The Impact of Child Support Enforcement Policy on Nonmarital Childbearing

A simple model of fatherhood and marriage choice implies that stricter child support enforcement will tend to reduce nonmarital childbearing by raising the costs of fatherhood. We investigate this hypothesis by examining nonmarital childbearing during 1980–1993, a period when child support policy and enforcement underwent enormous changes. We use a sample of women from the Panel Study of Income Dynamics, to which we add information on state child support enforcement. We examine childbearing behavior between the ages of 15 and 44, both before marriage and during periods of nonmarriage following divorce or widowhood. Discrete-time hazard models of nonmarital childbearing provide evidence that women living in states with more effective child support collection were less likely to bear children when unmarried. The findings suggest that policies that shift more costs of nonmarital childbearing to men may reduce this behavior. The Impact of Child Support Enforcement Policy on Nonmarital Childbearing Births outside of marriage have grown dramatically over the past three decades in the United States. In 1966, the nonmarital birth ratio—the percentage of births to unmarried mothers relative to all births—was approximately 8 percent. Since 1994, the figure has been between 32 and 33 percent. Among teenagers the ratio is much higher—79 percent in 1999 (Ventura and Bachrach, 2000). The nonmarital birth rate (number of births per 1,000 unmarried women) has also grown steadily. In 1965 it was 23.4 among all women aged 15–44 and 16.7 among teenagers aged 15–19 (U.S. Department of Health and Human Services, 1995). By 1994 it had risen to 46.9 among all women aged 15–44 and 46.4 among teenagers. The rate declined slightly, to 44.3, by 1998 among women aged 15–44 and to 41.5 among teenagers (Ventura and Bachrach, 2000). In 1996 Congress enacted new legislation designed to reduce welfare eligibility and increase the costs of single motherhood as part of the Personal Responsibility and Work Opportunity Reconciliation Act (PRWORA). State legislation with similar motivation both preceded and followed the PRWORA. These federal and state initiatives have included policies that lower welfare benefits, limit eligibility, impose stronger work requirements on welfare recipients, and place restrictions on benefits to unwed parents under age 18 who do not live with their parents (or in another adult-supervised setting) and attend school. Most of these policies were driven, at least in part, by the assumption that the availability of welfare to unmarried women was a major cause of nonmarital childbearing (Murray, 1984, 1993), despite empirical evidence to the contrary (Moffitt, 1992, 1998). The asymmetrical focus on women appears unreasonable. Decisions about sexual intercourse and marriage involve two persons rather than one. The same often is true for decisions about contraceptive use and abortion. Yet the research literature, as well as policy debates, has largely failed to recognize and critically analyze men’s role in nonmarital childbearing and how government policies may influence men’s behavior. The government’s poor record of establishing paternity and enforcing payment of child support by nonresident fathers (Sorensen, 1997) may partly be responsible for men’s failure to take 2 responsibility for contraception or to marry their sexual partners. Although efforts to establish paternity and enforce child support have intensified during the past decade and were strengthened by the PRWORA, they have generally been viewed as ways of reducing the financial costs of public welfare rather than as strategies for preventing nonmarital births. This paper presents a simple model of how the incentives of child support policy affect unmarried men’s decisions about fatherhood and marriage. The model implies that stronger child support enforcement reduces the likelihood of nonmarital childbearing. We test this hypothesis using data from the Panel Study of Income Dynamics during 1980–1993, a period when child support policy and enforcement underwent enormous changes. We find evidence in support of the hypothesis. U.S. CHILD SUPPORT POLICIES AND THEIR EXPECTED EFFECTS ON NONMARITAL CHILDBEARING Until very recently, financial responsibility for children born outside marriage rested primarily with the mother and her family and with government. Mothers who met the income test, which included the vast majority of unwed mothers, were eligible for Aid to Families with Dependent Children (AFDC), Food Stamps, Medicaid, and in many cases housing subsidies. In contrast, unwed fathers were more or less free to shirk their parental obligations, and most did so (Garfinkel, 1992). During the past quarter century, the federal government has taken a number of steps to prevent unmarried fathers from abandoning their children financially (Garfinkel, McLanahan, and Robins 1994; Garfinkel and McLanahan, 1986, 1994). In 1975, Congress created the Child Support Enforcement Program, which established local offices of