Shruti Agarwal
Welcome! I am a PhD candidate in Economics at the University of Warwick, interested in development economics, public economics, and political economy.
I study how public systems change the margins on which households and officials make decisions: which schools children can attend, how local leaders allocate effort, and whether information and social signals shift educational choices. My projects combine geocoded administrative data, quasi-experimental variation, and field and survey experiments to estimate how these margins affect sorting, service delivery, and human-capital outcomes.
Before the PhD, I was a Research Associate at J-PAL South Asia. I grew up in Meerut, in north India, and remain an avid listener of Nusrat Fateh Ali Khan.
Public Access, Feasible Choice, and Social Sorting: Evidence from Public Schooling in Rural India
Exposure to out-group peers in childhood can shape trust, cooperation, and shared norms. Schools are a primary site of such contact, but sorting across schools within local markets determines who meets whom. This paper estimates the causal effect of increased nearby public school access on caste segregation across primary schools in rural India. Using geocoded administrative data spanning more than a decade, I construct distance-based local education markets and measure within-market segregation and peer exposure. To identify the effect of access, I exploit a national reform that tightened proximity standards, instrumenting realized local access with baseline eligibility under the rule. IV estimates show that expanding nearby school access increases segregation: a 10 percentage point increase in within-1 km coverage raises the dissimilarity index by 7.3 percent of the mean and increases Scheduled Caste isolation by 7.1 percent. The evidence points to within-market resorting rather than a purely mechanical entrant effect. Effects exceed a nearest-school benchmark, survive restrictions to incumbent schools, and are not explained by private-sector entry, cross-sector reallocation, or observable school quality.
Relational Representation and Public Program Delivery: Evidence from Caste Reservations in India
Can political quotas improve public program delivery when authority is spread across multiple tiers of government? This paper argues that the effects of representation are relational: the same reserved leader may face very different constraints depending on the social position of officials above them. I study this question in rural Uttar Pradesh, where caste-based reservations assign leadership positions across nested lower-tier and higher-tier local governments. I combine local election data with administrative records from India's National Rural Employment Guarantee Act and exploit cutoff-based variation generated by reservation assignment rules. The design estimates how the effect of lower-tier caste representation varies with the caste environment at the higher tier. Cross-tier pairings shape both administrative implementation and distributive outcomes: employment generated, works completed, delays, bottlenecks, and the caste composition of program benefits. The findings suggest that quotas affect delivery not only by changing who holds office, but also by changing where that office sits within the hierarchy of local government.
Information, Experience, and Climate Belief Updating: Evidence from Students in Bangladesh
Do young people update climate beliefs in response to new information or direct exposure to extreme weather? We study this question using a three-wave panel of more than 5,000 students in Bangladesh, one of the world's most climate-exposed countries. The data measure climate knowledge, skepticism, anxiety, engagement, academic skills, and real-stakes willingness to donate to climate-related causes. We combine the panel with two sources of variation: a school-level randomized climate education intervention and quasi-experimental exposure to a major rainfall event during survey implementation. Neither channel produces detectable belief updating. Short informational materials, a detailed climate booklet with an incentivized quiz, and recent rainfall exposure have limited effects on knowledge, skepticism, anxiety, engagement, or willingness to pay. Instead, climate beliefs are highly persistent and strongly correlated with baseline academic skills. The findings suggest that adolescent climate beliefs in high-risk settings are not easily shifted by brief information campaigns or isolated weather shocks.
Status Signals and Educational Choice: Experimental Evidence on Vocational Training in Saudi Arabia
Educational choices may respond not only to expected earnings, but also to the social meaning of different schooling paths. We test this channel in Saudi Arabia, where technical and vocational education remains a policy priority but often carries lower social prestige than academic routes. We run a 2×2 randomized survey experiment that independently varies the prestige of a vocational career vignette and the provision of factual wage information. Status framing substantially increases willingness to consider vocational training and raises expected wages, despite containing no wage data. Wage information alone has no detectable effect on attitudes or behavior. Revealed engagement shows the same pattern: high-status framing increases post-survey demand for TVET information, while adding wage information weakens this response. The results point to social image, rather than earnings information, as a central constraint on vocational demand in this setting. Information campaigns may fail when they target the wrong margin of educational choice.
