Proxy access is the ability of certain privileged shareholders to have their own slate of director nominees included in the company’s proxy materials whether or not the board of directors (“Board”) approves. These materials include a proxy statement used to solicit shareholder votes and a voting card allowing shareholders to vote without having to attend the annual meeting.11.For a legal history of proxy access, see Bernard S. Sharfman, What Theory and the Empirical Evidence Tell Us about Proxy Access, 12 J.L. Econ. & Pol'y (forthcoming 2016). http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2757761. For many years, the default rules of corporate and securities law have provided the Board with exclusive authority to decide whether shareholder proposals seeking to implement proxy access are to be included in a public company’s proxy solicitation materials. Five years ago the Securities and Exchange Commission (SEC) amended its rules to require these proposals be included.22.17 C.F.R. §240.14a-8(i)(8) (2011).
Because of the difficulty of crafting a binding proxy access bylaw within the confines of the SEC’s 500 word limit on shareholder proposals,33.17 C.F.R. §240.14a-8(d) (2011). proposals are usually non-binding, requesting, not requiring, the Board to implement proxy access by amending the company’s governing documents. These proposals can be understood as the first step in the process of implementing proxy access on a company-by-company basis.
Roughly 200 companies received proxy-access proposals in 2016.44.Sidley Austin LLP, Proxy Access Update – Momentum Continues to Build in 2016 4 (2016), http://www.sidley.com/~/media/update-pdfs/2016/09/proxy-access-momentum-in-2016--september-22-2016.pdf. The proposals usually limit the availability of proxy access to large shareholders who have held at least three percent of company shares, individually or as an aggregation of 20 to 25 investors, for at least three years.
When voting on a proxy access proposal, shareholders need to be informed about the expected effect of proxy access on the market value of their shares. Boards also need to be informed about this expected change in value when considering if it should amend its governing documents to include proxy access, either for purposes of preempting a shareholder vote or considering its implementation subsequent to such a vote at the annual meeting. The SEC needs to be informed about the expected change in value on a market-wide basis prior to making any changes to its proxy access rules, including putting back on its agenda the idea of universal proxy access for all public companies (“universal proxy access”).55.Universal proxy access would automatically allow certain privileged shareholders to place their Board nominees into the proxy solicitation materials of almost all public companies without the need for a charter amendment or bylaw. The SEC adopted a universal proxy access rule that was to become effective on November 15, 2010. Prior to it being implemented, the D.C. Circuit Court of Appeals unanimously decided to vacate the rule after determining that the SEC had promulgated the rule in violation of the Administrative Procedure Act’s “arbitrary and capricious
One way to understand the value of proxy access is through empirical analysis of the shareholder proposals on proxy access that have already been submitted for inclusion in the proxy materials of public companies. Unfortunately, the empirical research on these proposals is limited to one empirical study. This study, even though well executed, leads to more questions than answers and thus cannot be relied upon as authority on a standalone basis. This is a critical point that shareholders, board members, and the SEC need to understand when such empirical evidence is used in support of or against proxy access.
I. The Bhandari Report
The available empirical study is a report prepared by Tara Bhandari, Peter Ilievy, and Jonathan Kalodimos.66.Tara Bhandari, et al., Public versus Private Provision of Governance: The Case of Proxy Access, (SEC Staff Working Paper, 2015), http://www.sec.gov/dera/staff-papers/working-papers/public-vs-private-provision-of-governance.pdf. For a review of empirical studies on universal proxy access, see Sharfman, supra note 1. The report was initiated when all three were employees of the SEC’s Division of Economic and Risk Analysis.77.Illievy and Kalodimos are no longer with the SEC. This report took the form of an “event study.” An event study investigates the impact of new information upon the expected stock returns of a targeted cross-section of firms.88.Roberta Romano, Less is More: Making Shareholder Activism a Valuable Mechanism of Corporate Governance, 18 Yale J. on Reg. 174, 187 n.37 (2001). In the report, the event was the Office of the Comptroller of New York City’s (“Comptroller”), the custodian and investment adviser to the New York City Pension Funds, unexpected announcement to the public that it had simultaneously submitted non-binding proxy access proposals to 75 public companies.99.Press Release, Office of the New York City Comptroller, Comptroller Stringer, NYC Pension Funds Launch National Campaign to Give Shareowners a True Voice in How Corporate Boards Are Elected (Nov. 6, 2014), http://comptroller.nyc.gov/newsroom/comptroller-stringer-nyc-pension-funds-launch-national-campaign-to-give-shareowners-a-true-voice-in-how-corporate-boards-are-elected/.
