Andrew Lo has written a few guest posts here before, and now he’s back with another excellent one — so excellent, in fact, that we asked if he’d join our corps of recurring guest bloggers. Happily for all of us, he accepted.
The recent proposal by the Fed to regulate bankers’ compensation practices is understandable given the events of the past two years, but setting caps on salaries and bonuses misses the fundamental problem of compensation on Wall Street. Despite the public resentment surrounding finance-industry payouts, the fact is that no one objects to paying for performance. We just want to make sure we’re not getting fleeced or paying for pure dumb luck, and this is where the problem lies.
To correctly link pay to performance, we need two key ingredients: a reasonably well-functioning labor market for talent and an accurate measure of performance. If either of these inputs is missing or broken, compensation levels can easily get out of whack, disrupting several industries as employers and workers respond to potentially misleading wage differentials.
To see how, consider first an industry with eye-popping compensation contracts that no one seems to complain about: professional sports. The process for recruiting new talent in every major spectator sport is certainly well-organized, and the labor market for professional athletes is relatively liquid and transparent compared to other labor markets. But what about those outrageous multi-million-dollar contracts? In fact, such contracts generate little public outrage because pay and performance are so directly linked in these instances, and because performance is not only easily measured, but is thoroughly analyzed by millions of fans as well as the franchise managers ultimately responsible for justifying these contracts to their shareholders. Is Tom Brady — the 2007 NFL MVP with a career record of .787, a playoff record of .824, 197 career touchdown passes, a 63.0 percent completion percentage, and a 92.9 percent passer rating — really worth the $8,001,320 he was paid in 2008? Absolutely. There is an even more direct connection between pay and performance in professional sports: ticket sales and advertising revenues. Brady’s star power can literally be measured in dollars and cents.
Now consider financial services: the usual performance measures used to structure Wall Street compensation — investment return, stock price appreciation, assets under management, and the dollar-value of deals closed, for example — may simply be too noisy to gauge the genuine value-added of even the most talented masters of the universe. The main reason is risk.
Virtually every aspect of financial decision-making involves making trade-offs between risk and reward, and even the most savvy investor is rarely correct more than 55 percent of the time. For example, a trader who generates a profit of 2 percent per day with 55 percent probability and -2 percent per day with 45 percent probability is wrong 45 percent of the time, but over the course of a year the expected compound return of his portfolio is 65 percent! However, the annual standard deviation — a common risk measure — of this trader’s compound return is a whopping 53 percent (for comparison, the annual standard deviation of the S&P 500 is currently around 25 percent). Such a high level of risk implies an 11 percent chance that this highly skilled trader’s annual return is negative in any given year (assuming normality), and a 30 percent chance that over a 10-year period, at least two years will show negative returns.
This example illustrates the challenge of determining who has skill and who got lucky, a considerably more difficult task in financial contexts than in professional sports, manufacturing, retail sales, entertainment, and many other industries. Therefore, we shouldn’t expect the traditional employment contracts and incentives to work the same way in financial services.
Now regarding the arcane probability calculations above, the shareholders of recently failed financial firms might say that this is exactly the kind of fast talking that got them in trouble in the first place, and sadly, they’re probably right. Far too often, sophisticated risk/reward computations have been used to justify enormous payouts to those who happened to be in charge when good outcomes occurred, only to leave shareholders holding the bag when the bad outcomes came along — a case of “Heads I win, tails you lose”.
The only way to meet this challenge is to restore the intellectual balance of power between executives and shareholders by: 1) constructing new compensation contracts that are as sophisticated as the business activities of the managers being compensated, and 2) educating shareholders, auditors, accountants, and lawyers so they can make more intelligent risk-based choices. Such contracts and decisions must be based on new, more accurate and more relevant measures of performance that explicitly account for risk, incentive effects, time horizon, and the uncertainties of the performance-attribution process. One size cannot fit all, and we should expect contracts to differ according to the risk/reward profiles of different financial services and their respective shareholders’ objectives.
Radical compensation reform will be no easy feat given the complexities of most financial institutions and the politics surrounding this effort. However, it will take more than bonus caps if we ever hope to create a more stable and robust financial system that will be free from the tyranny of bubbles, crashes, and credit cycles.