Sage Journals: Discover world-class research

Abstract

Chinese

The Housing Choice Voucher (HCV) provides rental housing assistance to millions of low-income households across the United States and plays a crucial role in shaping their exposure to concentrated poverty and racial segregation. While prior research has revealed that housing voucher recipients tend to face substantial constraints in terms of neighborhood location, less is known about the direct impact of entering and exiting such programs on individual locational outcomes. Employing a unique dataset that links between housing program records and census microdata between 2000 and 2018, this article examines the impact of entering and exiting housing assistance on the neighborhood context experienced by voucher recipients. Two-way fixed-effects models show that entering the housing voucher program has no statistically significant impact on the neighborhood poverty or racial composition experienced by recipients, while exiting the voucher program results in significant decreases in neighborhood poverty rates relative to both pre-voucher and voucher locations. However, there are substantial differences in these trajectories by race: while white households experienced significant post-voucher decreases in poverty relative to both their pre-voucher and voucher locations, non-white households did not experience post-voucher changes relative to their pre-voucher locations, and Black households experienced no statistically significant post-voucher poverty decreases at all. These findings point to the continued importance of race in shaping neighborhood outcomes: even households participating in the same housing assistance program experience racially disparate outcomes, both during and after their participation in the program.

Keywords

longitudinal analysis low-income housing policy neighborhoods poverty residential mobility

Introduction

Enabling low-income households to access low-poverty and high-opportunity neighborhoods has long been an important objective of federal housing policy in the United States. This consideration is particularly prevalent for the Housing Choice Voucher (HCV) program—currently the largest housing subsidy program directly administered by the US Department of Housing and Urban Development (HUD)—because program participants seek housing on the private rental market instead of living within site-specific housing projects. Notwithstanding well-documented constraints on the ability of voucher recipients to access many types of housing and neighborhoods, the notion that housing assistance programs may provide access to resource-rich neighborhoods remains a powerful and persistent narrative that motivates both a wide range of studies and continuous efforts to promote moves to opportunity via housing assistance programs. Empirical evidence regarding the tangible impact of the voucher programs on locational outcomes remains limited, however, focused primarily on the characteristics of voucher recipient neighborhoods or moves within the voucher program.

Although existing analyses provide a useful perspective on the neighborhoods in which voucher recipients reside, they do not contextualize how the neighborhood contexts experienced by those recipients change upon entering or exiting the voucher program, or the impact of program participation on long-term neighborhood trajectories. To that end, this article links multiple sources of microdata—comprehensive records of households that participated in the HCV program between 2005 and 2018 alongside 2000 decennial census records and 2015–2018 American Community Survey records—in order to identify the locations of individual households before, during, and after their participation in the voucher program. This novel dataset is used to address the following questions: to what extent do neighborhood characteristics change for households that (1) entered the voucher program, (2) exited the voucher program, and (3) entered and then subsequently exited the voucher program? How do these trajectories differ by race?

Comparing these long-term trajectories reveals the extent to which the voucher program or other housing assistance programs might matter—or fail to matter significantly—in terms of neighborhood attainment. If participation in the voucher program meaningfully contributes to upward mobility, we might expect to see sustained decreases in neighborhood poverty when comparing residential locations before, during, and after participation in the voucher program. By contrast, if vouchers are not a broadly effective locational attainment tool for low-income households, there would be little meaningful change in neighborhood conditions before, during, and after program participation. This article begins by considering the importance of neighborhood context for low-income households and the various mechanisms that constrain neighborhood choices within the HCV program, before examining the extent to which program entry and exit might shape locational outcomes. A combination of descriptive analysis and fixed-effect models is used to examine the neighborhood trajectories of voucher recipients, with particular attention to differences in locational outcomes by race. The article concludes by evaluating the implications of this study for the policy objectives of housing assistance programs with respect to housing stability and neighborhood attainment.

Literature review

Housing assistance and neighborhood effects

Households living in poor and segregated neighborhoods in the United States face persistent challenges in moving to different neighborhood contexts, often remaining stuck within a high-poverty neighborhood or cycling between a handful of similarly disadvantaged spaces (Sharkey, 2013). Even as poor households have become less spatially segregated over the past several decades, they have mainly remained in neighborhoods populated by other relatively low-income households (Dwyer, 2012). A lack of economic and social capital constrains the options available to low-income households. Financial precarity and housing precarity are closely entwined: a lack of stable income leaves low-income renters facing a high risk of eviction, which in turn impacts the search for future housing and job prospects and ignites a cycle of poverty (Desmond, 2012). Persistent exposure to concentrated poverty has been shown to have a range of potential negative effects for children in particular, for whom subsequent educational and economic prospects might be impacted by the quality of local services such as public schools, the structure of local social networks, physical isolation from the rest of the urban fabric, and exposure to higher levels of crime (Ellen and Turner, 1997).