An event study is used to determine “whether there is an abnormal stock price effect associated with an unanticipated event”1010.S.V.D. Nageswara Rao and Sreejith U, Event Study Methodology: A Critical Review, 3(1)A The Macrotheme Rev. 40, 44 (Spring 2014). (the Comptroller’s announcement) on a sample of firms that may have been uniquely affected by the event (the 75 firms to which the Comptroller simultaneously submitted proposals). The null hypothesis to be tested is whether the mean abnormal return (abnormal stock price effect on the targeted sample of firms) at the time of the event is equal to zero. That is, if there was no effect from the announcement, then the mean abnormal return at the time of the event will equal zero.1111.S. P Khotari & Jerold B. Warner, Chapter 1 – Econometrics of Event Studies, in Handbook of Empirical Corporate Finance 3 (B. Espen Eckbo ed., 2007). The event date was November 6, 2014. The authors found that the Comptroller’s announcement led to a positive, statistically significant, 0.53% abnormal return for the 70 firms1212.Five firms were removed from the sample because they had made earnings announcements on that day. used in their sample. In terms of hypothesis testing, the results meant that the null hypothesis had been rejected.1313.The Bhandari report was not exclusively focused on the Comptroller’s announcement. In total, it evaluated 158 proxy access proposals, including the Comptroller’s 75 proposals, at 133 firms over four proxy seasons. Bhandari, supra note 6, at 14. Moreover, they interestingly found a strong correlation between the returns generated on this event date and the returns of the sample on the date, approximately four years earlier,1414.The stay date was October 4, 2010. See Bhandari, supra note 6, at 11. when the SEC announced it was going to stay the implementation of its universal proxy access rule.1515.Bhandari, supra note 6, at 19.
II. Selection Bias and a Lack of Randomness
Even though the Bhandari report indicates that proxy access has value, this is far from the end of the story. The small sample size makes it very difficult to make inferences about why the Comptroller’s announcement had such a significant impact on the target firms. The sample cannot be further broken up to see if certain sub-groups are responsible for moving the numbers.
Moreover, the sample lacks randomness as a result of selection bias. Randomness means that each element of a population has an equal chance of being part of the sample. A random sample is required in order to make generalized claims about how the entire population of U.S. public firms would be affected by shareholder proposals on proxy access (i.e. external validity). The Comptroller’s selection process violated the requirement of randomness and, therefore, the results lack external validity.
The 75 companies were targeted for multiple reasons unrelated to enhancing shareholder value. Thirty-three were targeted because they were in industries directly related to climate change; 24 for a lack of board diversity; and 25 were cited for having received “significant opposition to their 2014 advisory vote on executive compensation.”1616.Sumberg, supra note 9. This resulted in 20 of the 75 target firms being from the gas and oil industry, nine from the utilities industry and another six from the coal industry.1717.Bhandari, supra note 6, at 43, tbl. 3. Such a weighting of companies either producing or consuming huge quantities of carbon-based fuels is not representative of the current universe of U.S. public companies.1818.Sharfman, supra note 1, at 15.
It is also reasonable to assume that the selection process was a function of how successful the Comptroller expected to be in either getting firms to implement proxy access prior to a shareholder vote or at least getting a substantial percentage of votes if a shareholder vote took place. The Comptroller would not have wasted its time selecting a firm where the expected probability of success was zero or close to it. According to Nell Minow, a leader in the shareholder empowerment movement, the Comptroller has “been very smart about picking companies where shareholders are looking to make a change.”1919.Jena McGregor, ExxonMobil Shareholders Just Approved a Powerful New Measure That Could Reshape investors’ Influence on Company Boards, Wash. Post, May 25, 2016, https://www.washingtonpost.com/news/on-leadership/wp/2016/05/25/big-investors-are-getting-a-powerful-new-right-in-record-numbers-if-they-ever-use-it/.
