Rewarding teachers on the basis of student performance is a growing trend in educational policy. This paper estimates the effects of a policy that ties payments with teachers’ pedagogical skills instead. We study a large-scale reform in Chile that introduced financial incentives tied to a teacher evaluation system. Using a unique administrative data set of over 500,000 studentteacher-year matches, we estimate the effect of the policy on student performance exploiting the program’s gradual roll-out through a differences-in-differences analysis. We document precise, null effects of the policy on student math and language standardized test scores. Estimating a structural model of teacher skills and student performance, we show that by making incentives more homogeneous across the distribution of teacher characteristics policymakers can improve the policy’s effects on student performance and overall welfare.

ACCESS PAPER