This paper explores how a conditional cash transfer program can influence students’ schooling decisions when program payments stop in the middle of their school career. To that end, I examine Mexico’s PROGRESA, which only covered students until the end of secondary school (at age 15) in its early years. The experimental setup of the program permits to study its impact on the probability to transition to high school. To account for compositional differences between treatment and control group as a result of the program, I employ a newly developed semiparametric technique that uses a combination of machine learning methods in conjunction with doubly-robust estimation. I find that exposure to PROGRESA during secondary school reduced the probability to transition to high school by 7.5 to 12.5 percentage points. Possible explanations for this effect include parents’ loss aversion, motivation crowding, price anchoring, and classroom peer effects.