Understanding Subjective Performance Pay

This project looks to understand the effect of incentivizing teachers based on an objective criteria, value-added, versus a subjective criteria, their principal’s evaluation of them. In addition, we vary how much exposure principals have to teachers in the lead up the evaluation to see how this affects principal rating accuracy and resulting teacher effort. We implement the RCT in 300 private schools in Pakistan and measure effects on student test performance and a variety of teacher behaviors including attendance, clock in and out, and 40 in classroom behaviors recorded from 5 hours of classroom footage. While subjective incentives are easy to implement and take advantage of the information principals have about their staff members, they could be susceptible to bias or collusion. Using information about social connections, we look to understand how effective the incentives are for biased versus fair principals.