In 2010, the Ministry of Education in Trinidad and Tobago converted 20 low-performing pilot secondary schools from coed to single-sex. I exploit these conversions to identify the causal effect of single-sex schooling holding other school inputs (such as teacher quality and leadership quality) constant. After also accounting for student selection, both boys and girls in single-sex cohorts at pilot schools score 0.14Ï? higher in the academic subjects on national exams. There is no robust effect on non-academic subjects. Additionally, treated students are more likely to earn the secondary-school leaving credential, and the all-boys cohorts have fewer arrests. Survey evidence reveals that these single-sex effects reflect both direct gender peer effects due to interactions between classmates, and also indirect effects generated through changes in teacher behavior. Importantly, these benefits are achieved at zero financial cost.