CII Letter to SEC on Proposed Amendments to Exemptions from the Proxy Rules for Proxy Voting Advice

Kenneth A. Bertsch is Executive Director and Jeffrey P. Mahoney is General Counsel at the Council of Institutional Investors. This post is based on their recent comment letter to the SEC in response to request for comments on the proposed rule regarding proxy advisors (discussed in posts here and here).

This letter focuses on claims by certain corporate representatives that there are pervasive factual inaccuracies in proxy advisors’ reports, claims that we believe were relied on in the Release and in the decision of a majority of SEC commissioners to support proposing a new regulatory regime for independent proxy advisors. We believe that the claims of pervasive errors are unfounded and misleading and do not provide a basis for rulemaking. As CII and other investor organizations and various investors have indicated, the paucity of evidence of pervasive factual errors by proxy advisors suggests that, in fact, no regulatory intervention is necessary or justified.

Shortly after the Commission issued the Release, CII requested that the SEC provide the underlying data for Table 2 in the Release (“Table 2”), which classifies what the table describes as company “concerns” on proxy advice drawn from certain supplemental proxy materials (DEFA14As) filed in the period 2016 to 2018…. These company “concerns” compiled in Table 2 clearly played a significant role in the SEC’s development and consideration of the Amendments, and therefore we believe access to the data underlying the concerns is important for meaningful public comment. We appreciate that on January 16, 2020, shortly before the February 3 close of the comment period, the SEC’s Division of Economic and Risk Analysis added to the comment file a memo providing some additional information (the “DERA Memo”). But it …. did not provide critical information that would be necessary to understand why the Commission believes its proposal is justified, such as the underlying data set that would tie the categories in Table 2 back to the underlying company statements that Table 2 purports to count. Any review or analysis of the table’s merits as a basis for rulemaking would require such information.

In the absence of the underlying data, we have done our best to describe in this letter some of the questions we have about the company “concerns” listed in Table 2 as well as our own concerns about Table 2. Specifically, we believe Table 2 does not demonstrate a high rate of errors. Only one of the categories provided in Table 2 even purports to count factual errors. It is a judgmental count, without underlying support to assess whether it is a fair judgment. It is also an extremely low number that does not justify rulemaking, especially rulemaking that presents a significant risk of promoting error by exacerbating time pressure on research and development of advice.

[We] have found that:

  • Most of the purported “concerns” actually are policy disputes or communications that a company has changed a proposal or is providing more or clarified information to address an issue raised in a proxy advisor report.
  • The number of claimed inaccuracies is a very small: 0.3 percent.
  • Matters classified in Table 2 as “Analytical errors” are disagreements on methodology, far from “factual errors” and not even mathematical errors.
  • Some of the filings cited by the SEC do not actually express any “concerns.” g., in some cases, a company merely disclosed additional information or amended its proposal, without criticizing a proxy advisor.
  • Company expression of concern about proxy advice would reflect personal self-interest. 67 of the filings covered by Table 2 made arguments on a question in which the CEO had a clear personal interest in the outcome. In many of those cases, the CEO had a direct financial interest.
  • Finally, some of the claims appear to be incorrect and/or misleading in their own right.

. . . .

In the Release, the SEC proposed. . . . Amendments [that would] impose a period for review of reports and feedback by company management ( “Review and Feedback Period”), an additional period during which company management is provided with the final report, analysis and voting recommendations ( “Final Notice Period”), and provision by the proxy advisor of a hyperlink to management’s statement on the advice ( “Hyperlinked Statement”).

The compelled Review and Feedback Period, Final Notice Period and Hyperlinked Statement provisions are largely premised on an assumed (but not substantiated) high rate of factual errors and methodical weaknesses in proxy voting advice.

. . . .

In fact, the evidence suggests the rate of factual errors in proxy advice is extremely low, and the mechanisms that proxy advisors have in place to correct any such errors are prompt and effective. Proxy advisors have strong incentives to provide clients accurate, high quality advice, and the absence of significant errors shows those incentives are working.

. . . .

