Can Performance-Based Pay Transform Educator Quality in Disadvantaged Schools?

Morgan et al. (2025) examine whether large, effectiveness-based financial incentives combined with improved hiring authority can attract and retain high-quality educators in low-performing schools. They analyze administrative data from the Dallas Independent School District, including student test scores, teacher evaluations, and staffing records. Using a difference-in-differences design, they find that the reform substantially increased teacher quality and student achievement. Math scores rose by about 0.46–0.54 standard deviations, while reading scores increased by roughly 0.22–0.34 standard deviations. They conclude that targeted pay and selective hiring can meaningfully improve outcomes in disadvantaged schools.

Why This Article Was Selected for The Policy Scientist

The distribution of teacher quality remains one of the most consequential determinants of inequality in educational outcomes, with implications for labor markets, social mobility, and long-run economic growth. This question is especially timely given renewed attention to post-pandemic learning loss and the limits of traditional school turnaround strategies. Morgan, Nguyen, Hanushek, Ost, and Rivkin build on a substantial body of prior work, particularly on teacher effectiveness and labor market sorting, to test a comprehensive staffing intervention.

The study uses rich administrative data linking students, teachers, and evaluations, which strengthens internal validity. The difference-in-differences design with careful attention to pre-trends represents a credible causal inference approach. The magnitude of the estimated effects is unusually large, increasing confidence in substantive significance.

Full Citation and Link to Article

Morgan, A., Nguyen, M., Hanushek, E. A., Ost, B., & Rivkin, S. G. (forthcoming). Attracting and retaining highly effective educators in hard-to-staff schools. American Economic Journal: Economic Policy. https://www.aeaweb.org/articles?id=10.1257/pol.20240356

Central Research Question

This article investigates whether a comprehensive staffing reform—combining large, effectiveness-based financial incentives with enhanced information and authority for school leaders—can attract and retain highly effective educators in chronically low-performing schools, and in turn, substantially improve student academic outcomes. The analysis centers on the Accelerating Campus Excellence (ACE) program implemented by the Dallas Independent School District (DISD), which sought to overcome longstanding labor market frictions that systematically sort higher-quality teachers into more advantaged schools. The core question is not simply whether higher pay matters, but whether pay tied explicitly to measured effectiveness, combined with strong administrative selection mechanisms, can transform the composition of a school’s workforce and generate meaningful gains in student achievement.

Previous Literature

The study builds on a well-established literature demonstrating the central importance of teacher quality for student outcomes, including foundational contributions by Rivkin, Hanushek, and Kain (2005) and Chetty, Friedman, and Rockoff (2014). Prior research has also documented that teachers systematically prefer more advantaged working environments, leading to persistent inequities in the distribution of instructional quality. Efforts to correct this imbalance through compensating differentials have produced mixed results. For example, Clotfelter, Ladd, and Vigdor (2011) find that salary premiums improve retention but do not necessarily increase teacher quality, in part because districts lack the information or incentives to selectively hire more effective educators.

The literature on school turnarounds similarly offers inconclusive evidence. Some interventions, such as those studied by Fryer (2014) and Schueler, Goodman, and Deming (2017), report modest improvements, often through multifaceted reforms that include extended learning time, tutoring, and leadership changes. However, these approaches typically do not center on large-scale, effectiveness-based compensation or wholesale restructuring of school personnel. The ACE program differs in its emphasis on transforming the educator workforce through both selection and incentives. In doing so, the study extends prior work by testing a policy design that directly integrates insights from labor economics—particularly the role of compensating differentials and information asymmetries—into a large-scale educational intervention.

Data

The analysis relies on detailed administrative data from DISD, spanning the 2011–2012 through 2018–2019 academic years. These data include student-level test scores from the State of Texas Assessments of Academic Readiness (STAAR), which are standardized by grade, subject, and year within the district. The dataset also contains extensive information on student demographics, including race, gender, and participation in programs such as free or reduced-price lunch, special education, and English language learning.

Critically, the authors link these student records to educator-level data, including teacher assignments, experience, and a rich set of evaluation metrics derived from the district’s Teacher Excellence Initiative (TEI). These evaluations incorporate multiple components, including value-added measures of student achievement, classroom observations, and student surveys. The dataset further includes detailed information on compensation, including the size of stipends associated with ACE participation, which vary by educator role and prior effectiveness rating.

