Research
Research
Abstract
While residential schools in North America have long been dismantled, India continues to expand its own residential school system, with a stated aim of "closing the gap" in education between Indigenous students and their peers. I provide the first causal evidence of the effect of enrollment in a residential school on the educational attainment of Indigenous women in India. Applying triple-difference and instrumental-variable strategies to a newly constructed dataset, I find that school exposure reduces educational attainment by up to 4 years. The result is driven by disruptions to family dynamics. Crowding out of day-school options by residential schools, along with mandatory residence at these schools, forces ST girls to cancel enrollment, raising dropout rates due to domestic work obligations by 10 percentage points.
Abstract
I study a conditional cash transfer (CCT) scheme that rewards below-poverty-line households for having a daughter who completes a minimum level of schooling and delays marriage. As the scheme was applied only in the Indian state of Karnataka, I use NFHS microdata and border regression discontinuity at Karnataka's state boundary. I find that the CCT scheme increases the presence of daughters mainly by expanding fertility: households are 15pp more likely to have a girl under 10 and 10pp more likely to have a boy under 10, with no change in sex ratios. The pattern is consistent with an 'insurance' mechanism: because the scheme mitigates future liability associated with having a girl (e.g., dowry), it becomes less financially risky for families to have one more child. Families continue childbearing to secure a highly desired son, consistent with their son preference, rather than substituting daughters for sons. Girls' education and health outcomes worsen despite the conditionality and financial incentives. Likely driven by tightened household resource constraints, the years of education decrease by 1.85 years, and secondary attendance decreases by 18 pp. Immunization rates remain unchanged. Survey responses indicate that the policy-induced increase in fertility places a significant burden on mothers, making existing constraints on health-care utilization, such as permission norms, distance to treatment, the need for accompaniment, and lack of a female provider, more binding.
Abstract
Anti-poverty programs can alter conflict dynamics, but their effects may depend on local political conditions. This paper examines whether political alignment between local representatives and the ruling state government shapes the effect of India’s National Rural Employment Guarantee Act (NREGA) on violence. Using a difference-in-discontinuities design that combines close state assembly elections with the phased rollout of NREGA, I show that in Naxalite-affected areas, aligned constituencies with NREGA experience more violence, driven by clashes between state forces and rebels. The evidence suggests that more effective implementation within aligned constituencies strengthens the state’s local reach and reduces rebel recruitment from poorer populations, thereby prompting insurgent escalation. In contrast, I find no such effects in Northeast India’s ethnic insurgencies, where recruitment is less tied to economic deprivation. The results show that political alignment changes NREGA’s effect on violence, and that the effect depends on the type of insurgency.
State-Sponsored Education and the Assimilation of Indigenous Peoples: SSHRC Insights Development Grant (Collaborator)