James Surowiecki recently published an excellent column in the New Yorker on why various reform attempts (such as say-on-pay, paying directors mostly in stock, increased disclosure on pay, and the shift to independent directors) have failed to curb CEO compensation.
Surowiecki identified two causes: the perception that selecting the CEO is critical to the company’s future and the belief that paying the CEO based on the company’s performance will improve the company’s bottom line. He sketched out a persuasive argument that both these beliefs are myth, and that the rise in pay is largely due to the increasing emphasis on the importance of corporate leadership and the shift to performance pay (especially stock options).
I agree strongly with both points. This will not come as a surprise to regular readers of this blog or to those who have read my book, Indispensable and Other Myths: Why the CEO Pay Experiment Failed and How to Fix It (University of California Press 2014). Surowiecki interviewed me (and read Indispensable) before writing the column, and he quoted me several times.
Compensation consultant Marc Hodak posted a sharply-worded critique of the column in his blog recently. (Hat tip to Stephen Bainbridge at professorbainbridge.com). The post’s rather strident tone is regrettable, but Hodak’s core argument is one compensation consultants and other defenders of the current system often make, so I think it deserves a reply.
I should note at the outset that I reject the temptation to dismiss Hodak’s arguments on the theory that his role as a compensation consultant provides a powerful incentive to justify the current system. Upton Sinclair observed, “It is difficult to get a man to understand something, when his salary depends on his not understanding it.” But Hodak’s arguments should stand or fall on their own merit, and not because of some ad hominem attack.
Hodak dismisses my argument that it is hard for boards to predict in advance which of several similarly-credentialed finalists will most help the company by asserting, “It would no doubt upset my client directors to think of all the time and energy they have spent fretting about the quality of their top management is pretty much wasted.” He then adds, “Seriously, to call this an ‘academic’ finding reinforces all the worst connotations of that term.” Nowhere does he discuss any of the studies I cite in Indispensable that support my finding, nor does he cite any contrary studies (although there are some, which I discuss in the book). His only argument is that his client directors would be disturbed to think they were misguided. The best face I can put on his argument is this: if smart people are investing lots of time and money in an idea, the idea must be right.
There is something intuitively appealing about this appeal to authority – it’s one I hear all the time – but it’s hardly evidence. Smart people have invested lots of time and money into all sorts of ideas that turned out to be wrong. In the 1960s, smart people invested a lot of money pursuing the idea that public companies should join together into large conglomerations to better leverage managerial expertise and provide one-stop-shopping for investors looking to diversify. In the 1980s, smart people invested more money breaking up those same conglomerates. In the late 1990s, smart people believed the internet would grow at an astronomical rate, justifying stratospheric valuations for companies that were not yet earning a profit, and they invested their money accordingly (until the crash came). In the early 2000s, smart people invested lots of money in financial instruments whose value depended on a housing market that would never stop rising in price. We know how that one turned out, too.
The point is not that smart, rich people always make mistakes, or even that they often do so. These examples just demonstrate that it’s far from unheard of for lots of rich, smart, sophisticated folks to all believe the same fallacy, and to invest accordingly. It’s therefore far from sufficient to point to the behavior of this group as conclusive evidence of empirical validity. We need to test propositions with actual empirical studies that look at how companies perform under different conditions. As I document in my book, the weight of the empirical research demonstrates that CEOs do not, on average, have much impact on enormous, well-established, companies. Some CEOs do (in both positive and negative directions), but on average, we should not expect the CEO a company hires to have much of an effect. And there is not yet a reliable way to tell ex ante whether a board is hiring a future superstar, a future disaster, or just a caretaker.
Hodak goes on to point out that even if, as Gabaix and Landier claim in their study, the best CEO will only add .016% to the market capitalization of a company when compared to the 250th best, that would still make paying a CEO $15 million a bargain. That was, of course, precisely Gabaix and Landier’s point. They are defenders of the current CEO pay system, and this was the very argument they made in their paper. I critiqued their study on a number of grounds in Indispensable, but the point Surowiecki was highlighting was simply that if the talent differences are actually that small, boards cannot reasonably be expected to discern them and be sure the company is hiring #1 and not #250. Hodak does not address this point, nor does he discuss any of the other critiques of this study I raised in my book, such as the study’s deeply problematic assumption that all companies are looking for CEOs with the identical skill set.
To undermine Jacquart and Armstrong’s work (a paper that found that higher pay fails to promote better performance and that incentive pay encourages unethical behavior), Hodak again resorts to the “rich, smart people can’t be wrong” argument. He writes, “So, all those directors with all those years of collective experience have been wasting not only immense amounts of time, but also all that money based on a mistaken believe in the power of incentives. Wow.” He also asserts this “may be one of the worst studies on executive compensation in recent years” without even hinting at why he holds this belief. The closest he comes is his statement that the authors conflate pay-for-recruiting with pay-for-performance, but he provides no evidence that they do any such thing.
The remainder of the post attacks a straw man. Surowiecki says that whether or not compensation committees can accurately predict CEO performance, they are willing to spend extra money to get the person they think will do the best job for the company. His point is that well-meaning boards who believe CEO leadership is critical and that they can discern which CEO candidate will best run the company might very reasonably pay a premium to get the best person. The problem is not that boards are “lazy, stupid or corrupt,” as Hodak mischaracterizes Surowiecki as implying, but that they are mistaken on the two points discussed above: that CEOs of large, established companies typically have a major impact on performance and that performance pay induces CEOs to work harder and smarter. Nor is Surowiecki’s point that boards believe that simply paying CEOs more will induce better performance. His point (and mine) is that boards believe that conditioning additional pay on performance will result in better performance. That is the entire rationale behind performance pay. If Hodak does not believe this is true, then we have little to argue about; he can join Surowiecki and me in understanding that performance pay is an extravagant waste of corporate resources.