If the north exposure, with its neutral light and attractive views, was the value buy in this market, the overhyped commodity was square footage. The difference between an office of 190 square feet and one of 210 square feet is not a noticeable difference. Most people who visit the school don’t even realize that offices differ in size. But if the only thing you are staring at on a spreadsheet is a list of offices with their measurements, this factor is bound to be overweighted. If there is a number, people will use it.
In hindsight I think that some of the furor created by explicitly ranking the faculty members could have been mitigated if the process had been a bit more transparent. For example, it might have been a good idea to make the number of bins public. This would have at least reassured Clyde that he had not been deliberately slotted into one of the later picks.
I also put a bit of the blame on the architect, Rafael Viñoly, and his team. Although they had dutifully spent hundreds of hours talking to students, faculty, and administrators about how the building would be used, and the result is a space both aesthetically pleasing and highly functional, no one told the architect how the offices would be assigned. Had he known, he might have avoided corner offices altogether. One small change he could have made, even late in the game, was to make the office that Doug Diamond took a bit smaller. Doug’s office is on the fifth floor, on the northeast corner, and, to rub salt in the wounds of the unlucky, it is the biggest office of them all. At the time I suggested that, if possible, the architect should chop some of his office off and give it to one of his neighbors, so that there would be a less obvious first choice. But he was only an architect; the term “choice architect” had not yet been invented.
________________
* John was also a big basketball fan who would regularly win the NBA fantasy basketball league. A few years after this episode, he ended up serving as agent for the 7-foot-6-inch basketball star Yao Ming.
† I did show John a draft of this chapter and asked for comments. He neither confirmed nor denied the details of my reconstruction of how things went, but he did concede that I had the basic facts right.
‡ When people in this chapter are identified by first name only, they are real characters but with fictitious names.
29
Football
Of the many unique aspects of the so-called job of being a professor at a top research university, the one that I most prize is the freedom to think about almost anything I find interesting and still get to call it work. You have already seen that I managed to write a paper about the mental accounting of wine drinkers. The next two chapters delve into other domains that on the surface may seem frivolous: player selection in the National Football League, and decision-making by contestants on television game shows. What the topics have in common is that they provided unique ways of studying decision-making at high stakes, and thus a reply to those critics who kept (and keep) bringing up the claim from the Gauntlet that behavioral biases go away when the stakes are high enough.
One version of this critique, which applies to the study of the National Football League, comes from Gary Becker, the most distinguished of the many practitioners of Chicago price theory.* I will call this critique the Becker conjecture. Becker believed that in competitive labor markets, only people who are able to perform their jobs like Econs are able to land the key positions. Becker made this conjecture when he was asked his opinion of behavioral economics. “Division of labor strongly attenuates if not eliminates any effects [caused by bounded rationality.] . . . [I]t doesn’t matter if 90 percent of people can’t do the complex analysis required to calculate probabilities. The 10 percent of people who can will end up in the jobs where it’s required.” In this chapter, we test the Becker conjecture. Does it apply to the owners, general managers, and coaches of the teams in the National Football League? Spoiler alert: it doesn’t.
My research about the National Football League was done with my former student Cade Massey, who now teaches at the Wharton School of Business. Similar to my experience with Werner DeBondt, I first met Cade when he was an MBA student, during my first year at the University of Chicago. I was impressed with his intuitive understanding of what makes people tick, and what makes a research project interesting. I encouraged him to continue his studies and pursue a PhD, and luckily for both of us, as well as for the students who are fortunate enough to take a class from him, he agreed to do so.
Our football paper is nominally about a peculiar institution called the NFL draft. In the NFL, teams get to pick players in a manner similar to the way we picked offices. And, not to worry: it is not necessary to care about American football to understand this chapter and its implications. In the end, this is a chapter about a problem that every organization faces: how to choose employees.
Here is how the NFL draft works. Once a year in late spring, the teams select prospective players. Almost all the candidates have been playing football at an American college or university, giving the professional scouts and general managers an opportunity to see how they play. The teams take turns choosing players, with the order of the picks determined by the teams’ records the previous year. The team with the worst record picks first and the team that wins the championship picks last. There are seven rounds of the draft, meaning each team starts out with seven “picks,” though there are additional picks handed out for reasons that are not important to our story. For the initial contract period, usually four or five years, an athlete can only play for the team that drafted him. When that contract runs out or the player is dropped from the team, the player is declared a free agent, and he can sign with whatever team he wants.
A key feature of this environment, which differs from the Chicago Booth office draft, is that teams are allowed to trade their picks. For example, the team with the fourth pick might agree to give up that pick in return for two or more later picks. There are a sufficient number of trades (over 400 in our sample) to make it possible to infer how teams value the right to pick earlier. Teams can also trade picks this year for picks in future years, which provides a way of examining the teams’ time preferences.
