Researchers in many disciplines have shown that lower overall firm performance motivates the termination of top managers. Some have also suggested that managerial dismissal stems particularly from shortfalls of performance relative to expectations, whether in relation to corporate executives or the managers of professional sports teams. But relatively few have investigated the role of expectations empirically, owing to difficulties in measuring them. In this article, we address this question theoretically by framing the interaction of performance, expectations, and the use (and potential dismissal) of managers as a labor demand problem, using applied microeconomic tools to demonstrate how variation in expectations and performance can alter how a National Football League (NFL) organization employs player and coaching talent. Empirically, we take advantage of an institutional change-the NFL's adoption of a salary cap and free agency in the early 1990s-to isolate an exogenous change in performance expectations. While the salary cap constrained the amount clubs could spend annually on player talent, it placed no constraint on expenditures for managerial (coaching) talent. This may have led clubs to allocate more resources toward coaching talent and consequently place higher expectations on head coaches, resulting in a greater probability of dismissal in the postcap period. Using data on NFL head coaches from 1979 to 2006 and probit models of turnover, we find that the probability of coach dismissal increased in the postcap era and that concrete measures of an organization's player talent, coaching talent, and on-field performance gained importance as determinants of turnover from the pre- to the postcap era, indicative of increased expectations.
