A Peer-Relative Perspective of the Contract Year Phenomenon Using Bayesian Analysis

This research examines the contract year phenomenon in a new context, by evaluating player performance relative to other players, rather than the same player over time. Since ordinary least-squares (OLS) methods assume independent observations, results may be biased as player performance is dependent on other players. As such, a Bayesian approach to allow for non-independence among player performance is used to evaluate players in contract years relative to their peers. Using novel performance data of National Football League (NFL) players and controlling for player salary, position, experience, and team effects, the results show that contract year players under-perform relative to players not in a contract year. Furthermore, increasing salary exhibits a greater negative effect on the performance of contract year players. The findings suggest managers would get more performance from players, on average, by not having players play during a contract year. Keywords: Contract year, player performance, NFL, Bayesian