The State, the Family and Redistribution

Columbia University, New York NY

Investigators

Abstract

The goal of this research project is to provide an economic analysis of the political ramifications of the decline in marriage. Arguably, the most significant development in the recent history of the family is the decline in marriage, and the rise in the proportion of single individuals in most Western countries. In the US, between 1964 and 1996, the proportion of currently married individuals fell by a quarter (from 84 to 58 percent) while the fraction currently divorced increased threefold (from 3 to 10 percent). In so far as marriage is one of the most important institutions for resource redistribution between individuals, and redistributive preferences are central to many political choices it is relevant to examine how the decline in marriage has affected political outcomes. Our research develops a theoretical model which demonstrates that under reasonable assumptions on the underlying income distributions, marriage patterns and income-sharing within households, changes in the extent of marriage alter the income distribution and thereby political outcomes. We use the model to identify how increased non-marriage affects (i) the divergence of male and female political preferences, the so called gender gap; (ii) the aggregate demand for redistribution; (iii) the demographic and economic profile of swing voters; and (iv) the divergence in preferences between married and not married voters, the so called marriage gap. A key assumption underlying our analysis is that in general women marry men who are richer then themselves. Hence, relative to being single, marriage makes men poorer and women richer. This suggests that the decline in marriage could explain the political gender gap. The change in the overall demand for redistribution is less clear since for each woman who turns left, a man turns right. We show that if non-marriage is first prevalent amongst poor individuals, and then ``creeps up'' the income distribution, then initially non-marriage reduces the support for redistribution as male erstwhile left-leaning voters shift to supporting the right (cf. Reagan Democrats). As non-marriage increases amongst relatively well off groups, women in this group will start to vote at variance with their socio-economic profile (cf. Soccer moms). This finding is consistent with the marriage gap, i.e. that, relative to unmarried individuals, married individuals support the right to a greater extent. We use US data to test our hypotheses. First, using longitudinal data from the Youth-Parent Socialization survey (1973 and 1982 waves) we investigate whether realized changes in an individual's marital status affect his/her political preferences in the hypothesized direction. Second, we examine how changes in a US state's ``marriage market'' over the period 1964-96 have affected male and female political preferences. We measure changes in the state marriage market in two ways: first, by the stock of currently divorced individuals in a state (computed from the annual Current Population Survey); and second, by state divorce law changes. Our information about individual political preferences in from the National Election Survey studies (1964-96).