Housing assistance programs in the United States have often been viewed as a potential policy tool to reduce the negative impacts of concentrated disadvantage for low-income households, either by increasing investment in under-resourced neighborhoods or by enabling moves to places with greater access to opportunity (Galster, 2013). Site-based public housing projects were historically concentrated in neighborhoods with lower incomes and higher levels of racial segregation (Massey and Kanaiaupuni, 1993; Newman and Schnare, 1997). The HCV program—created in 1974—allocates funding toward tenant-based vouchers that could be used on the private rental market, resulting in much greater dispersal of housing subsidy recipients relative to the prior model of primarily site-based public housing (Devine et al., 2003). The HCV program succeeded in reducing the probability that subsidized households would live in the most “distressed” neighborhoods relative to recipients of site-based subsidies (Newman and Schnare, 1997). The Moving to Opportunity (MTO) demonstration further demonstrated the potential for tenant-based subsidies to improve individual outcomes by providing households with vouchers to move to low-poverty neighborhoods, which meaningfully improved the future economic prospects of young children (Chetty et al., 2016). However, the scope of this program was ultimately limited to just five cities between 1994 and 1998 (de Souza Briggs et al., 2010). While MTO and similar demonstrations illustrate the potential for demand-side housing subsidies to meaningfully impact both locational outcomes and individual economic outcomes for low-income households, these interventions require substantial targeted investments that are difficult to sustain.

Housing vouchers and neighborhood access

Beyond policy demonstrations, most voucher recipients are largely left to their own devices in searching for and obtaining a unit with sufficiently low rent from a willingly participating landlord. These households face a combination of programmatic constraints, limited economic and social resources, and landlord practices that either implicitly or explicitly discriminate against voucher recipients searching for housing (DeLuca et al., 2013). Voucher recipients are often not able to move even if they would prefer to do so, and face substantial difficulties in obtaining housing in neighborhoods that meet their preferences (Wang, 2018). Some of these barriers are programmatic, including maximum Fair Market Rents (FMRs) that limit the availability of voucher-eligible units in more expensive neighborhoods (Collinson and Ganong, 2018). While the recent roll-out of Small Area Fair Market Rents (SAFMRs) theoretically enables households to move to more expensive neighborhoods by allowing FMR standards to vary by ZIP code (Dastrup et al., 2019; Reina et al., 2019), the persistence of racial discrimination and segregation may limit the effectiveness of such programmatic reforms for Black and Hispanic voucher recipients (McClure and Schwartz, 2019). Beyond programmatic limitations, voucher recipients may also experience source-of-income discrimination by landlords who are reticent due to social stigma associated with vouchers or financial concerns about rent levels and administrative costs (Ellen et al., 2023; Tighe et al., 2017). This discrimination is carried out not only through direct rejections but also increasingly through online rental listings, which feature language explicitly excluding voucher recipients (e.g. “No Section 8”; Hangen and O’Brien, 2023) or display voucher-friendly listings only in certain neighborhoods (Hess et al., 2023). The perception (and reality) of limited housing options may also make voucher holders more likely to remain in place, with voucher recipients in the most under-resourced neighborhoods being some of the least likely to move (Basolo and Yerena, 2017).

