But is He (or She) Lucky?
Quantifying the Role of Chance in Executive Performance
Executive compensation—a subject much in the news in recent years—tends to be measured by and based upon the performance of the company. In theory, a chief executive officer who manages a firm well will generate strong profits and stock returns, thereby justifying the large compensation package that he or she commands.
But what if much of a company’s performance, good or bad, is the result of luck, and how can that be measured? Seoyoung Kim, a new assistant professor of finance, has studied the subject in depth and reported her findings in a paper, “Measuring Luck in CEO Outperformance,” that was recently submitted for publication
"Some people will repeatedly be lucky or unlucky."
“We see the outcomes in hindsight,” Kim says of company performance, “but we don’t really know if they were the result of good decisions, based on what was known at the time the decisions were made. What we do know is that in a large pool of people, such as corporate CEOs, some people will be repeatedly lucky or unlucky.”
Her research resulted in two principal conclusions. The difference in company outcome under the stewardship of the great majority of CEOs, she said, is not much more than could be attributed to luck and to factors outside the executives’ control, as opposed to skill and competence. However, there is an elite group of top-performing executives whose results are so good that luck is highly unlikely to be the reason
“The very, very top performers in real life are doing so well they outperform what statistical simulations, adjusted for luck, show the top outcome should be,” she says
Kim says that questions of executive performance, luck and outcome have been studied before, but generally by defining luck as the industry or market-wide shocks that are outside a CEO’s control (such as sharp swings in oil prices). However, she contends that the remaining CEO-specific “skill” portion of performance may not reflect skill at all. Her paper adds to the existing research by looking at a broad range of CEOs and companies across a wide range of industry sectors over a long period of time—more than a decade and a half.
Using the Standard and Poor’s Executive Compensation Database, covering 1,500 companies during the period from 1992 to 2009, she was able to get a big-picture look at the performance of executives and companies over time and establish the range of outcomes. She then developed a statistical model based on the presumption that all CEOs are equally competent and that differences in corporate outcome are essentially the result of luck or other factors outside the control of the executive
Comparing the two sets of results, Kim was able to determine that most of the actual outcomes were not so different from the simulated ones as to be entirely attributable to skill, and that luck was likely to have been a factor. The difference, for instance, between actual outcomes of an executive in the 90th percentile, based on company performance, and one in the 50th percentile, was not starkly different from the difference that occurred in the statistical simulation. Only with the very top performers was there a substantial difference that was highly unlikely to be the result of luck alone.
“As the simulation results demonstrate,” she writes in the paper, “CEOs can be repeatedly lucky or unlucky, guaranteeing extreme differences in performance outcomes even if everyone is equally skilled and puts forth the same amount of effort.”
“It’s a hard question trying to determine what’s the result of luck and what’s the result of skill,” Kim says. “Evaluating how much variation in performances we can expect due solely to differences in luck, thus, has important implications for replacement decisions, for designing incentive contracts, and more broadly, for how CEO performance outcomes are interpreted in attempting to measure managerial effort or ability.”
