Racial Salary Discrimination in Major League Baseball: A Closer Look

This paper investigates whether salary discrimination on the basis of race exists in Major League Baseball. To answer this question, free agent contracts from 2012-2015 are analyzed. The dependent variable, which is the natural logarithm of annual salary, is constructed by dividing the total value of each contract by the number of years each contract specifies. Control variables include the average WAR (Wins Above Replacement) for each player under contract, a dummy variable signifying a player’s race, a player’s contract length. The fact that previous literature on the subject often fails to account for contract length is not to be overlooked; including contract length should add significant robustness, and the proper methodology under this presupposition involves using a two-stage least squares regression to solve the endogeneity issue arising from simultaneous causality between contract length and annual salary. First, though, it will be necessary to test whether there is an endogeneity problem in the first place. To answer this question, the results of a Hausman test will be provided in section IV of this paper. The motivation of this paper is grounded in the postulate that it is necessary to check for endogeneity between contract length and salary, and, if it exists, to use a two-stage least squares regression to account for the issue. The results of the Hausman test show that no endogeneity exists between contract length and the dependent variable, so an OLS regression is appropriate. Both an OLS regression and a TSLS regression show an insignificant coefficient on the race variable. Further, regressions for only long term contracts and for each tertile of both average WAR and the natural logarithm of average annual salary show the same results. It is thus concluded that no evidence of salary discrimination exists in major league baseball according to this study.