This study analyses all baseball arbitration cases from 1985 - 90 to determine if decisions differ among ethnic groups. Unlike earlier arbitration analyses, extensive controls for individual and team performance are used. Holding performance constant, the results indicate that Latin American players have 33% less chance and Blacks 19% less chance than does a White player of winning a case. These differences are not reflected by more risky salary demands by minority players and only marginally reflected by market salary practices toward minority players. Therefore, bias appears to emanate from arbitrator utility functions.