However, the selection bias discussed so far does not entirely explain how the Comptroller whittled down the number of targeted firms to 75 out of the over 3,000 public firms that it could choose from. It is reasonable to assume that it also targeted firms that had not been historically responsive to its engagement or the engagement of other like-minded shareholder activists on issues including board diversity, executive compensation, climate change, disclosure of political contributions, employee wages, etc. According to the Comptroller, “to effect true change, you need the ability to hold entrenched and unresponsive boards accountable and that is what we are seeking to do.”2020.Sumberg, supra note 9. Therefore, an additional targeting criterion may have been firms that had not adequately cooperated on one or more of these other issues. For those firms that had been cooperative, it would be counterproductive for the Comptroller to target them for proxy access. This additional criterion would have created more bias in the sample.
One counterargument is that the Comptroller’s sample was random with respect to the expected value of proxy access. Ironically, as described above, this may be true to some extent given that the Comptroller was not targeting firms based solely on the criterion of enhancing shareholder value. However, the abnormal returns found in the Bhandari report did not measure the value of proxy access per se, but the expected returns of proxy access as a function of both the market’s estimation of its value at a target firm (positive or negative) and the probability that proxy access could actually be achieved at the firm. Selection bias with respect to the second variable may have resulted in a lack of external validity.
As argued above, the Comptroller would have targeted firms where it believed it would have success, i.e. firms with dissatisfied investor bases, making the probability of success higher than it would be if the target firms were selected in a random fashion. Moreover, it is possible that the two variables are not independent, but are positively correlated. In other words, the greater the level of investor dissatisfaction means not only the higher probability of success but also the greater the likelihood the market will find the value of proxy access to be positive. If so, then the Comptroller’s selection process will yield more companies that the market feels will benefit from proxy access versus a randomly selected sample.
In sum, the Comptroller’s selection process excluded a vast sector of the universe of public companies. This adversely affected the ability of the analysis to accurately represent the expected benefits or costs of proxy access to all public companies, making the results biased, most likely in the upward direction. The study thus lacks external validity outside the boundaries of the Comptroller’s selection criteria.2121.Aswath Damodaran, Investment Valuation: Tools and Techniques for Determining the Value of Any Asset 121 (3d ed. 2012) (When a sample is “random, this does limited damage to the results of the study. If the choice is biased, it can provide results which are not true in the larger universe.”). The results of the Bhandari report may be able to inform us about how proxy access may have affected the firms in the small sample under study, but there is great uncertainty if it can be generalized to the three thousand plus other firms that also make up the universe of public companies.
III. Omitted Variable Bias
Even if the Bhandari report lacks external validity, one result that is still extremely interesting is the finding that the Comptroller’s announcement had such a large impact on the value of the target sample, a 0.53% average abnormal return. This result is puzzling given the proposals were non-binding and uncertainty existed over whether they would win approval by shareholders or be implemented by the Board even after shareholder approval. Moreover, there was uncertainty whether shareholders had the wherewithal or even desire to ever use their right to nominate if implemented, and if they did use their right to nominate, if any of their nominees would actually win election.
