Long-Term Effects of Teacher Performance Pay: Experimental Evidence from India

While the idea of teacher performance-pay is increasingly making its way into policy, the evidence on the effectiveness of such programs is both limited and mixed. The central questions in the literature on teacher performance pay to date have been whether teacher performance pay based on test scores can improve student achievement, and whether there are negative consequences of teacher incentives based on student test scores? The literature on both of these questions highlight the importance of not just evaluating teacher incentive programs that are designed by administrators, but of using economic theory to design systems of teacher performance pay that are likely to induce higher effort from teachers towards improving human capital and less likely to be susceptible to gaming. Also, while there is a growing body of high-quality empirical studies on the impact of teacher performance pay on education quality, most of these evaluations stop after two or three years, and so there is no good evidence on longer-term impacts (both positive and negative) of teacher performance pay on students who have completed most of their education under such a system. In this paper, the author contributes towards filling this gap with results from a five-year long randomized evaluation of group and individual teacher performance pay programs implemented across a large representative sample of government-run rural primary schools in the Indian state of Andhra Pradesh (AP). The main questions addressed in this paper are: 1) What is the impact of teacher performance pay (implemented for five years) on student test scores at various points of program exposure? 2) Are there any negative consequences of the teacher performance pay program? 3) What is the relative effect of group and individual teacher incentive programs? There are three main results in this paper. First, the individual teacher performance pay program had a large and significant impact on student learning outcomes over all durations of student exposure to the program. Students who had completed their entire five years of primary school education under the program scored 0.54 and 0.35 standard deviations (SD) higher than those in control schools in math and language tests respectively. These are large effects corresponding to approximately 20 and 14 percentile point improvements at the median of a normal distribution, and are larger than the effects found in most other education interventions in developing countries (see Dhaliwal et al. 2011). Second, the results suggest that these test score gains represent genuine additions to human capital as opposed to reflecting only "teaching to the test". Students in individual teacher incentive schools score significantly better on both non-repeat as well as repeat questions; on both multiple-choice and free-response questions; and on questions designed to test conceptual understanding as well as questions that could be answered through rote learning. Most importantly, these students also perform significantly better on subjects for which there were "no incentives"--scoring 0.52 SD and 0.30 SD higher than students in control schools on tests in science and social studies (though the bonuses were paid only for gains in math and language). There was also no differential attrition of students across treatment and control groups and no evidence to suggest any adverse consequences of the programs. Third, the authors find that individual teacher incentives significantly outperform group teacher incentives over the longer time horizon though they were equally effective in the first year of the experiment. Students in group incentive schools score better than those in control schools over most durations of exposure, but these are not always significant and students who complete five years of primary school under the program do not score significantly higher than those in control schools. However, the variance of student outcomes is lower in the group incentive schools than in the individual incentive schools. The authors measure changes in teacher behavior and the results suggest that the main mechanism for the improved outcomes in incentive schools is not reduced teacher absence, but increased teaching activity conditional on presence. Finally, the authors also measure household responses to the program--for the cohort that was exposed to five years of the program, at the end of five years--and find that there is no significant difference across treatment and control groups in either household spending on education or on time spent studying at home, suggesting that the estimated effects are unlikely to be confounded by differential household responses across treatment and control groups over time.