Why Do So Few Public School Districts Use Merit Pay?

This working paper presents a principal-agent model in the context of public schools to help explain the factors that affect district decisions about merit pay.

The general theory of incentives suggests that public school teachers will face weak performance incentives because of the nature of their work and workplace: teaching is complex and multidimensional work; schools have multiple goals that are vague and poorly observed; multiple (and mobilized) stakeholders are interested in public education. All of this is a recipe for low incentives. Others suggest that incentives in education are low not because of the nature of teaching, but because of the high degree of unionization among public school teachers. Despite the ostensibly slim odds suggested by both views, a few school districts and states nevertheless offer performance incentives for teachers in the form of merit pay.

In this paper we present a principal-agent model that includes the possibility that both the nature of teaching and the political costs of reform associated with unionization play a role in district decisions about merit pay. It predicts that districts will be more likely to offer merit pay when they have more information about teacher performance (in effect altering the nature of teaching) and less likely to do so when their teachers are unionized. The model also suggests that, all else equal, teacher salaries will be higher in districts that offer merit pay. To test the model’s predictions we analyze data from the U.S. Department of Education’s Schools and Staffing Survey for 1999-00 and Census 2000 School District Demographics. Our results suggest that the political costs of reform affect district merit pay decisions and that teachers in merit pay districts earn more than their counterparts in non-merit pay districts. We find little evidence, however, in support of the hypothesis that more information about teacher performance makes merit pay more likely.

A version of this paper was published in Journal of Education Finance, 33(3): 262-289. (2008).

Skip to content