The ability to link students, teachers, and schools over time enables the authors to construct a panel dataset that captures both compositional changes in the educator workforce and corresponding changes in student outcomes. This level of granularity is a notable strength, as it allows for a more precise assessment of the mechanisms through which the policy operates.

Methods

The empirical strategy is grounded in a difference-in-differences framework, supplemented by an event study design that traces the evolution of outcomes before and after the implementation of the ACE program. The authors compare changes in student achievement at ACE schools to those at a carefully constructed control group of similarly low-performing schools within the same district. To address concerns about selection bias, the control group is defined based on pre-treatment test scores, ensuring that treated and comparison schools exhibit similar baseline performance and pre-trends.

The event study specification allows the authors to assess both the timing and persistence of treatment effects, while also providing a diagnostic check for parallel trends prior to implementation. Importantly, the analysis accounts for the staggered rollout of the program across two cohorts of schools, adapting recent methodological advances in difference-in-differences estimation to ensure appropriate comparisons.

Given the small number of treated schools, the authors employ permutation-based inference to obtain p-values, rather than relying on conventional clustered standard errors that may perform poorly in such settings. This approach enhances the credibility of the statistical inference. The study also examines treatment dosage effects by distinguishing between one, two, and three years of exposure to the ACE program, and evaluates longer-term impacts by analyzing student outcomes in subsequent grades.

While the design constitutes a strong quasi-experimental approach, it does not rise to the level of a randomized controlled trial. The reliance on observational variation, even with careful controls and robustness checks, leaves open the possibility of residual confounding, though the consistency and magnitude of the results mitigate this concern.

Findings/Size Effects

The results indicate that the ACE program produced large and immediate improvements in student achievement. In the first year of implementation, math test scores increased by approximately 0.46 standard deviations, with effects rising to around 0.54 standard deviations in the second year. Reading scores also improved, though to a lesser extent, with gains of approximately 0.22 to 0.34 standard deviations. These effect sizes are substantial by the standards of the education policy literature, exceeding those typically observed in both compensatory pay experiments and broader school turnaround initiatives.

The analysis of teacher composition reveals that these gains were accompanied by a dramatic shift in the distribution of educator effectiveness. The share of highly rated teachers increased sharply, while the proportion of novice or lower-rated educators declined. Estimates of teacher value-added further indicate that incoming teachers at ACE schools were significantly more effective than those they replaced, particularly in mathematics, where differences approached 0.26 standard deviations.

The study also finds evidence of persistence, particularly for students exposed to the program for multiple years. Students with two or three years of exposure exhibit gains in later grades, with estimated effects on sixth-grade math scores approaching 0.3 to 0.4 standard deviations. However, these longer-term estimates are less precise and should be interpreted with caution.

A notable feature of the findings is the apparent sensitivity of outcomes to the continuation of financial incentives. When stipends were removed for a subset of schools, many high-performing teachers exited, and achievement gains partially reversed. This pattern reinforces the interpretation that the compensating differentials played a central role in both attracting and retaining effective educators.

Conclusion

This study provides compelling evidence that a policy design integrating large, effectiveness-based financial incentives with rigorous evaluation and selective hiring authority can substantially alter the distribution of teacher quality and generate meaningful improvements in student achievement. The magnitude of the estimated effects distinguishes this intervention from much of the prior literature, suggesting that earlier null or modest findings may reflect incomplete policy designs rather than fundamental limits of compensatory pay.

At the same time, the results highlight the importance of institutional context. The success of the ACE program appears to depend on the presence of a robust evaluation system, administrative capacity to implement large-scale personnel changes, and sustained financial commitments. These factors may limit the direct transferability of the findings to other jurisdictions, particularly those lacking similar infrastructure.

Methodologically, the study employs a well-executed quasi-experimental design that aligns with best practices in causal inference, including careful attention to pre-trends, control group construction, and inference with small samples. Nonetheless, the absence of random assignment suggests that future research could further strengthen the evidence base through experimental designs or natural experiments that more cleanly isolate causal effects.