These obstacles have concrete implications for the locational outcomes of participants in the housing voucher program. Voucher recipients tend to concentrate within certain types of neighborhoods and certain segments of metropolitan areas characterized by higher levels of poverty and racial segregation (Metzger, 2014; Park, 2013; Varady et al., 2010), with only a small proportion residing in neighborhoods with very low poverty rates (McClure et al., 2015). This tendency for voucher recipients to locate in neighborhoods with limited resources and opportunities is contingent not only upon receiving a voucher but also upon the specific demographic characteristics of the recipient household: Black voucher holders are particularly likely to live in neighborhoods with high poverty rates (Pendall, 2000), and Black and Hispanic voucher holders are less likely to live in neighborhoods with the lowest levels of economic distress (Schwartz et al., 2016). Residential mobility within the voucher program may improve neighborhood outcomes for certain households, contingent upon housing market factors and the demographic characteristics of individual recipients (Wang and Walter, 2018). However, residential moves within the voucher program rarely result in massive shifts and often only incrementally change a household’s neighborhood context (Feins and Patterson, 2005). The mobility trajectories of voucher recipients are also heavily conditioned by household demographic characteristics: voucher recipients are more likely to move to lower poverty neighborhoods during their time in the program if they are white, have larger households, and have higher incomes relative to other voucher holders, while households with disabilities face worse locational outcomes (Walter et al., 2015; Wang and Walter, 2018). A recent demonstration in King County, Washington shows that shifting the mobility patterns of voucher recipients to higher-opportunity neighborhoods requires both substantial financial resources as well as sustained support for housing search (Bergman et al., 2024). Beyond such targeted interventions, tenant-based housing assistance programs appear not to significantly improve the neighborhood context of participants and may even lead recipients to reside in places with higher concentrations of poverty and lower access to economic opportunity.

Entering and exiting housing assistance

Compared with the volume of research on household mobility within the HCV program, far fewer studies consider the household-level trajectories of households entering or exiting housing assistance. Do households that receive vouchers remain in the same neighborhood, or are they able to improve upon their neighborhood context because they are able to access a wider array of neighborhoods? While some versions of HCV program data include the ZIP code of participants prior to program entrance, the examination of pre-program neighborhood characteristics has often been constrained to smaller surveys or to households that participated in other housing assistance programs (Feins and Patterson, 2005). Many voucher recipients respond to receiving a voucher by not moving at all: in a study of 48 housing authorities, Finkel and Buron (2001) find that around one in five voucher households remained in the same unit upon receipt of the voucher, taking advantage of the rental subsidy to obtain a more affordable rent in their current location. Gubits et al. (2009) find in a randomized experiment that households living in higher poverty neighborhoods that used a voucher subsequently moved to lower poverty neighborhoods, suggesting that voucher receipt improved neighborhood access primarily for those that were located in higher poverty neighborhoods at the beginning of the period. Varady and Walker (1999) further find that voucher recipients improve their neighborhood context but remain in an area close to their initial neighborhood, possibly reflecting the durable impact of existing local social or economic ties. Comparing households that entered housing assistance programs with cross-sectional observations of households that would enter those programs within the following two years, Fenelon et al. (2023) find that vouchers lead to a reduced likelihood of living in higher poverty neighborhoods, particularly for non-white households, while entering site-based public housing developments leads to increased poverty exposure. Finally, Carlson et al. (2012) use propensity score matching to show that receipt of housing vouchers improved neighborhood quality for households in Wisconsin after several years in the program, pointing to a long-term benefit associated with receiving a voucher.

Although HUD rental assistance programs have an annual attrition rate between roughly 14% and 18% (McClure, 2018), empirical evidence regarding outcomes for those who exit each year is even less developed. A handful of prior studies find that households that are elderly or disabled are less likely to leave the program while households with higher or increasing incomes are more likely to leave, suggesting that a substantial share of exits from housing subsidy programs might be due to positive reasons such as greater economic self-sufficiency (Ambrose, 2005; Hungerford, 1996). However, other evidence suggests that exits from housing assistance result in a mix of positive and negative trajectories, with some attaining homeownership, others becoming homeless, and many simply continuing to rent similar types of housing in similar types of neighborhoods (Smith et al., 2015). A series of recent studies linking housing program data with administrative records from King County, Washington affirm this diversity of outcomes, and suggest that different types of exits are associated with different outcomes: “positive” exits such as wage increases are associated with greater economic self-sufficiency (Colombara et al., 2024), while “negative” exits such as falling out of the program due to eviction or relocation are associated with elevated rates of homelessness (Petrakos et al., 2023) and worse health outcomes (Matheson et al., 2023). In terms of neighborhood outcomes associated with program exit, Ramiller and Reid (2023) show that exits from site-based public housing result in significant reductions in neighborhood poverty, while exits from the tenant-based voucher program produce modest decreases in neighborhood poverty rates. This points to a relationship between program exit and neighborhood attainment; however, it is unclear if this indicates that subsidy recipients live in substantially higher poverty neighborhoods, if they move to lower poverty neighborhoods after exiting due to improved economic circumstances, or some combination of the two.

