Everything Is Obvious
Page 22
One increasingly popular approach is to pay bonuses that are effectively withheld by the employer for a number of years. The idea is that if outcomes are really random in the sense of a simple coin toss, then basing compensation on multiyear performance ought to average out some of that randomness. For example, if I take a risky position in some asset whose value booms this year and tanks a year from now, and my bonus is based on my performance over a three-year period, I won’t receive any bonus at all. It’s a reasonable idea, but as the recent real estate bubble demonstrated, faulty assumptions can appear valid for years at a time. So although stretching out the vesting period diminishes the role of luck in determining outcomes, it certainly doesn’t eliminate it. In addition to averaging performance over an extended period, therefore, another way to try to differentiate individual talent from luck is to index performance relative to a peer group, meaning that a trader working in a particular asset class—say, interest rate swaps—should receive a bonus only for outperforming an index of all traders in that asset class. Put another way, if everybody in a particular market or industry makes money at the same time—as all the major investment banks did in the first quarter of 2010—we ought to suspect that performance is being driven by a secular trend, not individual talent.
Delaying bonuses and indexing performance to peers are worthy ideas, but they may still not solve the deeper problem of differentiating luck from talent. Consider, for example, the case of Bill Miller, the legendary mutual fund manager whose Value Trust beat the S&P 500 fifteen years straight—something no other mutual fund manager has ever accomplished. Over this period, Miller’s success seems like a textbook case of talent trumping luck. He really did outperform his peers, year after year, for fifteen years—a winning streak that, as the investment strategist Michael Mauboussin has shown, would be extremely unlikely to have occurred in the historical population of fund managers if everyone were tossing proverbial coins.9 At the end of his streak, therefore, it would have been hard to dispute that Miller was doing something special. But then, in the three-year period from 2006 to 2008, right after his record streak ended, Miller’s performance was bad enough to reverse a large chunk of his previous gains, dragging his ten-year average below that of the S&P. So was he a brilliant investor who simply had some bad luck, or was he instead the opposite: a relatively ordinary investor whose ultimately flawed strategy just happened to work for a long time? The problem is that judging from his investing record alone, it’s probably not possible to say. Just as Michael Raynor explained for business strategies, like Sony versus Matsushita in the video war described in Chapter 7, investing strategies can be successful or unsuccessful for several years in a row for reasons that have nothing to do with skill, and everything to do with luck. Naturally, it won’t seem like luck, but there is no way to know that whatever story is concocted to explain that success isn’t simply another manifestation of the Halo Effect.
To be sure that we are not just falling for the Halo Effect, we really need a different measure of performance altogether—one that assesses individual skill directly rather than by inferring it from outcomes that might be determined by forces beyond the individual’s control. At the end of his streak, Miller was compared a lot with Joe DiMaggio, who had a famous hitting streak of fifty-six consecutive games during the 1941 baseball season. Superficially, the streaks are analogous, but in DiMaggio’s case, we also know that his career batting average was 0.3246, which is the 44th highest in baseball history, and that during the time of the streak it was an astonishing .409.10 So although there was still an element of luck in DiMaggio’s streak, his skill ensured that he was more likely to be “lucky” than most other players.11
Ideally then, we would like to have the equivalent of a batting average to measure performance in different professions. But outside of sports, unfortunately, such statistics are not so easy to put together.12 The reason is that outcomes in sports are generally repeated many times under close-to-identical conditions. A baseball player may have six hundred at-bats in a single season, and many thousands throughout his career, each one of which is, roughly speaking, an independent test of individual skill. Even for more ephemeral skills, like outstanding positional play in professional basketball, that are harder to measure directly but still help the team win, we have almost one hundred NBA games each season that we can watch to observe a player’s effect on his team and on the outcome.13 At first, it seems that an accomplishment like beating the S&P 500 for the year is a pretty good equivalent of a batting average for fund managers—and indeed fund managers with long streaks do tend to beat the S&P 500 more often than average, just like baseball players with high batting averages. By this measure, however, in a forty-year career a fund manager will get only forty “at-bats” total—simply not enough data to estimate the true value with any confidence.14
THE MATTHEW EFFECT
And finance is in many respects an easy case—because the existence of indices like the S&P 500 at least provide agreed-upon benchmarks against which an individual investor’s performance can be measured. In business or politics or entertainment, however, there is much less agreement about how to measure individual skill, and even fewer independent trials over which to measure it. Most important of all, serial accomplishments are usually not independent demonstrations of skill in the way that, say, each of Roger Federer’s grand slam victories in tennis are independent. One could argue that Federer’s reputation can intimidate opponents, thereby giving him a psychological edge, or that tournament draws are organized in such a way that the top seeds don’t play each other until the later rounds—all of which could be seen as an advantage deriving from his previous success. Nevertheless, every time Federer walks out on the court he has to win under more or less the same circumstances as the very first time he played professional tennis. No one would think it fair to give him, say, an extra serve, or the ability to overrule the referee, or any other advantage over his opponent, just because he’s won so often in the past. Likewise, it would be outrageous to give the team that wins the first of seven games in an NBA playoff series ten extra points at the start of the second game. In sports, that is, we place tremendous importance on making the playing field as level as possible and every test of skill independent from every other.
