Why do leaders matter? A revisit of player-turned-coaches from the NBA

ABSTRACTThis article re-estimates equations that have shown elite basketball players do better as leaders when they turn into coaches. Using comprehensive data from the NBA during the last three decades, we show that player-turned-coaches do not win statistically more matches in regular seasons while they do better in playoffs. The relatively better performance is driven by former guard players who lead to more defensive rebounds, free throws made, and points in the last moment of close matches during playoffs. In further analyses of direct channels, we show that guard-turned-coaches as a good example of expert leaders perform relatively better in strategic decision-making. For indirect channel analysis, we also provide statistical evidence that guard-turned-coaches are more creditable to both the insiders and outsiders of organizations.KEYWORDS: LeaderNBAplayer-turned-coachesstrategic decision-makingJEL CLASSIFICATION: J24M51 AcknowledgmentsWe acknowledge Yusen Kwoh, Weihong Zeng, Yi Chen, Chun Liu, Jie Ma, and Botao Qin for their constructive comments. We thank Qingyang Liu, Bingqian Li, Xuanrui Wu, Qian Liu for their excellent research assistance.Disclosure statementNo potential conflict of interest was reported by the author(s).Notes1 There are some discussions on this topic. For example, using data on British football managers and teams from 1994 to 2007, Bridgewater et al. (Citation2011) study substitution and complementarity between leaders and subordinates and show that managers who themselves played at a higher level raise the productivity of less-skilled teams by more than that of highly skilled teams.2 We do not report 2SLS results as both our full sample and the Goodall et al. (Citation2011) sample with season fixed effects do not provide significant results.3 To support our argument, we provide two types of results. First, we use t-test to check whether player-turned coaches are employed in better teams in terms of the total salary of a team. The results show that layer-turned coaches actually serve for teams with significantly less total salary. Thus, the sampled player-turned coaches are more likely to serve for teams with a lower level of team salary implying a lower level of competiveness. Second, we use the average winning percentage of the former coach as a proxy of team performance prior to the employment of current coach. We run simple OLS and Probit regressions which use the winning percentage as explanatory variables and use a dummy indicating whether a coach was a NBA player as the explained variable. Both results show that the average winning percentage of the former coach cannot explain whether the current coach is a former NBA player or not. Thus, there is no empirical evidence showing that the performance of NBA teams will lead to the choice of coach type in terms of his NBA career. The above results are available upon request.4 Due to the limitation of data, we do not have access to the statistics of close matches before 1997.5 We have checked the history of the NBA during 1997–2003 and found two potential reasons which can explain the insignificant results for the 1997–2003 period in Tables 4. First, the 1998–1999 NBA lockout lasted from July 1 1998 to January 20 1999 and forced the regular season to be a 'semi season' (50 games per team had been shorten). After the lockout teams played at a reduced pace and scoring declined on average by about four points per game per team, and turnovers increased while field goal percentage decreased. Second, the Hall of Fame coach Phil Jackson along with the two superstars Kobe Bryant and Shaquille O'Neal led the Los Angeles Lakers to three consecutive championships between 2000 and 2002. During the Lakers' 'three-peat', the rest NBA teams had less incentive to compete for the championship. This may explain why the impacts of coaches on the team performance became less significant during that period.6 Comparing guard-turned-coaches with non player-turned-coaches yields similar or even larger coefficients as those results in Panel D to Panel F.7 The results are available upon request.Additional informationFundingXiaobo He acknowledges financial support from the National Natural Science Foundation of China (Grant No. 71803128).