When North Carolina introduced a new principal performance-pay system based on school test-score growth in the 2017–18 school year, it became the country’s largest such system to date. Though performance-pay systems are common for teachers, with 25 states having implemented merit pay for teachers by 2007, they are far less common for principals. Opponents worried that pay cuts might force veteran principals to retire early, while supporters claimed there was “no more important investment” that could be made. In a new working paper, I see if concerns were warranted and determine the impacts of performance pay on principals’ labor market decisions.
How North Carolina’s System Works
In consultation with the North Carolina Principals and Assistant Principals Association and Business for Educational Success and Transformation North Carolina, North Carolina legislators invested $55 million over three years to overhaul the principal pay system. North Carolina moved from a traditional experience- and education-based pay system to a system in which the previous three years of a school’s test-score growth significantly influences the principal’s salary. Instead of receiving an experience raise of between $800 and $1,500 per year, with a $3,000 premium for a doctoral degree, principals now earn more by progressing through three test-score growth categories. Moving from the lowest to the highest growth category led to a pay increase of almost $15,000 in 2017–18.
As a measure of principal quality, school test-score growth is imperfect. It is helpful insofar as it correlates with true principal quality, but it does not capture the full spectrum of a principal’s impact on students at his or her school.
After the system was introduced, on average, new principals were more likely to receive raises relative to their pay under the old system than experienced principals. In contrast, experienced and highly educated principals were more likely to face pay cuts than their less experienced counterparts. The bite of this new principal-pay policy was limited, however. State legislators included a hold-harmless rule in their legislation that prevented salaries from falling, and school districts could bypass performance pay with local salary supplements. A reasonable interpretation of the policy is a weak pay raise in the short run and a significant change in financial incentives and salary trajectory in the long run. Overall, the new policy led to complex changes in principal salaries.
Evaluating the causal effects of North Carolina’s principal-pay policy is challenging. It is impossible to know what principals might have done in a counterfactual world or to run an experiment and assign principals, or even school districts, to performance pay randomly. A reasonable alternative is to find a group that adequately approximates the labor-market decisions principals would have made in a world without performance pay.
I therefore use a sample of experienced teachers as a comparison group to uncover the causal effects of performance pay on principals’ labor-market decisions. Most principals were formerly teachers. Teachers and principals operate in similar labor markets. There were no significant changes in North Carolina teacher-pay policies concurrent with the principal-pay policy rollout. Before the principal performance-pay policy was introduced, there were differences between the two in the levels of average labor-market transition rates, but the changes from year to year were strikingly similar. All of this supports the validity of using experienced teachers as a comparison group.
I compare how principals’ job transitions changed relative to those of experienced teachers using administrative payroll records from the North Carolina Education Research Data Center. Furthermore, I use principal characteristics in this data to impute how a principal’s salary will change due to the new salary system. I employ these predicted salary changes to understand the role of financial incentives in principals’ labor market decisions.
I find that performance pay changes principals’ labor-market decisions in nuanced yet desirable ways. Principals who are more likely to earn more under the new pay system are more likely to continue working as a principal—but at a different school. This increase in school transfers is driven by more effective and experienced principals moving to underperforming and Title I schools, where there is room to increase performance and reap the associated financial benefits. On the other hand, principals who would face stagnant wages are now slightly more likely to move into different jobs in traditional public education.
A concern about North Carolina’s policy was that it would cause veteran principals to retire early, which would be detrimental to students if the experienced principals were more effective. While I find highly experienced principals are more likely to leave their jobs because of performance pay, the principals who leave also tend to be less effective. These results suggest that performance pay pushes out principals who had previously failed to meet test-score growth expectations. At the same time, performance pay increases retention of principals with a history of exceeding expectations.
I do not assess the impacts of performance pay on principal effectiveness or on students. Understanding how performance pay affects these factors is essential for a thorough evaluation of North Carolina’s new policy. Although my scope is limited, the evidence I provide suggests that policymakers can use principal performance pay to change principals’ labor market decisions.
John Westall is a doctoral candidate in economics at North Carolina State University.