There are several other reasons why the result is perplexing. First, the proposals effectively excluded activist hedge funds from participating in proxy access because of the required three-year holding period.2222.Bernard S. Sharfman, Activist Hedge Funds in a World of Board Independence: Creators or Destroyers of Long-Term Value?, 2015 Colum. Bus. L. Rev. 813, 825 (2015). Second, the Bhandari report found that the Comptroller was not specifically targeting poor performing firms that could benefit the most from proxy access.2323.Bhandari, supra note 6, at 28. Third, the study controlled for abnormal returns generated by the industries where the target firms belonged.2424.Id. at 15 (“We control for industry in all of our tests.”). But most importantly, proxy access does not exist in isolation from the markets for corporate control (friendly and hostile takeovers through mergers and acquisitions)2525.Henry G. Manne, Mergers and the Market for Corporate Control, 73 J. Pol. Econ. 110 (1965). and influence (shareholder activism including hedge fund activism),2626.See generally Brian R. Cheffins & John Armour, The Past, Present, and Future of Shareholder Activism by Hedge Funds, 37 J. Corp. L. 51, 58 (2011); Paul Rose & Bernard S. Sharfman, Shareholder Activism as a Corrective Mechanism in Corporate Governance, 2014 BYU L. Rev. 1014 (2015). the primary means by which board members are replaced outside of board nominating committees.
In the market for corporate control, Doidge, Karolyi, and Stulz report that from 1997 to 2012, 4,957 firms were delisted from U.S. stock exchanges as a result of merger activity.2727.Craig Doidge et al., The U.S. Listing Gap 5 (Nat’l Bureau of Econ. Research, Working Paper No. 21181, 2015). This activity must have resulted in thousands of Board members losing their seats. In the market for corporate influence, shareholders are already getting significant board representation through engagement with the Board. From 2006 to 2013, a total of 1,128 dissident seats were granted to shareholders either through a proxy contest or private negotiation, with 179 in 2013 alone.2828.Shane Goodwin, Myopic Investor Myth Debunked: The Long-Term Efficacy of Shareholder Advocacy in the Boardroom 51, tbl. 1 (Harvard Bus. Sch., Working Paper, 2014), http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2450214. Moreover, of those 1,128 seats granted, 702 seats were gained through hedge fund activism.2929.Id. at 52.
Given the more powerful means by which to change the composition of a Board, this should put a significant cap on the value of proxy access as a means to reduce agency costs caused by the separation of share ownership from board management. In sum, the 0.53% average abnormal return found in the Bhandari report is counter-intuitive.
A possible explanation is that one or more independent variables, not specified in the event study’s regression equation, may be causing the abnormal return. If so, then there may be omitted variables that are correlated with both a company receiving a proxy access proposal from the Comptroller and the abnormal returns generated by the shares of the target firms on the event date.
So, what could these omitted or missing independent variables be if they indeed exist? For one, such a variable would describe the level of dissatisfaction the company’s shareholder investor base currently has with the Board and/or executive management. The identification of a dissatisfied shareholder base is critical to the workings of those who participate in the market for corporate control (takeovers) and hedge fund activism.<3030.See Damien Park, How Activist Investors Identify Their Targets, Director Notes (Conference Bd., N.Y.), June 2016, at 3, fig. 3.
If correct, then we should interpret the appearance of a proxy access proposal as a new or confirming signal to the market that there is a high level of shareholder dissatisfaction with the Board. Proxy access, unlike other shareholder proposals, makes a compelling statement that large activist institutional investors are extremely dissatisfied with the Board and would be happy to see a change.
Such a proposal is a clear signal to the market that the Board may be vulnerable to hedge fund activism or a takeover (friendly or hostile), especially when the stock price has been under pressure.3131.See generally Bernard S. Sharfman, A Theory of Shareholder Activism and its Place in Corporate Law, 82 Tenn. L. Rev. 791 (2015); Bernard S. Sharfman, The Tension Between Hedge Fund Activism and Corporate Law, J.L. Econ. & Pol’y (forthcoming 2016), http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2722124 (exploring the connections between proxy access proposals and market perception). In essence, the company has been put “in play.” This is valuable information for the market in its process of continually reevaluating the price of a company’s shares. We know from recent research on hedge fund activism that the up-front gains in a company’s stock price from such activism can be extremely rewarding to shareholders.3232.See, e.g., Lucian A. Bebchuk et al., The Long-Term Effects of Hedge Fund Activism, 115 Colum. L. Rev. 1085 (2015); Nicole M. Boyson & Robert M. Mooradian, Corporate Governance and Hedge Fund Activism, 14 Rev. Derivatives Res. 169, 175–78, 201 (2011); Alon Brav et al., Hedge Fund Activism, Corporate Governance, and Firm Performance, 63 J. Fin. 1729, 1731 (2008); Christopher P. Clifford, Value Creation or Destruction? Hedge Funds as Shareholder Activists, 14 J. Corp. Fin. 323, 324 (2008); Robin M. Greenwood & Michael Schor, Investor Activism and Takeovers, 92 J. Fin. Econ. 362, 374 (2009); April Klein & Emanuel Zur, Entrepreneurial Shareholder Activism: Hedge Funds and Other Private Investors, 64 J. Fin. 187, 217–18 (2009). Therefore, the increased potential for hedge fund activism or acquisition activity may be the true drivers of the abnormal returns found in the Bhandari report, not the market’s estimation of the value of proxy access as a stand-alone tool for enhancing corporate governance.
