely due to the

fact that index funds (which tend to have more managerialist preferences) own a larger proportion

of the shares of larger companies. The figure reveals, however, that there remains considerable vari-

ation in company shareholder scores even among the largest companies by market-capitalization.

Second, as we discussed in Section 6, there are reasons to expect investors that follow different

stock-picking strategies to also follow systematically different approaches to corporate governance.

In particular, the investment thesis for growth stocks generally entails expectations of future growth

that more than justifies the stocks’ rich valuation, and it is natural to think those expectations

correlate with faith in management and hence more managerialist voting behavior. To investigate

this, Figure 18 plots the average company shareholder score in each year by quintile of book-

to-market ratio, which is generally used to distinguish growth stocks from value stocks.7 For

almost every year, there is a monotonically-increasing relationship between the book-to-market

ratio and the average company shareholder score. This suggests that, on average, the mutual fund

shareholders of growth stock companies are systematically more managerialist than the mutual

fund shareholders of value stock companies.

We confirm this pattern using regression analysis reported in Table 14. The regressions in

columns (1) and (3) show that, controlling for year fixed effects, firms with higher book-to-market

ratios have systematically higher company shareholder scores.8 In columns (2) and (4) we ver-

ify the robustness of this finding by using dummies for quintiles 2-5 of the yearly distribution of

7We construct the book-to-market ratio following the algorithm described in Davis, Fama, and French (2000).8Since the distribution of the book-to-market ratio features a small number of extreme outliers, for the sake of

regression analysis we winsorize that ratio at the 1% and 99% values of the distribution.

29

book-to-market (quintile one is the omitted category).9 Moving from a firm in the bottom quintile

of book-to-market to a firm in the top quintile of book-to-market is associated with an average

increase of 5.49 (respectively, 4.86) units in company shareholder score on dimension 1 (respec-

tively, dimension 2). These changes amount to about one third and almost 40 per cent of the

standard deviation of the company shareholder score on dimensions 1 and 2, respectively.

We turn now to time trends in the preferences of public companies’ shareholder bases. The pe-

riod from 2002 - 2016 was marked by a shift of assets from actively managed funds to index funds

and increasing concentration in the fund industry. Figure 19 shows the resulting trends in average

company shareholder scores for sample companies in this period. After a steep drop in average

company shareholder scores on both dimensions between 2007 and 2008, both averages continue

to drop (albeit at a less steep rate) over 2008-2016. Between 2007 and 2016, the average company

shareholder score on dimension 1 drops from -3.87 to -12.87, while the corresponding average on

dimension 2 drops from -5.48 to -14.03. These drops amount to 40 percent and 60 percent of one

standard deviation in the distribution of the company shareholder scores for dimension 1 and 2,

respectively, as of 2007.

To gain further insight into how the distributions of company shareholder scores evolved in

this period, Figure 20 plots estimates of the density of company shareholder scores in 2002, 2007,

2012, and 2016. For both dimensions, the distributions become substantially less right-skewed and

more symmetric over the sample period, losing mass to the right of the mode (in line with the drop

in the evolution of the cross-company averages reported in Figure 19). In addition, the modes of

the distributions also become systematically higher over time. Finally, in the case of dimension 2,

the entire distribution shifts systematically to the left over time.

Figure 21 breaks down the aggregate time trend in company shareholder scores by company

size category. For each year we partition the sample into four market-capitalization buckets using

as cutoffs the 90th, 70th, and 30th percentiles of the distribution of market capitalization as of the

9For the specifications estimated in columns (2) and (4), for every pair of adjacent book-to-market quintiles, anF-test rejects the null hypothesis that the average company shareholder score in one quintile is the same as in theadjacent one (in every case, the p-values for the test are smaller than 0.01).

30

relevant year.10 For almost every year there is a monotonically decreasing relationship between

company size and the company shareholder scores on both dimensions. While the average com-

pany shareholder scores on dimension 1 across micro-caps and small-caps dropped quite steeply

during the 2007-2016 period, the corresponding scores for mid-caps and large-caps did not change

meaningfully. By comparison, for all groups of firms, company shareholder scores on dimension

2 dropped, on average, over 2008-2017. In fact, the average magnitude of the drop in the score

on dimension 2 experienced by micro-, small-, and mid-cap firms is comparable across the entire

period (while the time-series for the large-caps is hump-shaped).

8. CONCLUSION

In this paper we have systematically characterized the corporate governance preferences of mutual

funds. We show that a spatial model with just two latent dimensions of preference provides a highly

predictive model of mutual fund voting behavior. Those two dimensions in turn reflect the role of

the two leading proxy advisors and measure two different philosophies of corporate governance

and shareholders’ role.

Our parsimonious measures of mutual funds’ corporate governance preferences generate a

number of descriptive insights about the broader system of corporate governance and moreover

enable the quantitative testing of various hypotheses. In particular, we show that mutual funds

are clustered into three parties, a Managerialist Party and two shareholder rights parties. The two

shareholder rights parties follow two distinctive modes of shareholder engagement. The Share-

holder Intervention Party supports efforts to actively intervene in corporate affairs, including by

supporting shareholder-initiated reforms to basic corporate governance rules and activist investor

proxy contests for board seats. In contrast, the Shareholder Veto Party focuses on monitoring

corporate management and vetoing management’s proposed courses of actions when they raise

10For example, as of the end of 2016, the smallest Small-Cap company has a market value of equity of approximately$365M, the smallest Mid-Cap company has a market value of equity of approximately is $2566M, and smallest Large-Cap company has a market value of equity of approximately $11500M. In unreported analysis, we corroborate that thetime-series for our coverage variable is similar across market-cap buckets.

