doi:10.1016/j.jfineco.2005.04.003
Copyright © 2005 Elsevier B.V. All rights reserved.
Business ties and proxy voting by mutual funds
References and further reading may be available for this article. To view references and further reading you must
purchase this article.
Gerald F. Davisa and E. Han Kim
, a, 
aStephen M. Ross School of Business, University of Michigan, Ann Arbor, MI 48109, USA
Received 16 February 2005;
revised 28 March 2005;
accepted 5 April 2005.
Available online 19 October 2005.
Abstract
The magnitude of mutual funds’ business ties with their portfolio firms is documented and is linked to funds’ proxy votes at specific firms and to overall voting practices. Aggregate votes at the fund family level indicate a positive relation between business ties and the propensity to vote with management. Votes at specific firms, however, reveal that funds are no more likely to vote with management of client firms than of non-clients. Because the votes took place when funds knew their votes would be publicly scrutinized, fund families with a larger client base may have adopted voting policies that led to less frequent opposition to management at all firms.
Keywords: Conflicts of interest; Proxy voting; Mutual funds; Employee pension plans
JEL classification codes: G23; G34; L14
Table 1.
Fees and assets managed for Fortune 1000 clients
This table shows the mean and median dollar value of fees paid to, and 0.5% of assets managed by, six large mutual fund families by corporations in the 2001 Fortune 1000. Data are taken from IRS Form 5500 filings compiled by Judy Diamond Associates for calendar year 2001. Figures are aggregated by ultimate corporate parent such that pension plans of subsidiaries or different employee groups are summed to arrive at the final figure. “Fees” are fees paid out of plan assets to a provider by the plan sponsor, taken from Schedule C. “Assets/200” is calculated from assets invested with the fund as reported on Schedule D. “Combined” sums these two figures within corporate parent. Clients may pay fees, invest with the fund, or both; thus, “N of clients” for combined is not the sum for the two prior categories.
Table 2.
Ownership and client ties between Fortune 1000 firms and six large mutual funds
This table shows cross-tabulations of whether a Fortune 1000 firm was held in the fund's portfolio in 2001 and whether it was a pension client in 2001. The sample consists of 892 firms with valid 13F filings on institutional ownership for the second quarter of 2001 (from Spectrum). “Non-client”/“Client” indicates whether the firm reported any pension business with the fund in IRS Form 5500 filings for calendar year 2001. “Non-portfolio”/“Portfolio” indicates whether the fund held any shares in the firm as of June 30, 2001. The number of client firms tabulated differs from Table 1 because some Fortune 1000 firms report pension data but were not publicly traded at 2001 year-end (e.g., cooperatives such as Ace Hardware).
Table 3.
Fund voting policies and percentage of votes in favor of six shareholder proposals opposed by management at Fortune 1000 firms in 2003 to 2004
This table reports how the ten largest mutual fund families, CalPERS, and CREF voted on six shareholder proposals that appear on the proxy statements of their portfolio firms between July 1, 2003 and June 30, 2004 and that are opposed by the portfolio firm's board. Sampled firms are all portfolio firms among the 2001 Fortune 1000 whose proxy includes the shareholder proposal. Voting data are obtained from Form N-PX filings by the fund family and are aggregated across 10 or more funds within the family; CalPERS votes obtained from its website. Votes are aggregated across funds within a family such that each portfolio firm receives one vote per family when calculating means (even though many funds within a family may hold shares in that firm and thus cast votes). “For independent board chair” includes proposals requesting a formal separation of the positions of CEO and Chairman of the Board. “Require vote on pill” concerns proposals requesting that current or future shareholder rights plans (“poison pills”) be voted upon by shareholders. “Expense options” requests that stock options be expensed on the firm's financial statements. “Declassify board” entails holding annual (as opposed to staggered) elections for directors. “Vote on parachute” concerns proposals requesting that executive compensation premised on a change in control (“golden parachutes”) be voted upon by shareholders. “Allow cumulative voting” seeks to allow shareholders to cumulate votes in the election of directors. Entries are classified as “No” if the fund family votes against the proposal in more than 85% of portfolio firm, “Yes” if the fund family votes for the proposal in more than 85% of portfolio firms, and “Split” otherwise. The number in parentheses is the percentage of “yes” votes for the proposal among portfolio firms in the 2001 Fortune 1000. “Total” is the percentage of votes cast in favor of all six proposals combined.
