Tracking Investment: Trends in Housing Vouchers and FHA Mortgage Access

Stephanie Hawke, Economist, Housing Finance Analysis Division, Office of Policy Development and Research and Mariya Shcheglovitova, Extension Assistant Professor, University of Vermont

This series examines HUD's homeownership and rental assistance programs to:  

1. Compare household-level demographics among program participants (part 1 of this series),
2. Compare neighborhood-level data to understand the geographic distribution of households participating in these programs (this article), and  
3. Identify key household- and neighborhood-level drivers of program participation disparities (part 3 of this series, forthcoming).

The first article in this series describes the populations receiving tenant-based housing choice vouchers (TBVs) and Federal Housing Administration (FHA) mortgage insurance. That analysis finds that most recipients of FHA-backed mortgages are White homeowners, and most TBV program participants are Black renters. We also observe differences in the age distribution of federally assisted homeowners and renters, finding that most TBV program participants are over age 45 (58.5%), whereas most recipients of FHA-backed mortgages are under age 45 (60.3%). Although the data on the age of program participants indicate that FHA loans could be a pathway to homeownership for younger generations, persisting racial disparities in access to homeownership may reinforce the homeownership gap between Black and White households.

This article examines where households participating in federal rental and homeownership assistance programs are living. Location patterns have been a significant focus for researchers studying the Housing Choice Voucher (HCV) program. As of 2024, HUD has published three location reports (in 2003, 2015, and 2024) examining whether households receiving HCVs are located in "higher opportunity" neighborhoods, defined as lower-poverty census tracts. The latest report, which examined program data from 2010 to 2020, found no meaningful change in neighborhood poverty for program participants. Studies examining the intersection of race and income in the HCV program, however, indicate that Black and Hispanic voucher holders tend to live in neighborhoods with higher poverty rates than do White voucher holders. In addition, HUD's 2003 voucher location report stated that White voucher recipients were more likely to live outside of central cities than Black or Hispanic voucher recipients, indicating that the race and ethnicity of program participants may be determining factors for where they live.

The location of households receiving FHA-backed mortgages has received relatively little research attention. One notable analysis, however, examined the geographies of FHA lending during the 2008 U.S. mortgage crisis, a time when FHA mortgage insurance surged to stabilize local housing markets. This study found that in 2008, loans in majority-Black ZIP Codes were more likely to be FHA loans, suggesting that FHA played an outsized role in post-crisis homeownership in majority-Black neighborhoods. The study did not take into account information on the borrower's race and examined data only in metropolitan areas — a decision that undoubtedly influenced the results, as we show below.

In this article, we use census tract economic and demographic data to describe the neighborhoods where TBV and FHA program participants live. We begin with a brief history of how FHA loans financed suburban development and consider how these policies and practices shape the location patterns of FHA-backed mortgage recipients and TBV program participants. We then discuss the role of TBVs in facilitating the relocation of supported households to higher-opportunity neighborhoods. We conclude by discussing our place-based analyses, showing that geographic dispersion patterns continue for households participating in federal rental and homeownership assistance programs. We find that location trends for federally assisted homeowners and renters largely mirror national patterns of housing tenure, resulting in FHA-backed mortgage recipients living in wealthier, Whiter suburbs where homeownership rates are high and TBV recipients living in poorer, more racially and ethnically diverse urban areas where rental households are more prevalent.

Brief History of FHA and HCV: Suburban Development

Extensive research exists on the role of federal agencies in shaping the geographies of housing segregation. Much of this research has focused on the role of the Home Owners' Loan Corporation (HOLC) and its Residential Security maps, commonly referred to as redlining maps. Recent scholarship, however, has also has focused on FHA's Neighborhood Rating maps, which similarly assigned neighborhoods grades to indicate their value and marketability. FHA map grades were tied explicitly to neighborhood conditions and the race of neighborhood residents, with the highest-graded areas associated with White residency and the lowest-graded areas associated with Black residency. HOLC and FHA maps documented geographies of racial segregation and reinforced this segregation by directing the flow of capital and opportunity to White communities.