child support enforcement and authorized federal matching funds for states to help locate absent parents, establish paternity, establish child support orders, and obtain child support payments (U.S. House of Representatives, 2000, Section 8). The 1984 Child Support Amendments extended this legislation by requiring states to withhold child support obligations from the paychecks of delinquent fathers and to develop legislative guidelines to be used in determining child support awards. In 1988, the Family Support Act mandated that states adopt presumptive guidelines for 3 child support awards and initiate automatic withholding from fathers’ paychecks, regardless of delinquency. The Act also included provisions aimed at strengthening paternity establishment for children born to unmarried parents. Reforms in the 1996 PRWORA sought to further improve the child support system’s ability to establish paternity for children born outside of marriage, to locate nonresidential fathers, and to collect support payments. The results of this new legislation have been striking with respect to children born outside marriage. The proportion of never-married mothers with a child support award grew from 12 percent in the early 1980s to over 20 percent in 1994 (Hanson, 1995). Paternity establishment ratios (the number of paternities established in a given year divided by the number of nonmarital births) increased from 20 percent to 46 percent over the same period, and to 64 percent by 1998 (Nichols-Casebolt and Garfinkel, 1991; U.S. House of Representatives 2000, Table 8-22). A Theoretical Model of How Child Support Policy Affects Men’s Fatherhood and Marriage Choices Administration and enforcement of child support laws raise the likelihood that fathers who do not live with their children will nonetheless be required to make substantial financial contributions over many years to their support. Child support policy, therefore, increases the expected costs of fatherhood for absent fathers. This disincentive would, other things equal, make men more reluctant to father children outside marriage and, if a nonmarital pregnancy occurs, make them more likely to marry before the birth. For women, improved enforcement that leads to higher child support payments might appear to reduce the costs of children and create more incentive to have children outside marriage. However, given that a large proportion of both teenage and nonteenage women who give birth while unmarried are likely to go on welfare (Duncan and Hoffman, 1990; Foster, Jones, and Hoffman, 1998; Haveman and Wolfe, 1994) and given that welfare policy taxed child support payments during the time period we analyze For absent fathers who plan to provide support above the level required by child support policy, policy may not affect costs. But for no father does it lower the expected financial costs of fatherhood. 4 (1980–1993), this countervailing effect is likely to have been small. From 1980 to 1984, a mother on welfare retained none of the child support paid by the absent father. Rather, all payments were used to offset benefits paid from public funds. Between 1985 and 1993, a mother on welfare was allowed to keep only the first $50 of child support each month. All payments above $50 went toward reducing public spending on welfare and did nothing to increase her children’s standard of living. Indeed, from the mother’s viewpoint stricter child support enforcement may actually have increased the cost of raising children if she had been getting informal support from the father and if that support ended because of stricter enforcement (Waller and Plotnick, 1999). One may reasonably conclude that child support policy during 1980–1993 did little to affect women’s incentives regarding nonmarital childbearing. A simple utility maximization model captures the essence of how the incentives set up by child support policy in the 1980s and early 1990s would affect men’s behavior. We focus on male decision making because, as just argued, child support policy left women’s incentives largely unaffected. Consider an unmarried man romantically involved with an unmarried woman, but not cohabiting with her or sharing income to any significant extent. We assume he faces three fatherhood and marriage options. Under the “no-child” option, he avoids fatherhood and does not marry. Under the “marital-child” option a nonmarital pregnancy occurs and he marries (or cohabits with) his partner and becomes a father. He shares his income with both mother and child. Assuming no welfare case has opened, the child support agency does not become involved. In the “nonmarital-child” option, the man becomes a father but neither marries nor cohabits with the mother, while the mother and child go on welfare. Assume that the man’s utility depends upon his personal consumption, consumption of the mother and child (whether living with him or not), and the nonfinancial benefits of parenthood and of In response to the 1996 welfare reforms, most states have eliminated this $50 “pass-through.” Another option is nonmarital pregnancy followed by abortion. We