I present three essays on the economics of sport, examining salary discrimination in Major League Baseball (MLB), determinants of dismissals of college football coaches, and star player effects on attendance in the National Basketball Association (NBA).

New Evidence of Salary Discrimination in Major League Baseball. Salary discrimination in MLB has largely been discarded as a research topic. However traditional quantitative methods (particularly least squares regression) have concentrated on the effect of race for the average player. This is not where we should expect to find discrimination, as the relative cost of discriminating against a better player is surely higher. I use quantile regression to uncover salary discrimination against black players in the lower half of the salary distribution. Not only are the premia for white and Hispanic players statistically significant, but they are large: up to 25% for the bottom quintile of players. I also demonstrate that racial effects may be obscured in typical baseball salary regressions when they suffer from omitted variable bias, brought about by failing to properly consider speed and fielding ability.

Win or Go Home: Why College Football Coaches Get Fired. Models of dismissals of sports executives frequently ignore the development of expectations regarding performance. I explore the interplay between these expectations and the coachs tenure by examining dismissals of college football head coaches from 1983 to 2006. Using a discrete-time hazard model, I demonstrate that schools use prior performance in two ways: to evaluate the ability of the coach, and to establish performance standards for retention. As recent performance is more relevant for estimating ability, I show that stronger recent performances decrease the chance of dismissal, but stronger historic performances increase the chance of dismissal. Results describe a continual learning process on the part of schools. I also consider the effects of race, insider-ness, rivalries, and rules violations on retention.

Day to Day with the NBA Superstars. I use censored regression (Tobit) analysis on NBA game-level attendance data to examine the superstar externality effect identified by Berri and Schmidt (2006): star players increase road attendance, and since home teams (at least in the NBA) retain 100% of the gate revenues, this constitutes an externality. Game-level data has several advantages over the season-level data employed by Berri and Schmidt: the ability to directly control for sellouts, estimation of dynamic intra-season attendance effects, and the lack of sensitivity to star player distribution. A new intra-season effect estimated is a superstar substitution effect accounting for a star's negative effect on attendance at adjacent games, partially mitigating the superstar externality.