31

concerns. We show furthermore that funds’ preferences are shaped by fund characteristics that

influence their incentives. Index funds, growth stock funds, and larger funds are all more manage-

rialist than actively managed funds, value stock funds, and smaller funds, respectively.

We use our measures of mutual funds’ preferences to construct measures of the preferences

of public companies’ mutual fund shareholder bases. We show that, with the growth of large

passive management fund families, the shareholder bases of public companies have become more

managerialist over time. Moreover, we show that the shareholders of growth stock companies are

more managerialist than shareholders of value stock companies, consistent with the hypothesis that

the investment thesis for most growth stocks includes belief in the company’s current management.

We hope the introduction of our measures of mutual fund corporate governance preferences to the

literature will enable other researchers to test quantitatively a range of theories and hypotheses

about corporate governance.

REFERENCES

APPEL, I. R., T. A. GORMLEY, AND D. B. KEIM (2016): “Passive investors, not passive owners,”Journal of Financial Economics, 121(1), 111–141.

AZAR, J., M. C. SCHMALZ, AND I. TECU (2017): “Anti-Competitive Effects of Common Own-ership,” Journal of Finance.

BLACK, B. S. (1992): “Institutional investors and corporate governance: The case for institutionalvoice,” Journal of applied corporate finance, 5(3), 19–32.

BOLTON, P., T. LI, E. RAVINA, AND H. ROSENTHAL (2018): “Investor Ideology,” Unpublishedworking paper.

BRAV, A., W. JIANG, AND T. LI (2017): “Picking Friends Before Picking (Proxy) Fights: HowMutual Fund Voting Shapes Proxy Contests,” .

BUBB, R., AND E. CATAN (2018): “The Preferences of Companies’ Shareholder Bases,” Unpub-lished working paper.

CHOI, S., J. FISCH, AND M. KAHAN (2010): “The Power of Proxy Advisors: Myth or Reality?,”Emory Law Journal, 59(4), 869.

(2013): “Who calls the shots: How mutual funds vote on director elections,” Harv. Bus.L. Rev., 3, 35.

32

CLINTON, J., S. JACKMAN, AND D. RIVERS (2004): “The statistical analysis of roll call data,”American Political Science Review, 98(2), 355–370.

CONVERSE, P. E. (1964): “The Nature of Belief Systems in Mass Publics,” in Ideology and ItsDiscontents, ed. by D. E. Apter. New York: The Free Press of Clencoe.

DAVIS, J. L., E. F. FAMA, AND K. R. FRENCH (2000): “Characteristics, covariances, and averagereturns: 1929 to 1997,” The Journal of Finance, 55(1), 389–406.

ERTIMUR, Y., F. FERRI, AND D. OESCH (2013): “Shareholder votes and proxy advisors: Evi-dence from say on pay,” Journal of Accounting Research, 51(5), 951–996.

FICHTNER, J., E. M. HEEMSKERK, AND J. GARCIA-BERNARDO (2017): “Hidden power of theBig Three? Passive index funds, re-concentration of corporate ownership, and new financialrisk,” Business and Politics, 19(2), 298–326.

HECKMAN, J. J., AND J. M. SNYDER (1997): “Linear Probability Models of the Demand forAttributes with an Empirical Application to Estimating the Preferences of Legislators,” RANDJournal of Economics, pp. S142–S189.

ILIEV, P., AND M. LOWRY (2014): “Are mutual funds active voters?,” The Review of FinancialStudies, 28(2), 446–485.

ILIN, A., AND T. RAIKO (2010): “Practical approaches to principal component analysis in thepresence of missing values,” Journal of Machine Learning Research, 11(Jul), 1957–2000.

JOLIFFE, I. (2002): Principal Component Analysis. Springer-Verlag New York.

KAHAN, M., AND E. B. ROCK (2007): “Hedge funds in corporate governance and corporatecontrol,” University of Pennsylvania Law Review, pp. 1021–1093.

MALENKO, N., AND Y. SHEN (2016): “The role of proxy advisory firms: Evidence from aregression-discontinuity design,” The Review of Financial Studies, 29(12), 3394–3427.

MATVOS, G., AND M. OSTROVSKY (2010): “Heterogeneity and peer effects in mutual fund proxyvoting,” Journal of Financial Economics, 98(1), 90–112.

POOLE, K. T. (2000): “Nonparametric unfolding of binary choice data,” Political Analysis, 8(3),211–237.

POOLE, K. T., AND H. ROSENTHAL (1991): “Patterns of congressional voting,” American journalof political science, pp. 228–278.

POOLE, K. T., AND H. ROSENTHAL (2007): Ideology and Congress. Oxford University Press.

ROCK, E. B. (1991): “The Logic and (Uncertain) Significance of Institutional Shareholder Ac-tivism„” Georgetown Law Journal, 79, 445.

SHAPIRO-LUND, D. (2017): “The Case Against Passive Shareholder Voting,” Unpublished work-ing paper.

