Last year I posted on a paper by Joseph Price (BYU) and Justin Wolfers (Penn) that showed that NBA referees were racially biased. Price is again identifying biases amongst NBA refs, this time with coauthors Marc Remer (Johns Hopkins) and Daniel Stone (Oregon State).
The new working paper finds evidence that NBA referees are more likely to call fouls in favor of (1) home teams, (2) teams that are losing, and (3) teams that are down in the number of playoff games won. To account for the possibility that the three types of biases may be due to players playing differently when they are at home or behind, the authors look at play-by-play data which allows them to distinguish between discretionary turnovers (e.g., shooting fouls, charging) and non-discretionary turnovers (e.g., steals and bad passes). They show that the biases are due to the referees, not players.
The authors argue that each of these biases may increase league profits by (1) making home games more exciting, (2) making games more exciting in general, and (3) extending the number of games in the season. However, they find no evidence that the biases are explicit, and conclude that they are likely implicit.