It’s “Black Monday” in the NFL. The Vikings, Redskins, Lions, and Bucs have already fired their coaches today, and more firings are possible before the day is done. There are many variables that can affect the firing of a coach in professional sports. Of course, three easily observable factors are the performance of the coach (winning percentage, playoff appearances, and championships), the investment by the ownership (team payroll), and the sports league (NFL, MLB, NBA, and NHL). There are also intangible factors endemic to each city in America and Canada with a professional sports team that can influence the probability of a coach getting fired.
We decided to estimate a logistic regression model that could explain the probability of getting fired as a function of performance, investment by ownership, and professional league affiliation. We looked at data from all four professional sports leagues over the last twelve years. We then compared the predicted probability from our model of getting fired with the actual firings in each city. In theory, cities with intangible characteristics that make it more likely for a coach to get fired would have actual firings at a higher probability than predicted through our model of performance and investment. We tried several specifications of our model, and these rankings are robust.
Based on our study, the Top 8 Worst Cities (Highest probability for getting fired above predicted) are:
- San Francisco
- Oklahoma City
The Top 8 Best Cities (Lowest probability for getting fired below predicted) are:
- Salt Lake City
- Los Angeles
It’s interesting to note that the top 8 worst cities does not include big media markets like New York, LA or Chicago, where one might think there is large expectation for winning.
Mike Lewis & Manish Tripathi, Emory University 2013.