Mark Shroder
U.S. Department of Housing and Urban Development
Office of Policy Development and Research.

The Personal Responsibility and Work Opportunity Reconciliation Act (PRWORA) of 1996 significantly altered the U.S. system of public assistance for low-income families. The act replaced the best-known welfare program, Aid to Families with Dependent Children (AFDC), with a block grant to the states, eliminating entitlement to benefits, imposing work-related activity requirements as a condition of benefits, and imposing time limits for most recipients of not more than 5 years cumulative on the receipt of benefits.

In 1996 approximately 1.1 million households that received AFDC also lived in public housing or benefited from tenant-based or project-based Section 8 housing subsidies. PRWORA therefore had significant effects on the people and neighborhoods served by the U.S. Department of Housing and Urban Development (HUD). In some cases, it was clear that the law would provide stronger incentives for families to reduce their reliance on federal subsidies and become self-sufficient. Other families, it was feared, would face additional financial hardship.

Welfare reform has produced numerous surprises, both positive and negative, for poor families, government agencies, and researchers. We know that forecasts of huge increases in housing subsidy costs as families were forced off the welfare rolls were largely incorrect. Most families on welfare at least maintained their incomes—sometimes by working in an expanding economy, sometimes by finding other sources of support, and sometimes by being allowed under various exceptions to remain on or return to the rolls. As of this writing, the Temporary Assistance for Needy Families (TANF) program block grant is assisting somewhat more than half the number of public housing and Section 8 families assisted in 1996.

Whether on or off the rolls, people who were poor in 1996 are too often still poor, and children continue to be born into poverty in the United States. If poor, they will generally experience either some form of housing assistance or some form of worst-case need—inadequate shelter or unaffordable rent—for that assistance. Congress must soon reauthorize welfare reform, building on those changes that have acquired widespread support and continuing to reshape national programs to benefit the poor. Researchers, administrators, and policymakers should know how housing need and housing assistance affect the outcomes of families under welfare reform. This issue of Cityscape presents the latest evidence we have.

The articles in this issue, with two exceptions, were commissioned by HUD’s Office of Policy Development and Research (PD&R) in two grant competitions in 1998. One set of small grants funded research on the interaction of welfare reform with housing programs—the articles presented in the sections titled "Transition to Work" and "Homeless Families." A second set of grants was awarded in conjunction with a competition by the Office of the Assistant Secretary for Planning and Evaluation (ASPE) of the U.S. Department of Health and Human Services for surveys of outcomes among families leaving TANF. PD&R funded three ASPE grantees to compare the outcomes of TANF leavers who had and did not have housing assistance, and these grants were the genesis of three of the articles in the section titled "Life After Welfare."

The two remaining articles arose out of special circumstances. A Hudson Institute survey of Milwaukee families so closely resembled the methods and findings of the HUD grantees in Los Angeles, Cleveland and the surrounding Cuyahoga County, and three northern California counties that we asked its lead author to contribute an abridged version to the "Life After Welfare" section. The short concluding note, a response by PD&R staff to information requests from members of the Millennial Housing Commission, chartered by Congress to recommend new directions in housing policy, is related in theme but different in methods from the rest of this issue. The authors describe patterns of earnings among HUD-assisted tenants and length of stay in housing assistance.

Articles in This Issue

Transition to Work

The "Transition to Work" section comprises three articles that report careful statistical investigations of the relationships between housing assistance and the economic outcomes of welfare families, holding numerous other variables constant. Neil Bania, Claudia Coulton, and Laura Leete examine the geography of family success and failure in the Cleveland area. After controlling for personal and neighborhood characteristics, they find no association between employment of welfare leavers and residence in subsidized private housing, but they do find a positive association between employment and public housing. With the same controls, they find rather weak negative associations between earnings and the receipt of vouchers (a 20-percent reduction but only at the 90-percent confidence level) and between earnings and project-based assistance (not significant at conventional confidence levels). The magnitude of the public housing coefficient is sensitive to inclusion of the neighborhood poverty rate in controls. Vouchers also, confusingly, decreased welfare recidivism substantially (but at the same weak confidence level).

As in a Sherlock Holmes story, the most important clue in the findings of Bania and colleagues is the dog that does not bark: Distance to entry-level jobs is not a significant factor in family outcomes. They found that neighborhood poverty rate is negatively related to employment, earnings, and earnings growth and positively related to welfare recidivism and welfare spells. However, they found this aspect of neighborhood powerful after proving that (counter to their own hypothesis) an intuitively appealing aspect of location—job accessibility—had no effect on the same outcomes. After determining the worksites with entry-level jobs throughout Cuyahoga County, they found that the number of jobs accessible by bus within 40 minutes and by car within 20 minutes had no effect on employment.

Gregg G. Van Ryzin, Robert Kaestner, and Thomas Main report that housing assistance appears to have neither helped nor hindered the transition to work of a large sample of welfare-to-work participants in New York City. They consider the lack of clear program effects somewhat surprising, especially given the high cost of housing in New York City. They found the city’s own housing program had a substantively large but only marginally significant positive effect on self-sufficiency; their qualitative probing did not suggest any basis for replication in that program because its administrators do not see the assisted tenants as their clients—"our clients are the buildings." Like Bania, Coulton, and Leete, they found that neighborhood poverty is associated with welfare recidivism; they found a similar association between recidivism and neighborhood crime rates.

