It's a simple idea that's had a sweeping impact on welfare reform: Remove the financial incentive and welfare mothers won't have more children.
It began in 1992 when New Jersey enacted a "family cap" law barring additional payments to welfare mothers who have children 10 or more months after they apply for welfare.
Based on reports of the law's success, several states -- including Maryland -- implemented similar laws. The family cap also became an important part of the congressional push for welfare reform.
But a recent Rutgers University study indicates that New Jersey's family cap has had no impact on welfare mothers.
Michael C. Laracy, a senior program associate at the Annie E. Casey Foundation, of Baltimore, explains how the push for family cap legislation is being fueled by erroneous information. Before joining the foundation in August 1994, Mr. Laracy was assistant commissioner for policy, planning and program evaluation in the New Jersey Department of Human Services. The Annie E. Casey Foundation is the nation's largest philanthropy for disadvantaged children.
Q: Do the results of the Rutgers evaluation mean that the earlier claims of impacts were invalid? Why such a great difference?
A: Yes, the Rutgers findings render all previous claims invalid. Specifically, they show that the pronouncement by former Governor [James] Florio in November 1993 that the New Jersey child exclusion [family cap] law was an "obvious success" because two months of data suggested a 15 percent drop in birth rates was unwarranted and without merit.
Likewise, the Rutgers study refutes the testimony drafted by June O'Neill [then an economics professor at Baruch College and now director of the Congressional Budget Office] late last year -- which claimed to find a drop of between 19 and 29 percent in birth rates. Probable reasons for the erroneous findings in her analysis lie in that it used still-incomplete data, it did not use a long enough period of measurement and it disregarded the lags in reporting births that were known to be a problem in the sample.
Q: Did other states or the federal government make any significant public policy decisions based on the earlier erroneous analyses and claims about impacts in New Jersey?
A: Absolutely. Unfortunately, at least 10 states have enacted family cap laws, applied for federal waivers, or introduced legislation since the November 1993 Florio announcement. In almost every instance, the alleged results from the New Jersey law were cited in support. Moreover, in its welfare reform bill H.R. 4, the House of Representatives mandated that all states adopt such provisions, with many supporters citing the dramatic 29 percent figure. And proponents in the Senate continue to cite the now-discredited data.
Q: The Rutgers analysis shows no difference between the birth rates for the women in the treatment group, who were affected by the family cap, compared to an equal number of women in the control group who were exempt from the law's sanctions. What does this mean? What is its significance?
A: The Rutgers analysis reveals that the birth rates for both the treatment and control groups dropped virtually identical amounts from before the law took effect to afterward. The lack of any significant difference between the two groups shows that, at least thus far, the law has had no discernible impact -- no effect -- on the child-bearing practices of the women subjected to its penalties and incentives. The principal significance of this finding is that it supersedes and strongly refutes earlier, unofficial pronouncements that the law was reducing births to welfare mothers on the order of 15, 19, and even 29 percent.
Based on the Rutgers report, supporters of family cap laws -- in New Jersey, in the U.S. Congress, and in state capitals across the country -- no longer have any empirical evidence whatsoever for their claims that such laws will discourage out-of-wedlock births.
Q: Are the Rutgers findings surprising?
A: Not really. Over the last decade, a series of studies have shown very little relationship between welfare grant levels and birth rates of low-income women. These analyses generally compared the birth rates for welfare mothers in a range of states among which the monthly cash grant levels varied widely. Although the findings were not uniform, they collectively show only modest correlations between the number of children in a family and the states' benefit levels. For white and Hispanic families there tended to be a slight positive correlation; for black families, no relationship whatsoever. Thus, the Rutgers findings are quite consistent with the earlier scientific work in this area, all of which showed that reproductive decisions of low-income women are not very sensitive to welfare grants.
The Rutgers results are only surprising in the extent to which they so decisively refute supporters' claims over the last two years that the New Jersey child exclusion law had dramatically reduced birth rates among women on welfare.
Q: Is the Rutgers evaluation report definitive and final? Does it "prove" that the law will have no effect in the future?
A: The Rutgers study is not definitive, nor is it final; it is only the first of several analyses of the New Jersey package of welfare reform laws. The series of evaluations were required by the federal government as part of the waivers New Jersey was granted to implement the child exclusion law and FDP's [the Family Development Program's] other provisions. Subsequent reports will revisit the question of whether the birth rates change at later points in time. However, the findings released now are the first official and authoritative analyses of the law's effects; previous data and analyses were either unscientific, badly flawed or based on incomplete sample data. They were all
premature and unreliable, the Rutgers data documents.
