We use an admissions lottery to estimate the effect of a non-means tested preschool program on students' long-run earnings, employment, family income, household formation, and geographic mobility. We observe long-run outcomes by linking both admitted and non-admitted individuals to confidential...