The Long Shadow of Communal Violence: Childhood Exposure and Later-Life Mental Health
with Gaurav Dhamija and Sagnik Kumar Gupta
Abstract: We study the long-run relationship between childhood exposure to communal violence and women’s mental health in India. Using a newly constructed crosswalk spanning seven decades, we link historical Hindu--Muslim riot locations to respondents’ childhood districts in the Longitudinal Ageing Study in India (LASI). Exploiting variation in exposure across childhood districts and birth cohorts, we find that exposure to at least one riot between birth and age 14 is associated with a 2.4 percentage point higher probability of major depressive disorder in later life, or 29 percent of the sample mean. Greater cumulative exposure is similarly associated with worse later-life mental health. The association is strongest for exposure at ages 0-4 and 10-14, consistent with differences in vulnerability across developmental stages. Childhood riot exposure is also associated with a higher likelihood of child marriage, pointing to changes in life-course trajectories as a potential channel. These findings show that even localized, short-lived episodes of communal violence can have mental-health consequences that persist for decades.
with Nishith Prakash, Soham Sahoo and Deepak Saraswat
Abstract
This article examines the effect of electing criminally accused politicians on crime in India, considering the severity of their alleged offenses. Employing an instrumental variable approach that exploits the quasi-random variation in outcomes of close elections between candidates with and without criminal accusations, we find that a standard deviation (SD) increase in the share of criminally accused leaders in institutionally weaker states leads to a 0.05 SD rise in yearly reported crimes. Leaders accused of serious crimes have a more pronounced effect on crime outcomes, including crimes against women. Crucially, leaders accused of serious crimes also exert a negative influence on female labor force participation, underlining their detrimental impact on socioeconomic welfare. Our findings highlight the significance of considering the nature and severity of criminal accusations when evaluating the impact of criminally accused politicians on crime and society.
Abstract
Birth order shapes life outcomes, but do these differences persist across generations? This paper examines whether a mother's birth order is associated with her children's developmental trajectories. Using mother-child panel data from the National Longitudinal Survey of Youth, I find that children of later-born mothers score lower on cognitive tests than those of firstborn mothers. These gaps emerge early, persist into adolescence, and are associated with lower college completion. In magnitude, these intergenerational differences are about one-third the size of canonical first-generation birth order gaps. Suggestive evidence indicates that children of later-born mothers receive less cognitive and emotional stimulation at home than those of firstborn mothers.
Abstract
Extreme Risk Protection Order (ERPO) laws, or ``red flag'' laws, allow law enforcement and family members to petition a court to temporarily remove firearms from individuals who pose a credible risk of harm, on the basis of observable behavior rather than a prior conviction. Unlike most firearm policies, they are prospective, individualized, and civil, and by empowering those closest to a potential perpetrator to act, they are especially relevant to domestic violence, where danger is often visible within the household before any crime occurs. I estimate the causal effect of ERPO adoption on intimate partner homicide by exploiting the staggered rollout of these laws across U.S. states from 1999 to 2024 and applying modern difference-in-differences methods designed for staggered treatment adoption. Using agency-year homicide data from the FBI Supplementary Homicide Reports, I find that firearm intimate partner homicides decline by 10--15\% following adoption, with no evidence of differential pre-trends.
with Sunny Shankar Ladkat and Gaurav Dhamija
Abstract
Adolescent girls in India perform substantially more unpaid domestic work than boys. Using nationally representative time-use data, we examine how this gender gap changes when mothers are employed. Controlling flexibly for adolescent, parental, and household characteristics and a rich set of fixed effects, we find that maternal employment is associated with about 26 additional minutes of domestic work per day for girls, while boys are largely unaffected. The gap increases with mothers' work hours, is largest among mothers in casual wage employment, and widens as girls move through adolescence. Girls primarily accommodate this additional work by reducing leisure and personal care. The association is also stronger in more patriarchal districts and persists across most socioeconomic groups. Results are robust to an instrumental-variables strategy using a shift-share measure of district-level female labor demand.