We study the impact of autonomous schools – publicly funded institutions that operate more independently than government-run schools – on student achievement and school segregation, using data from 15 countries over 16 years. Our triple-differences regressions exploit between-grade variation in the share of students attending autonomous schools within a given country and year. We find that autonomous schools do not raise overall achievement, and our estimates are precise enough to rule out even modest positive effects in math and small positive effects in science. However, these aggregate results mask important heterogeneity, with consistently positive effects for high-socioeconomic-status students and natives, and negative effects for low-socioeconomic-status students and immigrants. In line with these results, we also find that autonomous schools increase segregation by socioeconomic and immigrant status. We conclude that autonomous schools have not generated the anticipated system-wide benefits.