Much of life, however, is characterized by what the sociologist Robert Merton called the Matthew Effect, named after a sentence from the book of Matthew in the Bible, which laments “For to all those who have, more will be given, and they will have an abundance; but from those who have nothing, even what they have will be taken away.” Matthew was referring specifically to wealth (hence the phrase “the rich get richer and the poor get poorer”), but Merton argued that the same rule applied to success more generally. Success early on in an individual’s career, that is, confers on them certain structural advantages that make subsequent successes much more likely, regardless of their intrinsic aptitude. In science, for example, junior scientists who land jobs at top research universities tend to experience lighter teaching loads, attract better graduate students, and have an easier time getting grants or publishing papers than their peers who end up at second- or third-tier universities. As a result, two individuals in the same field who may have been roughly comparable at the beginning of their careers may experience dramatically different levels of success five or ten years down the road on no more grounds than that they were hired at different institutions. And from there, it just gets more unequal still. Successful scientists also tend to receive a disproportionate share of the credit for anything with which they are associated, as when they write papers with unknown graduate students, who may have actually done most of the work or had the key ideas. Once someone is perceived as a star, in other words, not only can he attract more resources and better collaborators—thus producing far more than he would otherwise have been able to—he also tends to get more than his fair share of the credit for the resulting work.15
Merton was writing about scientific
careers, but as the sociologist Daniel Rigney argues in his recent book The Matthew Effect, the same forces apply to most other careers as well. Success leads to prominence and recognition, which leads in turn to more opportunities to succeed, more resources with which to achieve success, and more likelihood of your subsequent successes being noticed and attributed to you. Isolating the effects of this accumulated advantage from differences in innate talent or hard work is difficult, but a number of studies have found that no matter how carefully one tries to select a pool of people with similar potential, their fortunes will diverge wildly over time, consistent with Merton’s theory. For example, it is known that college students who graduate during a weak economy earn less, on average, than students who graduate in a strong economy. On its own, that doesn’t sound too surprising, but the kicker is that this difference applies not just to the years of the recession itself, but continues to accumulate over decades. Because the timing of one’s graduation obviously has nothing to do with one’s innate talent, the persistence of these effects is strong evidence that the Matthew Effect is present everywhere.16
Typically we don’t like to think that the world works this way. In a meritocratic society, we want to believe that successful people must be more talented or must have worked harder than their less-successful counterparts—at the very least they must have taken better advantage of their opportunities. Just as when we try to understand why some book became a bestseller, when we try to explain why some individual is rich or successful common sense insists that the outcome arises from some intrinsic quality of the object or person in question. A best-selling book must be good somehow or else people wouldn’t have bought it. A wealthy man must be smart in some manner or else he wouldn’t be rich. But what the Halo Effect and the Matthew Effect should teach us is that these commonsense explanations are deeply misleading. It may be true that abjectly incompetent people rarely do well, or that amazingly talented individuals rarely end up as total failures, but few of us fall into those extremes. For most of us, the combination of randomness and cumulative advantage means that relatively ordinary individuals can do very well, or very poorly, or anywhere in between. But because every individual’s story is unique, we can always persuade ourselves that the outcome we have witnessed was somehow a product of their unique attributes.
None of this is to say, of course, that people, products, ideas, and companies don’t have different qualities or abilities. Nor does it suggest that we should stop believing that quality should lead to success. What it does suggest, however, is that talent ought to be evaluated on its own terms. One doesn’t have to know Roger Federer’s ranking to appreciate that he is a great tennis player—you can tell that simply by watching him. In the same way, if everyone who knows Bill Miller agrees that he is an exceptionally smart and thoughtful investor, then he probably is. As Miller himself has emphasized, statistics like his fifteen-year streak are as much artifacts of the calendar as indicators of his talent.17 Nor is it even the case that his talent ought to be judged by his cumulative career success—because that too could be undone by a single unlucky stroke. As unsatisfying as it may sound, therefore, our best bet for evaluating his talent may be simply to observe his investing process itself.18 What we conclude may or may not correlate with his track record, and it is almost certainly a more difficult evaluation to perform. But whenever we find ourselves describing someone’s ability in terms of societal measures of success—prizes, wealth, fancy titles—rather than in terms of what they are capable of doing, we ought to worry that we are deceiving ourselves. Put another way, the cynic’s question, if you’re so smart, why aren’t you rich? is misguided not only for the obvious reason that at least some smart people care about rewards other than material wealth, but also because talent is talent, and success is success, and the latter does not always reflect the former.19
THE MYTH OF THE CORPORATE SAVIOR
