Abstract:

We investigate how gender composition of teams affects the decision-making process and the quality of decisions, using a novel database on the universe of collegial rulings from a sample of first- and second-instance criminal courts in Italy. Leveraging the quasi-random allocation of both judges and cases to judicial teams, we document three main findings. First, all-women teams have higher conviction rates than both mixed-gender and all-men teams. Second, all-men teams converge faster to a decision than any other team composition. Third, decisions made by all-men teams are more likely to be incorrect, as indicated by a higher likelihood of being appealed and subsequently overturned, whereas no significant difference in decision quality is observed between mixed-gender and all-women teams. To explain these results, we propose a framework of team decision-making where individuals differ in both preferences and diagnostic skills. The observed relationship between conviction and overturn rates, together with the analysis of written judicial opinions, supports the conclusion that variations in effort, rather than preferences, drive our results.