The patterns and recommendations in FHA Neighborhood Ratings maps were explicitly described and reinforced in the FHA's Underwriting Manual, which standardized methods for loan determination, including property appraisals. At the time FHA developed its appraisal standards, housing in cities was substantially older than housing in the suburbs, and these appraisal standards awarded higher valuations to new housing, which was disproportionately located in suburban areas. They also provided suburban developers and homebuyers with a higher share of FHA-insured capital, encouraging developers to expand the development of suburbs and homebuyers to locate in the suburbs. Nearly one-fourth of new houses developed in the United States from the 1940s to the 1960s received an FHA or U.S. Department of Veterans Affairs (VA) loan, reaching a peak of 40.7 percent in 1955.

The HCV program saw tremendous growth in the 1990s as federal initiatives to deconcentrate the residents of distressed public housing spurred HUD to reorient its programs. Because public housing had been sited predominantly in high-poverty urban neighborhoods targeted for urban renewal programs, policy officials, academics, and advocates concerned about the negative consequences of concentrated poverty saw vouchers as an ideal mechanism for increasing mobility across neighborhoods and expanding low-income renters' choices of housing units and development types. However, several factors constrained voucher recipients' ability to access housing and exercise neighborhood choice. These factors included programmatic and administrative hurdles such as limited time and support for tenants to search for housing and a lack of regulations to prevent landlord discrimination against vouchers as an income source. Structural barriers such as neighborhood segregation and economic isolation resulting from public housing and urban renewal programs further limited voucher holders' options.

A current report on HCV location patterns indicated that neighborhood poverty conditions for TBV households have seen little improvement over the past 10 years. In addition, racial and ethnic disparities exist in the relocation of households receiving vouchers to suburban areas that are shaped by FHA capital investment. For instance, one study found that most Hispanic families receiving federal rental assistance live in central cities. A Chicago-based study found that when households use vouchers to rent housing in suburban neighborhoods, these neighborhoods are characterized by higher poverty rates and higher numbers of Black residents than surrounding neighborhoods. Qualitative research on the perceptions of HCVs in the suburbs indicated that residents viewed them as a symptom of neighborhood decline. Overall, studies indicate that a complex relationship exists between the HCV program and neighborhood conditions that can limit choice and opportunity for program participants.

As we described in our previous paper, racial disparities in the populations served by federal rental and homeownership assistance programs indicate that the programs' evolution has struggled to overcome historical implementations of federal programs that have widened the homeownership gap between Black and White households. In this article, we present data on where households participating in federal rental and homeownership assistance programs are living. These data can help us contextualize how past geographies of lending practices and neighborhood disinvestment may continue to constrain housing locations for federally assisted renters and homeowners.

Data

We extracted TBV and FHA data from HUD's internal databases and combined these data into a novel dataset containing reported fields shared between the two programs. Our dataset included mortgage applicant and TBV program participant race and ethnicity, age, monthly household income, and household size as indicated by the number of dependents. The first article in this series offers more details and specifications about this dataset.

To integrate FHA and TBV program data with neighborhood data, we aggregated counts of TBV program participants and homeowners receiving FHA-backed mortgages to the census-tract level. Census-tract-level variables describing demographic, economic, and housing conditions were downloaded from the 2018–2022 5-Year American Community Survey (ACS) dataset using the Census API accessed through the tidycensus package in the R language for statistical computing. ACS variables used to describe neighborhood conditions in census tracts include housing tenure, poverty, race, and ethnicity. The ACS race variables used in this analysis describe individuals identifying as belonging to a single racial group (such as Black or African American Alone) but do not disaggregate race and ethnicity. As a result, individuals reporting belonging to a single racial group may also identify as Hispanic or Latino.

No single definition exists for the terms urban, suburban, and rural, and the definitions chosen can impact the results of an analysis. Many methods only distinguish between urban and rural areas. To help us integrate suburban areas into our analysis, we used the methods developed by Wilson et al. (2012). We identified counties as "urban," "suburban," and "rural" using the following criteria: urban counties were defined as central counties within metropolitan statistical areas containing the most populous principal city, suburban counties were defined as counties within metropolitan statistical areas that do not contain the most populous principal city, and all other counties were defined as "rural." Because census tracts are uniquely contained within counties, we were able to apply the county-level urban, suburban, and rural designations to census tracts.

Findings

Nationally, approximately 35 percent of occupied housing units are renter occupied and 65 percent are owner occupied. Owner-occupied households are more likely to be located in census tracts that are Whiter and have lower poverty rates (figure 1). If federal homeownership assistance programs follow national trends, we would expect FHA loans to be associated with Whiter and lower-poverty census tracts. We frame our findings comparing the neighborhood context for federally assisted homeowners and renters around three neighborhood variables: poverty; resident race and ethnicity; and location in an urban, suburban, or rural county.