If separating talent from success seems difficult, it is especially hard when performance is measured not in terms of individual actions, like an investment banker’s portfolio, but rather the actions of an entire organization. To illustrate this problem, let’s step away from bankers for a moment and ask a less-fashionable question: To what extent should Steve Jobs, founder and CEO of Apple Inc., be credited with Apple’s recent success? Conventional wisdom holds that he is largely responsible for it, and not without reason. Since Jobs returned in the late 1990s to lead the company that he founded in 1976 with Steve Wozniak in a Silicon Valley garage, its fortunes have undergone a dramatic resurgence, producing a string of hit products like the iMac, the iPod, and the iPhone. As of the end of 2009, Apple had outperformed the overall stock market and its industry peers by about 150 percent over the previous six years, and in May 2010 Apple overtook Microsoft to become the most valuable technology company in the world. During all this time, Jobs has reportedly received neither a salary nor a cash bonus—his entire compensation has been in Apple stock.20
It’s a compelling story, and the list of Apple’s successes is long enough that it’s hard to believe it’s all due to chance. Nevertheless, because Apple’s history can only be run once, we can’t be sure that we aren’t simply succumbing to the Halo Effect. For example, as, I discussed in Chapter 7, the iPod strategy had a number of elements that could easily have led it to fail, as did the iPhone. Microsoft CEO Steve Ballmer looks silly now for scoffing at the idea that consumers would pay $500 for a phone that locked consumers into a two-year contract with AT&T and didn’t have a keyboard, but it was actually quite a reasonable objection. Both products now seem like strokes of genius, but only because they succeeded. Had they instead bombed, we would not be talking about Jobs’s brilliant strategy and leadership that simply didn’t work. Rather, we would be talking about his arrogance and unwillingness to pay attention to what the market wanted. As with all explanations that depend on the known outcome to account for why a particular strategy was good or bad, the conventional wisdom regarding Apple’s recent success is vulnerable to the Halo Effect.
Quite aside from the Halo Effect, however, there is another potential problem with the conventional wisdom about Apple. And that is our tendency to attribute the lion’s share of the success of an entire corporation, employing tens of thousands of talented engineers, designers, and managers to one individual. As with all commonsense explanations, the argument that Steve Jobs is the irreplaceable architect of Apple’s success is entirely plausible. Not only did Apple’s turnaround begin with Jobs’s return, after a decade of exile, from 1986 to 1996, but his reputation as a fiercely demanding manager with a relentless focus on innovation, design, and engineering excellence would seem to draw a direct line between his approach to leadership and Apple’s success. Finally, large companies like Apple need a way to coordinate the activities of many individuals on a common goal, and a strong leader seems required to accomplish this coordination feat. Because this role of leader is by definition unique, the leader seems unique also, and therefore justified in receiving the lion’s share of the credit for the company’s success.
Steve Jobs may in fact be such an individual—the sine qua non of Apple. But if he is, he is the exception rather than the rule in corporate life. As sociologist and Harvard Business School professor Rakjesh Khurana argues in Searching for a Corporate Savior, corporate performance is generally determined less by the actions of CEOs than by outside factors, like the performance of the overall industry or the economy as a whole, over which individual leaders have no control.21 Just as with the hubs and influencers that I discussed in Chapter 4, Khurana concludes that conventional explanations of success invoke the power of inspirational leaders not because the evidence supports that conclusion, but rather because without such a figure we do not have any intuitive understanding how a large, complex entity functions. Our need to see a company’s success through the lens of a single powerful individual, Khurana explains, is the result of a combination of psychological biases and cultural beliefs—particularly in cultures
like the United States, where individual achievement is so celebrated. The media, too, prefers simple, human-centered narratives to abstract explanations based on social, economic, and political forces. Thus we both gravitate to, and are also disproportionately exposed to, explanations that emphasize the influence of special individuals in directing the course of incredibly complex organizations and events.22
Reinforcing this mentality is the peculiar way in which corporate leaders are selected. Unlike regular markets, which are characterized by large numbers of buyers and sellers, publicly visible prices, and a high degree of substitutability, the labor market for CEOs is characterized by a small number of participants, many of whom are already socially or professionally connected, and operates almost entirely out of public scrutiny. The result is something like a self-fulfilling prophecy. Corporate boards, analysts, and the media all believe that only certain key people can make the “right” decisions; thus only a few such people are considered for the job in the first place. This artificial scarcity of candidates in turn empowers the winners to extract enormously generous compensation packages, which are then presented as evidence that “the market” has valued the candidate rather than a small group of like-minded individuals. Finally, the firm is then either successful, in which case obviously the “right” leader was chosen, or it is not successful, in which case the board made a mistake and a new leader is sought out. Sometimes “failed” CEOs walk away with huge severance packages, and it is these instances that tend to get all the attention. In Khurana’s view, however, the outrage that is often expressed over these instances nevertheless perpetuates the mistaken belief that a firm’s performance can be attributed to any one individual—even the CEO—in the first place. If boards were more willing to question the very idea of the irreplaceable CEO, and if searches for CEOs were then opened to a wider pool of candidates, it would be more difficult for candidates to negotiate such extravagant packages at the outset.23