CalPundit’s Challenge

CalPundit looks at wage growth for workers and CEOs and issues this challenge:

Odd, then, that CEO pay rose 27% in 2003, isn’t it? Did the supply of CEOs shrink last year? Did demand skyrocket?

What’s more, compared to average workers, who remain stuck in the invisible grip of Adam Smith, CEO pay has increased about 3x since 1990 and about 7x since 1980.

Is this the free market at work? That’s what I’m told. So I have a contest in mind: a prize for the least laughable explanation for why CEO pay has gone up 7x since 1980 based on supply and demand. At a minimum, winning entries should explain the following:

  • Why the supply of CEOs has decreased.

  • Why the demand for CEOs has increased.

  • Why the elasticity of the CEO demand curve is apparently steeper than for any other commodity on the planet.

Please keep your entries under 100,000 words, and restrict your econometrics to fields no more complex than differential topology.

Grand prize to be announced at a future date.

I’ve always viewed skyrocketing CEO pay as the result of a variety of factors, some justified and some not, so I’ll give it a shot:

  1. Governance reforms at the end of the 1980s that lead to the substitution of salary-based pay with equity-based incentive pay. This came about mostly because of the wave of corporate takeovers: a company was ripe for a takeover when it was managed to accomplish something other than maximizing shareholder value. Since the CEO’s income was not tied directly to the company’s performance, they made a lot of bad decisions. So boards began using equity-based pay. This had several effects:
    • It replaced risk-free income with risky income, which entails increasing expected pay to compensate for the greater risk premium.
    • By design, equity-based compensation will encourage CEOs to (at least try to, and on average but clearly not always succeed) make better decisions, as viewed from a shareholder’s perspective. When CEOs make better decisions, profits go up. When CEO income is tied to profits, their income goes up.
    • Incentive pay, by design, induces more effort (it’s why commissioned sales people tend to work harder and usually make more than non-commissioned ones.) More effort lowers utility and additional income is required to compensate.
    • Evidence: Jensen and Murphy found in 1990 that on average, “CEO wealth changes by $3.25 per $1000 of shareholder value. This is $8.05 for small firms and $1.85 for large firms.” A 1998 follow up found that this increased to $18.00 per $1000 of added shareholder wealth over the course of the 1990s.

  2. Coincidence: CEO pay was tied to corporate performance and for a variety of reasons (possibly including better governance), the stock market boomed. Since pay was tied to stock-price and stock prices were skyrocketing, CEO pay necessarily soared. This explains much of the surge in CEO pay in the 1990s. Regarding 2003, that was the year that corporate profits and stock prices started recovering from the bust that started in 2000, so if pay is tied to profits and stock price, it’s not surprising that pay increased. (On the other hand, this implies that pay should have decreased in 2000-2002, which it did not.)
  3. Lake Woebegon: Every company wants to think it has an above-average CEO, and pay him or her accordingly. A typical compensation committee report will say something like, “we aim to pay our CEO in 60th-75th percentile for the industry.” When all companies try to keep their CEO pay in the top 25% to 40%, the result is predictable.
  4. A poorly functioning market for compensation consultants and poorly informed compensation committees. Compensation committees only get the information the CEO wants them to have, and are often chosen by the CEO. So they turn to consulting firms for information. These compensation consulting firms want to be hired right now, and they want to be rehired in the next year, and they are fully aware that the CEO will heavily influence the rehiring decision. Not surprisingly, the advice they give typically says to throw wads of cash at the CEO. Warren Buffet and Charlie Munger, in their typical colorful fashion, explained the situation a few months back:

    “The typical large company has a compensation committee,” said Buffett. “They don’t look for Dobermans on that committee, they look for chihuahuas.”

    He paused amid laughter, then added: “Chihuahuas that have been sedated.”

    Munger interjected: “I would rather throw a viper down my shirtfront than hire a compensation consultant.”

  5. Stock options encourage more risk taking because of their skewed payoffs (either zero or positive, but not negative.) As a result, we observe the pay of CEOs that gamble and win, while CEOs that gamble and lose are replaced. Imagine getting your impression of the lottery only from reading the newspaper: people win all the time and you never read that John Smith bought 200 tickets and didn’t win a dime.
  6. If firms are choosing their CEO from within more than in the past (which I’m not sure is the case) then there’s a tournament value to high CEO pay: it induces very high effort and careful decision making from all candidates within top management. Recent examples include Jack Welch’s successor at GE, Jeffrey Immelt, and Lou Gerstner’s successor at IBM, Sam Palmisano. In both cases, for several years leading up to each succession there was in fact sharp internal competition for the spot. (Incidentally, while there are many instances of CEO compensation levels that are patently absurd, it’s pretty tough to argue that Welch or Gerstner were overpaid. Though there was a flap over Welch’s retirement package.) Craig Barrett’s replacing Andy Grove at Intel is another example.
  7. Pure opportunism: CEOs have access to better information than the board, and they exercise such great control over the board, that they are able to transfer massive amounts of wealth from shareholders to themselves.

Within this list, I think that (1) plays a larger role than most people think. That is, I think that CEO pay should be much higher when it’s tied to performance (post-1990, roughly) than when it’s salary based. Since the switch to performance-based pay was much more dramatic for CEOs than other workers, it’s even justified that their pay increased much faster than their underlings. (6) is also a reason for higher pay, but I think its importance is second order. (2) is a historical accident, for the most part.

All that said, (3), (4), and (7) are genuinely big problems and combine to account for the bulk of the unjustified rise in CEO pay. With no real theory or data at hand to justify this, I’d say the split is about 50-50: roughly half the increase is the reasonable result of structural change in the market for and compensation of CEOs; the other half is due to informational asymmetries and other failures in the market for CEOs. Eliminating that second half would be a good thing — unless you’re a CEO.


UPDATE: Kevin points out that I failed to phrase my answer in the form of a question mention supply and demand in my answer. In general,things that make it less pleasant to be a CEO, like higher risk premia and higher effort costs, are tantamount to an inward shift of the supply curve: CEOs will require more pay for any given quantity of effort.

Similarly, the other factors — weak boards, asymmetric information, … — act as a shfit out in demand. In particular, as compared to firms without those problems, such firms are paying more for a given quantity of CEO effort and quality. E.g., with perfect information and correct incentives, a board might pay a CEO either $1m for 50 units of effort or $2m for 75 units of effort (draw that curve). Now consider a board plagued with bad information and bad incentives. It might pay $1.5m and $3m for the same respective amounts of CEO effort (draw that curve). The result is in fact a shifted out demand curve. Another way to see this is to note that a shift from a world of rational people to a world of chronic overpayers is just like a shift out of the demand curve.