Barbara A. Haley
U.S. Department of Housing and Urban Development Office of Policy
Development and Research Program Monitoring and Research Division

This special issue of Cityscape reports the findings of research on households who receive housing assistance from the U.S. Department of Housing and Urban Development (HUD).

Assisted housing is found in every metropolitan area and in every state: 18 percent are in rural and nonmetropolitan areas, 17 percent are in suburban areas, and 56 percent are in central cities.¹

Approximately 1.1 million households live in public housing units managed by some 3,200 public housing authorities. Another 1.4 million households live in HUD-subsidized, privately owned projects, including Section 8 and other multifamily-assisted programs. Approximately 1.9 million households receive assistance under the Housing Choice Voucher (HCV) Program, formerly known as tenant-based Section 8, in which households are expected to find individual housing units owned by private landlords.²

As of 2004, 31 percent of households receiving housing assistance were headed by elderly people,³ 20 percent were headed by people who were disabled but not elderly, and 39 percent were headed by people who had children. A small percentage (about 10 percent) of housing-assisted households did not have elderly or disabled household heads and did not have children.

Housing assistance programs serve large numbers of vulnerable people. Policymakers and the public want to know more about how these programs perform, and much can be learned from HUD's administrative records, the New York City Housing Vacancy Survey, the Survey of Income and Program Participation (SIPP), and qualitative interviews with participants in the Gautreaux Two Housing Mobility Study. HUD's Office of Policy Development and Research sponsored all but two of the research projects reported here.

The authors bring a variety of theoretical and methodological tools to the research questions posed. One set of questions in this issue relates to program dynamics: To what extent are recipients' rents an acceptable burden on their incomes? To what extent do HCV Program recipients find rental housing that is privately owned in better neighborhoods than where they formerly lived? How does housing assistance relate to household composition?

Another set of questions addresses housing assistance tenure: How long do households use this assistance? What kind of household tends to have the longest tenure? What circumstances are predictive of leaving these programs?

A third set of papers presents evidence regarding access to jobs with decent wages and the question of whether different programs are associated with different employment outcomes.

Articles in This Issue

Program Dynamics

Kirk McClure reports that, from 2000 to 2002, the program witnessed significant reductions in the incidence of high housing-cost burden. Households paying more than 40 percent of income for housing dropped about 6 percentage points, from 22.5 percent to 16.6 percent. About 38 percent of all households in the program paid more than 31 percent of income on housing in 2002, down from 47 percent just 2 years earlier. Suffering from a high housing-cost burden appears to result from the household having very low income rather than from market conditions or decisions by program administrators. It appears that this problem results from some households having very little or no income at the time their housing consumption was recorded.

Judith D. Feins and Rhiannon Patterson examine the geographic mobility of families with children that entered the HCV Program between 1995 and 2002. Using a specially constructed longitudinal dataset developed from HUD administrative records, they analyze the residential moves made by these families to see whether moves within the voucher program-particularly moves after the initial lease-up-are associated with improvements in the neighborhoods where the families live and/or with increases in their economic self-sufficiency. They find that subsequent to program entry (that is, after the moves to lease up), there is a small but consistent tendency for families making later moves to choose slightly better neighborhoods. The data show reductions across a number of indicators of concentrated poverty and improvements across a number of neighborhood opportunity indicators for households that moved.

Lance Freeman explores the relationship between housing assistance and household composition using data from the New York City Housing Vacancy Survey. The results show that, for New York City, household composition is related to the receipt of housing assistance. In particular, married and cohabiting partners are less likely to be recipients of housing assistance.

Duration of Housing Assistance Receipt

Brent W. Ambrose finds that individual characteristics and economic conditions play an important role in determining assisted housing tenure. The mean census tract poverty rate for households receiving housing assistance is 22 percent, and, as the proportion of the population that does not speak the majority language increases, the less likely the household is to leave assisted housing. Households headed by an elderly or disabled individual are significantly less likely to leave assisted housing programs. Households headed by teenagers in public housing, receiving tenant-based vouchers, or in multifamily housing are more likely to exit than other households. A one-point increase in household income relative to area median income greatly increases the odds that a household will leave a tenant-based assisted housing unit or a public housing unit. Households in public housing with income from wage or salary have a significantly higher probability of leaving public housing, but this was the case only for public housing, not among households residing in multifamily or tenant-based programs. In addition, households are more likely to leave assisted housing during periods of economic expansion and are less likely to leave during periods of economic uncertainty.

Edgar O. Olsen, Scott E. Davis, and Paul E. Carrillo use administrative data of families who participated in the HCV Program between 1995 and 2002, combined with data from other sources, to estimate the differences in attrition rates from the program. They find that large decreases in the program's payment standard and increases in the tenant contribution to rent would have small effects on program attrition. They also find that age and disability are by far the most important influences on the likelihood that the family will exit the tenant-based voucher program. Disabled families are about 37 percent less likely to exit, and elderly families are around 23 percent less likely to exit each year than are otherwise similar families. Differences in attrition rates based on other family characteristics are much smaller.

