The Return of Moneyball
It’s opening night for the movie Moneyball, the big screen’s adaptation of Michael Lewis’ celebrated book. I say celebrated, which is accurate, and deserved. It is easily the best book I read in 2004, and the only book that has ever prompted me to write two academic papers (more on that later). The book triggered an immense amount of discussion among baseball fans, and people in business and finance, along with no small amount of controversy. My favorite symbol of the controversy was the hilarious “Fire Joe Morgan” blog, created in 2005 to poke fun at “ridiculous” commentary from sports journalists. The FJM crew chose Morgan as the figurehead for the site “because of his penchant for ignorantly slamming Michael Lewis’s “Moneyball,” but he was far from alone.
Accompanying the movie’s release has been another large slug of discussion regarding the merits of Lewis’ argument. New critics have emerged as well, and Allen Barra gives them exposure in the provocatively titled WSJ piece, “The Moneyball Myth.” Seven years later, the A’s have returned to mediocrity and the draft picks highlighted by Lewis mostly failed to pan out. Further, Barra points to data on a key statistic in the Moneyball strategy, “on base percentage.” Barra quotes author Alan Hirsch: “Consider Beane’s most revered stat, on-base percentage. In 2002 the American League average was .331, the year before that it was .333. Last season the league’s OBP had dropped to .327, and this year it’s down to .323.” These stats are meant to buttress the Hirsch-Barra claim that “Moneyball did not … change the game very much.”
At some level that is true. Indeed, the game of baseball hasn’t changed much in 100 years, and one alleged strategic innovation and it’s description by a financial author really shouldn’t change the game very much.
Nevertheless, beneath the surface, I believe the “Moneyball” strategy and its diffusion had quite an impact, and I have the data to back it up. When Jahn Hakes and I embarked on our first academic paper on the subject, we thought there was a decent chance that we could refute the economic claims in Moneyball, in particular that players with high OBP were under-priced in the labor market. Any card-carrying economist knows this is inconsistent with equilibrium in a well-functioning, competitive labor market, and were not baseball teams intensely competitive? But instead, Jahn and I found that high OBP players did come cheap, relative to the contribution of their skill to winning baseball games. Intriguingly however, we found that the “OBP discount” vanished in 2004, the year that Moneyball was published. The likely reason: other teams, like Michael Lewis, had looked into what was going on in Oakland, and hired people out of the A’s front office. Now there were multiple bidders for high-OBP players in baseball’s labor market, thus driving up their price.
The multiple bidder story seemed a plausible conjecture, and Hakes and I wanted to look deeper into it. This resulted in our second paper on the topic, which had several goals. First, we wanted to know if the mis-pricing was of short or long duration. We extended our sample back to the mid-1980s, and found that high OBP players were under-priced for the entire 1986-2003 period. Second, the 2004 “correction” was sustained in the subsequent two seasons. It was not a mirage. Finally, we separated walks from base hits in the OBP statistic, to further refine our sense of the source of the mis-pricing. We found the source of the mis-pricing lay in “plate discipline,” i.e. the ability to avoid swinging at bad pitches and accept taking a walk. Moreover, the changes in the returns to plate discipline over the entire period were huge. Compare two figures. In the early, pre-expansion period of 1986-1993, the estimated percentage boost in salary from a one standard deviation increase in the ability to take walks was a measly 2.8%. Post-Moneyball, the figure was 14.0%. The financial returns to the overlooked skill increased by a factor of five. Is that not indicative of a fundamental change in the game? (Intriguingly, we found that the returns increased fairly steadily throughout the sample, to 5% in 1994-97, and 7% in 1998-2003, before doubling after the A’s success. Make of that what you will,)
Strategic innovations on the playing field can cancel each other out. Greater attention to plate discipline, i.e. not swinging at balls out of the strike zone, will be counteracted with adjustments by pitchers. The “low” OBP referred to by Hirsch matches the OBP in the 1986-1993 era. Although recognition of the importance of plate discipline might not be reflected in data on the field of play, the data on player compensation can hardly be shouting any louder on the subject. Yes, the game has changed, if you know where to look for signs of it.
Among the many pieces on Moneyball out there, Chris Gorski of Inside Science “get’s it” and his story is worth a read. As for the movie, the WSJ’s (Fire!) Joe Morgenstern is a fan: ‘Moneyball': Stars, Stats and Perfect Pitch. I’m looking forward to it.