SPECIALISTS, GENERALISTS AND TEAM PERFORMANCE:
EVIDENCE FROM RANDOM ASSIGNMENT IN E-SPORTS
by Kenny Ching
UCL School of Management, University College UK
Team production is ubiquitous in the economy, but creating the right team remains a challenge for many organizations. This study examines how team composition influences performance, focusing on how familiarity amongst teammates shifts the returns to specialist and generalist teams. Applying theories of team production to contexts where team members coordinate interdependent activities extemporaneously, we develop predictions about the main and marginal effects of team composition. We test our assumptions and hypothesis in the context of e-sports, where individuals choose their level of specialization endogenously, but are randomly assigned to five-person teams. After analyzing nearly 8.4 million matches, we find a strong specialist effect: specialists win more often, specialist teams outperform generalist teams, and teams tend to improve as they add specialists. Notably for the theory we advance, the main effects are magnified when teammates have played together more often in the past, suggesting that familiarity is a substitute for generalist skills, and that specialization and familiarity are complements. We discuss implications of these results in the context of changes to the labor market and formation of entrepreneurial teams.