Economists love using football to test their theories. Call it Gridiron-omics 101. What have they learned? Businesses are bad at judging talent, performance, profit, and strategy.
Pity Mike Smith. The Atlanta Falcons coach infuriated fans on Sunday when he called an overtime fourth down play that backfired in a loss against the New Orleans Saints. The controversial decision -- the Falcons were inches from a first down, but back at their own 29-yard line -- failed spectacularly, setting up a game-ending field goal from the Saints. Yet according to one economics study, Smith's well-documented tendency to go for it on fourth will probably end up winning his team a lot more games than it'll lose them.
Yes. An economics study.
Pro football has become one of academia's favorite labs for testing out economic theories. The reasons are pretty plain. It's a high-stakes game with richly paid professionals. There are clear winners and losers. There's a cut-throat market for talented labor. Each game involves an intricate web of choices (ah, incentives!) that yield reams and reams of data.
Economists think they have a lot to teach football. But what can football teach us about economics and business?
Here are four lessons for business from Gridironomics 101.
(Or: Why Companies Are Bad at Maximizing Profits)
In 2003, Berkely professor David Romer released a classic paper asking a simple question: Do companies really maximize their profits? You would think so. But testing that idea in the real world is tricky. Who can really say if GE makes the ideal decisions about how to market is microwaves?
So to get an answer, Romer turned the gridiron. Specifically, he looked at the same dilemma the Falcons faced: Whether or not to go for it on fourth down. Given the incentives to win, Romer reasoned that coaches would make the optimal choice. Turns out, that wasn't the case.
Using a probability analysis, Romer looked at data form more than 700 NFL games to measure the value of getting a first down at any point on the field against the value of punting the ball or kicking a field goal. Going for it on fourth was often the best choice, even in some unexpected situations. If a team was on its own side of the field and needed four yards or less for a first down, going for it was the right move. At the opponent's five yard line, the chance of going for it and scoring a touchdown was more valuable than the guaranteed points from a field goal. Overall, being more aggressive on fourth would improve a team's chance of winning each game by 2%. And yet, coaches rarely made the right call.
Why were coaches risk averse at the cost of winning? Romer reasoned that there might be outside pressures forcing them to be conservative, such as anger from fans. Just ask Mike Smith. More likely, coaches really thought they were doing the right thing. They just didn't understand the statistics.
2) MARGIN OF VICTORY MATTERS MOST
(Or: Why Companies Are Bad at Assessing Strategy
Psychologists have found that people are pretty bad at assessing decisions in hindsight. We tend to suffer from "outcome bias." If a decision works out, we think it was the correct call, even if the probability suggested it would fail -- and vice versa.
A group of professors from Brigham Young University wanted to see if that bias occurred in a real-world business setting. Again, pro football, where every play involves an great amount of deliberate decision making, seemed like the natural choice. The team focused its analysis on the most basic element of offense: whether to run the ball or pass it.
The team wanted to know if coaches change the balance of their running and passing attacks based on games they won or lost essentially by chance. If they did, that would demonstrate outcome bias. But how do you tell if a game was won on luck? By the margin of victory.
According to the study, games won by less than a touchdown don't mean much about a team's future success. And yet, even close games had a big impact on coaches' decision making. For instance, teams were twice as likely to keep up the same basic run/pass strategy after a win than a loss. Coaches tended to switch things up after a defeat even if the numbers suggested the result was a fluke. In the end, the researchers wrote, NFL coaches simply seemed to believe their own decision making was far more important to the game's result than it actually was. Hence, chance outcomes impacted long term strategy choices.
(Or: Why Companies Are Bad at Firing People)
If coaches aren't particularly good at evaluating their own performance, is management any better? Not particularly. In another study, BYU found that NFL teams were more or less inept when it came to firing their head coach.
The study started with this premise: Teams would fire a coach either to punish poor coaching or to motivate the team to work harder. In reality, firings didn't seem to address either problem. According to the data, teams tended to fire coaches based on their record during the prior season. But in doing so, they failed to take into consideration the margin of those games, which are the best way to predict long-term performance. A coach who lost a lot of close games would be just as likely to have a good followup season as a coach who lost a lot of close games. Meanwhile, teams didn't improve much after a firing. The next season they tended to fare worse, then eventually climbed back to their old record.
The authors argue that their findings have implications outside of football. Previous studies have found that corporate CEOs are often credited for stock movements that are beyond their control, and that politicians tend win or lose elections based on economic factors that have nothing to do with them. But coaches can't control their owners the way CEOs can capture a board, and unlike politicians, they're hired by sophisticated organizations. In the end, their employers aren't any better are assessing their performance.
(Or: Why Companies Are Bad at Picking Talent)
The team that finishes the season with the poorest record this year will have the privilege of selecting Andrew Luck, the phenom Stanford quarterback, with the first pick of the 2012 draft. (The media have dubbed this race to the bottom "Suck for Luck.") But according to Yale economists Cade Massey and Richard Thaler, that prize might actually be a curse.
In 2005, Massey and Thaler used data from the NFL draft to test whether labor markets effectively price talent. In short: they don't. Not in the NFL, anyway.
The pair looked at every draft class from 1991 to 2004, and compared each player's on-the-field value (measured by factors such as Pro Bowl appearances and games started) to their paycheck. Because NFL teams operate under a salary cap, it would be in management's interest to find not only the best player, but also the best value.
And when it came to value, the first pick of the draft turned out to be worst. Rookie salaries are determined by when they're drafted. The higher they're picked, the more they're paid. Yet, each athlete only had a 52% chance of outperforming the next player picked at their position over the long term. And once salary was factored in, late picks offered much more bang for their buck. Nonetheless, few teams made a concerted effort to trade away their pricey early draft picks for higher value late picks. Why?
Possibly because they were too confident in their own ability to spot talent. At the end of the paper, the two economists suggested that the same trend might be present in corporate America, where boards of directors pick CEOs based on significantly less information than NFL scouts gather on prospects. As they wrote: "Perhaps innovative boards of directors should start looking for the next Tom Brady (pick number 199) as CEO rather than this year's hot young prospect."