We use administrative data and a randomization design to examine the long-term educational impacts of a large-scale vocational training program for disadvantaged youth in Colombia on trainees and their relatives. Up to eleven years after randomization, trainees were more likely to enroll in...