A Market Test for Discrimination in the English Professional Soccer Leagues Stefan Szymanski Imperial College This paper proposes a market test for racial discrimination in salary setting in English league soccer over the period 1978–93 using a balanced panel of 39 clubs. If there is a competitive market for the services of players, the wage bill of the club will reflect their productivity and hence the performance of the club in the league. Discrimination can be said to exist if clubs fielding an above-aver- age proportion of black players systematically outperform clubs with a below-average proportion of black players, after one con- trols for the wage bill. Statistically significant evidence of discrimi- nation in this sense is found. I. Introduction Statistical testing for the presence of racial discrimination remains one of the most controversial fields in economics. Discrimination in the sense of unequal pay for equal work is conventionally identified through the construction of an earnings function. This relates per- sonal characteristics that influence productivity to earnings so that any residual differences in the earnings of two groups can be attrib- uted to discrimination. A fundamental criticism of this test is that omitted-variable bias may have created a mistaken impression of dis- crimination. If unobserved characteristics that affect productivity are I am indebted to Michael Crick, Izzet Agoren, Lamin Sabally, Tim Kuypers, and Andrew Craven for their help in compiling the data set used in this paper. I have benefited considerably from the advice of Ron Smith, Steve Machin, and Ian Pres- ton. I would like to thank seminar participants at Imperial College and University College London and an anonymous referee for helpful comments. [Journal of Political Economy, 2000, vol. 108, no. 3]  2000 by The University of Chicago. All rights reserved. 0022-3808/2000/10803-0008$02.50 590