New Findings on Welfare and Children's Development: Summary of a Research Briefing (1997)

The immediate focus of most reform efforts is adult employment. But concerns about the detrimental effects of welfare for children are also a prominent aspect of today's debates, particularly insofar as they involve long-term and intergenerational effects. Speculations about the mechanisms that account for possible long-term effects of welfare receipt focus primarily on the transmission of expectations regarding adult roles and access to resources that facilitate the attainment of economic independence. For example, children who grow up in families that have histories of sporadic or no work are thought to acquire low aspirations regarding their own employment prospects and thus to make life choices about schooling and childbearing that may actually constrain the economic options available to them in young adulthood. Research into this complex set of questions that was presented at the briefing addressed the timing of welfare receipt and of its effects, the conditions associated with long-term effects, and the factors that contribute to intergenerational welfare receipt.

Although there is no typical pattern of welfare receipt during the childhood years, most children's exposure to welfare is episodic, and the years of receipt tend to occur together within stages of development. Greg Duncan of Northwestern University presented findings (Yeung et al., 1996) that explore the effects of the timing in childhood of AFDC receipt on total years of schooling.⁴

Duncan and his colleagues found that the relative influence of welfare receipt on years of schooling completed and on the risk of nonmarital births varied according to the timing of that receipt—whether it took place in early, middle, or late childhood. Receipt during the early childhood years had no discernable effects on these later outcomes. However, for blacks, taking into consideration a wide range of selection factors (e.g., age of mother at time of birth, mother's education, mother's work hours, average family income), welfare receipt in middle childhood (ages 6 to 10) was associated with relatively few years of schooling completed and a higher risk of teenage childbearing compared with children who did not live in AFDC families during this period of development. For black children, spending middle childhood in a heavily welfare-dependent as opposed to a welfare-free home was associated with roughly one-half year less schooling and a three times higher risk of teenage childbearing.

Duncan speculates that such effects may be attributable to the fact that middle childhood and early adolescence are stages at which children, having begun to perceive of welfare receipt as a viable future life course, may start to engage in behaviors—such as early sexual activity, disengagement from school, and problem behavior—that foreclose other options. It is during the middle and high school years that children form aspirations for the future and model parents' experiences in education and employment. This is also a time when parents can begin to connect children to activities that foster career aspirations, which may not be as accessible to families on welfare. Early in childhood, welfare receipt may supply needed income and little stigma and not lead to negative outcomes; in contrast, welfare receipt in middle childhood and adolescence may be detrimental.

It is critical to note that the Panel Study of Income Dynamics (PSID) lacks measures of the home environment, parenting skills, and parents IQ and mental health. As a result, it is not possible to distinguish statistically between the effects on child development of welfare receipt and those of these other conditions. If, for example, it could be established that welfare undermines families' efforts to provide rich learning environments for their children and that these learning environments are the "true" cause of children's later educational attainment and fertility, then any efforts to interpret the PSID findings as supporting a causal link between welfare, schooling, and childbearing would be inappropriate. At a minimum, Duncan's and his colleagues' findings suggest that concerns about how children are affected by welfare need to be balanced with considerations of how older children are affected as they begin to aspire to and prepare for adult roles.

Findings presented by Mark Rosenzweig of the University of Pennsylvania (Rosenzweig, 1995) also bear on efforts to specify more carefully those for whom welfare is most likely to have negative long-term effects. Rather than inquiring about the stage of childhood in which welfare receipt occurs, Rosenzweig examined whether young AFDC recipients' access to other sources of financial support affects their initial decisions to have children without marrying.

Rosenzweig first examined the relationship between state AFDC benefit levels and rates of nonmarital childbearing among young women on AFDC, using data from eight birth cohorts of women in the National Longitudinal Survey of Youth. Like other investigators who have studied this relationship, he found little evidence connecting average AFDC benefits to differential rates of nonmarital fertility in general.

Rosenzweig then examined whether young women on AFDC who differed with respect to their parents' earnings at age 14 were more or less likely to have children without marrying by the time they were 22. He found that higher AFDC benefit levels increased the probability of nonmarital childbearing for those women whose parents' in-

comes were less than $10,000 per year (about one-quarter of his sample). For women whose parents had higher incomes, variation in state benefit levels did not affect the likelihood of nonmarital childbearing.

Rosenzweig concluded that early childbearing is most likely to be associated with higher AFDC benefit levels when other potential or actual sources of support, such as monetary donations from parents, are meager. At the same time, Rosenzweig cautioned that welfare generosity alone cannot explain contemporary trends in nonmarital births, and the magnitude of the relationship between welfare and out-of-marriage childbearing is a matter of active debate. (For a discussion about the determinants of early childbearing as a selection factor, see Geronimus and Korenman, 1992; Furstenberg et al., 1993; and Hoffman et al., 1993.)

Research by Mark Rank of Washington University (Rank, 1996) extended the consideration of long-term effects to an intergenerational framework. Policy makers and researchers alike have long questioned whether welfare receipt in childhood predisposes welfare receipt in adulthood. Rank sought to determine whether welfare dependency or poverty is the mechanism that underlies patterns of intergenerational welfare receipt.⁵

Rank found that 75 percent of people on welfare did not report growing up in households that received welfare (though under-reporting of parental welfare receipt is quite likely). However, he also found that children who did grow up in families that received welfare were at least three times more likely to receive welfare when they become adults than children not raised in welfare families; see Table 2.

Rank's efforts to identify the reasons for this intergenerational pattern of welfare receipt highlight the effects of poverty on families' ability to help their children succeed in school, work, and society as a whole, rather than the direct effects of welfare dependence. The effects of parental welfare receipt disappeared when family income was taken into account. Rank concluded that the financial and economic constraints that accompany poverty, regardless of welfare receipt, set in motion influences that place children at risk of future welfare dependency. But there was one important exception to this conclusion: Rank found that welfare receipt as a child had a modest effect on the person's probability of having children as a single parent, even when parents' economic backgrounds were taken into account.

Considered together, the findings reported at the briefing point to welfare receipt during the preteen and adolescent years, and to consequences for school completion and premarital childbearing, as areas of concern and hence, future study. Additional questions for further study include: What devel-

opmental processes during the middle childhood and adolescent years may leave children vulnerable to long-term negative effects of welfare receipt? As welfare reform proceeds, how should policy makers and researchers reformulate questions about how families' economic resources affect the course of children's development and future prospects? Do different mixes of financial support (e.g., public benefits, earnings, child support payments, donations from parents) have differential effects on children's long-term development? What are the noneconomic mechanisms over time, such as role modeling, that may help explain links between young children's economic circumstances and later outcomes? What measures of which outcomes need to be developed to expand efforts to understand the long-term effects of families' economic well-being, including their reliance on public assistance?