CEO Pay Levels in the U.S. Are Converging Amid Increased Benchmarking

The pay levels of CEOs of publicly listed firms in the U.S. have converged sharply over the last two decades, amid the increasing use of compensation benchmarking against other firms with a similar profile.

In a new paper, we looked at a wide sample of publicly listed U.S. firms over the period 1996 to 2023, tracking the variation in pay across chief executives. After peaking around the year 2000, we find the cross-sectional variation in CEO pay levels has declined almost 40% since 2007. Pay has converged toward the median at both tails of the distribution curve, and that is the case at an economy-wide level, across different industries, within groups of similar-sized companies in the same industries, and within compensation peer groups.

The increased use of benchmarking, especially to industry peers closest in size, appears to be driving this reduction in the dispersion of pay levels. Three institutional developments in recent decades have encouraged the industry-size benchmarking: the 2006 Securities and Exchange Commission (SEC) rule mandating the disclosure of peer groups used for compensation benchmarking; the rising influence of proxy advisors that favor industry-size benchmarks for CEO pay; and the introduction of say-on-pay (SOP) regulations. Amid these developments, boards have experienced increased pressure to adopt industry-size benchmarking to conform to a standardized pay-setting process, which has led to increasingly closed “pay clusters.” As firms typically target the median level of pay of such peer clusters, pay compresses in the cross-section.

We conducted three sets of analyses to test our explanation for the increased convergence of CEO pay levels. In our first set of tests, we created a simulation in which all firms adopted an industry-size-based benchmark and then traced the evolution of pay dispersion in these conditions. Pay dispersion declined substantially in the simulation, thus supporting the idea that adoption of industry-size benchmarking could result in a decline in the cross-sectional dispersion in pay levels.

In a second set of tests, we use regression analyses to explore whether the decline in pay dispersion goes hand in hand with an increasing tendency to benchmark CEO compensation to industry peers closest in size. Over time, compensation peer groups show less dispersion in terms of size, are less likely to deviate from the industry-size criterion, and show greater compliance with the industry-size-based peer selection methodology adopted by Institutional Shareholder Services (ISS), the largest proxy advisory firm. In both in the time series and the cross section, these patterns are strongly associated with our results on pay dispersion.

Finally, in a third set of tests, we provide causal evidence on how three recent institutional developments – the 2006 SEC rule that mandated the disclosure of peers used for compensation benchmarking, the growing influence of proxy advisors, and the introduction of SOP regulation – have together contributed to the rise of industry-size benchmarking.

The timing of the trend is telling. The decline in pay dispersion starts soon after the implementation of the 2006 SEC disclosure mandate and not before. The 2006 mandate was relatively unexpected, and among firms whose compensation peer groups exhibited larger deviations right before the regulation was introduced, the tendency to adopt industry-size benchmarking was stronger.

Additionally, we tested our conclusions against companies and executives in which SEC disclosure mandates have limited influence. That includes non-U.S. public firms, unlisted U.S. firms, and senior executives outside the top five managers in public U.S. firms (because their pay is typically not determined via compensation benchmarking). In all of these settings, there was no evidence of the same decline in pay dispersion as for the top executives in public U.S. firms. For example, for CEOs in private U.S. firms the dispersion trajectory of pay was entirely different over the study period, peaking in 2014 and 2018 before settling back to very similar levels as they were two decades ago.

Proxy advisors also played a role. Using peer rankings produced by ISS’s publicly disclosed methodology, we found that a firm’s likelihood of including a potential peer varied with the degree of conformity the peer has with ISS’s policy for compensation benchmarking. This relationship strengthens significantly over the study period, amid the increase in industry-size benchmarking and it disappears right around an arbitrary peer group size cutoff used in ISS’s peer selection methodology. Further, in placebo tests, this discontinuity is absent in years prior to the introduction of ISS’s policy.

Also pointing to the influence of proxy advisors, we find that firms are more likely to add or drop peers in line with ISS’s recommendations from the prior year. Industry-size benchmarking becomes more prevalent as passive ownership increases, which is consistent with the view that passive investors rely more on proxy advisor guidance.

A final factor is SOP regulation, which requires publicly listed firms to regularly submit the compensation of their top executives to shareholders for an advisory vote. Firms are then required to disclose the voting outcome, and to respond to potential shareholder discontent. We find that following weak support in SOP votes, firms bring compensation peer groups closer in line with the industry-size benchmarking criterion.

To sharpen identification, we looked at the frequency of SOP voting, since more frequent SOP votes can be seen as a way to step up pressure on firms to align their compensation practices with the standardized benchmarking criteria. We compare firms whose shareholders narrowly approved an annual SOP vote with those that narrowly supported a vote every three years. In the years after the vote (but not before), firms whose shareholders favored annual SOP voting exhibited greater conformity with the industry-size benchmarking criterion than those with less frequent voting. This supports a causal role for SOP advisory votes in encouraging the adoption of industry-size benchmarking.

The implications of applying industry-size benchmarking in the executive pay-setting process are complex. Standardized criteria and increased transparency may, on the one hand, make evaluating pay practices easier and promote greater accountability. Uniformity in pay-setting practices may also enhance perceptions of fairness in CEO compensation design. But on the other hand, pressure from regulators and investors to make CEO pay more transparent and standardized may lead boards to prioritize compliance over customization. Suboptimal incentive structures and inefficiencies may be the byproducts of firms adopting approaches that sacrifice alignment with firm- and executive-specific needs in favor of greater standardization.

The trend extends beyond pay levels to the structure of compensation contracts, which also have become more similar across firms in recent decades. For example, another study finds that, since 2006, variety in salaries, bonuses, stock awards, options, non-equity incentives, pensions and other elements of a CEO’s full compensation package has fallen by about one quarter.

While a one-size-fits-all pattern seems clear, there’s no simple answer as to whether benchmark-driven CEO pay convergence is positive or negative for companies and their shareholders. But it is an important, if unintended, consequence of governance reforms and the growing presence of institutional investors and proxy advisors in publicly listed companies.