33

APPENDIX

Table 1: Proposal Categories by Meeting Year

2010 2011 2012 2013 2014 2015 totalMP-Compensation 729 1360 1075 1108 1077 761 6110MP-Corporate Finance 114 91 75 76 55 51 462MP-Corporate Governance 43 35 30 45 33 63 249MP-Elect Director (Contested) 9 54 45 64 95 91 358MP-Elect Directors 4029 3891 4050 3498 3463 3018 21949MP-Merger / Acquisition Related 102 113 91 170 149 99 724MP-Other 119 75 64 60 63 35 416MP-Ratify Auditors 40 54 47 38 36 42 257SP-Corporate Governance 207 188 182 172 181 251 1181SP-Elect / Remove Directors (Contested) 6 47 45 103 71 66 338SP-Executive/Director Compensation 108 33 61 82 64 70 418SP-Other 42 9 13 18 8 6 96SP-Social Proposal 94 102 91 121 159 137 704

Table 2: Corporate Governance Shareholder Proposal Categories by Meeting Year

2010 2011 2012 2013 2014 2015 totalSP-Proxy Access 0 0 6 7 10 84 107SP-Independent Chairman/Lead Director 40 28 55 59 63 62 307SP-Written Consent 18 33 19 28 28 36 162SP-Special Meetings 45 30 17 11 14 21 138SP-Declassify Board 16 4 4 9 7 4 44SP-Reduce Supermajority Requirements 14 14 16 18 9 9 80SP-Majority Vote for Directors 33 36 24 25 17 2 137SP-Cumulative Voting 20 27 20 2 6 2 77SP-Eliminate Dual Class Shares 0 4 4 4 3 6 21SP-Subject Pill to Shareholder Approval/ 3 0 5 2 4 2 16SP-Other Corporate Governance Proposals 18 12 12 7 20 23 92

34

Table 3: CRSP Coverage

Year Number ofCRSP Funds

Number ofMerged Funds

Fraction TNA CRSP TNA Merged Fraction

2010 5,763 3,224 0.56 5,252,852 4,080,021 0.782011 5,970 3,326 0.56 4,990,577 3,890,089 0.782012 5,833 3,357 0.58 5,586,311 4,356,507 0.782013 6,024 3,342 0.55 7,652,467 5,992,014 0.782014 6,207 3,312 0.53 8,603,009 6,725,080 0.782015 6,420 3,237 0.50 8,551,900 6,723,619 0.79Notes: Number of CRSP Funds is the number of unique crsp_portno’s in CRSP in the respective year that holdU.S. common stock. Number of Merged Funds is the number of such crsp_portno’s that were merged with afundid from ISS Voting Analytics for which we estimated an ideal point. TNA CRSP is the sum of total netassets invested in common stock for crsp_portno’s in CRSP that hold U.S. cmmon stock. TNA Merged is thesum of total net assets invested in common stock for crsp_portno’s in CRSP that were merged with a fundidfrom ISS Voting Analytics for which we estimated an ideal point. All TNA in $ millions.

35

Table 4: CRSP Summary Statistics

N Mean St. Dev. 25th %-tile median 75th %-tileTNA ($ millions) 8143 882 5284 17 98 458Index fund 8143 0.12 0.33 0 0 0Growth fund 8143 0.23 0.42 0 0 0Value fund 8143 0.11 0.31 0 0 0Exp. ratio (bps) 3478 104 84 65 101 134Number funds in family 8143 95 107 20 70 123Family TNA ($ millions) 8143 98954 229798 2122 18894 59342Frac. family indexed 8143 0.16 0.28 0 0.021 0.15Frac. family growth 8143 0.26 0.22 0.083 0.24 0.39Frac. family value 8143 0.12 0.15 0.017 0.08 0.16Family avg. exp. ratio (bps) 7930 70 62 37 67 98Notes: Sample is all crsp_portno’s that appear in the CRSP Mutual Fund dataset from 2010 - 2015 that helddomestic equity. Time-varying variables are created by averaging over the years the fund is in the sample.TNA is total net assets in common stock. Index fund, Growth fund, and Value fund are indicators for funds thatCRSP identifies as "pure" index funds, growth funds, and value funds, respectively. Exp. ratio is the expenseratio of the fund. Family TNA is total net assets in domestic common stock held by mutual funds in the fund’sfamily. Frac. family indexed / growth / value are the fractions of common stock held by funds in the fund’sfamily that are in funds in the respective category. Family avg. exp. ratio is the average expense ratio acrossfunds in the fund’s family that own domestic equity, weighted by TNA of each fund. Number funds in familyis the number of crsp_portno’s in CRSP for the management company that manages the fund.

Table 5: Merged Sample Summary Statistics

N Mean St. Dev. 25th %-tile median 75th %-tileTNA ($ millions) 3611 1519 7394 67 258 857Index fund 3611 0.17 0.38 0 0 0Growth fund 3611 0.27 0.44 0 0 1Value fund 3611 0.15 0.35 0 0 0Exp. ratio (bps) 1761 98 46 65 97 127Number funds in family 3611 115 113 40 85 132Family TNA ($ millions) 3611 128937 259684 9009 29643 80759Frac. family indexed 3611 0.19 0.29 0 0.058 0.17Frac. family growth 3611 0.28 0.21 0.13 0.28 0.39Frac. family value 3611 0.12 0.13 0.046 0.091 0.17Family avg. exp. ratio (bps) 3574 65 38 37 66 93Number of votes 3611 1452 2592 377 641 1326Notes: Sample is all crsp_portno’s that appear in the CRSP Mutual Fund dataset from 2010 - 2015 that weremerged with a fundid from ISS Voting Analytics for which we estimated an ideal point. Common variables areas defined in notes to Table 4. Number of votes is the number of proposals voted on by the fund used to estimateits ideal point.

