Performance pay can bring stronger teachers into the classroom

Kindergarden student Misko (4) plays with puzzles next to a teacher and volunteer in an empty school, closed during a one-day strike by local teachers seeking higher salaries and better work conditions.

Former Education Secretary Arne Duncan once remarked: “Money is never the reason why people enter teaching, but it is the reason why some people do not enter teaching.” Education reformers should begin to address Duncan’s point by investigating whether some folks who currently avoid careers in public education do so, in part, because they do not want to be paid solely based on their career longevity and degree credentials. Specifically, policymakers should ask whether many of the prospective employees that public education covets most (i.e., academically accomplished, higher aptitude) currently “vote with their feet” by choosing occupations and organizations that link pay to performance.

Despite a large research literature on teacher pay-for-performance (PFP), including several randomized experiments, most of this work has failed to address the core question of whether PFP could lead to positive “sorting” effects on the composition of the teaching workforce through attraction, selection, and attrition processes.

There are strong theoretical reasons to anticipate that higher-ability job seekers may consciously search for employment opportunities that offer PFP. Labor economists have found significant causal evidence for positive sorting effects whereby high-ability individuals respond to PFP opportunities. On the whole, these studies actually suggest that the sorting effects of performance pay trump the basic incentive effects. In fact, these researchers found that performance pay exhibited both an effort and a sorting/selection effect on participants with the sorting effect “[making] up a little more than half of the total increase in effort.”

Although education researchers have not yet fully engaged with this research literature, a handful of studies combine to suggest that more academically accomplished teachers and even prospective teachers express attitudinal preferences for PFP over the current fixed-pay system. Some of this survey research reveals that recent college graduates and early career teachers in the U.S. with the highest cognitive abilities—as measured by performance on SAT/ACT exams—are more likely to support PFP than their less academically accomplished peers. Most notably, Dan Goldhaber and colleagues find evidence that the highest-performing teachers—as measured by value-added test score gains—elected to participate in a large-scale PFP program (Denver’s Pro Comp) at a higher rate than the district’s lower-performing teachers.

Although these existing studies offer important insights into the preferences of prospective and current teachers about PFP, policymakers still need to know whether adopting PFP is likely to result in districts being able to recruit more accomplished candidates. To shed light on that important question, Michael Jones and I recently conducted a study that examined the net consequences of districts adopting PFP programs on the quality of their teacher recruitment efforts.

Our study—which appears in a recent issue of Public Administration Review—relies on data from two waves of the Schools and Staffing Survey. We use this data to determine whether school districts that adopted PFP were subsequently able to attract and hire early-career teachers who graduated from more-selective colleges and universities, a widely used proxy for teacher academic aptitude that we elaborate more on in the paper.

We find that, on average, school districts that implemented PFP to reward excellence in teaching secured new teacher hires who graduated from colleges and universities with average incoming SAT scores that were about 30 points higher than the new teacher cohorts hired by districts that did not adopt PFP. Importantly, these findings could not simply be explained by pre-trend differences in school districts’ teacher recruitment patterns prior to the adoption of PFP. Moreover, these results persist even after accounting for districts using other pay recruitment incentives including: market-pay, hard-to-staff schools pay, and more generous baseline salaries. In other words, the adoption of PFP appears to play its own role in enabling school districts to recruit early-career teachers with higher levels of academic aptitude.

It is also important to point out that most existing PFP programs are quite small relative to the average teacher’s overall salary. Therefore, our evidence for a sorting effect may actually underestimate the potential policy impact that more ambitious PFP programs could have on the composition of the education workforce. For example, Auguste and colleagues estimate that “offering a 20% performance bonus to the top performing 10% of teachers would induce roughly an 11% increase in the number of top-third students becoming teachers.” In short, it is quite possible that our empirical findings are estimating the mere floor that properly designed and well-funded PFP programs could have on a district’s ability to hire smarter teachers.

Our findings should encourage policymakers and researchers to pay greater attention to investigating whether and how various pay-reform initiatives shape recruitment and hiring outcomes in public education. For example, one might ask whether PFP influences employee sorting along important personality dimensions that have long been of interest to scholars and practitioners in the closely related field of human resources management, such as employee risk-aversion. In a series of laboratory experiments, Bowen and colleagues find evidence that teachers are generally more risk-averse than comparable non-teacher professionals and that the most risk-averse teachers prefer working under fixed (e.g. rigid salary schedules), versus variable pay (e.g. pay-for-performance) systems. In cases where school districts may wish to attract and hire employees who are more willing to take risks, understanding whether PFP—and variable pay more generally—can improve attraction and/or reduce attrition among less risk-averse employees would seem to be a particularly fruitful line of inquiry.

Finally, future research could improve significantly upon this current study by employing measures of employee quality that go beyond academic aptitude/college selectivity, such as an employee’s actual on-the-job effectiveness. In a particularly novel study, Matt Chingos and Marty West found evidence that highly effective teachers (as measured by value-added test scores) who leave the teaching profession typically enter directly into more financially lucrative post-teaching careers. To that end, researchers might build on existing work that examines whether school districts that enact PFP programs can use PFP to improve retention efforts by minimizing attrition among the district’s most effective employees.

Experimentation with teacher PFP remains controversial. Some suggest that PFP has failed to gain traction because of a lack of political will. Opponents routinely tout an array of well-designed studies, including gold-standard random assignment studies, that fail to demonstrate that PFP incentive schemes result in higher levels of teacher performance and, by extension, student performance outcomes. Unfortunately—though not always—these debates about the promise or peril of implementing teacher PFP fail to consider whether performance incentives might have powerful selection or “sorting” effects on the composition of the future education workforce. To that end, our findings offer some of the first evidence that, across a national sample of school districts, the adoption of PFP was accompanied by a tangible increase in the ability of school leaders to hire more academically accomplished teachers.