Published Papers:
The Marginal Returns to Distance Education: Evidence from Mexico’s Telesecundarias, with Gabrielle Vasey. 2024. American Economic Journal: Applied Economics, 16(1): 253-85.
Abstract: This paper analyzes a large-scale and long-running distance education program in Mexico. We estimate marginal treatment effects (MTEs) for learning in math and Spanish in telesecundarias relative to traditional Mexican secondary schools using an empirical framework that allows for unobserved sorting on gains. The estimated MTEs reveal that school choice is not random and that the average student experiences significant improvements in both math and Spanish after just one year of attendance in telesecundarias. We find that the existing policy reduces educational inequality, and our policy-relevant treatment effects show that expanding telesecundarias would yield significant improvements in academic performance.
Working Papers:
The Heterogeneous Effects of Changing SAT Requirements in Admissions: An Equilibrium Evaluation, R & R at Journal of Political Economy
Abstract: Many universities are reducing emphasis on standardized exam scores in admissions out of concern that the exams limit college access for students from disadvantaged backgrounds. This paper analyzes how such a policy change would affect enrollment patterns and graduation rates at four-year colleges in the United States. To do so, I build an equilibrium model in which colleges rebalance their admissions criteria towards other measurements of students' human capital in the absence of standardized exam scores. The model allows high school students' application decisions and human capital investments to respond endogenously to the admissions policy, while colleges adjust admissions thresholds to maximize their objectives. I estimate the model using data from the Education Longitudinal Study of 2002. I find that banning the SAT causes a small increase in college attendance for low-income students but has a negligible effect on the enrollment of under-represented minority (URM) students, despite estimating that many universities have substantial preferences for diversity. The reason for this result is that endogenous human capital investment and equilibrium responses by capacity-constrained colleges completely offset the diversifying effects of relying more on grades and allowing non SAT-takers to apply to college. Elite colleges are worse off after banning the SAT, as they enroll students with lower skills and see graduation rates drop by 3 pp, while completion rates rise at less selective schools. A separate policy that requires all students to take the exam raises college completion for URMs by 1.8 pp by helping schools to identify stronger students.
Welfare, Workfare, and Labor Supply: A Unified Evaluation, with Francesco Agostinelli and Giuseppe Sorrenti, R & R at Journal of Econometrics
Abstract: We analyze the impact of U.S. Earned Income Tax Credit (EITC) and welfare reforms on single mothers' labor supply. Our labor supply model, which incorporates detailed program structures and quasi-experimental policy changes, shows that the EITC contributed 50% to the employment growth caused by reforms to both programs between 1992 and 1996. The model also suggests that the 1993 EITC reform's effects would have varied significantly under a different tax system. Finally, we show that labor supply trends during the 1990s depended heavily on tax rates, wage offers, and welfare benefits, and that populations that differ along these dimensions are unlikely to satisfy the parallel counterfactual trends assumption in commonly used difference-in-differences designs.
If at First You Don't Succeed: A Dynamic Evaluation of Grade Retention, with Hugo Reis and Petra Todd
Abstract: In many European countries, it is common for secondary school students to be retained. This paper evaluates the impact of grade retention on learning, dropout, and educational attainment in Portugal, which has some of the highest retention rates in the world. An extended Roy model is developed and implemented to model retention's cumulative effects on test scores, dropout, and educational attainment for different subgroups followed over multiple grades. Results show substantial heterogeneity in test score impacts. In math, 82\% of students retained experience test score gains, with an average impact of $0.46$ sd among those who graduate; in Portuguese, only 42\% of students experience gains, with an average impact of $-0.17$ sd. Also, retention increases dropout by $7$ pp for the average retained student. We validate our model by simulating retention effects on dropout for 12th grade students at the margin of being retained for whom we can compare model-based estimates to RDD estimates. Lastly, the estimated model is used to solve for the optimal retention policy that maximizes average lifetime earnings, considering that retention yields cognitive gains for some students who stay in school while also decreasing educational attainment for students who drop out. Portuguese retention rates are found to be too high for plausible estimates of cognitive skill returns.
Heterogeneity and Endogenous Compliance: Implications for Scaling Class Size Interventions, with Karun Adusumilli and Francesco Agostinelli
Abstract: This paper examines the scalability of the results from the Tennessee Student-Teacher Achievement Ratio (STAR) Project, a prominent educational experiment. We explore how the misalignment between the experimental design and the econometric model affects researchers' ability to learn about the intervention's scalability. We document heterogeneity in compliance with class-size reduction that is more extensive than previously acknowledged and discuss its consequences for the evaluation of the experiment. Guided by this finding, we implement a new econometric framework incorporating heterogeneous treatment effects and endogenous class size determination. We find that the effect of class size on test scores differs considerably across schools, with only a small fraction of schools having significant benefits from reduced class sizes. We discuss the challenges this poses for the intervention's scalability and conclude by analyzing targeted class-size interventions.