36

Table 6: Goodness of Fit by Number of Dimensions

# of dims CP APRE1 0.84 0.362 0.87 0.493 0.90 0.584 0.91 0.655 0.93 0.706 0.94 0.747 0.94 0.778 0.95 0.819 0.96 0.83

10 0.96 0.85

Table 7: Fraction of Votes Cast in the Party’s Minority

proptype All Man. Party SI Party SV PartyMP-Compensation 0.24 0.14 0.05 0.04MP-Corporate Finance 0.22 0.24 0.04 0.09MP-Corporate Governance 0.30 0.22 0.06 0.08MP-Elect Director (Contested) 0.29 0.29 0.04 0.09MP-Elect Directors 0.23 0.16 0.03 0.08MP-Merger / Acquisition Related 0.15 0.18 0.10 0.02MP-Other 0.17 0.25 0.03 0.02MP-Ratify Auditors 0.13 0.05 0.01 0.12SP-Corporate Governance 0.32 0.19 0.06 0.17SP-Elect / Remove Directors (Contested) 0.28 0.28 0.06 0.08SP-Executive/Director Compensation 0.37 0.10 0.05 0.05SP-Other 0.33 0.16 0.07 0.07SP-Social Proposal 0.37 0.08 0.13 0.08All 0.25 0.16 0.04 0.08Notes: For each party we determine for each proposal the majority vote among members of the party.We then calculate the fraction of party member votes that are cast against the party’s majority.

37

Table 8: Fraction of Yes Votes by Proposal Category, Party

All Man. Party SI Party SV PartyMP-Compensation 0.68 0.83 0.59 0.46MP-Corporate Finance 0.68 0.72 0.67 0.52MP-Corporate Governance 0.57 0.72 0.44 0.42MP-Elect Director (Contested) 0.40 0.61 0.20 0.59MP-Elect Directors 0.72 0.83 0.64 0.52MP-Merger / Acquisition Related 0.83 0.80 0.81 0.92MP-Other 0.21 0.33 0.08 0.07MP-Ratify Auditors 0.85 0.95 0.93 0.16All MP 0.70 0.82 0.62 0.50SP-Corporate Governance 0.53 0.26 0.82 0.70SP-Elect / Remove Directors (Contested) 0.54 0.36 0.71 0.36SP-Executive/Director Compensation 0.44 0.10 0.94 0.34SP-Other 0.42 0.18 0.84 0.26SP-Social Proposal 0.39 0.09 0.83 0.31All SP 0.47 0.19 0.84 0.50

Table 9: Fraction of Yes Votes by Corporate Governance Proposal Category, Party

All Man. Party SI Party SV PartySP-Proxy Access 0.69 0.46 0.92 0.64SP-Independent Chairman/Lead Director 0.35 0.13 0.55 0.64SP-Written Consent 0.55 0.15 0.91 0.89SP-Special Meetings 0.57 0.23 0.96 0.61SP-Declassify Board 0.89 0.83 0.96 0.81SP-Reduce Supermajority Requirements 0.83 0.68 0.94 0.94SP-Executive/Director Compensation 0.41 0.09 0.94 0.26SP-Majority Vote for Directors 0.71 0.42 0.96 0.97SP-Cumulative Voting 0.36 0.10 0.85 0.25SP-Eliminate Dual Class Shares 0.88 0.75 0.99 0.95SP-Subject Pill to Shareholder Approval/Redeem Pill 0.85 0.70 0.98 0.99SP-Other Corporate Governance Proposals 0.43 0.14 0.84 0.48

38

Table 10: Fund Characteristics by Party

Man. Party SI Party SV PartyFraction Index 0.34 0.13 0.15

TNA ($ millions) 2815 599 1064Family TNA ($ millions) 274326 34284 66865

Expense Ratio (bps) 45 79 80Fraction Growth 0.31 0.26 0.17

Fraction Value 0.1 0.15 0.22Fraction MF TNA in Party 2015 0.68 0.11 0.054Notes: Index is the fraction of funds (by assets) in the party that are index funds; TNAis the average total net assets in domestic equity of funds in the party; Family TNA isthe average total net assets of the fund family of the funds in the party; Expense Ratiois the weighted average expense ratio of funds in the party in basis points (weightedby the fund’s TNA). Fraction Growth and Fraction Value are the fraction of funds (byassets) in the party that are growth funds and value funds, respectively.

Table 11: The Determinants of Mutual Fund Preference: Fund-level Covariates

Score 1 Score 2

(1) (2) (3) (4)

Index fund −8.638 −11.698 −16.955∗∗ 2.211(7.046) (15.275) (7.281) (4.820)

Growth fund −3.504 −10.159∗∗ −5.215∗ −6.492∗

(2.921) (4.232) (3.007) (3.323)Value fund −1.224 −1.028 7.008∗∗ 8.832∗∗

(3.467) (5.080) (2.916) (3.447)Exp. ratio (bps) 0.079 0.232∗∗∗

(0.144) (0.089)2nd quintile TNA −5.797 −2.174 −1.054 2.560

(5.816) (6.769) (2.338) (3.716)3rd quintile TNA −7.518 0.399 −3.787 1.819

(6.143) (8.073) (2.622) (4.111)4th quintile TNA −13.353∗∗ −8.415 −6.719∗∗ −3.234

(6.325) (8.924) (3.302) (4.718)5th quintile TNA −26.448∗∗∗ −19.461∗ −17.697∗∗∗ −11.834∗∗

(6.872) (10.107) (4.628) (4.621)Constant 11.582∗ 2.765 7.435∗∗ −24.117∗∗

(6.862) (21.417) (3.523) (12.228)

Observations 3,611 1,761 3,611 1,761R2 0.055 0.125 0.122 0.252

Note: ∗p<0.1; ∗∗p<0.05; ∗∗∗p<0.01Regression is weighted using the funds’ TNA.

