When CEOs Are Accidentally Overpaid
Why are chief executive officers paid so much? Although skyrocketing CEO pay is not a substantial factor behind the increase in income inequality, it is certainly startling. The average U.S. CEO now makes about 350 times what the average worker makes. That means CEOs typically make in one day what their employees make in one year.
There's no reason why that number disparity should, by itself, be annoying. After all, star entrepreneurs make a lot of money, and few people mind. Star basketball players make a lot of money, and few people mind. So why should we mind if star managers -- which is what CEOs really are -- make a lot of money, too?
I strongly suspect that it's not inequality itself that makes people mad: It's unfairness. Basically, many people suspect that high CEO pay is unfair. Star entrepreneurs get paid when they create valuable companies. Star basketball players get paid when they make lots of points, rebounds or assists, or otherwise help their team produce wins. But do CEOs get paid when they make their companies succeed?
Some suspect the answer is no. For example, a 2014 study by economists Michael Cooper, Huseyin Gulen and P. Raghavendra Rau found that the companies with the highest-paid CEOs tend to perform worse, in terms of stock returns, than other firms. That could just be related to the fact that CEOs are mostly paid according to stock performance these days, and stock performance tends to reverse itself over a three-year period. But it's not encouraging. Then there are all those stories of CEOs who walked away with huge pay packages after their companies tanked.
Paying CEOs for performance was supposed to fix this. Indeed, tying CEO pay to companies' stock performance, especially through stock options, has been responsible for much of the run-up in CEO pay. For some reason, the solution seems to have made the problem worse.
Economists are starting to ask why. One of the most important of these economists is Kelly Shue of the Chicago Booth School of Business. Shue has written several papers showing that the way we compensate executives might be broken.
In 2011, Shue found that having the right friends matters a lot for executive pay. She looked at MBA graduates of Harvard Business School who were randomly assigned to different class sections. The class section assignments naturally created human networks -- people know the people they went to school with. She then looked at HBS graduates who became executives later on, and found something startling. The difference in pay between different sections was larger than the difference within different sections.
Remember that these sections were randomly created. So Shue's finding means that human networks were a very important determinant of pay levels. She also found that when one executive's industry experienced a boom, the compensation of that executive's former HBS section-mates in totally unrelated industries would go up as well.
In other words, Shue found that when it comes to executive pay, it's often who you know, rather than what you do, that determines how much you earn. Somehow, these human networks seem to persist from business school all the way to the boardroom. The fact that the pay boost holds across unrelated industries means that some of it is almost certainly not based on the real productive power of business networks, but on humans' natural tendency to pay their friends more.
Now, Shue has another paper that reveals a second sobering fact about executive pay. Along with co-author Richard Townsend of Dartmouth, Shue looked at the stock options granted to CEOs and found that there are quite a lot of years in which the number of options granted doesn't change at all. Here, from Shue and Townsend's paper, is a graph of the change in the number of stock options granted to CEOs in various years:
Notice the big spike at zero. That means that, in a whole lot of years, companies simply grant their CEOs the same number of new stock options that they gave them last year.
That will tend to make CEO pay grow over time. If we keep giving CEOs the same number of new stock options every year, then what happens is that a run of good recent performance increases the amount that CEOs get paid for good future performance. That means some CEOs are getting paid too much for performance. And if the stock market goes up -- as it usually does, and has in recent decades -- that excess compensation will be larger. As Shue and Townsend also show, salaries aren't being adjusted to offset this.
In other words, many financially unsophisticated companies have been paying CEOs too much, simply by accident.
So Kelly Shue's research shows there are several reasons why many CEOs in the U.S. are indeed overpaid. Correcting that overpayment will not reduce U.S. inequality by a substantial amount, but it will improve companies' bottom lines a bit. And it will increase the average American's confidence that the system works -- that people get paid for what they produce.
This column does not necessarily reflect the opinion of Bloomberg View's editorial board or Bloomberg LP, its owners and investors.
To contact the author on this story:
Noah Smith at email@example.com
To contact the editor on this story:
Paula Dwyer at firstname.lastname@example.org