Picking Friends Before Picking (Proxy) Fights: How Mutual Fund Voting Shapes Proxy Contests

Posted by Alon Brav (Duke University), Wei Jiang (Columbia University), and Tao Li (University of Florida), on Monday, February 12, 2018

Over the past two decades the frequency of proxy contests for board representation or control has increased as shareholder activism has become both an established investment strategy and an important form of corporate governance. Since dissidents are typically minority stockholders, a successful campaign, such as Trian Partners’ intervention at Procter and Gamble Co., requires support from their fellow shareholders. The general apathy of retail investors towards voting matters implies that it is usually necessary that dissidents win the support of a majority of institutional shareholders. Hence, “picking friends”—the selection of a target with a pro-activist shareholder base—is a crucial element in an activist’s decision on whether to launch a proxy contest.

In our paper, Picking Friends Before Picking (Proxy) Fights: How Mutual Fund Voting Shapes Proxy Contests, publicly available on SSRN, we study the determinants of mutual fund voting in proxy contests, in which voting decisions are arguably more informative as compared with uncontested meetings where investor votes are mostly precatory. Further, we explicitly model a simultaneous system consisting of both activists’ target selection and mutual fund voting, which allows us to uncover the funds’ voting rules for all potential proxy contests based on the subset of voting records of materialized contests.

Our study relies on a unique and comprehensive data set of mutual fund voting records for proxy contests, extracted from the mandatory N-PX filings by U.S. registered management investment companies. We find that mutual funds are more likely to support a dissident when the target firm experiences poor recent stock price or accounting performance. Support for the dissident is higher when two leading proxy advisory firms, Institutional Shareholder Services and Glass, Lewis & Co., issue a “For” recommendation for the dissident rather than when either of the advisory firms supports the management. We further identify large differences in fund families’ tendency to follow advisory firms’ recommendations. The families that are most responsive to proxy advisors are mainly smaller fund families that lack resources to conduct independent proxy research. We also find that mutual funds are more likely to vote for hedge fund activists rather than other types of dissidents, consistent with the notion that investors believe that hedge funds are more economically driven than other types of activists and are therefore an effective force of governance. Mutual funds, however, do not support a dissident’s slate of directors when the dissident has been a “frequent” activist targeting many companies in the past, but tend to support those activists whose targeting signals a high commitment in the past (e.g., seeking board representation). In other words, institutional investors favor focused and determined activists and are not necessarily impressed by an activist’s length of a hostile track record.

Fund characteristics also predict differential support for dissidents. One salient pattern is that passively-managed funds are less likely than active funds to vote for dissidents. The gap between active and passive votes has been persistent across years, and is larger for small capitalization stocks. To the best of our knowledge, this is the first study reporting direct evidence that passive funds are more “friendly” towards management than active funds. This is also confirmed by our family-level study, which shows that the most pro-dissident fund families typically have a low fraction of passive funds, while the least pro-dissident groups tend to have a disproportionately high number of passive funds. One potential reason is that unlike actively-managed funds, passive funds—index and exchange-traded funds—are not rewarded by “beating the index.” Instead, they are usually rewarded by low expense ratios and small tracking errors. We also find that a mutual fund is significantly more likely to support a dissident when the abnormal returns of same-industry firms in the fund’s portfolio are higher. This is consistent with the idea that mutual funds make voting decisions based on the overall performance of their portfolios. In addition, funds earning a positive basis-adjusted return (return net of the cost of investment) on the target stock are less likely to support the dissident than a fund earning a negative return. This suggests that “unhappy” shareholders, who have lost capital investing in the stock, tend to favor the changes proposed by the dissident.

Since investors’ voting decisions and dissidents’ target selection are jointly determined an analysis of shareholder voting behavior in materialized proxy contests may not fully reveal the underlying “voting rules” adopted by institutional investors. We therefore employ a parsimonious system of equations to model the joint contest-voting dynamics. This system yields a positive coefficient of correlation between the propensity to target by an activist and the propensity to support the activist by investors, consistent with the notion that activists tend to target firms with unobservable characteristics that predict strong shareholder support, beyond the predictive power of observable attributes. We then proceed to construct two proxies to capture investors’ “inherent” pro-activist stance. The first is a fund’s voting tendency to support the dissident, retrieved from their voting outcome relative to their peers within a given event, while the second measure is constructed using pair-wise fund ranks based on the funds’ support for dissidents on the common events that each pair of funds participated in. Both measures are normalized and orthogonalized to be unrelated to observable fund characteristics. We find that both proxies strongly predict activist targeting.

Finally, we propose two measures of mutual funds’ degree of “persuadability,” that are meant to capture mutual funds’ willingness to learn from and be persuaded by a dissident. The first proxy is the tendency of a company’s institutional shareholder base to be swayed by proxy advisors’ recommendations. The second measure is based on the idea that a fund willing to carefully assess the merit of each case is likely to have high variation in the votes cast over time, and we therefore proxy for a fund’s “persuadability” using the variation in the votes it cast in the past prior to the proxy contest. For both measures we find that activists are more likely to target companies whose shareholder base exhibits a high degree of “persuadability,” consistent with the idea that shareholders’ willingness to learn about the merit of the intervention is an important factor in activists’ target selection.

The full paper is available for download here.