Abstract
This paper analyzes how marital sorting has affected household income inequality. Using CPS data for 1976 to 2019 we examine how wage inequality has evolved for both husbands and wives in households in which both spouses are working full time. We explore whether marriage sorting patterns have changed over this period and find evidence of increased assortative matching on wages. To identify the sources of these changes we introduce a multivariate distribution regression model with multiple sources of selection. The estimation of the model is complicated by the large number of outcomes corresponding to the characterization of a single sample observation. This is captured by a four dimensional multinomial probit model which we estimate via simulation methods. We conduct a number of counterfactual exercises and find the observed increase in positive sorting largely reflects the correlation across unobservables influencing both members' of the marriage wages. We also decompose the changes in sorting patterns into structural, composition and selection effects and find the increase in positive sorting primarily reflects the increased return to skill premia.
Invited by: Giulio Zanella