Data and methods

This article identifies recipients of Tenant-Based HCVs between 2005 and 2018 using national records from HUD Annual Longitudinal Files.¹ These subsidy recipients are linked with individual records from the 2000 decennial census and records between 2015 and 2018 from the American Community Survey (ACS) using a Personal Identification Key that serves as a common identifier across all three datasets. This linkage enables the construction of residential trajectories across multiple periods: (1) individuals observed in the 2000 decennial census; (2) individuals observed in the PICTRACS dataset between 2005 and 2014; and (3) individuals observed in either the PICTRACS or ACS microdata between 2015 and 2018. Each of these periods is aligned with estimates of poverty rate and racial composition at the census tract level: records from 2000 are matched with tract-level estimates from the 2000 decennial census, records from 2005 to 2014 are matched with 2005–2009 and 2010–2014 5-Year ACS estimates, and records from 2005 to 2018 are matched with 2015–2019 5-Year ACS estimates.²

Households are separated into groups based on the time at which they entered or exited housing assistance.³ Program entry dates are provided within the PICTRACS dataset, while exit dates are defined by the last year in which the individual was observed receiving housing assistance in the PICTRACS dataset. This results in four groups: (1A) individuals that entered housing assistance at some point between 2001 and 2014 and were observed leaving housing assistance before 2018; (1B) individuals that entered housing assistance between 2001 and 2014 and that remained in housing assistance through 2018; (2A) individuals that were receiving housing assistance on or before 2000 that were observed leaving housing assistance before 2018; and (2B) individuals that received housing assistance for the entirety of the time period between 2000 and 2018 (Table 1). There are clear differences in the demographic characteristics between these four groups, reflecting the differing conditions under which each group entered or exited the voucher program (Table 2). Individuals from the group that received temporary assistance (1A) were far more likely to be non-Hispanic white and in younger age groups, while members of the group that continuously received assistance (2B) were more likely to be Black and in an older age group.

Table 1.

Description of analysis groups.

		2000 (Decennial)	2005–2014 (HCV)	2015–2018(HCV or ACS)
Newcomers	Group 1A: Temporaryassistance	No assistance	Assistance	No assistance
	Group 1B: Entered assistance	No assistance	Assistance	Assistance
Long-term assistance	Group 2A: Exited assistance	Assistance	Assistance	No assistance
	Group 2B: Continuous assistance	Assistance	Assistance	Assistance

Table 2.

Household characteristics in 2000 by group prior to matching.

	Group 1A:Temporaryassistance	Group 1B:Enteredassistance	Group 2A:Exitedassistance	Group 2B:Continuous assistance
Age (18–35)	28.7%	22.6%	20.7%	15.7%
Age (35–49)	39.4%	36.8%	49.7%	40.9%
Age (50–64)	26.6%	31.6%	24.8%	33.2%
Age (65+)	5.3%	9.0%	4.8%	10.2%
Children	69.7%	61.5%	75.6%	60.4%
White	54.2%	37.3%	37.5%	31.2%
Black	29.0%	40.5%	39.3%	43.2%
Hispanic	11.0%	16.0%	15.8%	18.7%
Other	5.8%	6.2%	7.5%	7.0%
Female	75.8%	78.5%	80.1%	81.8%
Household size	3.5	3.3	3.4	2.9
N	211,000	10,940,000	77,000	5,333,000

Source: Decennial Census (2000). DRB Approval Number: CBDRB-FY25-P2137-R1184.

These groups are separated into two sets of comparisons: first, households that were not participating in the voucher program as of 2000 and for which program entry may be directly observed (Groups 1A and 1B); and second, households that were already participating in the voucher program as of 2000 (Groups 2A and 2B). This enables a direct comparison of the different trajectories of these groups over time, highlighting whether program entry and exit have meaningful impacts upon neighborhood trajectories. Given the demographic differences between these groups, propensity score matching is used to construct populations that are demographically and economically similar. Matching produces a subset of individuals from Group 1B that resemble Group 1A, and a subset of individuals from Group 2B that resemble Group 2A. These observations are constrained to match within the same county and are matched based on individual age, race, sex, as well as neighborhood poverty and racial composition in 2000. Given the large number of observations in each control group (Groups 1B and 2B), 25 control observations are matched to each treatment observation with replacement and variable matching weights. Matching diagnostics indicate that the resulting sets of comparison groups are highly balanced (see Appendix Figure A1).