Before we started this project, Cade and I had a strong hunch that there was some serious misbehaving going on in this environment. Specifically, we thought that teams were putting too high a value on the right to pick early in the draft. Part of this feeling was based on observing a few extreme examples. One of the most famous involved a larger-than-life character named Mike Ditka, a legendary former player who became the coach of the New Orleans Saints.
In the 1999 draft, Ditka decided that the only thing stopping the Saints from winning a championship soon was the acquisition of one player, a running back named Ricky Williams. The Saints owned the number twelve pick, and Ditka was worried that Williams would be snapped up before their turn came, so he announced publicly that he would be willing to trade away all of his picks if he could get Williams (not the smartest negotiation strategy). When it was the Washington Redskins’ turn at the fifth pick and Ricky Williams was still available, the Saints were able to complete the trade Ditka wanted, although at a very steep price. Specifically, to move from the twelfth pick to the fifth pick, the Saints gave up all the picks they had in the current draft plus their first- and third-round picks the following year. Those latter picks turned out to be particularly costly to give away, because the Saints ended up as the second worst team in the league in 1999, meaning they gave away the second pick in the entire draft in 2000. Clearly, snagging Williams was not enough to turn the team around, and Ditka was fired. Williams played four years for the Saints and was a very good but not transformative player, and the team could have used the help of all the players they might have acquired with the draft picks they traded away. Cade and I wondered: why would anyone make such a trade?
The Saints’ trade was just an extreme example of the general behavior we thought we would find, namely overvaluing the right to pick early. Five findings from the psychology of decision-making supported our
hypothesis that early picks will be too expensive:
1. People are overconfident. They are likely to think their ability to discriminate between the ability of two players is greater than it is.
2. People make forecasts that are too extreme. In this case, the people whose job it is to assess the quality of prospective players—scouts—are too willing to say that a particular player is likely to be a superstar, when by definition superstars do not come along very often.
3. The winner’s curse. When many bidders compete for the same object, the winner of the auction is often the bidder who most overvalues the object being sold. The same will be true for players, especially the highly touted players picked early in the first round. The winner’s curse says that those players will be good, but not as good as the teams picking them think. Most teams thought that Ricky Williams was an excellent prospect, but no one loved him as much as Mike Ditka.
4. The false consensus effect. Put basically, people tend to think that other people share their preferences. For instance, when the iPhone was new I asked the students in my class two anonymous questions: do you own an iPhone, and what percentage of the class do you think owns an iPhone? Those who owned an iPhone thought that a majority of their classmates did as well, while those who didn’t thought iPhone ownership uncommon. Likewise in the draft, when a team falls in love with a certain player they are just sure that every other team shares their view. They try to jump to the head of the line before another team steals their guy.
5. Present bias. Team owners, coaches, and general managers all want to win now. For the players selected at the top of the draft, there is always the possibility, often illusory, as in the case of Ricky Williams, that the player will immediately turn a losing team into a winner or a winning team into a Super Bowl champion. Teams want to win now!
So our basic hypothesis was that early picks were overvalued, meaning that the market for draft picks did not satisfy the efficient market hypothesis. Fortunately, we were able to get all the data we needed to rigorously test this hypothesis.
The first step in our analysis was just to estimate the market value of picks. Since picks are often traded, we could use the historical trade data to estimate the relative value of picks. If you want to get the fifth pick and you have the twelfth pick, as Ditka did, how much do you normally have to throw in to make that trade? The outcome of that analysis is shown in figure 18 below. The dots are specific trades that we used to estimate the curve. There are two things that jump out from this figure. The first is that it is very steep: the first pick is worth about five times as much as the thirty-third pick, the first one taken in the second round. In principle, a team with the first pick could make a series of trades and end up with five early picks in the second round.
FIGURE 18
The other thing to notice about this figure is how well the curve fits the data. The individual trades, represented by the dots, lie very close to the estimated line. In empirical work you almost never get such orderly data. How could this happen? It turns out the data line up so well because everyone relies on something called the Chart, a table that lists the relative value of picks. Mike McCoy, a minority owner of the Dallas Cowboys who was an engineer by training, originally estimated the Chart. The coach at the time, Jimmy Johnson, had asked him for help in deciding how to value potential trades, and McCoy eyeballed the historical trade data and came up with the Chart. Although the Chart was originally proprietary information only known by the Cowboys, eventually it spread around the league, and now everyone uses it. Figure 19 shows how highly the chart values first-round picks.
