Identifying an individual worker's contribution to firm production can be difficult in a team setting where spillovers in labor productivity exist among team members. This paper studies a model of labor productivity where workers have heterogeneous abilities, and can differently affect the productivity of their fellow teammates. Applying the model to the setting of the National Basketball Association (NBA), I can identify the marginal value that a basketball player brings to a particular team lineup, both through his own individual contributions and his complementary contribution to teammates' productivity. Estimates from the model imply that teammates have a significant impact on individual player productivity, and that taking into account spillovers across teammates is important to assessing both overall team productivity and an individual player's contribution to team productivity. I then evaluate whether player complementarities are valued in the NBA labor market in terms of higher salaries, and find that they are undervalued, and that players are instead paid mainly for their individual offensive production. This creates an asymmetry between player incentives and the team objective. To assess the size of this inefficiency, the top trading cycle algorithm of Shapley and Scarf (1974) is used to identify a Pareto optimal matching between players and teams, that accounts for the complementarities between heterogeneous players' skill sets. (JEL J30, L25, L83, M51)