Research

I investigate whether an improvement in school accessibility for young girls decreases their share of unpaid care work. I work in collaboration with a local NGO in Jaisalmer, India, that opened a new girls’ school designed to reduce key barriers to attendance. I collect two rounds of household survey data, including detailed time-use measures for girls in schooling age and their household members. I find that girls offered access to the new school spend more time on education and less time on domestic duties, alongside increased commuting time and reduced leisure. I find no evidence of a reallocation of unpaid care work across household members.

Low-intensity urban flooding is widespread in urban centers of the Global South, but its welfare effects are not well understood. Our study exploits spatial spillovers in flood exposure from drainage infrastructure upgrades in Barishal, Bangladesh, as a natural experiment to study the welfare costs of these low-intensity events.  Using a hydrodynamic flow model, we show that these spillovers are widespread, non-linear, and of meaningful magnitude. Combining these predictions with a spatially representative household survey of 2,649 households conducted after the 2025 rainy season, we show that an additional hour of flooding increases work absences by 0.3 days, raises income losses by 13-15 percent, and increases the number of days children miss school by 0.12-0.14 days. 

Across the developing world, women face mobility barriers that constrain access to jobs, education, and social networks. Gender-targeted transport policies, such as Delhi's Pink Ticket Scheme, which grants women free bus travel, aim to ease these constraints but may generate unintended consequences at the point of delivery. We study whether free fares lead bus drivers to discriminate against female passengers, raising the effective cost of travel for the very group the policy targets. Combining a mystery shopper field exercise with a lab-in-the-field experiment with bus drivers and conductors, we causally identify discriminatory non-stopping behavior and decompose its underlying mechanisms, distinguishing among gender norms, fairness concerns, economic incentives, and safety considerations. Descriptive data already show buses failing to stop for women roughly 50% more often than for men. The findings will inform the design of inclusive public transport systems in low- and middle-income countries.

We study whether early-life exposure to the Kumbh Mela, one of the world’s largest religious gatherings, affects infant mortality and child growth. Preliminary estimates suggest that Kumbh exposure increases infant mortality by 1.9 percentage points among urban births within 60 km and increases the probability that a child is stunted or underweight by 5.6 percentage points. Our back of the envelope calculations imply approximately 2,300 additional infant deaths and 16,900 additional children under five who are stunted or underweight per Kumbh Mela event.

I study how drought affects the timing of conceptions in India by combining high-resolution rainfall data with retrospective birth histories from the two most recent rounds of the Demographic and Health Surveys. Exploiting spatial and temporal variation in moisture shortfalls, I find that severe droughts during the main agricultural season are followed by a sharp decline in conceptions in the short run, with partial catch-up within the subsequent year.