Welfare and the well-being of children.

Currie, Janet

There is broad public consensus that welfare programs should benefit poor children. Yet we know remarkably little about whether they actually do. Most research on welfare programs, as well as much of the debate about welfare reform, has focused on the way that parents respond to the incentives created by the system, rather than on the effects of these programs on children. My recent research begins to fill this gap.

The fundamental question is whether any of these programs benefit children. If it can be shown that they do, a second question is: which types of programs are most effective? For example, do cash or in-kind programs produce bigger benefits for children? Finally, do welfare programs have differential effects on different groups, and if so, why?

Cash Versus In-Kind?

The oldest and most important program providing cash benefits to single mothers with children is Aid to Families with Dependent Children (AFDC). This program has been attacked by critics who argue that participation in AFDC promotes maternal behaviors that are bad for children. Nancy Cole and I investigated this claim, using the birthweight of infants born to mothers who participated in AFDC during pregnancy as a marker for child well-being.(1) We found that AFDC mothers were indeed more likely to delay obtaining prenatal care, to smoke, to drink, and ultimately to have low-birthweight babies than other mothers. However, when we used either fixed effects or instrumental variables methods to control for unobserved as well as observed characteristics of the mothers, we found that there was no statistically significant association between birthweight and participation. Thus, AFDC does not seem to induce negative behaviors associated with low birthweight; however, conditional on income, participation in the program does not improve birth outcomes either.

I address the larger question of whether increases in income per se improve child outcomes in a study with Duncan Thomas. We show that maternal income is related significantly to children's test scores, which in turn are significant predictors of schooling attainment, even after controlling for maternal test scores.(2) However, analysis of individual maternal achievement tests shows that those skills that are most highly rewarded in the labor market are not always the same skills that are associated with improved child outcomes.

Finally, in two studies I show that although the available evidence is incomplete, it suggests that in-kind programs that target benefits directly to children have much larger measurable effects on specific outcomes than do equivalent cash transfers.(3) Thus, if the public has in mind that there are certain specific services such as basic medical care and high-quality childcare that every child should receive, then in-kind programs designed to provide those services are likely to be more cost-effective than traditional income support programs. This finding provides an economic rationale for the fact that over the past 20 years an increasing proportion of total welfare dollars has been allocated to such programs.

Still, even these programs do not always have the intended effects on all groups of children, a point that can be illustrated using recent research about two important programs: Medicaid and Head Start.

Differential Effects of Medicaid

Medicaid is the main system of public health insurance for poor women and children. In further work with Thomas,4 I use panel data that follow a single child over time, and show that children covered by Medicaid are more likely to have had any doctor visits in the past six months than other children. Moreover, the effect of being covered by Medicaid is larger than the effect of being covered by private health insurance, which probably reflects the fact that Medicaid has no copayments or deductibles. This effect is the same for black and white children. However, white children also receive more visits for illness when they are covered by Medicaid than when they are uninsured, and this is not true for blacks. Thus, equivalent coverage does not guarantee equal care.

In two studies with Jonathan Gruber, I look at the effect of becoming eligible for Medicaid on the utilization of medical care and on child health.(5) We identify the effects of Medicaid eligibility using recent, dramatic, federally mandated expansions of the Medicaid program to pregnant women and children who were not covered previously. We construct an index of the generosity of Medicaid in each state that depends only on state rules.[6] We then impute eligibility to each woman or child in our sample, and use this index as an instrument for our imputed eligibility measure.

We find that expansions of eligibility to pregnant women increased the fraction of women eligible for Medicaid from 12 percent to 43 percent. This increase was associated with an 8.5 percent decline in the infant mortality rate. However, we calculate that the early extensions of Medicaid eligibility to very poor women who were already income-eligible for other welfare programs were much more cost effective than later expansions to higher-income women. That is because the higher-income women were less likely to become covered. Hence, they did not take advantage of the provision of free preventive prenatal care. But once they arrived at the hospital to deliver, the hospital enrolled them in the Medicaid program, so that costly services received by unhealthy newborns were paid for by the program.

We use the same methodology to look at the effects of extending eligibility to additional groups of low-income children. We find that, although many newly eligible children did not take up coverage, becoming eligible for Medicaid reduced the probability that a child went without a doctor's visit in the past year, and also improved the quality of care as measured by the fraction of these visits that took place in doctor's offices rather than in hospital outpatient clinics or emergency rooms. These changes were linked to significant reductions in child mortality from internal causes, and had no effect on mortality from external causes (such as accidents).

I explore the complex relationship between formal takeup and benefits received in a study that focuses on differences between children of immigrants and children of the native born.[7] I show that recent expansions of Medicaid eligibility had negligible effects on coverage among immigrant children, but increased the utilization of basic services more for immigrants than for nonimmigrants. Moreover, many immigrants in border states appear to have dropped private health insurance coverage for their children when they became eligible for Medicaid coverage, leading to an increase in the fraction uninsured among this group.

Together, these results are consistent with evidence from other countries showing that extensions of insurance coverage alone will not eliminate socioeconomic differences in health care utilization or health.[8] They suggest that outreach programs designed to improve takeup could increase the cost-effectiveness of the Medicaid extensions to pregnant women, although at the risk of increasing the extent to which private insurance is "crowded out."[9] On the other hand, lack of information cannot explain the observed differences in patterns of takeup and utilization between black and white children, or between immigrants and nonimmigrants. Cultural explanations that posit that blacks or immigrants value medical care less than other parents also are difficult to reconcile with evidence showing that they are as likely as other parents to bring their children in for free preventive care when they become eligible for Medicaid.

