Research

In many developing countries, women marry young, and married women are far less likely to be enrolled in school than single women of the same age. This paper studies how the timing of marriage affects women’s educational attainment in India. I exploit variation in marriage timing generated by a common social practice in which daughters marry in birth order. Instrumental variables estimates imply that a one-year increase in age at marriage raises women’s probability of completing high school by 6.1 percentage points.

I then develop and estimate a dynamic equilibrium model in which households choose schooling and marriage timing in a directed-search marriage market. Estimates imply that although grooms prefer younger brides, the value they place on women's education more than offsets the associated age penalty. Even so, the age penalty, schooling costs, a finite search horizon, and birth-order constraints limit the gains from postponing marriage, thereby sustaining early marriage in equilibrium. I use the estimated model to evaluate an increase in the minimum legal marriage age. The policy raises schooling primarily among high-income women but generates substantial welfare losses by worsening marriage prospects. These findings show the importance of household resources and marriage-market equilibrium responses for policies aimed at increasing women's education. 

Firearms are the most common weapon used in intimate partner homicide and sharply raise the probability that intimate partner violence turns lethal. Beginning in the 1990s, many states enacted laws restricting firearm access for individuals subject to a domestic violence restraining order (DVRO). We use the Federal Bureau of Investigation’s Supplementary Homicide Reports to measure homicide outcomes and compile data on civil protection order filings from the National Center for State Courts and individual states’ annual statistical reports. Using a staggered difference-in-differences design, we find that firearm dispossession laws reduce male-perpetrated firearm homicides of women by approximately 35 percent and intimate partner homicides by roughly 20 percent, with no corresponding increase in homicides involving other weapons. These findings are consistent with the laws operating by disarming the small group of high-risk individuals already subject to restraining orders, rather than by changing aggregate firearm ownership or the demand for court protection. We find no clear evidence that mandatory surrender laws outperform those that leave removal to judicial discretion.

How does the birth of a child alter the allocation of resources within the household? Using panel data from the Japanese Panel Survey of Consumers (JPSC), we document that women’s share of household consumption expenditures falls sharply after the birth of the first child and remains persistently lower for at least ten years. Time-use data show a parallel reallocation: women's housework and childcare rise substantially while market work and leisure fall, while husbands' time allocation changes much less.

We  then test whether this decline reflects greater maternal altruism or a decrease in women's bargaining power within the household. To do so, I exploit a 2007 Japanese pension reform that improved the outside option of women married to husbands in permanent employment. Using a difference-in-differences design, we test whether the post-birth decline in women’s resource shares becomes smaller for affected wives after the reform.

This project studies how women’s age and educational attainment affect marriage prospects in India. Using a randomized field experiment on an online Indian matchmaking platform, I create fictitious female profiles that vary in age and education and use male users’ views and invitations to connect to measure interest from prospective grooms. The design allows me to estimate whether older women face a penalty in the marriage market, whether education increases women’s desirability, and whether higher education can offset any penalty associated with age. More broadly, it examines how marriage-market returns may shape women’s incentives to delay marriage and invest in schooling.

This paper examines whether childhood exposure to the Food Stamp Program (FSP) improved intergenerational health mobility using Panel Study of Income Dynamics data and the program’s 1961–1975 rollout as a natural experiment. We find that full childhood access to food stamps reduced the persistence of poor health between parents and children by roughly half and increased upward health mobility, measured as the implied change in the child’s expected adult health rank when parental health is low, by about 14 percentile ranks when moving from zero to full childhood exposure. These effects, while large, are estimated with large confidence intervals due to a modest sample size. Our results suggest that early-life nutrition assistance can help break the cycle of poor health across generations. We discuss potential mechanisms, placebo and robustness checks, and limitations, and highlight implications for health equity and social safety net policy. 

I study whether having a first-born son affects a mother's exposure to domestic violence in India. Using data from the National Family Health Survey (2019-21), I exploit quasi-random variation in the sex of the first child, and estimate the causal effect of first-born child sex on multiple dimensions of violence, including physical, sexual and emotional abuse.  I further examine heterogeneity across birth cohorts with varying exposure to prenatal sex determination technology. Preliminary findings indicate that having a first-born son reduces mothers' exposure to severe physical and sexual violence.