We present new evidence on the effectiveness of large-scale reforms of public teaching careers towards improving student outcomes in Latin America. Since 2002, Colombia’s public teacher candidates are screened through an entry competition featuring a standardized examination, and their subsequent career progression is subject to performance evaluations; the fact that educators hired before the reform remained exempt from the new rules provides a unique research setting for policy evaluation. We are able to exploit unusually detailed data on the universe of secondary schools and teachers in Colombia; our empirical strategy exploits variation in student performance and teacher types across different subjects, within any given school and year. We pin down a sizeable performance premium of quality-screened teachers with respect to their traditional colleague: the former improve student achievement by about 7% of a standard deviation within school and year. We are able to exclude various types of selection biases, cohort effects and other confounders. We show that the Colombian entry examination is an effective candidate screening device and that teachers’ achievement on the exam is predictive of their performance on the job. We provide suggestive evidence on how the reform affected other policy-relevant aspects such as teacher candidate quality and career attrition.