Lance Freeman uses event history methods to describe and explain the dynamics of housing assistance exits between 1995 and 2002. His results show that, except for the first year, the likelihood of exiting housing assistance is greatest in the earliest years. The probability of a household receiving housing assistance beyond 5 years is 58 percent and beyond 10 years is 36 percent. Being White, younger, not disabled, and/or not having children are personal characteristics associated with shorter spells of housing assistance receipt. These results suggest that life-cycle factors that predict residential mobility in general play an important role in determining exits from housing assistance. In addition, a higher vacancy rate in the local housing market and the availability of housing alternatives for low-income minorities also appear to be important determinants of housing assistance exits. Compared to families who receive housing assistance in the Northeast, those residing elsewhere are more likely to exit assisted housing in a given year.

Housing Assistance and Employment

Peter A. Tatian and Christopher Snow track income and earnings for households who received assistance for at least 8 consecutive years, from 1995 to 2002. Income and earnings during that period rose by 34.1 and 93.1 percent, respectively. They find that income trajectories are highest for households that are non-Hispanic Black or Hispanic; have a household head aged 18 to 25 years; have a single working-age adult; have children; are neither disabled nor elderly; have a youngest child less than 3 years old; have no spouse or cohead present; have an income level in the lowest deciles; receive welfare; do not receive Supplemental Security Income, Social Security, or pension income; were homeless at time of admission to housing assistance; live in high-poverty census tracts; and/or live in the central city or outside a metropolitan area. Steepest increases in income are for households in project-based Section 8 units and the flattest for vouchers and other site-based programs, indicating that the income trends are explainable by differences in household characteristics between the programs. After controlling for household characteristics, the odds of being employed for this group of long-term program recipients are essentially the same for residents of Section 8 site-based, voucher-assisted, and public housing.

Edgar O. Olsen, Catherine A. Tyler, Jonathan W. King, and Paul E. Carrillo examine the relationship between different types of housing assistance and earnings and employment. They use HUD's administrative data for nonelderly, nondisabled households who received rental assistance between 1995 and 2002, combined with data from other sources. The results indicate that each broad type of housing assistance is associated with receipt of lower wages than received by households that are not assisted, but the extent of the difference depends on which program is under consideration. Participation in the Family Self-Sufficiency Program, an initiative within the public housing and housing voucher programs to promote self-sufficiency, is associated with significantly higher wages than those received by assisted households who are not part of this initiative.

Scott Susin merged the 1996 panel of the longitudinal SIPP with HUD's administrative data. The merged data enabled him to accurately identify recipients of housing assistance and construct a valid comparison group. The 4 years of the SIPP panel coincided with the 1990s economic boom. Poverty and receipt of welfare decrease for households in both subsidized housing and the comparison group. Households receiving housing assistance and the comparison group all show strong gains in employment, earnings, and income. Families in public housing, however, have substantially lower incomes than their comparison group, and poverty rates are 8 percentage points higher. For recipients of vouchers and project-based subsidies, the differences are smaller, with none statistically significant. The patterns for earnings from employment are similar to that for income. Families in public housing and households with project-based subsidies have lower earnings than the comparison group. No statistically significant differences occur for voucher recipients. Public housing residents live in census tracts with poverty rates averaging 8.8 percentage points higher than tracts with the comparison group, so differences between these groups may be partly due to neighborhood effects.

Joanna M. Reed, Jennifer Pashup, and Emily K. Snell conducted indepth interviews of voucher holders who participated in the Gautreaux Two Housing Mobility Study. Their respondents are women who used vouchers to move out of segregated, highly concentrated poverty neighborhoods into more affluent areas. The researchers connect life priorities dictated by motherhood and membership in a contingent labor market to labor force participation. They compare movers' and nonmovers' labor market experiences before they moved, finding similar employment experiences and histories of holding low-wage service jobs, interrupted by periods of welfare receipt. The primary obstacles to working are childcare; illness and health problems, including pregnancy; transportation; and layoffs from temporary jobs. Respondents have positive attitudes toward employment. Moving to more affluent neighborhoods has little or no impact on the employment situations of most study participants.

Conclusion

Changes in the legislation regulating the federal housing assistance programs occur regularly
and not always in an atmosphere of clarity and understanding. The Office of Policy Development and Research is pleased to present these papers to the public in the belief that they can contribute to informed debate about programs that serve 4.4 million households.

Notes

1. The author thanks Mark Perdue for his assistance in producing these estimates. The author also thanks Robert W. Gray, Hal Holzman, and Mark Shroder for helpful comments in response to an earlier draft.

2. For more information about these programs, see the following websites:
http://www.hud.gov/offices/pih/programs/ph/index.cfm.
http://www.hud.gov/offices/pih/programs/hcv/project.cfm.
http://www.hud.gov/offices/pih/programs/hcv/tenant.cfm.

3. Elderly is defined here as 62 years and older.