The paper estimates the effect of delayed school enrollment on student outcomes, using administrative data on Chilean students that include exact birth dates. Regression-discontinuity estimates, based on enrollment cutoffs, show that a one-year delay decreases the probability of repeating first grade by two percentage points, and increases fourth and eighth grade test scores by more than 0.3 standard deviations, with larger effects for boys. The paper concludes with implications for enrollment age policy. (Contains 4 figures, 6 tables and 27 footnotes.)