This was a major research paper conducted by my group member (Shadi Chamseddine) and I for our ECON 5029 F Methods of Economic Research course.
The data to create this analysis is found in the 'Data' folder.
Our paper explores the optimal roster composition in terms of salary dispersion and within-team Marginal Revenue Product (MRP) on team performance in the NBA within the leagues roster composition constraints. Competing economic theories pit hierarchical pay structures against more homogeneous pay structures to determine which elicits the greatest incentive to boost worker productivity. Within the branch of the salary dispersion analysis, we use bootstrap salary variance and Theil’s index as our measures for dispersion. Both measures show there is no causal relationship between salary dispersion and team performance in the NBA, and therefore support neither tournament theory nor pay compression theory. Turning our attention to the MRP dispersion analysis, we find a one percent increase in MRP dispersion is associated with on average an 11.0 percent increase in team performance. The result is statistically significant at the one percent level. Our results suggest MRP dispersion is a better indicator of team performance than salary dispersion. Given salary cap constraints and the high impact of star players on team performance, it is optimal to fill your team with a few top productive players and the remainder with minimum salary players.