Robert Reich recently posted an article on Salon (http://bit.ly/1o6tkjq) arguing that workers — and CEOs — are not paid what they’re worth. Instead, workers’ pay depends on the extent of unionization in the industry, Reich says, because unions enable workers to extract a greater percentage of the joint gains of production. CEOs are paid some 300 times more than typical workers not because they’re worth that much, Reich continues, but because they appoint the directors on the compensation committee or because boards don’t want to be perceived as having hired a second string CEO.
Reich’s broad point — that CEOs are not necessarily worth what they are paid — is quite correct. In the details, though, Reich is entirely wrong. He adopts two common theories of CEO pay: the corruption theory and the advertisement theory. Neither has strong empirical support. The corruption theory — most closely associated with Harvard professors Bebchuk and Fried — predicts exactly the opposite of what the data shows. CEO pay was flat in real terms during the 1940s, 1950s, and 1960s, then rose somewhat in the 1970s, climbed dramatically during the 1980s then skyrocketed during the 1990s. CEO power over boards follows precisely the opposite pattern. Companies began insisting on independent directors during the 1980s and by the 1990s most boards were dominated by directors from outside the company. In other words, CEO pay was stagnant while CEO power over directors was strongest, and CEO compensation climbed fastest while CEO power was weakest.
The advertisement theory doesn’t work either. Are directors really spending millions of dollars to create a perception that their CEO is better than s/he is? That seems like an entirely irrational use of money. Wouldn’t it be more effective to just hire a CEO who was worth that sort of money? And this explanation also fails to match up with the data. If this is what motivates boards in paying their CEOs, we’d expect to see a one-way ratchet in CEO pay, with companies constantly striving to outdo one another to advertise their CEOs’ quality. But that’s not what’s happened. For decades, CEO pay basically stood still in real terms. More recently, in the years following the internet bubble and again during the financial crisis, CEO pay actually shrank. That shouldn’t ever happen if boards are primarily focused on advertising their companies’ fitness when they set CEO pay.
The real problem with CEO pay is the assumption that CEO pay should be tied to corporate performance. The empirical studies (both of actual companies and in the laboratory) show that tying CEO pay to corporate performance does not improve corporate profits. It does, however, boost CEO pay. During the period when CEO pay remained flat, boards used little if any performance pay (such as stock options or restricted stock). The performance pay trend began in the late 1970s and accelerated during the 1980s, precisely when CEO pay began its dramatic rise. Then, after the stock market crashed in 2000, the value of CEOs’ pay packages — tied, as they were, to stock prices — shrank too, then recovered with the stock market. Ironically for Reich, then, the core reason CEO pay has risen so dramatically is precisely boards’ efforts to ensure that CEOs are worth what they’re paid. The problem is that they’re doing it wrong. More on that another time.