In-Group Favoritism and Peer Effects in Wrongful Acquittals: National Basketball Association Referees as Judges

We provide the first analysis of racial in-group bias in decision errors. Using player-referee matched data from the National Basketball Association, we show that there is no racial bias or in-group bias in foul calls regardless of whether the call was correct or incorrect. But a player is 27 percent more likely to get away with a foul if the referee is of the same race. Thus, racial in-group bias exists not in the actions of referees but in their inactions, and this effect is driven by black referees. Although higher rates of decision errors impact referees’ playoff assignments, month-to-month adjustment in referees’ behavior occurs only for foul call errors, which are more noticeable. Black referees tend to make errors in favor of black players unless they have two white peer referees. In-group favoritism in white referees’ errors arises only when there are two black peers.