Skip to main content

Teacher Performance Pay

Researchers: Karthik Muralidharan (UC San Diego) and Venkatesh Sundararaman (World Bank)
Location: India

A fundamental question in education policy around the world is that of the relative effectiveness of input- and incentive-based policies in improving the quality of schools. While the traditional approach to improving schools has focused on providing them with more resources, there has been growing interest in directly measuring and rewarding schools and teachers on the basis of student learning outcomes. However, the empirical evidence on the effectiveness of incentive-based policies is quite limited, with identification of the causal impact of teacher incentives being the main challenge.

This study contributes towards filling that gap with evidence from a large-scale randomized evaluation of a teacher performance pay program implemented across a large representative sample of government-run rural primary schools in the Indian state of Andhra Pradesh. Researchers studied two types of teacher performance pay: group bonuses based on school performance and individual bonuses based on teacher performance, with the average bonus calibrated to be around 3 percent of a typical teacher’s annual salary. The study was conducted by randomly allocating the incentive programs across a representative sample of 300 government-run schools in rural Andhra Pradesh, with 100 schools each in the group and individual incentive treatment groups, and 100 schools serving as the comparison group.

The study found that the teacher performance pay program was effective in improving student learning. After two years of the program, students in incentive schools performed significantly better than those in control schools in math and language tests. In a parallel initiative, additional school inputs (an extra contract teacher or a cash grant for school materials) were found to also be effective in raising test scores, but the teacher incentive programs were three times as cost effective, and incentive-based schools performed significantly better than schools receiving input programs.

There are several unresolved issues and challenges that need to be addressed before scaling up teacher performance pay programs. In particular, it may not be feasible to scale up such programs without first investing in the ability to independently measure student learning and track student learning growth trajectories over time. Nevertheless, the study provides a "proof of concept" that a well-implemented teacher performance pay program may be able to significantly improve learning outcomes in developing countries.

Results

The study found that the teacher performance pay program was effective in improving student learning. After two years of the program, students in incentive schools performed significantly better than those in control schools in math and language tests. No evidence of any adverse consequences of the program was found. Students in incentive schools do significantly better not only in math and language, for which there were incentives, but also in science and social studies, for which there were no incentives. This suggests positive spillover effects. There was no difference in student attrition between incentive and control schools and no evidence of any adverse gaming of the incentive program by teachers. School-level group incentives and teacher-level individual incentives perform equally well in the first year, but the individual incentive schools outperformed the group incentive schools after two years of the program.

In a parallel initiative, additional school inputs (an extra contract teacher or a cash grant for school materials) were found to also be effective in raising test scores, but the teacher incentive programs were three times as cost effective, and incentive-based schools performed significantly better than schools receiving input programs.

Related Publications

Muralidharan, Karthik and Venkatesh Sundararaman. 2011. "Teacher Performance Pay: Experimental Evidence from India.” Journal of Political Economy 119 (1): 39–77.