FIGURE 19
When Cade and I tracked down Mr. McCoy, we had a nice conversation with him about the history of this exercise. McCoy stressed that it was never his intention to say what value picks should have, only the value that teams had used based on prior trades. Our analysis had a different purpose. We wanted to ask whether the prices implied by the chart were “right,” in the efficient market hypothesis sense of the term. Should a rational team be willing to give up that many picks in order to get one of the very high ones?
Two more steps were required to establish our case that teams valued early picks too highly. The first of these was easy: determine how much players cost. Fortunately, we were able to get data on player compensation. Before delving into those salaries, it is important to understand another peculiar feature of the National Football League labor market for players. The league has adopted a salary cap, meaning an upper limit on how much a team can pay its players. This is quite different from many other sports, for example Major League Baseball and European soccer, where rich owners can pay as much as they want to acquire star players.
The salary cap is what makes our study possible. Its existence means that each team has to live within the same budget. In order to win regularly, teams are forced to be economical. If a Russian oligarch wants to spend hundreds of millions of dollars to buy a soccer superstar, one can always rationalize the decision by saying that he is getting utility from watching that player, as with buying an expensive piece of art. But in the National Football League, acquiring an expensive player, or giving away lots of picks to get a star like Ricky Williams, involves explicit opportunity costs for the team, such as the other players that could have been hired with that money or drafted with those picks. This binding budget constraint means that the only way to build a winning team is to find players that provide more value than they cost.
The league also has rules related to rookie salaries. The compensation of first-year players, by draft order, is shown in figure 20. The figures we use here are the official “cap charge” that the team is charged, which includes the player’s salary plus an amortization of any signing bonus paid up front. Figure 20 shares many features of figure 18. First of all, the curve is quite steep. High picks are paid much more than lower-round picks. And again, the estimated line is a very good fit for the data because the league pretty much dictates how much players are paid in their initial contracts.
FIGURE 20
So high picks end up being expensive in two ways. First, teams have to give up a lot of picks to use one (either by paying to trade up, or in opportunity cost, by declining to trade down). And second, high-round picks get paid a lot of money. The obvious question is: are they worth it?
Another way of asking this question is: what would have to be true to make the price of early picks rational, and is it in fact true? The price says that, on average, the first player taken in the draft is five times better than the thirty-third player. That fact alone does not tell us anything, since players’ values can vary by much more than a 5:1 ratio. Some players are perennial all-stars who can transform a team. Others are complete busts that cost the team a lot of money and provide little in return. In fact, high-profile busts actually hurt performance because the teams are unable to ignore sunk costs. If a team is paying a high draft pick a lot of money, it feels under a lot of pressure to put him in the game, regardless of how well he is playing.
The key appears to be how good a team’s managers are at distinguishing between stars and busts. Here is a simple thought experiment. Suppose you rank all the players taken at a given position (quarterback, wide receiver, etc.) by the order in which they were picked. Now take two players drafted consecutively, such as the third running back and the fourth. What is the chance that the player taken earlier is better by some objective measure? If the teams were perfect forecasters, then the player taken first would be better 100% of the time. If the teams have no ability, then the earlier pick will be better half the time, like flipping a coin. Take a guess at how good teams are at this task.
In reality, across the entire draft, the chance that the earlier player will be better is only 52%. In the first round it is a bit higher, 56%.† Keep that thought in mind, both as you read the rest of this chapter and the next time you want to hire someone and are “sure” you have found the perfect candidate.
Although this result gives a st
rong hint of how our analysis would come out, it is worthwhile to provide an outline of our more thorough evaluation. We followed the performance of each player drafted during our study period for the duration of his initial contract. Then, for each player-year, we assigned an economic value to the performance of that player; in other words, we estimated the value the player provided to the team that year. We did so by looking at how much it would cost to hire an equivalent player (by position and quality) who was in the sixth, seventh, or eighth year of his contract, and was thus being paid the market rate, because after his initial contract ran out he became a free agent. A player’s performance value to the team that drafted him is then the sum of the yearly values for each year he stays with the team until his initial contract runs out. (After that, to retain him, they will have to pay the market price or he can jump to another team.)
In figure 21, we plotted this total “performance value” for each player, sorted by draft order, as well as the compensation curve shown in figure 20. Notice that the performance value curve is downward-sloping, meaning that teams do have some ability to rate players. Players who are taken earlier in the draft are indeed better, but by how much? If you subtract the compensation from the performance value, you obtain the “surplus value” to the team, that is, how much more (or less) performance value the team gets compared to how much it has to pay the player. You can think of it like the profit a team gets from the player over the length of his initial contract.
Misbehaving: The Making of Behavioral Economics Page 30