Voting Through Agents: How Mutual Funds Vote on Director Elections
Stephen J. Choi
New York University School of Law
Jill E. Fisch
University of Pennsylvania Law School - Institute for Law and Economics
New York University School of Law; European Corporate Governance Institute
August 17, 2011
University of Pennsylvania, Institute for Law & Economics Research Paper No. 11-28
NYU Law and Economics Research Paper No. 11-29
Shareholder voting has become an increasingly important focus of corporate governance, and mutual funds control a substantial percentage of shareholder voting power. The manner in which mutual funds exercise that power, however, is poorly understood. In particular, because neither mutual funds nor their advisors are beneficial owners of their portfolio holdings, there is concern that mutual fund voting may be uninformed or tainted by conflicts of interest. These concerns, if true, hamper the potential effectiveness of regulatory reforms such as proxy access and say on pay. This article analyzes mutual fund voting decisions in uncontested director elections. We find that mutual funds use a variety of strategies to economize on the costs of making voting decisions, including having funds in the same fund family vote in lockstep, voting virtually always in accordance with management recommendations, and voting virtually always in accordance with recommendations of ISS. Smaller fund families employ these strategies to a greater extent than larger families.
We further adduce evidence on how ISS recommendations affect fund voting. Funds that account for less than 10% of the assets in our sample exhibit a strong tendency to follow ISS recommendations, a much smaller percentage than funds that virtually always follow management recommendations (approximately 25% of assets). A much larger percentage (36% of the assets) votes in accordance with ISS withhold recommendations in approximately 50% of the cases. We conclude that the influence of ISS is due more to funds’ measured evaluation of the ISS recommendations rather than to funds blindly following these recommendations.
We find no evidence that funds in families that are affiliated with commercial banks, investment banks, or insurance companies have a stronger proclivity than independent funds to vote in accordance with management recommendations or to shield their votes from criticism in order to maintain good business relations or generate new business for their affiliates.
The largest fund families - Vanguard, Fidelity, and American Funds, each of which individually accounts for roughly 11% of total mutual fund assets - vote substantially differently both from each other and from ISS recommendations. This is strong evidence of heterogeneity in the voting behavior of mutual funds in director elections.
Finally, we examine the factors associated with high (in excess of 30%) withhold votes in director elections. An ISS withhold recommendation, in conjunction with at least one of four factors - a withhold vote by Fidelity, the director missing 25% of board meetings, the company having ignored a shareholder resolution that received majority support, and a Vanguard withhold vote on outside directors with business ties to the company - is associated with a 49% probability of receiving a high withhold vote. Directors in these groups account for 48% of all directors who received high withhold votes. By contrast, an ISS withhold recommendation that is not combined with one of these factors is associated with only a 21% probability of a high withhold vote, and the general probability of a high withhold vote is a mere 2%. These findings suggest steps that companies and directors should take to try to avoid high withhold votes. They are also evidence that not all ISS recommendations have the same impact on voting outcomes.
Number of Pages in PDF File: 64
Keywords: Corporations, securities, mutual funds, shareholder voting, corporate governance, ISS, Institutional Shareholder Services, proxy advisor recommendations, voting patterns, mutual fund families, empirical research
JEL Classification: D72, G34, K22
Date posted: August 23, 2011 ; Last revised: May 15, 2012