Mary Corcoran and Colleen Heflin examined employment and earnings among a random sample of welfare mothers in a Michigan metropolitan area. They found no association between housing assistance and either employment or earnings, controlling for barriers to work among respondents. In fact, most women in their sample were working at followup, averaging more than 35 hours per week. Although public housing residents were more likely to lack access to a car, differences in other barriers to work between HUD-assisted and unassisted populations were largely insignificant. The evidence does not support their initial hypothesis that assisted renters face more barriers than the average welfare recipient.

Life After Welfare

The "Life After Welfare" section is composed of four articles that compare various outcomes among assisted and unassisted welfare families. The first three, by the ASPE grantees, are consistent in approach despite the diversity in the population of welfare leavers surveyed. The final report from Milwaukee encompasses current TANF recipients, but it tends to confirm the patterns in the other studies.

These four articles independently report that their samples show more "other" adults in the households among unassisted welfare leavers in Massachusetts, Los Angeles County, and three Northern California counties, as well as among current and former welfare recipients in Milwaukee, than among their HUD-assisted counterparts. Because it is the policy of George W. Bush’s administration to reform the means-tested programs to support the institution of marriage, this finding is of special interest. However, unassisted leavers are more likely to be living in crowded conditions with extended families or unrelated adults. The articles generally report that assisted respondents’ own earnings are equal to or greater than those of the unassisted; but largely as a consequence of the other adults, the articles tend to report either no difference in material hardship or higher household incomes and less poverty among the unassisted.

The consistency of this association between assistance and household composition is striking. However, the reader is cautioned that the assisted are statistically different from the unassisted on a number of other demographic variables and that both observed and unobserved differences in the two groups may account for this pattern.

Nandita Verma and Rick Hendra report on the post-TANF experience of families in Los Angeles County, approximately 9 percent of whom were assisted by HUD, mostly with vouchers. (This low level of HUD assistance is typical of California but not of the United States as a whole. The welfare population of Los Angeles County, however, exceeds that of many states and is of independent interest.) Assisted households are more likely to be from minority groups and to be older than unassisted households. Earnings and employment after welfare are somewhat higher among assisted households, although they are low for both groups. Assisted households are more likely to return to welfare, more likely to be living in poverty, and more likely to suffer food insecurity. Conversely, assisted households are less likely to show signs of housing hardship or unmet medical need. It is of particular concern that families receiving vouchers show the least satisfaction with the neighborhoods in which they live, when in principle they should evince the most.

David C. Mancuso, Charles J. Lieberman, Vanessa L. Lindler, and Anne Moses compare assisted and unassisted TANF leavers in three affluent northern California counties (San Mateo, Santa Clara, and Santa Cruz). Nineteen percent of the sample was assisted, principally with vouchers. As in Los Angeles, assisted leavers were more often minorities, had longer average welfare histories, and tended to have more children and older heads of household. At the 18-month followup, assisted heads of household had a substantially higher employment rate than the unassisted, but because they were less likely to live with other adults, their households had lower total earnings and family income was more likely to fall below the federal poverty level. Assisted leavers experienced a much lower incidence of worst-case housing problems, and the authors found that the receipt of housing assistance is not related either positively or negatively to a family's probability of suffering bad outcomes after leaving TANF.

The families in Gloria Nagle's Massachusetts study are much more likely to be assisted than their California counterparts. In fact most are assisted, partly because of the greater frequency of public housing in Massachusetts and partly because of the state's own housing programs. Assisted families were much more likely to be from minority groups. Assisted household heads were more likely to hold jobs, and those jobs were more likely to offer paid sick and vacation leave. These advantages were significantly offset, however, by the absence of income from other adults. Assisted families had much lower rents but were nevertheless more likely to have had the utilities turned off for nonpayment.

Rebecca J. Swartz summarizes the experience of a sample of Milwaukee mothers who were on AFDC in the month that Wisconsin implemented its welfare reform; about a quarter of them reported receiving housing assistance. Respondents with housing assistance had lower incomes than those without assistance, largely because they had lower average earnings and were less likely to have other adults contributing to the household. According to other indicators, however, respondents with housing assistance reported having a higher standard of living than their unassisted counterparts.

Homeless Families

The existence of homelessness in the United States has engendered much anguish but not much clarity. The simplest facts are in dispute. The articles in the "Homeless Families" section seek to answer two basic questions about homeless families with children:

Dennis P. Culhane, Stephen R. Poulin, Lorlene M. Hoyt, and Stephen Metraux tracked admissions to Philadelphia homeless services by families with children. Their time-series analysis clearly shows that increases in rent and unemployment have driven up the number of homeless families in Philadelphia, but welfare reform has had no measurable effect.

Donna Haig Friedman, Tatjana Meschede, and Michelle Hayes report on a survey of families leaving a transitional shelter program with intensive services in Boston. They found that 6 to 12 months after leaving the shelter with a voucher, 90 percent of these families had retained the subsidy. Their findings are a welcome suggestion that housing instability is a solvable problem for the most disadvantaged children in the nation.

The issue closes with a short descriptive article by Jeff Lubell, Mark Shroder, and Barry Steffen on work participation and duration of assistance among HUD-assisted tenants.

Conclusion

This issue of Cityscape presents findings on the joint outcomes of welfare reform and housing assistance. The successes of welfare reform have not included the elimination of child poverty. The public policy challenge to relieve human distress and promote self-sufficient families remains. U.S. housing and welfare laws have changed and will change again, and legislators can benefit from the findings of the social sciences. The articles in this issue may help to clarify the choices before us.