Figure 1. National trends at the census-tract level for owner-occupied households indicate that homeowners live in census tracts that are Whiter and have lower poverty rates.

Source: 2018–2022 5-Year ACS dataset.
Note: Each dot represents a census tract.

As with HUD's HCV location reports, we find that nearly half of TBV program participants live in census tracts with poverty rates exceeding 20 percent (the generally accepted "high" poverty cutoff; table 1). By contrast, more than half of homeowners receiving FHA-backed mortgages live in census tracts with poverty rates of less than 10 percent. With respect to poverty, trends in the proportion of federally assisted homeowners and renters tend to mirror national trends for homeowner and renter-occupied households (figure 2). Federally assisted homeowners (FHA) make up 64 percent of all assisted homes in the lowest-poverty census tracts (<= 10%). Owner-occupied households are close in magnitude, making up 74 percent of all occupied units in these tracts. Although renters and owners make up relatively equal proportions of occupied households in tracts with poverty rates of between 20 and 30 and 30 and 40 percent, federally assisted renters (TBV households) represent most of the assisted homes in these tracts (76% and 88%, respectively). In "extreme poverty" census tracts, which have a poverty rate exceeding 40 percent, renters represent 65 percent of all households and 92 percent of assisted households.

Table 1. The distribution of renters participating in the TBV program and homeowners receiving an FHA-backed mortgage by census tract poverty rate.

	Less than 10%	Between 10% and 20%	Between 20% and 30%	Between 30% and 40%	More than 40%
TBV	21.40%	29.35%	24.82%	15.7%	8.72%
FHA	51.93%	33.62%	10.69%	2.79%	0.97%

Source: Authors' analysis of HUD data and 2018–2022 5-Year ACS dataset.

Figure 2. The proportion of all renters and owners and federally assisted renters and owners by census tract poverty rate.

Source: 2018–2022 5-Year ACS dataset and authors' analysis of HUD data.

Because of the historical role of agencies such as FHA and HOLC in perpetuating segregation and promoting investment in White communities, investigating whether investment from federal housing assistance programs continues to flow to majority-White communities is important. We find that census tracts with higher shares of White residents also have higher shares of homeowners receiving FHA-backed mortgages (figure 3). Notably, nearly half (45%) of FHA-backed mortgages are issued to homeowners in census tracts where 75 percent or more of the residents are White. Conversely, the share of TBV households remains relatively constant in census tracts where 25 percent or more of the residents are White, indicating that the racial composition of a neighborhood might relate more strongly to participation in federal homeownership programs than participation in the TBV program.

Figure 3. The prevalence of FHA loans is higher in communities with more White residents.

Source: 2018–2022 5-Year ACS dataset and authors' analysis of HUD data.

We also find distinct patterns in TBV program participation and in households who receive FHA-backed loans when we partition census tracts by the share of the population identifying as Black (figure 4) and the share of the population identifying as Hispanic or Latino (figure 5). Notably, we observe that more than 40 percent of TBV program participants live in census tracts where less than 10 percent of the population is Black or Hispanic, and more than half of homeowners receiving FHA-backed loans are located in these census tracts. In census tracts where more than 40 percent of residents are Black, we observe that the prevalence of TBV program participation is nearly three times higher than the prevalence of households receiving FHA loans. If we consider the location of federally assisted households, we can conclude that FHA continues to support majority-White communities, whereas TBV housing assistance supports more racially and ethnically diverse communities.

Figure 4. Prevalence of renters participating in the TBV program and homeowners receiving FHA-backed mortgages by the share of the population identifying as Black.

Source: 2018–2022 5-Year ACS dataset and authors' analysis of HUD data.

Figure 5. Prevalence of renters participating in the TBV program and homeowners receiving FHA-backed mortgages by share of the population identifying as Hispanic.

Source: 2018–2022 5-Year ACS dataset and authors' analysis of HUD data.

Although historical accounts indicate FHA's importance in suburban development, we find only a slight (2 percentage point) overrepresentation (figure 6) in homeowners receiving FHA loans in suburban counties when compared with the total number of suburban households (renter and owner). Conversely, we observe that TBV households are overrepresented in urban counties by approximately 10 percentage points when compared with the total number of households located in urban areas.

