Leadership1 publisher2 min readPublished
The new mark-to-market pay number tests 30 years of pay-at-risk orthodoxy
Average pay at risk for S&P 1500 chief executives reached 78% in 2023, up from 46% in 1993, while target pay converged on the industry median. Compensation Actually Paid is the first figure that lets a committee check whether that design pays for performance.
The Board Room · Leadership desk
What happened
- Average pay at risk for S&P 1500 chief executives rose from 46% in 1993 to 78% in 2023, according to a Shareholder Value Advisors memorandum by its president, Stephen O'Byrne.
- Competitive pay policy, now widely adopted, sets target pay at a market percentile, usually the 50th, regardless of what the company's past performance was.
- ISS reported in its 2024 proxy review that failed say-on-pay resolutions had dropped to a record low of under 1% for the S&P 500.
Compiled by The Board RoomSomething wrong?How this is made
Why it matters
- constraint A committee that sets target pay in dollars cannot argue the size of the grant rewards last year's results, because the share count falls when the price rises. The incentive has to come entirely from movement after the grant date.
- capability Directors can now run the pay-for-performance question on their own filings instead of debating it in principle, because CAP puts equity outcomes and shareholder return in the same units.
- contradiction Pay Governance and ISS describe the past two decades as improved design; O'Byrne describes the same period as producing pay for performance at few companies. A committee writing this year's proxy has to pick one of those readings.
- decision With at-risk share near its practical ceiling, the levers still under a committee's control are the target percentile and the peer group used to define market pay. Those are what an investor challenge will land on.
When target pay is set in dollars, the share count on an equity grant is the dollar target divided by the price. A price increase therefore cuts the number of shares granted, and a decline raises it. O'Byrne calls that a systematic performance penalty [11]. The consensus answer is that the incentive arrives after the grant, because grant value moves up and down with performance from that point on [12].
Average S&P 1500 CEO pay was 78% at risk in 2023 and 46% in 1993 [1], so the portion not at risk fell from 54% to 22% [17], a reduction of about 59% [18]. Other top-five executives went from 41% to 70% [2]. CEOs added 32 points of at-risk pay across the three decades, and the executives below them added 29 [19].
A quarter of the variation in pay mix across companies has disappeared since 2006, and one study cited in the memorandum found pay dispersion within industry-size groups down 45% since 2007 [3][4]. The theory is that a high at-risk percentage supplies the incentive while a 50th percentile target caps both retention risk and shareholder cost [6].
The institutions that vet these plans have called the trend an improvement. Pay Governance said in 2018 that "corporate governance in general and of executive compensation has improved dramatically over the past 20 years" [7]. ISS, in its 2024 proxy review, said that "many compensation committees appear to be doing a better job at addressing investor concerns" following a low say-on-pay vote [9].
Proxy-reported CEO pay is close to target pay, target pay is designed to be independent of performance, and so a weak correlation with results proves nothing [10]. That is a serviceable answer to the correlation critique, and it depended on equity never being reported on a mark-to-market basis [13]. Compensation Actually Paid includes the year-end value of current-year grants plus the change in value during the year of unvested grants from earlier years [14]. Relative pay and relative shareholder return can now be put on the same basis.
O'Byrne writes that there is pay for performance at some companies but not many [15]. The published text stops as it begins to describe plotting relative pay against relative total shareholder return, so it omits the company counts behind that judgment [20]. What a committee can settle without them are the two settings it still chooses: the percentile it targets, and the peer definition behind it, which is usually median pay for the same position at companies of similar size in the same industry [5].
What to watch
- Whether the full Shareholder Value Advisors memorandum publishes the company counts behind its relative pay versus relative total shareholder return plot.
- Whether say-on-pay failures stay below 1% of the S&P 500 once investors can run CAP-based correlations themselves.
- Any committee that moves from a dollar-denominated target to a fixed share count, which would remove the grant-size penalty O'Byrne describes.