Are you ever surprised that more companies don’t fail their say-on-pay votes? Say on pay was adopted by the SEC under a Dodd-Frank mandate signed into law against the backdrop of the 2008 financial crisis. The mandate was enacted largely in reaction to the public’s railing against the runaway levels of compensation paid to some corporate executives despite poor performance by their companies, especially when those firms were viewed as contributors to the crisis itself. Say on pay was expected to help rein in excessive levels of compensation and, even though the vote was advisory only, ascribe some level of accountability to boards and compensation committees that set executive compensation levels. But, while say on pay may have driven more investor engagement—certainly a good thing—the anticipated say-on-pay challenge by shareholders to out-of-line pay packages did not really materialize. From the get-go, the average failure rate has only been about 2%. Instead, say-on-pay votes have served largely as confirmations of board decisions regarding executive compensation and not, in most cases, as the kind of rock-throwing exercises that many companies had feared and some governance activists had hoped. According to a 2011 Business Week article, Robert A.G. Monks, who founded ISS in 1985, concluded that say on pay was “‘at best a diversion and at worst a deception….You only have the appearance of reform, and it’s a cruel hoax.’” This paper, Failed Say on Pay: How Do Companies Course Correct after to a ‘No’ Vote?, from the Rock Center for Corporate Governance at Stanford University, with authors from Stanford and Equilar, looked at the 2% that fail the vote and what they did in response to pass muster with investors. But the underlying message is reflected in the authors’ questions: “Does this process reflect a healthy dynamic of the market correcting egregious practices, or does it simply reflect a standardization process whereby observed outlier practices are brought in line with industry norms? Do the changes companies make in response to a failed vote lead to substantive improvement in the managerial incentives of their pay programs?”
The authors report that, since the rule went into effect, “companies in the Russell 3000 Index received average support of 91 percent for their pay programs,” with that percentage “fluctuating narrowly between a low of 89.2 percent (in 2022) and a high of 91.7 percent (in 2017). Meanwhile, the annual failure rate (companies receiving less than 50 support) averaged a mere 2 percent.” (Note that the policies of major proxy advisory firms set thresholds for weak say-on-pay votes—and reviews of comp committee responses—at 70% for ISS and 80% for Glass Lewis.) For that 2%, however, the board usually takes some heat and determines to make changes to plan design and disclosure going forward—according to Equilar, an average of 2.5 changes. The changes are designed not only to “bring pay packages more in line with shareholder preferences,” but also to “reflect the preferences of proxy advisory firms,” which, research has shown, wield significant influence over the vote.
In the study, the authors looked at a sample of the 77 companies in the Russell 3000 that received less than 50% support for their say-on-pay proposals in 2022. The authors point out that the rejections of the proposals were resounding, with average support of only 35%. As described, companies in the sample were “highly diverse” with regard to revenue (average sales of $8B, ranging from zero to $131B), market cap (average of $23.5B, ranging from $213M to a high of $468B), and prominence, including both highly recognized name brands and relatively unknown names. The authors examined how these companies sought to course correct in 2023 after their failed votes: “Do they primarily reduce pay levels, or do they make ancillary changes to plan design (performance targets, measurement periods, and mix) without committing to a lower overall level? Are these changes substantive (i.e., lead to a closer alignment between pay and performance), and do shareholders subsequently approve of them?”
What was the reason these companies received failed votes? The paper concludes that high pay levels were probably “a main determinant of the voting results,” with stock-price performance likely making a modest contribution to the negative votes. The authors identified grants of “large one-time special award[s] to the CEO” as the “leading factor triggering a failed say-on-pay vote.” In the sample, the CEO pay levels were deemed high (average CEO comp was $23.1 million; median $16.8 million) compared to the Russell 3000 median of $6.2 million and even compared to the amounts the same CEOs received the previous year (average CEO comp was $14.2 million; median $6.7 million). (Note that the authors acknowledge that this is a relatively “rough comparison” and that a more fair comparison “would measure each CEO’s pay against that of a comparably sized peer in the same industry.) With regard to stock price performance, the authors report that the companies in the sample had a total shareholder return of only 4%, while, in the same year, the Russell 3000 increased by 26%.
Another pivotal factor appears to be the influence of the proxy advisory firms. According to the paper, in 2022, ISS recommended against all but one of the proposals that failed, with average support of only 36%. The authors note that, where ISS had given a favorable recommendation in the prior year but changed to a negative recommendation, “say-on-pay support fell an average of 54 percentage points year over year, compared with only a 15 percent reduction in support when ISS maintained a negative recommendation both years.”
In response to the failed votes, companies typically increased their proxy disclosure, describing the extent of their engagement with institutional investors and proxy advisors, the feedback they received and reasons offered for the negative votes. According to the authors, among the reasons, 30% criticized the features of a special award, 18% cited as a negative factor “too little (or no) performance-based awards in the long-term incentive program,” and 13% cited high overall pay levels. Ten percent took issue with “a discretionary action the company took to award a bonus payment (short- or long-term) when it would not otherwise have been merited.” To address the criticisms, the authors found that companies made an average of 2.5 changes, with 66% changing the performance metrics or weightings in their incentive programs, 36% increasing disclosure, 19% changing the performance measurement period and 18% reducing overall pay, among other changes. The authors raise the question of whether these changes “represent an honest effort by compensation committees to improve poorly structured incentives, or are they cosmetic moves to appease shareholders and proxy advisory firms?”
What was the impact? The study found that, although vote outcomes improved substantially to the levels achieved prior to the failed vote (average support increasing from 35% to 76%), those outcomes were still well below average support (90%) in the Russell 3000. Curiously, the authors found some correlation between the level of support and the number of changes made: companies that made two or three changes received more support than those that made only one or none. Making more than four changes did not seem to help much more. However, regardless of the number of changes, if those changes were sufficiently “in line with ISS criteria” to convince the proxy advisory firms to make positive recommendations, the companies “almost uniformly experienced larger say-on-pay increases” than those that did not—a finding consistent with prior “research showing that proxy advisory firm recommendations contribute to standardization in CEO compensation.” The authors also raise the issue of whether proxy advisory firms can even adequately digest and analyze all the data regarding pay structure for thousands of companies to identify those companies with pay practices that don’t measure up. Whether standardized compensation structures are good or bad for shareholders remains to be seen. The authors ask whether “the companies that fail their say-on-pay votes the most egregious offenders, or ‘unfairly’ caught up in somewhat arbitrary compensation guidelines?”
At the end of the day, did the say-on-pay vote have much impact? The authors ask whether there is “any evidence that say on pay or proxy advisor recommendations on say on pay create better managerial incentives and increased shareholder value?” Apparently, the authors don’t seem to think so. Most notably, the authors conclude that, notwithstanding “a strong rebuke from shareholders and making significant changes to pay structure and disclosure, companies overall did not substantially reduce pay,” especially in comparison to the pay levels in the year before the failed vote. While average compensation declined to $11.8 million from $23.1 million, it was still “well above Russell 3000 averages ($6.2 million),” with the median pay of $10.4 million still about 50% higher than the $6.7 million in the year prior to the vote. In addition, the authors attribute the decline primarily to a reduction in long-term equity awards. Nor did shareholder return increase significantly—median returns of negative 18% were almost comparable to negative 19% for the Russell 3000.