assume the man can unilaterally choose to try to avoid fatherhood by use of condoms or vasectomy. If contraceptive failure leads to an unplanned nonmarital pregnancy, he must choose between the other two options. Fatherhood is possible only if his partner agrees or an unplanned pregnancy occurs (and she does not seek an abortion). If she carries to term, we assume her own utilitymaximizing calculus led her to expect she would be better off with a child whether he chooses the second or third option. If she refuses to bear children, the effect of child support incentives on the man’s behavior is moot. 5 marriage (or cohabitation). Let Cp = personal consumption, Ck = consumption of the mother and child, Y = his own income, V = the implicit value to the man of being married to his current partner and of parenthood, V = the sum of the increase in his earnings produced by marriage (Gray, 1997) plus savings from economies of scale from living as a family unit, and G = the welfare benefit received by the mother and child if they do not live with the father. Let K = 1 if he fathers a child, 0 if not; and let M = 1 if he marries, 0 if not. If a child is born, the woman is the primary parent and earns nothing. Ignore, for the moment, the existence of a child support program. The man chooses among the combinations K = 0, M = 0; K = 1, M = 0 and K = 1, M = 1 to maximize his utility: (1) Max U(Cp, K*Ck, M*V) s.t. Y + (1-M)*K*G + M*V = Cp + K*Ck Given his preferences, the man determines the maximum utility of each combination and makes a global utility-maximizing choice. Without information on those preferences, the model does not predict which option he will choose. A policy that requires child support payments is easily incorporated into the model. Suppose the father must pay S in child support to the mother and that S < G. Assume that paying S does not affect his own income and that welfare policy taxes child support payments 100 percent, so the mother’s welfare benefit falls by S. The new budget constraint is (2) Y + (1-M)*K*G + M*V (1-M)*K*S = Cp + K*Ck Note that under the choices K = 0, M = 0 or K = 1, M = 1, the constraint is the same as in (1) and so is the maximum utility obtainable from both choices. If, however, he chooses K = 1, M = 0, the leftBy allowing Ck to enter the utility function only if there is a child, we assume he gets no utility from financing consumption of the woman if she is not the mother of his child. Child support policy does not appear to affect men’s earnings (Klawitter 1994; Freeman and Waldfogel, 1998). 6 hand side of (2) falls by S. The maximum utility obtainable from this choice must fall, and it will fall monotonically in S. Hence, the likelihood that an unmarried man would choose the nonmarital-child option will decrease as S grows. Figure 1 illustrates the budget constraints under each choice. The vertical axis measures the man’s personal consumption; the horizontal one measures consumption of the mother and child. Under the no-child option, all income is devoted to personal consumption. The constraint collapses to a point on the vertical axis equal to his income, Y. (The line with slope –1 starting at Y is for reference.) To analyze the marital-child option, we draw the constraint under the assumption that V + V is positive. To show a situation where the marital-child option does not necessarily provide more utility than the no-child option, the figure assumes the man’s consumption must be less than Y. Then the marital-child option’s constraint starts at Z and has slope of –1. Because mother and child cannot consume from V, the constraint does not extend to the horizontal axis. Under the nonmarital-child option in the absence of a child support system, the mother and child receive a welfare benefit of G, while V = V = 0. If the father contributes nothing to their support, he consumes Y. Recall that the mother earns nothing. The budget constraint therefore starts at point W. If the parents agree, the father may provide under-the-table support payments to the mother and his child. In that case, his personal consumption declines by one dollar for each dollar of support. Again the constraint has slope of –1. We draw the constraint under the assumption G > V + V so that the marital-child option does not necessarily provide more utility than the nonmarital-child option. To incorporate child support payments into the figure, suppose the father must pay S in child support and that S < G. Under the assumption that S does not affect his own income, his maximum personal consumption shifts down to Y – S. Given that welfare policy taxes child support payments 100 If V + V 0 the man would choose the marital-child option only if the value of parenthood (the utility contributed by Ck) at least compensated for the loss of utility due to V + V . If G < V + V the man would never choose the nonmarital-child option. This will be the case for some men. To illustrate where choice may be affected by child support policy, the figure assumes G > V + V . FIGURE 1 Budget Constraints for Different Fatherhood and Marriage Options