In a recently released study (“Are CEOs Paid for Performance?”), MSCI ESG Research addresses the longstanding question about whether CEO pay actually reflects company performance. MSCI’s study reviewed cumulative pay and performance over 10 years and reported that “average 10-year total shareholder returns for the lowest pay quintile companies were 39% higher than for the highest pay quintile.” In short, according to MSCI, companies that paid their CEOs the least over the 10-year reporting period outperformed those that paid their CEOs the most. Based on these findings, MSCI concludes that CEO pay does not reflect long-term stock performance.

Following the study’s release, many in the press were quick to cite MSCI’s findings in articles suggesting that CEOs aren’t worth the money they’re paid. “Top-Paid CEOs Aren’t Very Good at Their Jobs” (CNN) and “Best Paid CEOs Run Some of the Worst-Performing Companies” (Wall Street Journal) were just a few of the more prominent headlines. While such coverage might be effective in fueling further cynicism toward executive pay (and generating more page clicks for the various news organizations), they fall short of interpreting the MSCI study accurately — nor do they advance the discussion of what constitutes proper pay for performance.

At first glance, MSCI’s findings and, more specifically, the resulting headlines and articles from the mainstream press seem to run counter to an analysis we conducted to isolate specific pay practices of high-performing companies. In 2014, Willis Towers Watson conducted a detailed study of executive pay design among 50 U.S.-based S&P 1500 companies with sustained high performance over a 15-year period in order to understand what these companies do differently in how they design their pay programs. (For details of our study, see “Enduring high-performing companies take the road less traveled in executive compensation design,” Executive Compensation Bulletin, July 15, 2014.) 

Our analysis revealed that CEO target pay opportunities (i.e., target total direct compensation, reflecting base salary plus target bonus plus grant-date value of long-term incentives) were very similar between the high performers and the overall market median when adjusted for company size. However, the actual realizable pay in the high-performing companies exceeded market median levels, and often significantly — by 43% among large companies and 28% for small companies.

This upside was a result of leverage in these companies’ bonus and long-term incentive programs coupled with positive performance results, which supports our conclusion that effective program design can ensure appropriate rewards for high performance. Our study also raised questions about the need for companies to adopt above-median target pay philosophies. Reviewing the MSCI study carefully, those findings also raise questions about the appropriateness of higher target pay opportunities. However, linking performance outcomes to pay opportunity can be misleading.


The MSCI findings most often cited by the media are based on “total summary pay,” which is the total pay figure reported in the Summary Compensation Table (SCT) that appears in all publicly held U.S. companies’ annual proxy filings. This total includes annual salary and annual bonus paid for the previous fiscal year, as well as the grant-date value of long-term incentive awards, the change in pension value and certain other compensation (e.g., perquisites and other personal benefits).

As MSCI’s study points out, about 70% of CEOs’ total SCT pay comes from long-term incentives, which in contrast to annual incentives are valued using the grant-date accounting value, not what’s eventually earned. Most companies, however, do not adjust the pay opportunity (i.e., SCT pay) based on company or individual performance. According to Willis Towers Watson’s Global Executive Incentive Design Survey, 57% of U.S. respondents make annual adjustments to target long-term incentives based on competitive market data, while just 19% adjust target values due to company performance. (For more on that survey, see “Survey findings offer unique insights on incentive plan design issues globally,” Executive Pay Matters, April 4, 2016.)

While some may argue that the pay opportunity should vary more based on prior performance, current market practice leans more heavily toward basing final payout outcomes on performance subsequent to grant. Consequently, while SCT pay is one important lens to use in evaluating executive pay, it’s not the most meaningful number to evaluate on an absolute basis year over year if you want to truly understand how an executive’s pay changed with a company’s performance.

So, if not pay opportunity, what definition of pay would be more appropriate to measure against company performance? When we reviewed proxy disclosures of Fortune 500 companies that included pay-for-performance details in their Compensation Discussion and Analysis, we found that most companies used earned or realizable pay or some other pay definition focusing on pay outcomes. (See “Explaining pay for performance — an inexact science, for now,” Executive Compensation Bulletin, March 25, 2015 for more details.) Only about a third (36%) used SCT pay in telling their pay-for-performance story. Why? As the MSCI study itself points out, “In reality, CEOs almost never realized the exact amount reported in the summary table.”


The Securities and Exchange Commission (SEC) has proposed new disclosure rules that would require companies to provide a clear description of the relationship between “compensation actually paid” to the named executive officers and the company’s cumulative total shareholder return (TSR). As proposed, compensation actually paid would be SCT pay modified to include a different measure of the change in pension value and the value of equity awards at vesting, rather than when granted. Thus, the primary difference from the SCT definition is that compensation actually paid focuses more on pay outcomes, rather than on the pay opportunity.

While the tendency is to assume that all compensation programs are structured the same, the reality is that various forms of pay, performance horizons and required hurdles for payout vary widely depending on the unique strategic and operational goals of each company. We believe the proposed SEC regulations, if finalized largely in their proposed form, will provide stakeholders important information on the relationship between executive compensation actually paid and the TSR performance of the company, while maintaining the flexibility needed for companies to enhance their disclosures to tell their unique stories in a manner that best matches their specific business objectives and compensation structures. (For a closer look at the SEC proposal, see “Q&As on the SEC proposed pay-for-performance rules,” Executive Compensation Bulletin, May 14, 2015.)


Robert Newbury

Robert Newbury

Willis Towers Watson

Melissa Costa

Melissa Costa

Willis Towers Watson
New York

Robert Newbury is a director based in Columbus, Ohio who leads Willis Towers Watson’s Executive Compensation Resources unit. Melissa Costa is a senior executive compensation consultant in the firm’s New York office. Email, or