Standard errors are clustered at the fund family level.

39

Table 12: The Determinants of Mutual Fund Preference: Family-level Covariates

Score 1 Score 2

(1) (2) (3) (4)

Frac. family indexed −30.453∗∗ −31.009∗ −32.280∗∗ −23.459∗∗

(12.117) (16.848) (12.882) (10.740)Frac. family growth −14.291 −15.723 −7.470 −4.376

(17.421) (18.308) (14.161) (14.097)Frac. family value −7.513 −8.459 52.887∗ 54.031∗∗

(27.289) (28.296) (27.666) (27.356)Family avg. exp. ratio (bps) 0.002 0.158

(0.133) (0.102)2nd quintile family TNA 7.145 7.325 −5.552 −4.856

(9.052) (9.216) (9.010) (9.385)3rd quintile family TNA 20.100∗∗ 21.930∗∗ −3.542 0.965

(9.048) (9.643) (8.071) (8.744)4th quintile family TNA 9.531 9.649 −7.723 −3.462

(9.580) (10.441) (8.205) (9.179)5th quintile family TNA −7.566 −7.223 −15.715∗∗ −6.085

(8.763) (11.529) (7.935) (9.943)Constant 2.354 2.630 5.790 −13.957

(9.140) (20.031) (9.091) (15.331)

Observations 3,611 3,574 3,611 3,574R2 0.120 0.122 0.215 0.234

Note: ∗p<0.1; ∗∗p<0.05; ∗∗∗p<0.01Regression is weighted using the funds’ TNA.

Standard errors are clustered at the fund family level.

40

Table 13: Summary of Company Shareholder Base Preference Sample

Year

NumberofCompanies

Avg. %SharesOwned byMFs

Avg. % SharesOwned by MFswith Ideal Points

Avg. Number ofFundShareholderswith Ideal Points

2002 2,827 19 11 59

2003 3,161 20 11 60

2004 3,268 22 13 73

2005 3,265 21 14 75

2006 3,310 22 14 79

2007 3,158 24 16 104

2008 2,571 28 21 153

2009 2,744 27 22 157

2010 2,782 27 23 149

2011 2,637 27 23 152

2012 2,683 27 23 147

2013 2,842 27 23 136

2014 2,995 26 22 124

2015 3,020 25 22 117

2016 3,023 26 23 122

41

Table 14: Company shareholder scores: Growth vs Value companies

Company Shareholder Score 1 Company Shareholder Score 2

(1) (2) (3) (4)

Book-to-Market 5.043∗∗∗ 3.841∗∗∗

(0.385) (0.385)2nd Quintile Book-to-Market −0.078 1.183∗∗∗

(0.322) (0.235)3rd Quintile Book-to-Market 1.021∗∗∗ 2.436∗∗∗

(0.358) (0.260)4th Quintile Book-to-Market 3.343∗∗∗ 3.878∗∗∗

(0.399) (0.280)5th Quintile Book-to-Market 5.486∗∗∗ 4.862∗∗∗

(0.473) (0.329)Constant −9.910∗∗∗ −8.625∗∗∗ −9.663∗∗∗ −9.668∗∗∗

(0.429) (0.430) (0.429) (0.276)

Year Fixed effects Yes Yes Yes YesObservations 41,360 41,360 41,360 41,360R2 0.061 0.062 0.095 0.099

Note: ∗p<0.1; ∗∗p<0.05; ∗∗∗p<0.01Standard errors are clustered at the company level

42

0

300

600

900

1200

0 10 20 30

k

valu

e

Figure 1: Scree Plot

43

Columbia FundsCharles Schwab Pax World Manag

Calvert Investm

Domini Social I

Green Century C

American Funds

Vanguard Group

Institutional S

Fidelity Manage

T Rowe Price As

William Blair &

BlackRock IncState Street Ba

Glass Lewis

Management−50

0

50

100

−50 0 50

Dimension 1

Dim

ensi

on 2

Figure 2: Dimension 1 vs. Dimension 2

44

ISSMgmtGL

ISSMgmt GL

0

50

100

150

200

250

0

200

400

−50 0 50 100

−50 0 50 100

score1

score2

coun

tco

unt

Figure 3: Univariate Histograms of Scores on Dimensions 1 and 2

45

0

1000

2000

3000

4000

0

1000

2000

3000

−0.03 −0.02 −0.01 0.00 0.01 0.02

−0.03 −0.02 −0.01 0.00 0.01 0.02

A1

A2

coun

tco

unt

Mgmt Rec.

Vote No

Vote Yes

Mgmt Rec.

Vote No

Vote Yes

Figure 4: Histograms of a1 and a2 by Management Recommendation

46

0.00

0.25

0.50

0.75

1.00

0.00

0.25

0.50

0.75

1.00

MP

−Com

pens

atio

nM

P−C

orpo

rate

Fin

ance

MP

−Cor

pora

te G

over

nanc

e

MP

−Ele

ct D

irect

or (C

onte

sted

)M

P−E

lect

Dire

ctor

s

MP

−Mer

ger /

Acq

uisi

tion

Rel

ated

MP

−Oth

erM

P−R

atify

Aud

itors

SP

−Cor

pora

te G

over

nanc

e

SP

−Ele

ct /

Rem

ove

Dire

ctor

s (C

onte

sted

)