Table 4.
Tobit models for percent of votes cast by fund families in favor of six shareholder proposals opposed by management in Fortune 1000 companies
This table reports Tobit regressions of the percentage of votes cast in favor of six shareholder proposals opposed by management by 23 funds or fund families at Fortune 1000 firms at annual meetings between July 1, 2003 and June 20, 2004. The first model includes observations on all six proposals; thus, each of the 23 fund families contributes six observations. Fixed effects by proposal are coded by a set of five dummy variables. The remaining models include votes on each of the six separately. Votes are aggregated across funds within a family such that each portfolio firm receives one vote per family when calculating means (even though many funds within a family may hold shares in that firm and thus cast votes). “Independent Chair” includes proposals requesting a formal separation of the positions of CEO and Chairman of the Board. “Require Vote on Pill” concerns proposals requesting that current or future shareholder rights plans (poison pills) be voted upon by shareholders. “Expense Options” requests that stock options be expensed on the firm's financial statements. “Declassify Board” entails holding annual (as opposed to staggered) elections for directors. “Vote on Parachute” concerns proposals requesting that executive compensation premised on a change in control (golden parachutes) be voted upon by shareholders. “Allow Cumulative Voting” seeks to allow shareholders to cumulate votes in the election of directors. The sample includes AIM Invesco, Alliance Capital Management, American Century, American Funds, Ariel, Barclays (iShares), CalPERS, CREF, Dreyfus, Fidelity, Franklin-Templeton, Janus, Legg Mason, Merrill Lynch, Morgan Stanley, Oppenheimer, PIMCO, T. Rowe Price, Putnam, Schroder, Scudder, State Street Corporation, and Vanguard. T-statistics are in parentheses.
n=138 in first model; 23 in all others.
Table 5.
Shareholder proposal Tobit regressions with alternative measures of business ties
This table reports results for Tobit regressions of the percentage of votes cast in favor of six shareholder proposals opposed by management in Fortune 1000 firms between July 1, 2003 and June 30, 2004, using alternative measures of client ties. Each model includes observations of 23 fund families’ votes on all six proposals; thus, each of the 23 fund families contributes six observations. Fixed effects by proposal are coded by a set of five dummy variables. Votes are aggregated across funds within a family such that each portfolio firm received one vote per family when calculating means (even though many funds within a family may hold shares in that firm and thus cast votes). “Log(fees)” is the natural logarithm of the total dollar value of fees and 0.5 percent of assets managed via pension fund business. “Fees/size” is the total value of fees divided by the dollar value of the fund family's holdings in Fortune 1000 firms in 2003. “# recordkeeper contracts” is the count of contracts with Fortune 1000 firms in 2001 to act as recordkeeper on the firm's pension plan(s). “Clients” is the number of Fortune 1000 pension clients in 2001. “Recordkeeper dummy” is a dummy variable for whether the fund family or its parent has any recordkeeper contracts with Fortune 1000 firms. The sample includes AIM Invesco, Alliance Capital Management, American Century, American Funds, Ariel, Barclays (iShares), CalPERS, CREF, Dreyfus, Fidelity, Franklin-Templeton, Janus, Legg Mason, Merrill Lynch, Morgan Stanley, Oppenheimer, PIMCO, T. Rowe Price, Putnam, Schroder, Scudder, State Street Corporation, and Vanguard. T-statistics are in parentheses.
We thank Bernard Black, Dana Muir, Adair Morse, Paige Ouimet, Amit Seru, Clemens Sialm, Lu Zheng, and participants of the Journal of Financial Economics/Ohio State University Conference on Agency Problems and Conflicts of Interest in Financial Markets and the finance seminar at the University of Michigan for helpful comments and suggestions. We are grateful for outstanding research assistance from Julian Atanassov, Sandra Gomez, and Amrita Nain in compiling the data on mutual fund proxy voting and we thank Joyce Buchanan for editorial help.
Corresponding author. Fax: +1 734 763 3117.
Abstract
Purchase PDF (189 K)
E-mail Article
Add to my Quick Links
Cited By in Scopus (1)
Purchase PDF (134 K)