The results for pregnant women indicate that it is important to analyze the effects of social insurance programs on the incentives of providers. One additional factor that may be important in explaining differential utilization of care among those eligible for Medicaid is the difference in availability of physicians willing to accept Medicaid payments. With Gruber and Michael Fischer, and using state-level data, I show that increases in Medicaid fee ratios for obstetrician/ gynecologists are associated with significant declines in infant mortality, presumably because of increases either in effective physician supply or in the quality of services provided.[10]

Medicaid fee policy also can have an impact on the availability of clinics providing some types of services. In work with Lucia Nixon and Nancy Cole, I show that while restrictions on the Medicaid funding of abortion have no direct effect on birthweights (that is, there is no evidence that pregnancies that would result in unhealthy babies are disproportionately likely to be aborted), the availability of abortion services does affect birthweight.[11] Restrictions on the Medicaid funding of abortion therefore may have an indirect effect on infant health by reducing the number of abortion providers.

Differential Effects of Head Start

Head Start is a federal-local matching grant program that aims to improve the skills of poor pre-schoolers so that they can begin schooling on an equal footing with their more advantaged peers. Unlike Medicaid, it is not an entitlement program, and only about one-third of eligible children are served. Head Start has enjoyed widespread bipartisan support over a long period, although the evidence that the program has long-term effects is inconclusive. Experimental studies that focus primarily on inner-city black children typically find an initial positive effect on children's cognitive achievement that fades out in two or three years. Supporters of the program argue that a narrow focus on cognitive test scores is inappropriate, given that Head Start is intended to affect a range of outcomes.

In work with Thomas, I find that siblings who were in Head Start have higher test scores at the end of the program than either stay-at-home siblings, or siblings who went to other preschools.[12] The effects are of the same magnitude for both black and white children, and indicate that Head Start closes one-third of the gap between these children and others. But consistent with the experimental studies, we find that the effects on black children fade out rapidly. In contrast, the effects on the test scores of white children do not fade out. Moreover, white children age 10 and above are significantly less likely to have repeated a grade if they attended Head Start, and are thus less likely to have experienced the age/grade delay that often leads to not completing high school. Both black and white children who attended Head Start were more likely to be immunized than stay-at-home siblings, although we found no effect on height-for-age, a measure of long-term nutritional status.

In related work, we find that Head Start has large and lasting effects on the test scores of Latino students.[13] The effects are greatest on the test scores for which the greatest gap exists between the average Latino student and the average non-Latino student. A closer inspection of the data reveals that these positive effects are confined to children of native-born mothers: on average, children of foreign-born mothers do not experience any changes in test scores when they attend Head Start. Even among children with foreign-born mothers, there are significant differences in the effects of Head Start, which are related to family structure and maternal test scores: children of foreign-born mothers with relatively high scores, and those in households with grandparents present, are better off at home than in Head Start. Children without grandparents present or whose mothers have relatively low scores, gain from participation in the program.

[1] J. Currie and N. Cole, "Welfare and Child Health: The Link Between AFDC Participation and Birthweight, " American Economic Review (September 1993).

[2] J. Currie and D. Thomas, "Nature Versus Nurture? The Bell Curve and Children's Cognitive Achievement," NBER Working Paper No. 5240, August 1995.

[3] J. Currie, "Welfare and the Well-Being of Children: The Relative Effectiveness of Cash Versus In-Kind Transfer," in Tax Policy and the Economy, Volume 8, J. M. Poterba, ed. Cambridge, MA: MIT Press, 1994. See also J. Currie, Welfare and Well-Being of Children, Chur, Switzerland: Harwood Academic Publishers, 1995.

[4] J. Currie and D. Thomas, "Medical Care for Children: Public Insurance, Private Insurance, and Racial Differences in Utilization, " Journal of Human Resources (Winter 1995).

[5] J. Currie and J. Gruber, "Saving Babies: The Efficacy and Cost of Recent Expansions of Medicaid Eligibility for Pregnant Women, " NBER Working Paper No. 4644, February 1994; and "Health Insurance Eligibility, Utilization of Medical Care, and Child Health," NBER Working Paper No. 5052, March 1995, and Quarterly Journal of Economics, forthcoming.

[6] For a detailed description of these expansions, see A. Yelowitz, "The Medicaid Notch, Labor Supply, and Welfare Participation: Evidence from Eligibility Expansions," Quarterly Journal of Economics, forthcoming.

[7] J. Currie, "Do Children of Immigrants Make Differential Use of Public Health Insurance?" NBER Working Paper No. 5388, December 1995.

[8] J. Currie, "Socioeconomic Status and Health: Does Universal Eligibility for Health Care Reduce the Gaps?" Scandinavian Journal of Economics 4 (1995).

[9] For a detailed discussion of crowdout, see D. Cutler and J. Gruber, "Does Public Insurance Crowd Out Private Insurance?" NBER Working Paper No. 5082, April 1995, and Quarterly Journal of Economics, forthcoming.

[10] J. Currie, M Fischer, and J. Gruber, "Physician Payments and Infant Mortality: Evidence from Medicaid Fee Policy," NBER Working Paper No. 4930, November 1994, and American Economic Review (May 1995).

[11] J. Currie, L. Nixon, and N. Cole, "Restrictions on Medicaid Funding of Abortion: Effects on Pregnancy Resolutions and Birthweight," Journal of Human Resources, forthcoming.

[12] J. Currie and D. Thomas, "Does Head Start Make a Difference?" American Economic Review (June 1995).

[13] J. Currie and D. Thomas, "Head Start and Cognition Among Latino Children: Interactions with Language and Culture, " mimeo, February 1996.

Janet Currie, Janet Currie is a professor of economics at UCLA, and a research associate in the NBER's Program in Health Economics. A brief biographical story about her appears in the Profiles Section of this issue.