SP

−Exe

cutiv

e/D

irect

or C

ompe

nsat

ion

SP

−Oth

erS

P−S

ocia

l Pro

posa

l

proptype

proptype

coun

tco

unt

LoadingDim1

Extremely Positive

Non−Extreme

Extremely Negative

LoadingDim2

Extremely Positive

Non−Extreme

Extremely Negative

Figure 5: Signs of a1 and a2 by Proposal Category

47

ISS Supports

ISS Opposes

MP

−Com

pens

atio

nM

P−C

orpo

rate

Fin

ance

MP

−Cor

pora

te G

over

nanc

e

MP

−Ele

ct D

irect

or (C

onte

sted

)M

P−E

lect

Dire

ctor

s

MP

−Mer

ger /

Acq

uisi

tion

Rel

ated

MP

−Oth

erM

P−R

atify

Aud

itors

SP

−Cor

pora

te G

over

nanc

e

SP

−Ele

ct /

Rem

ove

Dire

ctor

s (C

onte

sted

)

SP

−Exe

cutiv

e/D

irect

or C

ompe

nsat

ion

SP

−Oth

erS

P−S

ocia

l Pro

posa

l

0.00

0.25

0.50

0.75

1.00

0.00

0.25

0.50

0.75

1.00

proptype

coun

t

LoadingDim1

Extremely Positive

Non−Extreme

Extremely Negative

Figure 6: Sign of a1 by Proposal Category, ISS Rec.

48

GL Supports

GL Opposes

MP

−Com

pens

atio

nM

P−C

orpo

rate

Fin

ance

MP

−Cor

pora

te G

over

nanc

e

MP

−Ele

ct D

irect

or (C

onte

sted

)M

P−E

lect

Dire

ctor

s

MP

−Mer

ger /

Acq

uisi

tion

Rel

ated

MP

−Oth

erM

P−R

atify

Aud

itors

SP

−Cor

pora

te G

over

nanc

e

SP

−Ele

ct /

Rem

ove

Dire

ctor

s (C

onte

sted

)

SP

−Exe

cutiv

e/D

irect

or C

ompe

nsat

ion

SP

−Oth

erS

P−S

ocia

l Pro

posa

l

0.00

0.25

0.50

0.75

1.00

0.00

0.25

0.50

0.75

1.00

proptype

coun

t

LoadingDim2

Extremely Positive

Non−Extreme

Extremely Negative

Figure 7: Sign of a2 by Proposal Category, GL Rec.