A counterargument is that the report’s finding of a strong correlation between the returns of the target firms on the date that the Comptroller announced its proxy access initiative and the returns that the target firms yielded on the date the SEC stayed its universal proxy access rule confirms the primary role of proxy access as being the cause in the change in value. Yet, this interesting finding does not negate the potential for omitted variables as an explanation for the counter-intuitive results. The potential for omitted variables still needs to be researched. If such a variable is found, then the strong correlation discussed above is just that: a correlation between two events, and only two events, that occurred four years apart, and no more.
In sum, the mean abnormal returns are much too high to be explained simply by the disclosure of the Comptroller’s proxy access proposals. Proxy access is just a very small part of the story of how Board composition is influenced by market forces. The potential for omitted variables is great and needs to be explored in future empirical studies.
The Bhandari report, which focused on one event at one point in time, must also be understood in the context of non-stationarity: the potential for the stock market to react differently to the same events at different points in time.3333.See Rao & Sreejith U, supra note 10. If non-stationarity exists, then the stock market may provide “one result for a period and a diverse outcome for another period” as the perception of investors change over time.3434.Id. This is consistent with an efficient market where the market price is an unbiased estimate of the true value of the investment, but is not necessarily a correct one at any point in time.3535.See Damodaran, supra note 21, at 112.
It is easy to see how non-stationarity may play a role in the results of future event studies on proxy access. At this time, the stock market has zero practical experience with proxy access. Investors have yet to use proxy access to nominate candidates for the Board. Therefore, there is no data to evaluate how the performances of those nominees who have been elected to the Board have affected shareholder value. As a result, it is possible that as the market becomes more informed about the real value of proxy access, future event studies, including studies on the value of shareholder proposals on proxy access, may provide different results based on changed perceptions.3636.See Yaniv Konchitchki & Daniel E. O'Leary, Event Study Methodologies in Information Systems Research, 12 Int’l J. of Acct. Info. Systems 99, 108 (2011).
To overcome the perception that the Bhandari report may be tainted by the potential for non-stationarity, a number of event studies would need to be conducted on various event dates over a number of years. Hopefully, they will generate results that are consistent. Until then, the issue of non-stationarity will need to be acknowledged by those who utilize the Bhandari report.
The Bhandari report is an important first step in the process of trying to understand the value of proxy access based on shareholder proposals. However, even though the authors appear to have done the best job possible with a limited data set, it is not possible to use the report as support for the proposition that proxy access is an enhancement to the corporate governance of a public company, either generally or at a targeted company. More specifically, the results of the report lack external validity resulting from a sample that is not randomly generated; there is the strong possibility of omitted error bias; and the issue of non-stationarity limits the significance of the results.
Bernard S. Sharfman is an associate fellow of the R Street Institute and a member of the Journal of Corporation Law’s editorial advisory board. Mr. Sharfman would like to thank Jonathan Cohn, Shane C. Goodwin, John G. Matsusaka, and Tara Bhandari for their helpful comments and suggestions. Mr. Sharfman is dedicating this article to his wife, Susan David, and his daughter, Amy Sharfman.