Research

Working Papers

          "Caste Residential Segregation and Children's Education Outcomes in Rural India" (with Alok K. Bohara) Link

In many villages in India, the Scheduled Castes and Scheduled Tribes live in physically separate hamlets. This paper investigates whether such caste residential segregation lowers the educational outcomes of marginalized children. Using the Indian Human Development Survey (2011-2012), we estimate the association between caste residential segregation and children's learning in reading, writing, and mathematics among children aged 8-11 years. The SC children’s reading and writing learning assessment in segregated villages are 0.170 and 0.164 standard deviations lower than those of SC children in non-segregated villages of the same district, respectively. The result remains consistent under the coefficient stability test and robust to unobserved confounding under the Oster (2019) criteria. The own-group specificity results confirm that the learning outcome for SC children falls in the villages where SC communities are segregated and not in the villages where only ST communities are segregated, a pattern which is inconsistent with segregated villages being merely disadvantaged in general. The estimates for ST children are directionally similar but not significant. The results suggest that caste residential segregation is a significant indicator of educational disadvantage and that expanding educational inputs alone may be insufficient.

Marrying Up, Paying the Price:  Caste Hypergamy, Domestic Violence, and Health Injuries in India (with Alok K. Bohara and Richard Santos) Link

Caste stratification in India not only governs social hierarchy but also influences the welfare of women within the household. This paper studies whether caste hypergamy, Scheduled Caste/Scheduled Tribes women marrying General or Other Backward Class men (marrying up), increases domestic violence (DV) and the subsequent physical health injuries (HI) it causes, and whether caste hypogamy, higher-caste women marrying down, produces a symmetric effect. Using pooled data from two rounds of the National Family Health Survey (NFHS 2015–2016 and 2019–2021) and a simultaneous IV-ordered probit framework that addresses the endogeneity of domestic violence, we find that caste hypergamy significantly increases domestic violence, which in turn significantly increases health injuries. By contrast, caste hypogamy has no significant effect on domestic violence, showing a sharp asymmetry rooted in the directional nature of the caste hierarchy. The DV–HI relationship is non-linear, and heterogeneity analysis shows that the hypergamy effect is strongest in early marriage and decreases with marital duration. These results carry clear implications for India's inter-caste marriage promotion policies: incentivizing inter-caste marriages without simultaneously strengthening the legal and economic protections available to lower-caste wives may expose the most marginalized women to increased domestic violence.