Rich CEOs Get Richer
The new study begins with an old query: To what extent does the pay of CEOs reflect how well—or badly—they have performed? Dozens of researchers have probed this precise question only to find the association to be modest, even tenuous, thereby fueling doubts and suspicions about the caliber of corporate governance in an era of skyrocketing CEO pay.
But is the association between CEO pay and company performance really as tenuous as so much of the research suggests?
In the words of the paper in the current issue of the Academy of Management Journal, almost all the research "take[s] snapshot views, typically focusing on performance in a given year." Yet, such leading scholars as Eugene Fama of the University of Chicago argue that executive compensation is best viewed as a multiperiod phenomenon in which boards of directors can adjust CEOs' pay to reflect their entire records, a process that Fama has labeled "settling up."
Calling an absence of evidence on settling up a "major void ... in understanding of contemporary business practices," the new study sets out to fill this void by analyzing the relationship of pay and performance among 590 big-company CEOs as it evolves over as long as a decade.
It finds that, while some amount of settling up does take place, it tends to be relatively minor. In the main, CEOs who are overpaid tend to become increasingly so over the course of time.
In the words of the study's authors, Adam J. Wowak of the University of Notre Dame, Donald C. Hambrick of the Penn State Smeal College of Business, and Andrew D. Henderson of the University of Texas at Austin: "When all other factors are controlled for, CEOs who have been prevailingly overpaid tend to receive the biggest current raises (or smallest pay cuts). This is quite at odds with settling up ... Instead of observing settling up, we observed that the rich get richer and the poor get poorer—relatively speaking, of course."
What accounts for this anomaly? The authors surmise that board members "succumb to a sociocognitive trap, believing that the amounts they previously paid their CEOs were simply in line with their CEOs'
inherent talent levels. ... As long as this belief holds, boards who have been overpaying their CEOs will continue with their generosity, giving outsized raises to CEOs who are already paid well above market norms, in the hopes of keeping them motivated and in office."
Still, overpaid CEOs cannot continue to push their luck indefinitely, the study finds: If they stumble, they are significantly more likely than modestly paid chiefs to fall. As the authors put it, boards "cling to their prior assessment of ... CEOs whom they have deemed supervaluable until current performance seriously drops, at which point these CEOs may become especially vulnerable to dismissal."
Such persistently overpaid CEOs, the professors add, "may provoke strong, essentially retaliatory responses from their boards. Or, to adopt a more economics-based interpretation, perhaps boards conclude that generously paid CEOs who deliver poor performance will be unable to ever rebalance the employer-employee ledger and therefore should be replaced."
Comments Hambrick: "Our findings appear to reflect two major trends in corporate governance over the past few decades—a growing incidence of outrage-inducing CEO compensation for what proves to be poor-to-middling company performance and a marked increase in CEO dismissals. But is overpaying CEOs and then firing them when they falter the best way to run companies? Probably not."
The new study's findings, he believes, should impart momentum to the growing movement for increased shareholder say on pay. "By virtue of their distance from companies' top management, shareholders could bring an element of detachment to pay decisions that this study suggests is now often lacking," he says.
The paper's findings emerge from data on 590 CEOs who served as chiefs for four years or more in public companies with at least $10 million in sales and assets during the decade 1996 through 2005. Compensation includes the aggregate value of all salaries, bonuses, restricted stock grants, long-term incentive plan payouts, and stock-option grants (valued using a standard formula). Determination of overpayment or underpayment is based on "an array of well-known predictors of CEO pay," including company size, recent performance (measured by total shareholder return and return on equity), CEO age and tenure, and firms' industry affiliations.
Examining pay revisions starting at the third year of CEOs' tenures, the professors found that 44 percent were pay cuts, averaging about $710,000, and 56 percent were pay increases, averaging about $725,000. The rough balance between increases and decreases notwithstanding, pay adjustments were something less than rational models of settling up.
Thus, the authors provide a hypothetical case of two fourth-year CEOs for whom "all available predictors (e.g., performance records, company attributes, industry factors) point to 10 percent pay raises." Yet, both made 50 percent more than the market warranted during the past year, so that, "according to our results this recent overpayment will exert a strong downward effect on this year's raises."
And indeed it does, but in a rather odd way. CEO A, who was overpaid by 50 percent in years one and two as chief, will now receive a 1 percent pay cut; in contrast, CEO B, who was paid "exactly at market norms" for those first two years as chief will now receive an 11 percent pay cut.
In short, the two hypothetical CEOs "both experienced a downward correction from last year, but the halo surrounding CEO A exerted an upward effect. The rich got richer, relatively speaking."
Further analysis suggests that by year ten the halo dims only by a bit: now a consistently overpaid CEO sustains a cut that is still 7 percentage points less (11 percent vs. 4 percent) than that of a chief who has all along been paid at market rates.
The new study, entitled "Do CEOs Encounter Within-Tenure Settling Up? A Multiperiod Perspective on Executive Pay and Dismissal," is in the August/September issue of the The Academy of Management Journal. This peer-reviewed publication is published every other month by the academy, which, with about 19,000 members in 103 countries, is the largest organization in the world devoted to management research and teaching. The academy's other publications are the The Academy of Management Review, The Academy of Management Perspectives and Academy of Management Learning and Education.