Following these descriptive trajectories, a two-way fixed-effects model accounts for both individual and year fixed effects and measures the impact on neighborhood outcomes ( $y_{i}$ ) of participation in the voucher program ( $V_{i t}$ ) and status after exiting the program ( $P_{i t}$ ) for each group of households that exited the program (Groups 1A and 2A) compared with each corresponding group of similar households that did not exit the program (Groups 1B and 2B).

\begin{matrix} y_{i} = β_{0} + β_{1} V_{i t} + β_{2} P_{i t} + β_{3} H_{i t} + α_{i} \\ + δ_{t} + ϵ_{i t} \end{matrix}

(1)

The model incorporates fixed-effects for individual ( $α_{i}$ ) and year ( $δ_{t}$ ), and controls for the time-variant household characteristics ( $H_{i t}$ ) of household size and presence of children. For each set of comparison groups, trajectories are assessed for the overall sample as well as four mutually exclusive race groups: non-Hispanic white, Black, Hispanic, and Asian/Other Race.

Results

Neighborhood trajectories

Figure 1 displays neighborhood trajectories for each of these groups in terms of neighborhood poverty rate, the share of the neighborhood population that does not identify as non-Hispanic white (hereafter, “non-white”), and distance moved from initial location. Comparing each pair of groups (Group 1A vs. Group 1B; Group 2A vs. Group 2B) reveals the extent to which the trajectories of those that exited the program (1A, 2A) compare with similar households that did not exit (1B, 2B). These trajectories show that households in the sample are consistently impacted to some degree by macroeconomic conditions and demographic trends regardless of program participation timing and race. Every group experienced poverty increases between 2005–2009 and 2010–2014 followed by corresponding decreases during the 2015–2019 period, for example, reflecting increases in economic disadvantage associated with the Great Recession that affected all groups in a similar manner. However, there are also important differences in the trajectories of these groups: households that exited assistance (1A, 2A) experienced larger decreases in average poverty rates between 2010–2014 and 2015–2019 than their counterparts that did not exit housing assistance (1B, 2B). Exposure to neighborhood poverty also differed substantially based on the race of the voucher recipients. Non-Hispanic white voucher recipients were consistently exposed to the lowest levels of neighborhood poverty, while Hispanic and Black households consistently experienced the highest average poverty rates over time regardless of program participation status.

Figure 1.

Neighborhood trajectories by analysis group and by race.

The effects of program entry and exit also vary substantially by race. White households entering the voucher program (1A, 1B) saw average neighborhood poverty rates increase by around four percentage points between 2000 and 2005–2009, while white households that were already receiving vouchers experienced more modest increases of around 2% during the same period. By contrast, people of color tended to experience minimal changes or slight decreases in poverty rates between the two periods, regardless of whether they were already receiving rental assistance in 2000. This suggests that while the constraints of the voucher program altered the neighborhood choices available to white households, households of color were already constrained to higher poverty neighborhoods notwithstanding program participation. The effects of program exit are more similar by race: all groups experienced decreases in neighborhood poverty between 2010–2014 and 2015–2019, with larger decreases for households that exited the program (1A, 2A) compared with those that remained (1B, 2B).

However, there are important differences in the relative scale of these effects. White households that exited the program experienced decreases in average neighborhood poverty that were 1.7% greater compared with those that remained in the program. By contrast, Hispanic households that exited the voucher program only experienced additional decreases of 1% relative to those that remained in the program, and Black households experienced additional decreases of only 0.7%. Program entry and exit are thus both associated with major changes in exposure to neighborhood poverty for white households but had a much smaller impact on Black and Hispanic households. By comparison, voucher recipients experienced relatively minor changes in neighborhood racial composition over time regardless of their program participation status. Households tended to live in neighborhoods with a growing share of non-white residents over time, reflecting increasing racial diversity at a national scale. There was a sharp divide in the racial composition of neighborhoods occupied by white and non-white voucher holders, however, reflecting persistent racial segregation even amongst low-income subsidized households.

Program entry and exit also had large impacts on the distance that households moved away from their original locations. Households that entered the program after 2000 (1A, 1B) moved an average of five to six miles away from their original locations by the subsequent period, while those that were already in the program moved an average of less than one mile. Households that were already receiving rental assistance in 2000 (1B, 2B), meanwhile, continued to live a similar distance away from their origin points. This indicates that entering the program necessitated major shifts in residential location, while those already in the program experienced a high degree of residential stability. Program exits led to further increases in the distance from the origin location: those that received temporary assistance (1A) ended up in post-voucher locations that were on average 10 miles away from their pre-voucher locations, while those that exited assistance after having received it for an extended period (2A) moved an average of four miles from their original location in 2000. Notably, white households experienced the highest rates of mobility among those that entered or exited assistance during the study period, with those that received temporary assistance moving nearly 14 miles on average between 2000 and 2015–2019. Those that received continuous assistance throughout the study period (2B) moved the shortest average distance across the entire study period, reflecting the greater stability afforded by continual housing support.

