Life on the Red Carpet: Star Players and Referee Bias in the National Basketball Association

Abstract

Fans of the National Basketball Association (NBA) have long considered the idea that NBA referees are biased in various ways, such as when certain “star players” benefit from so-called “phantom fouls” committed against them or are sheltered from calls against fouls they commit. Using two data sets, the first based on Wallace, Caudill, and Mixon (2013), and a second based on all player games during the 2011 NBA Playoffs series, including dozens of NBA All Stars, the present study empirically investigates this potential form of referee bias by examining both early- and late-game free throw shooting in these most critical games of any season for NBA players, coaches, owners, and fans. The empirical results suggest that marquee NBA players are the beneficiaries of referee bias that occurs near the end of NBA Playoffs contests. More specifically, through regression models with various fixed effects, we find that NBA All Stars are awarded with an additional 0.32 free attempts per minute during the fourth quarter of NBA Playoff games.

Key Words:

• Sports Stars
• Referee Bias
• Professional Basketball.

JEL classifications:

• D03
• L83.

The authors thank three anonymous referees of the Journal and H.E. Frech for valuable comments on a previous version. The usual caveat applies.

Notes

1. Caudill and Mixon (1998) show that mid-1980s changes in the NBA’s playoffs format led to a lengthier playoff series.

2. Across the entire sample of 48 NBA players, 25 were named to one or more of these two All Stars teams (2010 and 2011), while 23 were not. The difference between the ratio of All Stars (0.521) and that of common players (0.479) is, at 0.042, relatively small. In fact, this difference is not statistically significant (one-tailed z-prob = 0.386) at any conventional level, indicating that this sample is evenly divided between all-star caliber NBA players and common NBA players.

3. We do recognize the mechanical relationship wherein if a player is fouled while shooting and the shot attempt is unsuccessful, then the play is not counted as a field goal attempt. This relationship should, however, affect all players in all quarters of a game. It is also worth noting here that there are perhaps two types of end-of-the-game fouls from which stars might benefit. The first we refer to as “protective” fouls. These are the subject of our study, and they represent fouls called by officials to minimize defensive effort and allow the stars to perform. The other type of foul we refer to as “strategic.” The fouls are intentional and made with the hope that sending the other team to the free throw line will result in missed free throws and regaining possession of the ball. Whereas protective fouls are called against defenders of the star players, the strategic fouls may be committed against any player, with a strong preference for a poor foul shooter. The main empirical difference between the two is that strategic fouls are committed (usually) by players of teams that are losing, while protective fouls may be called against players whose team is leading. This “fouls typology” represents an interesting empirical issue, investigation of which requires much more detailed data. In the meantime, we do provide some circumstantial evidence that what we are finding in this paper is protective fouling.

4. Inclusion among this group means that player i’s regular season free throw shooting percentage was among the top 20 free throw shooting percentages during the NBA’s regular season, provided that player i also attempted 200 or more free throws during the NBA’s regular season.

5. To restate the latter result, the insignificant coefficient estimate for SHOOTER suggests that good free throw shooters are not taking appreciably more free throws in the fourth quarter of NBA Playoffs games. Anecdotally, the top two players in terms of number of free throw attempts during the season are Dwight Howard and Blake Griffin, both of whom are being fouled because they are poor free throw shooters.

6. The All Stars in the sample attempt an average of 0.394 field goals per minute and an average of 0.213 free throws per minute, both during the fourth quarter, for a ratio of 1.850. Our dummy variable coefficient on ALLSTAR indicates an additional 0.320 free throw attempts per minute. Using free throws attempted during quarters one through three as a base yields 0.460 free throw attempts per minute, which, if multiplied by 1.850, indicates a required field goals attempted per minute (during the first three quarters) of 0.850. Only 5 of 255 cases of All Star player-games reach this level: Derrick Rose (1), Kobe Bryant (2), and Russell Westbrook (2). These calculations provide some circumstantial evidence that this study’s findings relate to “protective” fouling.