In January 2016, Tanzania implemented a fee-free secondary school reform. Using variation in district and cohort exposure to the reform, we employ a difference-indifferences strategy to estimate the short-term impacts of the reform. Despite a relatively small drop in user costs, the reform substantially increased enrolment into secondary education. While these enrolment effects were predominantly driven by an increase in public school enrolment, there was also a delayed positive effect on private school enrolment. Districts most exposed to the reform experienced a significant drop in exam scores relative to less-exposed districts, which cannot be explained by academic abilities of new students. These findings are in line with a theoretical school choice model, where fee elimination loosens enrolment constraints, and increased enrolment harms the quality of public education. (JEL I21, I24, I28) * We wish to thank Youdi Schipper, Sam Jones, Thomas Markussen, Patricia Justino, and Lant Pritchett for providing excellent comments on the paper. Further thanks go to the National Examinations Council of Tanzania for providing public access to a large and useful database on student exam records.