Proxy voting decisions are characterized by strong strategic complementary relationships in the sense that the utility to a shareholder from voting in a particular manner increases with the extent that other shareholders vote in the same manner. Such complementarities arise because the likelihood that the vote outcome goes in a certain direction depends on the fraction of shareholders that vote in that direction. Shareholders thus have an incentive to condition their votes on those of their peers to make their aggregate votes count more. By coordinating their voting behavior, individual institutions can increase the likelihood that the outcome of the vote is in their favor and hence attain a higher utility from their votes.
In my recent paper, Thy Neighbor’s Vote: Peer Effects in Proxy Voting, I examine peer influence among geographically proximate institutional investors in the context of proxy voting. Influencing peers to vote in a certain way involves a trade-off. The benefit of an increase in the likelihood of a desired vote outcome must be weighed against the costs incurred to exert positive influence, such as information costs involved in searching and identifying like-minded peers as well as bargaining and monitoring costs associated with reaching and enforcing agreements. Geographic proximity can facilitate the collection of soft information and monitoring through repeated interactions and face-to-face meetings, thereby reducing the costs associated with peer influence. Therefore, institutions are likely to derive a higher expected utility from influencing the voting behavior of their geographically proximate peers.
To disentangle peer effects from confounding effects, I exploit variation in the voting behavior of publicly traded financial institutions due to close-call votes on shareholder-sponsored governance proposals at these institutions (hereafter referred to as focal institutions). Investment managers, as agents of their beneficial investors, may undersupply monitoring effort, because they bear the costs of monitoring activities, but capture only a fraction of the benefits (Bebchuk, Cohen, and Hirst, 2017). Improved internal corporate governance at these institutions can better align the interests of shareholders and investment managers and force investment managers to allocate more effort to monitoring activities. Therefore, the passage of governance proposals and the ensuing improvement in governance at a publicly traded financial institution can make the institution more active in monitoring its portfolio companies through proxy voting. Since the result of a close-call vote around the majority threshold at a focal institution is unlikely to be correlated with contextual or correlated effects faced by the institution (Cuñat, Gine, and Guadalupe, 2012), this setting enables me to isolate peer effects from the effects of common unobserved factors. If institutions influence their neighbors in voting, the shock to the voting behavior of these focal institutions should be transmitted to their neighbors.
Using a sample of shareholder-sponsored governance proposals at publicly traded financial institutions that pass or fail within ±2.5 percent around the majority threshold, I first examine the effect of close votes on the voting behavior of focal institutions (the own effect). Using a difference-in-differences specification, I find that, relative to publicly traded institutions that narrowly reject a shareholder-sponsored governance proposal, those that narrowly pass such a proposal become more likely to oppose management in their proxy votes. The economic magnitude of the own effect is non-trivial: For example, focal institutions that narrowly pass a shareholder-sponsored governance proposal become 2.5 to 2.8 percentage points more likely to oppose management after the vote than before the vote, relative to focal institutions that narrowly reject such a proposal. These results suggest that improved governance of a focal institution makes the institution more active in monitoring its portfolio companies.
If institutions condition their voting decisions on those of their neighboring institutions, the change in the voting behavior of focal institutions induced by close votes should lead to changes in the voting behavior of their neighbors. I thus examine the effect of close votes at focal institutions on the voting behavior of their neighbors (the peer effect). Importantly, the granular nature of the voting data enables me to identify peer effects within a neighboring institution’s portfolio at a given time. Since the passage of a governance proposal at a focal institution makes the institution more likely to oppose management, the neighbors of the focal institution that narrowly passes a governance proposal may become more likely to oppose management in stocks that are heavily exposed to the influence of the focal institution than in stocks that are not exposed. I use a triple-difference specification to test this prediction. The first difference is between neighbors of focal institutions that narrowly pass a close vote and those of focal institutions that narrowly fail a close vote (Pass). The second difference is from before to after the close vote (Post). The third difference is between stocks with high exposure to the influence of the focal institutions and those with low exposure (Exposure). I use two measures to capture the exposure to the influence of focal institutions. The first is the fractional ownership of the focal institutions in the stock, and the second is an indicator for whether the focal institution is a top 10 institutional shareholder in the stock. The triple-difference term (Pass × Post × Exposure) captures how close-vote outcomes (pass or fail) at focal institutions affect their neighbors’ votes in stocks that have high or low exposure to the focal institution after the close vote than before. If peer effects exist in institutions’ proxy voting behavior, the triple-difference term should be positive and significant.
I find evidence consistent with peer effects. Specifically, relative to the neighbors of a focal institution that rejects a close vote, the neighbors of a focal institution that passes a close vote become more likely to vote against the management of a firm after the close vote than before the vote, when the firm is held more heavily by the focal institution. The economic magnitude of this effect is significant. For instance, the coefficient on the triple-difference term is 1.2 to 1.8 percentage points when exposure is measured using the indicator for whether the focal institution is a top 10 institutional shareholder. This result is obtained after controlling for a host of fixed effects, which allows me to eliminate many potential sources of omitted variable bias that can confound inferences.
The peer effects in voting behavior at the institution level may have aggregate effects on actual vote outcomes. That is, a focal institution that just passes a governance vote may make its neighbors more active in monitoring the management of firms that are heavily held by the focal institution, thereby increasing the likelihood that the management of such firms loses votes on governance proposals. I find evidence consistent with this prediction. Specifically, relative to stocks that do not have a focal institution as a top 10 institutional shareholder, those that do experience an increase of around 3.5 to 4.8 percentage points in the likelihood that management loses a vote when the focal institution passes a close vote, relative to when it fails one, after the close vote than before. This result suggests that peer influence can create social multiplier effects and have large impacts on vote outcomes.
The findings on peer influences in proxy voting have important policy implications. The corporate voting mechanism is one of the most important and commonly used channels through which institutional investors influence corporate governance and decision-making. The amount of effort institutional shareholders put into voting decisions depends on the payoff to the decisions. If institutions vote independently without taking into consideration the voting decisions of other shareholders, they may underinvest in producing information that can help them improve their voting decisions because the probability that an institution casts a decisive vote is generally negligible. Therefore, regulations that aim to facilitate peer influence (e.g., increasing the ease of coordination among shareholders) are likely to enhance institutional shareholders’ incentive to produce information in corporate voting and improve corporate governance. Moreover, the findings that close votes at focal institutions lead to changes in the voting behavior of their neighbors and affect actual vote outcomes suggest a social multiplier effect in proxy voting. Shocks to the voting behavior of one shareholder can transmit to nearby shareholders, leading to large effects on governance outcomes.
Bebchuk, Lucian A., Alma Cohen, and Scott Hirst, 2017, The agency problems of institutional investors, Journal of Economic Perspectives 31, 89–112.
Cunat, Vicente, Mireia Gine, and Maria Guadalupe, 2012, The vote is cast: The effect of corporate governance on shareholder value, Journal of Finance 67, 1943–1977.
This post comes to us from Professor Jiekun Huang at the University of Illinois at Urbana-Champaign. It is based on his recent paper, “Thy Neighbor’s Vote: Peer Effects in Proxy Voting,” available here.