Empirical Analysis of the Impact of Universal Proxies on Proxy Contests

It seems fitting to talk about elections today, albeit of a very different kind.  Our memo on the SEC proposal to mandate universal proxies is here.

A paper by the Harvard Law School Program on Corporate Governance offers the first empirical analysis of proxy contests and the potential impact of universal proxies, and concludes that a universal proxy rule is unlikely to strongly favor either companies or dissidents.  In fact, it would have slightly favored management nominees if it had been used at the proxy contests reviewed in the paper.

Dissidents put forth competing slates at 0.5% of elections every year.  In more than 90% of situations, a settlement is reached.  Most proxy contests take place at very small companies.  In the last eight years, there were only 105 proxy contests at Russell 3000 companies compared to 241 overall.  The study focuses only on the results of proxy contests at large companies.

Although attempts have made by both sides, a universal proxy has never been used in a proxy fight at a major U.S. company.  The paper speculates that each time it has been suggested by either party, it has been mostly to gain a public relations advantage with investors.

A shareholder using a company card could not have voted for a dissident, and instead the best they could do if they preferred a dissident would be to withhold votes on management nominees.  Similarly, a shareholder using a dissident card that wanted to vote in favor of management nominees who are not named on that card could only withhold votes from those dissidents and perhaps those management nominees who are named on the dissident card.  As a result, the paper determines that 22% of proxy contests, or 17 contests, at large U.S. companies between 2008 and 2015 may have had what it calls “distorted outcomes,” meaning that another candidate may have been elected if universal proxies were used.

These distorted results come from a system where shareholders voting on a unilateral card would prefer to vote for a nominee not included on that card, where the only option they have is to vote to withhold from the nominees on the card that they are using.  A nominee’s election may be a distorted outcome depending on the number of votes withheld from nominees on the other side’s card.  For example, the number of votes for director candidate A, plus the number of votes withheld from opposing nominee B that would have otherwise gone to candidate A, is greater than the number of votes for candidate B plus the number of votes withheld from opposing nominees that would have gone to candidate B.  On that basis, the most likely distorted outcome is that the last-elected nominee (the nominee who received the lowest number of votes necessary to be elected) would be replaced by the nominee with the next highest number of votes who was not elected.

The paper points out two types of distortions.  There may be a distorted choice between the company and dissident sides elected in a proxy contest or there may be a distorted choice within the same side in a proxy contest.  Of the 17 contests examined, the study shows that 14 involved distorted choices between sides, and that eight favored dissidents.  This means, using the study’s assumptions, that a universal proxy rule would have instead favored management nominees in those cases.  The imbalance is slight, however, which leads the study to conclude that universal proxies are unlikely to overwhelmingly benefit one side or the other.  In 11 of the 17 contests, there may have been a distorted choice within sides, and a number of those seem to involve either the CEO or the chairman.

The study also does not believe that universal proxies are likely to result in many more contests where at least one dissident is elected, as the data shows that less than 5% of contests where no dissidents were elected may have had one elected if using universal proxies.  If the overall effect of universal proxies would have been largely balanced between the two sides, then, the study concludes, there should not be any increased frequency in the number of contests.  There could be an impact on how shareholders vote, if they are currently voting with a particular split between management and dissidents in mind.   Ultimately, a universal proxy rule may result in greater focus on the actual nominees, rather than on the competing management and dissident sides.