If you play well they will come—and vice versa: bidirectional causality in major-league baseball



The Granger-causality test is applied to the annual attendance and win-percentage data for 29 major-league teams. It is shown that bidirectional causality exists for these teams and that there are some essential differences between the original 10 of 16 franchises that comprised the majors in 1903 and the six that relocated between 1953 and 1961. Some differences and some similarities are also seen in the parameter estimates for both blocs of teams, the relocated teams, and seven long-lived expansion franchises. Finally, the parameter estimates are manipulated to yield noise-free equilibrium estimates for both attendance and performance. In tandem, these two sets of estimates provide fodder for speculation as to the futures of each of the extant 23 franchises considered here. Copyright © 2007 John Wiley & Sons, Ltd.