Explaining the Female Labor Participation Gap Across Religion and Caste in India

Abstract: Why does labor share vary for married women across religion and caste in India? The gaps in employment shares across groups are not trivial, ranging from 17% for Hindu Brahmins to 32% for Christians in 2011. I estimate a static female labor supply model that incorporates variations in individual demographic characteristics from the Indian Human Development Survey and two sets of parameters characterizing the labor behavior of specific groups. Religion and caste enters the model through a preference component and an "institutional" component embedded as a fixed cost for women entering the labor force. Given model specification and estimates, I document group labor supply behavior in the intensive and extensive margin and quantify the relative importance of each model component in generating behavioral differences. In a counterfactual exercise without group heterogeneity---every group faces the same preferences and fixed-labor costs---the variance of participation across groups reduces by 0.75 relative to patterns seen in data. This suggest that inequality in terms of demographic characteristics, such as education, non-labor income, and children, alone does not generate large participation gaps. When group differences enter the model through preferences, the variance of participation does not change significantly. Only when heterogeneity in labor costs is introduced, the variance of participation gaps starts to increase. I further investigate possible sources of variations in labor cost and study how groups react to policy changes in terms of wages and a subsidy to non-labor income.