We study whether differences in management can explain variation in productivity and how more effective managers can be recruited in absence of high-powered incentives. To investigate this, we first extend the canonical teacher value-added model to account for school principals, and we document substantial variation in their ability to improve students' learning. Teachers' survey responses and quasi-experimental designs based on changes in school management validate our measure of principal effectiveness. Then, we leverage the timing of adoption of a civil service reform and show that despite having relatively rigid wages, public schools were able to attract more effective managers after increasing the competitiveness and transparency of their personnel selection process.