49

Columbia Funds

Charles Schwab Pax World Manag

Calvert Investm

Domini Social I

American Funds

Vanguard Group

Institutional S

Fidelity Manage

T Rowe Price AsWilliam Blair &

BlackRock IncState Street Ba

Glass Lewis

Management

−25

0

25

50

−50 −25 0 25

Dimension 1

Dim

ensi

on 2

2010 − 2011

Columbia FundsCharles Schwab

Pax World Manag

Calvert Investm

Domini Social IAmerican Funds

Vanguard Group

Institutional S

Fidelity ManageT Rowe Price As William Blair &

BlackRock Inc

State Street Ba

Glass Lewis

Management−30

0

30

60

−40 −20 0 20 40

Dimension 1

Dim

ensi

on 2

2012 − 2013

Columbia FundsCharles Schwab

Pax World Manag

Calvert Investm

Domini Social I

Green Century C

American Funds

Vanguard Group

Institutional S

Fidelity ManageT Rowe Price As

William Blair &BlackRock Inc

State Street Ba

Glass Lewis

Management−20

0

20

40

60

−50 −25 0 25

Dimension 1

Dim

ensi

on 2

2014 − 2015

Figure 8: Preference Estimates Over Time50

Dimension 1

Dim

ensi

on 2

25

25

−100 −50 0 50

−50

050

100

50

50

50

75

75

75

75

GL

Mgmt

ISS

Figure 9: Density of Mutual Fund Preferences

51

Shareholder Intervention Party

Managerialist Party

Shareholder Veto Party

−50

0

50

100

−50 0 50

score1

scor

e2

Figure 10: The Parties

52

Ma

In

Na

Am

VaVa VaVa

VaVa

BlAX

JP

Di

JPJP

FAUS

In

Fi

JP

TI TI

We

Pr

FiFi

Fi No

Pi

Fi

VaVa

Am

Am

Le

Fi FiFi

T

Fr

AmAm

Al

T

Fi

Pu

Pu

Am

Pu

MF

Am

MF

Fi

Am

Fi

Ja

Fi

Le

AX

Mu

Ja

Va

T

T

FiFi

Fi

Fi

Va

Dr

AX

Bl

Su

Co

Co

Mo

Fi

US

Am

FiFi

DW

Fi

T

Am

USWe

T

Dr

Va

Pi

Fi

Pu

Co

AmAmT

Ja

Fe

Ma

Ma

Va

Ja

De

Ma

Ma

In

In

Pu

Di

Mu

Am

Dr

Le Fi

MaMa

SuWe

Mu

JP

Va

VaSu

BlFi

Fi

Ma

Pu

T

T

T T

T

MF

Fi

Th

JPJP

Pi

MFMF

T

T

T JP

PuPu

Ca

Fi

MFMF

Lu

We

Su

AmAmAm

No

Mo

Bl

De

Fi

JPAX

AX

Co

MF

Mu

In

Am

Fi

PrPr

DW

Di

Lu

Ma

Pi

T

Ma

DeDeDe

Wi

Pu

In

PN

Pr

Fi

Fi

Co

Pu

Ma

Mu

AX

Pr

RSSuSu

TI

Na

Pu

Na

Ma

BlBl Bl BlBl

BlBlBl

Bl

BlBl

Co

In

PrPr Va

StBl

Bl

StSt

StSt

PrNo

Pr

Pr

Pr

Ma

Va

Di

T

MF

Ma

T Gu

Va

ThMa

MF StPr Pr

Bi

Va

TI

PrPr

Ry

WiDe

De

Sc

Va

Fi

T

Fi

Ha

Bl

Ma

Va

Bl

PrPr

Di

Co

Ma

Ma

GoSt

Am

DiDi

JP

Di

AX

In

Go

Ma

In

Va

Di

Bi

Ma

Ma Ma

Am

Ry

MF

PrPrFi

Fr

TI

In

In

Bl

VaVa

In

Bl

T

Pr

Di

MF

TIBl

Fe

Fi

Ry

AX

Su

AX

Va

Di

Va

IN

Pa

IN

Fi

Pr

PuPu

IN

Co

IN

Pr

AXBl BlBlAXBl

Co

JP

Pu

Ry

AX

MF

Am

DiPr

Al

Al

Pu

AmFi

Ra Ra

Bl

AQ

MaSt

Co

FiFi

Mo

Ma

Fi

StSt

Fi

DiCh

Ex

JP

TD

BlBlRE

Di Ex

Go

Fi

AX

Fi

Fi

FiFi

In

In

AL

JP

St

Dr

St

Ex

12

St

In

Am

We

JP

Co

−50

0

50

−40 0 40 80

Dimension 1

Dim

ensi

on 2

Correct N

Correct Y

Incorrect N

Incorrect Y

Figure 11: Dissident Nominee for Director at DuPont

Gl

Ma

In

Ma

Ru

Na

Ha

SE

VaVa VaVa

VaVa

AX

Di

JP

FAUS

In

Go

TITI TI

Ma

Ca

Pr

Gu

Ch

FiFiFiFi Jo

No

Am

VaVa

AmLe

Fr

Am

Op

Al

Pu

Ry

Am

Ga

Ry

Am

AX

Mu

VaGa Va

Dr

AXAX

Pr

Co

Sh

MoUSOp

Am

Fi

Am

USWe

T

Dr

Va

Ry

AmAm

Ja

Ch

Fe

Ma

Ga

Pa

Ea

Va

Wi

Wi

Ru

Ru

DeAd

Ma

Ma

MaMa

GuSe

Va

Mu

Am

Dr

LeSe

Fi

MaMa

Go

Pu

Am

WeMu

Pr

Br

Ma

Ma

Co

Go

Th

UB

Ca

AmAmAm

No

Ja

He

Pa

Mo

Ch

Bl

De

Co

AXAX

Ch

Mu

In

Am

Fi

PrPr

DW

AX

Lu

Ru

Ma

T

Ma

DeDeDe

Wi

Pu

SE

In

PN

Va

Co

Ch

Se

Mu

AX

Pr

RSSuSu

Pr

TI

Na

TINa

Jo

Va

Ja

Ma

BlBl Bl BlBl

BlBlBlBl

BlBl

Va

Co

In

Pr

StSt StSt

Ma

Fr

PrPr Pr

Ru

Pr

Pr

Va

Di

T

Br

Ma

Pr

Gu

Gu

Op

Va

Ha

Pr

Bi

Tr

TI

PrPrJa

Ry

WiSc

Va

Ha

Fi

Br

Bl

IN

Bl

Ma

Va

Bl

PrPr

Di

Ja

Ma

In

Pa

Hi

Ma

St

In

DiDiDi

AX

Ry

In

Go

MaJa

Fi

Di

Bi

Am

Ry

Pr

Ja

Fr

Op

TIBl

VaVa

In

Bl

Di

TI

Ja

As

Bl

Va

AX

VT

Al

No

Na

Pr

AX

VaVa

IN

Sc

Ha

Ma

PuPu

ININ

AQAQPr

AXBlBlBlAX

ChChCh

Al

AlAl

Al

Bl

Pu

Ry

Pr

AX

Bl

Tr

In

Ha

In

DiPr

Al

Al

Pu

Bl

Pu

Bl

AmAm

Ra

BlSt

Cu

SE

JoJo

Co

Co

ALAQ

Hu

Ru

Co

FiJa

Gr

In

StSt

DiEx

JP

TDAQ

PuPu

BlBl

Di

Ex

Ex

Ch

Ch

Fi

Fi

MF

Ca

In

SE

Ea

Bl

St

Ru

Dr

Dr

St

Ga

SE

Al

SE

ET

In

ET

Am

−50

0

50

100

−40 0 40 80

Dimension 1

Dim

ensi

on 2

Correct N

Correct Y

Incorrect N

Incorrect Y

Figure 12: A Shareholder Proposal on Exec. Comp. at Waste Management, Inc.