Because the current analysis is based on impact data that most researchers would still consider early or preliminary, no one should conclude that it is the final word. It is certainly conceivable that future evaluations, which will incorporate subsequent years of data, will show slightly different results. However, the fact that this evaluation shows no impacts thus far, strongly suggests that we should not expect dramatic results in the out years, certainly nothing on the order of 15 percent, 19 percent or 29 percent.
Q: Does Rutgers' evaluation say anything about possible negative consequences of the child exclusion law? What about the children who were born to families on AFDC and were excluded from their mothers' benefits?
A: Unfortunately, this initial evaluation is very narrowly focused and does not address the vital question of how the law affects children who are born and are penalized by being excluded from their mothers' cash grants. Considerable research has shown strong correlations between child poverty and short-term negative outcomes for children. That is, the poorer the child's family, the worse off the child tends to be. One inherent effect of the child exclusion law is that children who are born will be in deeper poverty than they would otherwise have been because their mothers' grants are not increased on their behalf.
A reasonable hypothesis to test, therefore, is that the law might cause increased negative outcomes for infants penalized by it. However, for political and budgetary reasons, the Florio administration had decided to minimize the evaluation of "social" impacts -- such as abortions, homelessness, hunger, incidence of child abuse or neglect, low birth weight and/or failure to thrive among infants, maternal or child stress, criminal behavior, or other negative child or maternal outcomes -- that plausibly might unintended consequences of family cap.
Q: What is the significance of the fact that the birth rates for both the treatment and control groups dropped from the "pre" period to the post period? Is it possible that the law discouraged births in the control group -- even though they were exempt from the provisions of the child exclusion law -- as well as in the treatment group?
A: This is very hard to assess. Some proponents of child exclusion laws have already claimed -- without any basis whatsoever -- that the law actually caused the drop in the control group, perhaps because the women mistakenly thought it would affect them despite the fact that they were exempt. However, Rutgers, NJDHS [New Jersey Department of Human Services], and the county welfare agencies had instituted a series of procedures to ensure that all members of the control group understood that they would not be affected by the provisions; that they would be treated as if the laws were never enacted.
Q: What does the Rutgers study mean with regard to the recent announcement by NJDHS that there was an increase in the number of Medicaid-financed abortions among New Jersey welfare mothers since the law was implemented? If the law is not causing any reductions in the birth rate, how could it be causing an increase in abortions?
A: This is another difficult question; one, thus far, with no certain answer. Many opponents of abortion have long worried that child exclusion laws would force welfare mothers to terminate pregnancies they couldn't afford to bring to term because of the laws' penalties. These critics have correctly pointed out how difficult it is to change sexual behavior or contraceptive practice, at least in the short and mid-term, and that women who do become pregnant would opt to take advantage of the accessible, free, and safe abortions funded through New Jersey's Medicaid program. They thus fear that the reported drop in births of AFDC mothers was attributable to increased abortions, rather than abstinence or better birth control.
However, the 4 percent increase in the abortion rate occurred over a relatively short period; and -- like birth rates -- unexplained variations in the rates of abortions are normal. Moreover, while this increase in Medicaid abortions appears to have reversed a downward trend over several years prior to the enactment of the child exclusion law, rigorous analysis is difficult due to a switch in the systems NJDHS used to reimburse Medicaid procedures. Finally, they are simple "pre and post" comparisons for the whole AFDC caseload; they are not from the treatment and control group samples.
At this point, the data on Medicaid abortions is comparable to the initial birth rate data that precipitated former Governor Florio's claims in late 1993: suggestive, but still unreliable and subject to major revisions.
Q: How should policy-makers in Washington, D.C., and in state capitals interpret these preliminary Rutgers findings? Should the results of the Rutgers evaluation affect their decisions about enacting child exclusion laws.?
A: Responsible public officials should pay serious attention to the Rutgers evaluation. It starkly shows the danger of rushing to judgment on complicated and controversial public policy issues based on inadequate information. (It should be noted that New Jersey's current governor, Christie Whitman, has consistently reserved judgment about the merits, or lack thereof, of the state's child exclusion law. She has said she wants to await the results of the overall Rutgers evaluation.)
Proponents of family cap laws no longer have any empirical basis for their views. They can argue their case only on normative grounds, based on their values and beliefs. And given the negative consequences of such laws for the poor children, normative arguments are tenuous, at best. Policy-makers would be well advised to exercise caution in making decisions about adopting versions of New Jersey's law: As was the case a year ago -- the jury is still out.