The State, the Family and Redistribution

这个研究项目的目的是对结婚率下降的政治后果进行经济分析。可以说，近代家庭历史上最重要的发展是婚姻的减少，以及大多数西方国家单身人口比例的上升。在美国，从1964年到1996年，已婚人士的比例下降了四分之一（从84%降至58%），而离婚人士的比例增加了三倍（从3%升至10%）。就目前而言，婚姻是个人之间资源再分配的最重要制度之一，再分配偏好是许多政治选择的核心，因此研究婚姻的下降如何影响政治结果是相关的。我们的研究开发了一个理论模型，该模型表明，在对潜在收入分配、婚姻模式和家庭内部收入分享的合理假设下，婚姻程度的变化会改变收入分配，从而改变政治结果。我们使用该模型来确定不婚的增加如何影响(i)男性和女性政治偏好的差异，即所谓的性别差距；（二）再分配的总需求；（iii）摇摆选民的人口及经济情况；（四）已婚和未婚选民的偏好差异，即所谓的婚姻差距。我们分析的一个关键假设是，一般来说，女性嫁给比自己更富有的男性。因此，相对于单身，婚姻让男人更穷，让女人更富有。这表明，结婚率的下降可以解释政治上的性别差距。再分配总需求的变化不太清楚，因为每有一个女人向左转，就有一个男人向右转。我们表明，如果不结婚首先在穷人中普遍存在，然后在收入分配中“爬升”，那么最初的不结婚会减少对再分配的支持，因为男性以前的左倾选民转向支持右翼（参见里根民主党）。随着未婚人口在相对富裕的群体中增加，这一群体中的女性将开始根据其社会经济状况进行投票（参见足球妈妈）。这一发现与婚姻差距是一致的，即相对于未婚个人，已婚个人更大程度上支持这项权利。我们使用美国的数据来检验我们的假设。首先，使用青年父母社会化调查（1973年和1982年）的纵向数据，我们调查了个人婚姻状况的实现变化是否会影响他/她在假设方向上的政治偏好。其次，我们研究了1964年至1996年期间美国一个州“婚姻市场”的变化如何影响男性和女性的政治偏好。我们通过两种方式衡量州婚姻市场的变化：第一，通过一个州目前离婚的个人数量（根据年度当前人口调查计算）；其次，各州离婚法发生了变化。我们关于个人政治偏好的信息来自全国选举调查研究（1964-96）。

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).