53

Gl

Ma

In

Ru

Am

Ry

O'

VaVaVa

VaVaVa

Ch

Di

Co

FA

TI

Am

TI

Ch

Ha

Fe

Jo

In

Mu

Pr Pr

Fi

Ch

Na

Fi Fi

Ev

MuGo

Fi

Va

Ch

Al

T

Al

Va

AXAX

Fi

Fr

Pr

Ca

Jo

MF

Fe

Fi

Am

Pr

Bl

Ch

Dr

Co

FA

SE

Va

Go

Lu

FrFr

Ru

Ma

FrPu

Co

Mu

PuPuPu

Ha

Lu

SE

In

To

Ch

Ch

In

Va

AmAmAm

Ha

AmHS

AX

Ic

Co

Al

FrFr

T

Mu

MF

Ch

Ad

No

Go

Na

Fi Fi

DW

AX

Va

T

T MF

Dr

Am

Go

Ha

Fi

VaDi

Le

Pa

DiDi

Ch

Tr

Ma

Ma Na

Ca

No

Su

TI

Ma

PrPr Pr

TI

Co

Jo JoJo

Ca

Ja

Na

Na Bl

BlBl

BlBl

Bl

Bl

BlBl

MF

In

Pr

Di

US

Va

SE

St

Ha

Pr

In

Pr

Th

Mu

AX

Pu

Va

Di

Co

USPr PrPrPrPrPrPr

Co

Dr

AX

Ja

FAPrPr Fr

Fe

IrSu

Jo

Fi

Va

Ch

In

IN

Bl

Gl

Bl

Fi

Ir

No

Ja

ThThThThThTh

StSt

St

PN

Jo

DiDiDi

Ma

Su

Ry

Pr

Ha

Wi FA

TI

HaIn

In

In

Di

MF

Fr

AX

Fi

VT

Di

Va

INIn

Fi

PrPr

Ja

Pu

AX

AX

AXAX

Al

SE

Ch Ch

CaRy

DiPrFi

Fi

Pu

Wi

Fi

AQJo

JoIn

Ma

Ma

StSt

Ea

Di

Ch ChDi

Al

St

Ch

Ch

Go

Fi

PN

PNGo

SE

SE

AX

SE

−50

0

50

−40 0 40

Dimension 1

Dim

ensi

on 2

Correct N

Correct Y

Incorrect N

Incorrect Y

Figure 13: A Say-on-Pay Proposal at PolyOne Corporation

Gl

Ma

InJo

Ru

AmAm

Am

Le

VaVaVaVa

Va

Gr

Gu

Sc

JP JP

JP

Fi

Di

Fi

Le

TI

Lo

Jo

Va

PrPrPr

Wi

Wi

Wi

Ch

FiFiFi

LS

JP

Va

Ro

EU

Va

SE

FiCo FiFi

Fi

Di

In

Co

US

Va

AX

Me

Ch

Co

Va

Va

Ru Ru

Ru

AS

Tr

Tr

Co

Fi

Va

WaLe

JP

Di

Di

Ch

Ja

Ja

Ro

Co

Ic

AS

AX

AX

US

Ch

Ad

Na

Fi FiAX

AS

Ru

Al

IN

T

T

Wi

SE

IN

Va

Di

Va

Di

Ch

Ma

Ca

Su

TI

Pr

Co

Me

Va

Ja

NaAXBl

Bl

Bl

BlBlBl

Bl

Bl

Le

In

Su

US

SE

St

Ma

Ar

AX

Al

Pa

Va

Di

Bl

Al

BlBl

PrPrPrPrPr

Gu

Va

AX

AX

Co

Pr

Tr

Su

Li

RySe

Wi

Va

Fi

Bl

Va

Bl

Bl

Ma

Ir

No

In

Sc

Co

St StSt

Jo

Jo

DiDiDi

AX

In

Ar

Ry

Gu

AS

AS

Ma

Le

Pr

JP

FA

TI

AX

Ro

Bl

Bl

Di

JP

Bl

Su

Pr

Va

Ja

Di

Na

Na

Na

Wi

IN

RaUS

In

Jo

Fi

Pr

Ra

AXAl

ChChCh

RB

Gu Gu

Jo

Jo

AlBl Am

Fi

Le

RyRy

Pr

Ch

Gu

Fo

Fo

Fo

Di Fi

Sc

Ry

Fi

−50

0

50

−40 0 40

Dimension 1

Dim

ensi

on 2

Correct N

Correct Y

Incorrect N

Incorrect Y

Figure 14: An Uncontested Director Election at Avnet, Inc.

54

−30

0

30

60

4 6 8 10 12

log(Market Value of Equity)

Com

pany

Sha

ehol

der

Sco

re 1

−50

0

50

100

4 6 8 10 12

log(Market Value of Equity)

Com

pany

Sha

reho

lder

Sco

re 2

Figure 15: Company shareholder scores vs. market capitalization (2012)

55

−50

0

50

−40 0 40 80

Dimension 1

Dim

ensi

on 2

Figure 16: Company shareholder scores (2002) and density of mutual fund scores

−50

0

50

−40 0 40 80

Dimension 1

Dim

ensi

on 2

Figure 17: Company shareholder scores (2016) and density of mutual fund scores

56

−15

−10

−5

0

1 2 3 4 5

Sco

reDimension 1

−16

−12

−8

−4

1 2 3 4 5

BE/ME quintile

Sco

re

Dimension 2

year20022005

20072009

20112013

2016

Figure 18: Average company mean shareholder scores by Book-to-Market quintiles - selectedyears

57

−12

−9

−6

2004 2008 2012 2016

year

Sco

re

Dimension Dimension 1 Dimension 2

Figure 19: Yearly average of company shareholder score

0.00

0.01

0.02

0.03

0.04

0.05

−60 −30 0 30 60

Dimension 1

dens

ity

0.00

0.02

0.04

0.06

−50 0 50 100

Dimension 2

dens

ity

year 2002 2007 2012 2016

Figure 20: Company shareholder score distributions: evolution over time

58

−10

0

2004 2008 2012 2016

year

Sco

re

Dimension 1

−15

−10

−5

0

2004 2008 2012 2016

year

Dimension 2

cap.bucket Large−Cap Mid−Cap Small−Cap Micro−Cap

Figure 21: Average company-level score by market-cap category

59