Trends for TBV program participants and homeowners receiving FHA-backed mortgages, however, roughly follow national trends for the location of homeowners and renters, with more than 50 percent of renters overall (56.04%) and TBV recipients (59.03%) located in urban counties and more than 50 percent of homeowners (54.62%) and homeowners receiving FHA mortgages (52.06%) located in suburban and rural counties (table 2). Although the data indicate that TBVs are primarily an urban program and FHA-backed mortgages are slightly more represented in suburban areas, these trends generally follow existing patterns of where renters and homeowners live, and receiving an FHA-backed mortgage or participating in the TBV program may not necessarily drive homeowners and renters to locate in specific areas.

Figure 6. Prevalence of households receiving FHA loans and TBV-supported households in rural, suburban, and urban counties relative to total occupied households.

Source: 2018–2022 5-Year ACS dataset and authors' analysis of HUD data.

Table 2. Share of households receiving federal rental and homeownership assistance and overall share of owner- and renter-occupied households in rural, suburban, and urban counties.  

	Rural counties	Suburban counties	Urban counties
TBV	9.48%	31.49%	59.03%
FHA	12.72%	39.34%	47.94%
Renter-occupied households	11.02%	32.94%	56.04%
Owner-occupied households	15.80%	38.82%	45.38%

Source: 2018–2022 5-Year ACS dataset and authors' analysis of HUD data.

In addition, we find that the TBV program and FHA-backed mortgages serve relatively equal shares of households regardless of location. As a share of renter households, the TBV program performs very similarly in rural, suburban, and urban areas, serving 3.1 percent, 3.5 percent, and 3.8 percent of renter households, respectively. Like the TBV program, FHA loans reach relatively equal shares of owner-occupied households regardless of location, serving 0.9 percent of homeowners in rural counties, 1.1 percent in suburban counties, and 1.2 percent in urban counties.

Discussion

We find that census-tract demographics and poverty have a relationship with the prevalence of FHA-insured mortgages and TBV-supported households. FHA-insured mortgages are issued primarily to homeowners living in census tracts where more than half of the population is White and census tracts with a poverty rate of less than 20 percent. Whether White residents make up most of the census tract population has less of an impact on where households receiving TBV rental assistance are located; however, nearly half of renters supported by TBVs live in high-poverty census tracts (defined as census tracts with a poverty rate of at least 20%). The difference between poverty trends in census tracts with a high prevalence of federal rental assistance and those with a high prevalence of homeownership assistance programs is notable. A TBV-supported renter is twice as likely to live in a high-poverty census tract and nearly nine times as likely to live in an extreme-poverty census tract (defined as a census tract with a poverty rate exceeding 40%) than a homeowner receiving an FHA-backed mortgage.

We find that the location of renters participating in the TBV program and homeowners receiving FHA-backed mortgages generally follow overall trends in areas where renter- and homeowner-occupied households are located. Although we find a higher concentration of TBV program participants in urban areas and a slightly higher representation of homeowners receiving FHA-backed mortgages in suburban areas, we find that the TBV program and FHA-backed mortgages serve relatively equal shares of households regardless of location. The TBV program serves approximately 3 percent of renter households in rural, suburban, and urban counties, and approximately 1 percent of homeowners receive an FHA-backed loan in rural, suburban, and urban counties.

The results we presented in our previous PD&R Edge article indicate clear differences in the populations that federal rental and homeownership assistance programs serve; our analysis at the census tract level also suggests that differences persist in where participants in these programs live and the neighborhoods that benefit from federal housing assistance dollars. However, trends for federally assisted households largely follow existing national trends for homeowners and renters. Without intervention, the private mortgage and rental markets could perpetuate a status quo that widens racial and economic disparities. Federal rental and homeownership assistance programs provide housing market interventions that can increase housing stability, promote home equity gains, and build intergenerational wealth. Present-day disparities in access to quality, stable, and affordable housing punctuate the need to continue to advance racial and economic equity through federal housing policy.

In the final paper in this series, we will (1) review policy interventions in federal rental assistance and homeownership programs that aim to overcome racial and economic barriers in access to housing, and (2) model participation in federal homeownership and rental assistance programs as an outcome of participant demographics and neighborhood characteristics to identify key household and neighborhood-level drivers of program participation disparities.