"A comment on Dincecco et al. (2022): Pre-Colonial Warfare and Long-Run Development in India" (
published at the Canadian Journal of Economics)
with
Rachel Forshaw,
Tim Olkers and
Manali Sovani [paper]
[abstract]
We test the reproducibility and replicability of Dincecco et al. (2022), which reports a positive relationship between pre-colonial interstate warfare and long-run development patterns across India. Overall, we confirm that all of the study's estimates are computationally reproducible by using both the provided replication package in Stata and code written by the present authors in R. We test for and find no evidence of data manipulation in the final datasets. Concerning direct replicability, we consider different ways of measuring distance to conflicts and also alternative proxies for both the dependent variable and variables which capture channels by which the main effects operate. We are able to replicate the magnitude and significance of the estimated coefficient on conflict exposure in most of the tests, noting that while most estimates are substantively in line with the original study, some alternative measures of distance to conflict imply different magnitudes for estimates, and proxy estimates are sensitive to both the time period and type of conflict considered.
"Reproducibility and Robustness of Economics and Political Science Research"
(
published in Nature, 2026)
with
Abel Brodeur,
Derek Mikola,
Nikolai Cook and others
[paper]
[abstract]
We assess the reproducibility and robustness of 110 articles published in leading economics and political science journals with mandatory data and code sharing policies. More than 85% of published claims are computationally reproducible. Across robustness checks, 72% of statistically significant estimates remain significant and in the same direction, while the median reproduced effect size is 99% of the originally published estimate. Six independent research teams also examine prespecified hypotheses about the determinants of robustness. More experienced research teams find lower levels of robustness, while robustness is not systematically related to author characteristics or data availability.
"The Enjoyment Paradox: College-Educated Mothers Invest More in Their Children's Learning and Enjoy It Less"
(
Revise and Resubmit, Review of Economics of the Household)
with Ariel Kalil and Haoxuan Liu
[paper]
[abstract]
College-educated mothers invest more time in children's learning than non-college-educated mothers, yet they report substantially lower enjoyment of those same activities. Using purpose-built survey data from 1,552 mothers of kindergarteners in Chicago Public Schools—the first dataset to measure parental time investment, enjoyment, stated motivations, and well-being in a single sample—we document this enjoyment paradox and trace its internal structure. The raw enjoyment gap is 28 percentage points on a peak-enjoyment measure (10/10) and 0.87 points (0.62 standard deviations) on a continuous scale; it survives rich controls for demographics and psychological barriers. The gap is concentrated among non-employed, above-median-income college-educated mothers, who report the lowest enjoyment and invest less than their employed college-educated counterparts at the same income level, even though non-employment is associated with more investment among less-educated mothers. Below the income median, the pattern reverses. We interpret these findings through a framework suggesting that the hedonic cost of intensive investment falls most heavily on the group facing maximum norm salience without a competing professional identity.
"Beyond a Free Lunch: Evaluating the Trade-Offs of an Out-of-School Nutrition Program on Adolescent Girls' Human Capital" (
Revise and Resubmit, Journal of Economic Behavior & Organization)
with
Tanmay Devi,
Ajinkya Keskar and
Dibya Mishra [paper]
[abstract]
We study how the design of in-kind transfers can impact adolescent girls’ schooling. India’s SABLA scheme made nutrition support available outside school for girls ages 11–14. We use variation across rollout districts, birth cohorts, and gender in a difference-in-differences and triple-difference framework. The 2011 rollout increased girls’ dropout by 11% and widened the gender gap in dropout by 9%. The dropout increase is larger among girls whose mothers did not attend school and in districts where girls’ schooling was weaker before the program. A later redesign that encouraged school re-entry does not show an adverse dropout effect. Evidence on the program’s health effects suggests a 2.8% lower anemia rate among exposed cohorts measured 8–10 years after rollout. The results show a transfer-design tradeoff: support delivered outside school can protect vulnerable girls while weakening school attachment for girls near the dropout margin.
"Pennies for Pads: Subsidized Menstrual Health Products and Adoption in India" with
Dibya Mishra [paper]
[abstract]
We estimate the effect of India’s Suvidha program, which sells sanitary napkins at ₹1 per pad through the Jan Aushadhi Kendra pharmacies. Exploiting the staggered rollout of Kendras across subdistricts, we apply a doubly robust difference-in-differences estimator to nearly 500,000 women aged 15–24 from two National Family Health Survey rounds. Subsidized availability raises sanitary napkin use by 9.6 percentage points, a 16% gain over a 61% baseline. A falsification test using Kendra openings before subsidized pads existed is consistent with the subsidized product, rather than Kendra presence per se, as the operative channel. Effects are larger where a Kendra opens near hospitals or supermarkets, and at woman-owned outlets.
"Ignorance is Bliss? Class Rank Disclosure and Course Choice in High School" [paper]
[abstract]
High schools are moving away from class rank. Does withholding rank information from students change the coursework they choose? We find that withholding rank changes who takes advanced coursework. Participation rises in atypically strong classes, where the disclosed rank understates a student's standing in the broader population, and falls in atypically weak classes, where only a disclosed rank shows students how high they stand among their peers. We obtain these results from a dynamic model of course selection in which students learn about their abilities from course grades and about their relative standing from disclosed rank, holding fixed differences across schools and entering cohorts and the observable makeup of a student's peers. We estimate the model on administrative records for roughly 70,000 high school students and validate it on held-out cohorts, one of which faced a different disclosure regime.
"Unraveling Teacher Steering and Student Sorting in Between-School Segregation's Impact on Within-School Segregation" [paper]
[abstract]
This study develops a framework to analyze the impact of between-school segregation on within-school segregation. We model students' course choices as a game of incomplete information, incorporating inter-student and student-teacher interactions. Using idiosyncratic variation in student composition across cohorts in Texas high schools, we find that Black and low SES students receive less encouragement from teachers as their representation increases, suggesting weaker student-teacher interactions. Conversely, Hispanic and female students receive more encouragement with increased representation. White and academically high-performing students are less likely to enroll in college-prep courses as their representation grows, suggesting higher inter-student competition. Additionally, higher expected college-prep enrollment in a cohort discourages students from enrolling, though stronger within-group coordination incentives exist among White, Hispanic, and female students. Policy simulations using an entropy index support the segregation paradox, indicating that racial disparities in course-taking widen in integrated schools. These findings suggest merit in exploring policies like all-girls schools, race-separate courses, and SES desegregation to promote equity in course-taking.
"If You Build It, They Will Come: A Demand Analysis of Household-Level Childcare Choices" with
Flávio Cunha and
Jung Youn Lee [paper]
[abstract]
This study investigates the phenomenon of childcare deserts by integrating three components. First, we introduce a social planner model that captures the trade-off between maximizing market shares for high-quality programs and prioritizing economically vulnerable locations. Second, we harness a unique dataset that links families with childcare providers, enabling the establishment of proximity-based choice sets and facilitating the estimation of demand for childcare seats. Third, we explore interventions undertaken by two local governments to identify the social planner’s preferences. Our framework offers insights into expansion locations for high-quality programs, while considering potential cannibalization effects.
"Investigating the Impact of Quality Ratings in Child Care Provision on Long-Term Educational Outcomes" (with Flávio Cunha)
"Redrawing Boundaries: An Algorithmic Approach to Fairer School Attendance Zones" (with Dibya Mishra)
"Negotiating Work: The Role of Marriage Matching in Women's Labor Force Participation in India"