Table 2 Analysis

CII tried but was unable to replicate the SEC analysis of management concerns as expressed in 2018 Forms DEFA14A.

. . . .

Table: Summary of Analysis on Table 2 Data for 2018

Company Claims on Proxy Advisors in 2018 DEFA14As
Number of registrants Number of filings Number of filings expressing “concerns” Factual errors Analytical errors General or policy dispute(s)* Amended or modified proposal Other Additional disclosure**
SEC 78 84 84 17 28 58 6 2 NA
CII 78 84 71 7 13 50 7 9 21

Both the SEC and CII classified some supplemental filings as fitting in more than one category.

* In Table 2, the SEC labels “General or policy dispute” as a singular, while in its definition, it refers to “General or policy disputes” (that is, plural).

** SEC does not use this category.

. . . .

The SEC defined its categories as follows:

  1. Factual errors: “when the registrant identifies what it considers to be incorrect data or inaccurate facts that the proxy voting advice business uses in some part as a basis for its negative recommendation”
  2. Analytical errors: “when the registrant identifies what it considers to be methodological errors in the proxy voting advice business’s analysis that it used as a basis for its negative recommendation”
  3. General or policy disputes: “when the registrant does not dispute the facts or the analytical methodology employed but instead generally espouses the view that specific evaluation policies or the evaluation framework established by the proxy voting advice business are overly simplistic or restrictive and do not adequately or holistically capture the merits of the proposal”
  4. Amended or modified proposal: “when the registrant responds to a current or prior year negative recommendation from a proxy voting advice business by indicating that it has amended or modified proposals or existing governance practices prior to the annual meeting and requests investor consideration of these facts in making their vote”
  5. Other: “where the registrant objects to the proxy voting advice business’s negative recommendation but does not specifically cite nor respond to the rationale for the negative recommendation and instead makes a generalized argument in favor of the proposal” (emphasis added)

. . . .

CII added a sixth category not used by the SEC: that the company provides “Additional disclosure.” Provision of additional disclosure occurred with some frequency, at least in comparison with total number of supplemental filings cited by the SEC. We believe the DERA categories appear to be unsatisfactory in several respects, but particularly in not noting that some filings simply provide additional information. It is conceivable that the SEC considers a company’s provision of additional information to be modification of a “practice,” and therefore within the SEC’s definition of “Amended or modified proposal.” But this is not clear in the SEC’s description of Table 2. We designated this category (sometimes in addition to one or more other category) where provision of additional information was a key aspect of the registrant filing, intended to help satisfy concerns raised in one or more proxy advisor reports. In 10 of the filings identified by the DERA Memo, we see additional disclosure but no other “concerns” that fit in any of the SEC’s categories.

We have these observations:

  • Most of the purported “concerns” actually are policy disputes or communications that a company has changed a proposal (or governance practice) or is providing more or clarified information to address an issue raised in a proxy advisor report. We would expect complicated questions of analysis and opinion to involve differing approaches to analysis and differing opinions. The only categories used by the SEC that appear to be relevant for purposes of the Commission’s regulation based on frequency of factual errors are, at most, the categories on company assertions of “factual errors” and “analytical errors.”
  • The number of claimed inaccuracies is very small: If we try to use the SEC’s classifications to the extent information is available, we count 57 ISS and Glass Lewis reports criticized for either “factual error” or “analytical error” in 2018, out of more than 11,000 published by those firms in the year on U.S. registrants. This would constitute an “error rate” on a report basis of 0.5%. And this appears to be exaggerated, both because some of these company assertions appear to be in error, and because the SEC’s counts for these categories appear high. We count 26 filings that claim error in 38 reports, an “error rate” of 0.3% on a report basis.
  • “Analytical errors” are disagreements on methodology, not “factual errors.” The SEC defines “analytical error” as “methodological error,” and does not seem to include in the “analytical error” category actual analytical mistakes, such as math errors. (We assume therefore that the SEC classified the latter as “factual errors.”) Institutional investors retain proxy advisors to do their own analysis, relying on their own methodologies, not to simply validate every company according to each company management’s preferred methodology for evaluating its pay, or other matters. The SEC identifies only 17 filings that claim “factual errors.” Below, we identify the 12 filings that the SEC was most likely to have found as asserting factual error, but we think no more than seven actually do so (0.06%). And the company assertions of factual errors appear to be incorrect in most of those seven filings.
  • Some of the filings cited by the SEC do not actually express any “concerns.” Table 2 indicated that issuers expressed “concerns” on proxy advice in 84 filings in 2018. The data file linked with the later DERA Memo conceded that one of these was an error by DERA (in that filing, the company cited Glass Lewis support for management in a proxy fight, which in no way was a management complaint about Glass Lewis). In at least 12 other filings, the SEC appears to have attributed to the company, which usually provided additional information, expression of “concern” that was not actually expressed by the company in the filing. In a few other filings, one could impute concern. Most of these amended proposals or provided more disclosure.
  • Company expression of concern about proxy advice would reflect personal self-interest. CII identified 67 filings that made arguments on a question in which the CEO had a clear personal interest in the outcome. In many of those cases, the CEO had a direct financial interest. In all 67 of those filings, the company argued in favor of the position that served the CEO’s personal interest. In no cases (zero) did the company argue against the CEO’s interest. This is relevant in that the SEC is seeking to compel proxy advisors to show their work to management of the companies that are subject of the research, so those companies can provide “feedback.” From the precedent offered by supplemental proxy filings, it appears that the “feedback” provided will be one-sided if the CEO’s interests are at stake. It is unrealistic to believe that company management will correct “errors” or critique methodologies where the errors or misguided methodologies serve the interests of the CEO.
  • Finally, some of the claims appear to be incorrect and/or misleading. [These are discussed specifically in the full letter.]

. . . .

In conclusion, we would emphasize that the SEC’s own evidence suggests the rate of factual errors in proxy advice is extremely low, and the mechanisms that proxy advisors have in place to correct any such errors are prompt and effective. Proxy advisors have strong incentives, through the market for their research and analysis, to provide clients accurate, high quality advice, and the absence of significant errors shows those incentives are working.

SEC staff involved in the rulemaking have suggested, in meetings with us, that what matters most is that the regulatory approach ensure the lowest possible incidence of error, whatever the actual rate of errors in recent reports. This implicitly is an argument for a “perfection standard,” which is neither economically justified nor likely to be attainable given the nature of the requirements the SEC has proposed

In our view, the SEC staff’s suggestion that the point is to reduce errors, no matter how low their present frequency, is a fallback argument that belies awareness that error rates are low and that many of the complaints are, instead, differences of opinion and preferences on methodologies. As others have said, this makes the Amendments a clear example of a solution in search of a problem. But as important, there is no evidence that the proposal is calculated to attain a lower rate of error.

Rather, what the proposal seems designed to do is ensure that management will “be comfortable” that its perspectives will be reflected in the proxy advice that investors procure. That is the point of the hyperlink, which the Release’s economic analysis treats as indispensable, albeit disruptive to the proxy advisors. We submit that the goal of management comfort is both itself unattainable and inappropriate, given that investors privately order the advice specifically to obtain independent, critical analyses of management proposals, just as they might have sought advice from independent, buy-side stock analysts before purchasing their shares and might read independent financial analyst reports when considering whether to continue to hold the shares. Moreover, the Release ignores strong reasons to expect that the Commission’s proposals will lead to more errors, and lower quality proxy advice—even setting aside clear costs from jeopardizing the independence of advice and introducing major new conflicts of interest. With the time for proxy voting decisions already compressed and subject to very large seasonal burdens, the SEC Review and Feedback and Final Notice requirements will subtract substantially from the time that proxy advisors and/or investor clients have to review proxy voting research, analysis and decisions.

That said, we think that the argument for new, costly and intrusive regulation that compromises free speech rights must, at a minimum, be based on evidence beyond a contention that proxy voting advice is not perfect or error-free.

The complete letter, including footnotes, is available here.

Both comments and trackbacks are currently closed.