We use population data on capital income and wealth holdings for Norway to measure asset positions and wealth returns before individuals marry and after the household is formed. These data allow us to establish a number of novel facts. First, individuals sort on personal wealth rather than parental wealth. Second, people match also on their personal returns to wealth and assortative mating on returns is as strong as that on wealth. Third, post-marriage returns on family wealth reflect in equal part the pre-marriage returns of the two spouses. However, for households at the top of the wealth distribution at marriage, family wealth is largely managed by the spouse with the highest potential to grow it, providing a microfoundation for the scale dependence in wealth returns documented in several empirical papers. Fourth, marriage lowers the degree of wealth inequality as well as the heterogeneity in returns relative to the counterfactual case of no marriages. We use a simple framework to illustrate how the inequality-attenuating role of marriage is affected by assortative mating on wealth and returns and by wealth management task allocation between spouses.