Research

Abstract: How has the degree of educational and racial assortative mating changed in the U.S. in the past few decades? Are marriages more favored by people now compared to the past? I use a multi-dimensional matching model to study the change in marital preferences on education and race in the United States, with Census data in 1960 and 2020. The model relaxes the “Separability” assumption in Choo and Siow (2006) and assumes match-specific idiosyncratic marriage matching shocks. I find the degree of educational and racial assortative mating decreased and people favor marriages less. My counterfactual analysis shows that changes in marital preferences increased household income inequality.

Abstract: We construct a nonstationary, equilibrium model of the marriage market in the presence of adverse selection. We show that an equilibrium exists where marriages exhibit negative duration dependence and stocks change over time to reduce the quality of the marriage market. We show that divorce provides a better signal to other singles than never having been married. Our nonstructural empirical results support our theoretical results. 

Abstract: We examine whether a woman from a violent family is more likely to match with a man from a violent family in India. The violent family here means whether a woman’s (man’s) father ever beat her (his) mother. Domestic violence is prevalent among women in India and has been associated with poor mental and physical health. If people are assortatively mating on violence, there will be intergenerational transmission of family values. We use a two-sided multi-dimensional matching model of the Indian marriage market to structurally estimate the marriage preferences on violence. Other critical drivers of marital sorting such as education, caste, and age, are also considered. A counterfactual analysis is designed to see how the health outcome of children will be improved if there is a policy to reduce domestic violence.