The Illusion of Leadership: Are CEOs Paid For Luck?


I think some CEOs are overpaid. No, I'm not going to give you some leftist mantra about how no one should make $50 million a year because I don't believe that. What I believe is that there are other people capable of doing the job of CEO that would do it for less money. I think boards spend too much time looking for proven superstars than for up can coming managers who would love to take the top job for less than the current CEO makes. Now a recent research paper backs up my view, but for a different reason. This paper argues that there is a significant component of CEO pay that is based on luck.

According to the standard principal-agent theory CEO pay should not respond to exogenous forces beyond CEO control, yet the data shows that CEO pay substantially responds to exogenous shock casting doubt on the appropriateness of the principal-agent model in the CEO pay context.

In other words, this paper argues that the CEO gets credit for things that are beyond his or her control, creating what the author calls the illusion of leadership. Here is the author summarizing his findings.

An interesting and previously not documented result emerges from my empirical analysis. I show that there was no pay for luck in the 80s, a period of the US corporate history distinguished by active market for corporate control (large number of hostile takeovers and leveraged buyouts(LBOs)). It seems that when the incumbent management could be (and often were) challenged by shrewd outside investors, management and the board of directors were getting the compensation schemes right. On the other hand after the end of 80s when court decisions and legislation in the US brought the hostile takeover market to virtual halt (Jensen, Murphy and Wruck, 2004, page 23), economically large and statistically significant pay for luck emerges. The interpretation of the result seems clear. When the market for corporate control works, the internal governance of the corporation also works. When the market for corporate control is impeded by managerial lobbying for protection and by ill-advised court decisions and legislation(Jensen, Murphy and Wruck, 2004, page 28), pathologies in CEO compensation (part of the internal corporate governance) develop.

In other words, the market for corporate control is broken because CEOs don't want to be held accountable. At first glance, this may seem to contradict the current evidence that lousy CEOs are booted from power quicker than ever before, but it doesn't. This paper doesn't focus on that case. It focuses on the situation where random good things happen to the company and the CEO gets credit and additional compensation as a result.

I wonder if the author tested this hypothesis in a way that factored out the changes in accounting and tax treatment of CEO compensation. Those changes encouraged compensation via stock options, which is one reason CEO pay since the early 90s has tracked stock prices more closely. Regardless, it makes logical sense that CEO pay could be affected by some form of "survivorship bias," where their string of successes is due much more to luck than anyone realizes.

*Click on the picture for a larger version of MC Escher's "Ascending and Descending"

  • Rob, Great post but scary. You remind me of that kid in the “Emperor’s New Clothes” (Look, He’s naked… and really lumpy!)
    You’re going to get into trouble if you keep this up.

  • Eisenhower said “I’d rather have a lucky general than a smart one.” In a complex, multivariable, nonrepeatable situation, it’s not always possible to distinguish luck from skill, and I’d be suspicious of any statistical methodology claiming to reliably do so.

    I do agree that boards are too susceptible to the susperstar syndrome, often overlooking perfectly good talent in order to pay twice as much for someone who is hot at the moment. But this isn’t a matter of luck vs skill; it’s a matter of mindless fashion-following.

  • David is right. “In a complex, multi-variable, non-repeatable situation, it’s not always possible to distinguish luck from skill.” But because something is “not always possible” it’s not necessarily “not at all possible or even “mostly not possible.” What’s the old saying about the “unexamined life?”

  • Larry

    Is this not essentially the same argument as in Moneyball?

    “Superstars” (baseball players, movie stars, CEO’s) are handsomely rewarded for results they have very little to do with (RBI’s, box office returns). In baseball, the question has focused in on questions such as “what is a pitcher truly responsible for?”, and it is generally accepted that the homerun, the strikeout and the walk are all areas in which the pitcher exerts complete control.

    so, then, what area’s does a CEO control explicitly? More importantly, how does the average CEO perform in these capabilities? I don’t know, but smart boards should figure this out and play smarter than their competitors.

  • Sunny Smith

    It is pretty clear that the trend has been for CEO compensation to be on the rise. According to Forbes, in 2004 the heads of America’s 500 biggest companies received an aggregate 54% pay raise.
    These are the most current stats that I have, but I think that it is safe to say that this trend has not drastically changed since then. I guess the key to making CEO’s accountable is linking their pay to controllable factors, and utilizing pay and compensation as an incentive to influence desirable behaviors and performance. I guess as far as hiring the superstars is concerned, I guess that most companies that follow this trend are betting that a track record of success lowers their risk, as well as perhaps earns their company some free press and PR once the star is signed; these benefits for them outweigh the added costs.