New Jersey's experiment in welfare reform

Public Interest, Fall, 1996 by Ted G. Goertzel, Gary S. Young

In the wake of the welfare-reform plan passed by its legislature in January of 1992, the state of New Jersey has seen a significant reduction in birth rates and welfare dependency in its poorest cities. And, paradoxically, the new evidence of these changes, from two independent and reliable sources, suggests that the apparent mechanism of change was quite different from that anticipated by the plan's architects.

The New Jersey reform was controversial because of the so-called "family cap" provision that denies an additional cash benefit to a woman who conceives an additional child while receiving Aid to Families with Dependent Children (AFDC) benefits. Welfare advocates waged war on this provision, attacking the "myth" that welfare recipients have additional children simply to receive increases in their benefits. They fought, and lost, a bitter political and legal struggle with New Jersey Assemblyman (now state Senator) Wayne Bryant, the architect and chief proponent of the reform.

Bryant, an African-American liberal Democrat from Camden, the state's most welfare-dependent city, wanted the "family cap" provision as a way to send welfare recipients the message that welfare must be temporary, not a way of life. He took steps to ensure that word of this change would spread quickly through inner-city communities, in the hope that it would encourage welfare mothers to take advantage of other features of the welfare reform which were designed to strengthen families and to facilitate their entry into mainstream society.

Thus, contrary to claims of opponents, the New Jersey reform was not an experiment designed to test the emotionally charged hypothesis that women have babies just to get an additional $64 per month in welfare benefits. The "family cap" was part of a comprehensive package of reforms intended to change the culture of inner-city communities. The new rules required recipients to meet with social workers to formulate a "family plan" to improve their situation through education, work experience, or marriage. Rules were changed to encourage two-parent families, and additional educational and child-care resources were offered.

Data now available show that, since the implementation of the reform, births to welfare mothers and welfare caseloads have declined significantly. In addition, state officials have found no increase in abortions or in reports of suffering by poor children. And significant numbers of welfare mothers are going to college, an option not available under the previous policy.

Confusing early findings

Much of the confusion about the New Jersey reform was because of a reliance on data from an experimental research project that was required by the federal government as part of the waiver process whereby states are allowed to opt out of AFDC rules. The experiment was rigorous and scientific in form and intent, but it was unable to capture the changes that the New Jersey reform actually caused. A group of women were exempted from the provisions of the new law. This group, labeled the "control" group, was contrasted to an "experimental" group that followed the same rules as all other New Jersey recipients.

Results of this experiment were eagerly anticipated as a test of whether or not the "family cap" worked. In response to congressional requests, an early report was prepared by June O'Neill, then a professor of economics at the City University of New York, which found a slightly lower birth rate in the "experimental" group than in the "control" group. A later report, by Michael Camasso, a professor of social work at Rutgers University, found no statistically significant difference between the two groups.

But data from the experiment conflicted with the monthly records kept by the New Jersey Division of Family Development which showed that birth rates and case loads were dropping. Supporters of the New Jersey reform used these data to argue that the reform had been a success. Opponents insisted that, even though birth rates had gone down, the reform had not "really" been successful because there was little difference between the "experimental" and the "control" groups. Each side to the debate selected the data that supported its point of view, while discounting the data that did not. As one New Jersey official observed: "Some are yelling, 'Hooray, it works' and some are yelling 'Hooray, it doesn't work.' Please recognize that it is too soon to recognize a trend."

While opponents of the reform focused their attention on the lack of a statistically significant difference between the "experimental" and "control" groups, supporters observed that, in both the O'Neill and Camasso studies, birth rates went down sharply in both the "experimental" and the "control" groups. There was also a much smaller decline in births among the general population of New Jersey welfare mothers. The decline in both the "experimental" and "control" groups may be due to the well-known "Hawthorne effect," in which participants in an experiment change their behavior simply because they know researchers are watching them. Or it may be because births are more common among women who have recently gone on welfare, and none of these women were included in the "experimental" or "control" groups in the second year.