CEO Salary Justification Season Is Open

Proxy season is over. Then comes the annual compilation of executive compensation data. Equilar and the Times, for example, reported that the compensation of the median CEO at a large public company was more than $15 million in 2012.

This means that now we are into the season of justifying these stratospheric numbers—and particularly the high rate of growth of those numbers. (2012 median compensation was 16 percent higher than in 2012.) For example, there was Steven Kaplan’s unconvincing attempt to justify high CEO pay by comparing it to . . . high pay among the top 0.1% (see Brad DeLong for a summary).

This ground has been trodden over a million times, and there’s little new that anyone can say about the issue. The common defense of high CEO pay is that it’s a justifiable investment given the market for talent. This is how Robert Shiller put it in Finance and the Good Society (p. 22):

“So it is plausible in turn that the board of Corporation B would offer a really attractive package to lure the CEO—a package that might, say, offer options on the company’s stock potentially worth $30–50 million if he is successful. That amount is not enormous relative to the earnings of a large company. A diligent board . . . might consider a highly qualified, proven CEO worth all of this.”

Let’s pause for a moment to note that “$30–50 million if he is successful” is actually a pretty miserly compensation package by today’s standards. The 2012 median compensation of $15 million reflects the current value of stock and option grants, not their value “if he is successful,” which is much higher. And $15 million is the median annual compensation, not the total for the CEO’s tenure.

Moreover, the standard argument that the market forces you to pay people what they are worth to your company is simply wrong. A very good developer can be worth millions of dollars a year to a software company. But she can’t command that much in salary because there are plenty of almost-just-as-good developers (and probably some just-as-good developers) who will work for, say, $150,000 per year. When you buy anything, you compare its value to that of the next best available alternative. Or, at least, that’s what you’re supposed to do.

How does this work for CEOs? Let’s assume for the moment that there is some potential CEO who, on an expected basis, can make the company worth $100 million more than it is worth under the current CEO. Should you be willing to pay her up to $99 million to work for you? No—because there are probably lots of other people out there who can also make the company worth $100 million more, or at least some large fraction thereof. It’s not the $100 million that matters—it’s $100 million minus the value of the next best available alternative.

Now, you might think that only one person in the whole world—let’s call him Ron Johnson—can increase the value of your company by $100 million, and no one else can come close. But unless Ron already has some deep connection to your company (e.g., Steve Jobs returning to Apple—and even in that case, his success was hard to foresee), you are almost certainly wrong. The marginal impact of a CEO is extremely hard to estimate in advance, and any expected value you come up with will be swamped by the standard deviation. The only honest answer is to say that there are a bunch of people who could probably help your company a lot, and that implies that you should hire the one who will do the job for the least money.

Instead, however, the directors manage to convince themselves that Ron is the only person who can save their company, and saving the company is worth $100 million, so he should get $99 million. They do this by making all sorts of basic errors of thinking, like converting their vague, irrational intuitions into certainties. Then they justify a specific transaction—overpaying Ron—by referring to a conceptual possibility.

Sure, CEOs are important, and some are highly valuable to their companies. But we’re not talking about LeBron James or Leo Messi here—there are a lot of people who can do the same job roughly as well as each other. There’s no reason the rules of ordinary labor markets should be suspended for them.