Fixed-effect model

Fixed-effect models compare within-individual changes in neighborhood trajectories between comparison groups while controlling for temporal effects and time-variant individual characteristics. Compared with the group that entered and remained in rental assistance through 2018, the group that entered and exited experienced significant decreases in neighborhood poverty relative to pre-voucher locations (Table 3, column 1). This suggests that households receiving temporary assistance do experience long-term benefits from program participation. However, these effects are concentrated among white households exiting the program, who experienced average additional decreases in neighborhood poverty approximately 1.4 percentage points between their pre-voucher and post-voucher locations compared with those that entered and remained (Table 3, column 2). These relative decreases in neighborhood poverty were not statistically significant for non-white households (Table 3, columns 3–5). Similar patterns are observed for neighborhood racial composition, with white households experiencing a substantial additional decrease in the share of non-white residents between pre-voucher and post-voucher neighborhoods upon exiting, while Black and Hispanic households did not experience statistically significant changes.

Table 3.

Fixed effect model coefficients.

	Entered vs. temporaryassistance					Continuous vs.exited assistance
	Overall	White	Black	Hispanic	Asian/other	Overall	White	Black	Hispanic	Asian/other
Poverty rate
Current voucher	0.04	0.38	−0.38	−0.55	−0.06	—	—	—	—	—
Post-voucher	−0.72**	−1.36***	−0.13	−1.12	−0.02	−0.77**	−1.57***	−0.06	−0.82*	−1.38*
Household size	−0.33***	−0.41**	−0.33**	−0.26***	−0.34***	−0.09***	−0.11*	−0.14***	−0.13***	−0.13*
Children	0.00	0.01	-0.11	−0.03	0.14	−0.28***	−0.19*	−0.45**	−0.26*	−0.2
N	5,398,000	2,438,000	1,946,000	697,000	317,000	2,488,000	841,000	1,056,000	408,000	184,000
R²	0.63	0.65	0.56	0.64	0.64	0.69	0.73	0.61	0.68	0.71
Percentage non-white
Current voucher	0.32	0.32	0.54	0.13	1.37	—	—	—	—	—
Post-voucher	−1.13**	−2.93***	0.93	−0.35	2.64*	−2.09***	−3.81***	−0.67	−2.56***	−2.03**
Household size	0.08**	−0.32**	0.4*	0.12*	−0.01	0.26***	0.02	0.33***	0.18**	0.03
Children	−0.37**	−0.27**	−0.55*	−0.13	−0.22	−0.34*	−0.56***	−0.45	−0.34	−0.27
N	5,398,000	2,438,000	1,946,000	697,000	317,000	2,488,000	841,000	1,056,000	408,000	184,000
R²	0.86	0.82	0.70	0.81	0.85	0.89	0.88	0.76	0.83	0.89

***p < 0.001. **p < 0.01. *p < 0.05.

Sources: HUD Longitudinal Files (2005–2018); Decennial Census (2000); 1-year ACS (2015–2018); 5-yearACS (2005–2009; 2010–2014; 2015–2019). DRB Approval Number: CBDRB-FY25-P2137-R11842.

While the analysis of households that entered housing assistance allows for the comparison of pre-voucher and post-voucher outcomes, comparing households that received assistance throughout the study period with those that exited indicates that the impact of exiting for long-term housing subsidy recipients is more substantial (Table 3, column 6). For this group, exiting housing assistance results in statistically significant decreases in poverty rates for white, Hispanic, and Asian/Other Race voucher recipients, with the largest decreases in both poverty rates and share non-white for white voucher recipients (Table 3, columns 7, 9 and 10). Black households, however, did not experience statistically significant decreases after exiting the voucher program (Table 3, column 8). This indicates that while most long-term recipients of housing assistance move to different neighborhood contexts after exiting the voucher program, Black households largely remain in neighborhoods with similar poverty rates after their participation in the voucher program. These patterns are again replicated for neighborhood racial composition, with all except Black voucher recipients experiencing decreases in the share of non-white residents upon exiting the program.

Discussion

Key findings

These results highlight neighborhood outcomes associated with three household trajectories within the HCV program between 2000 and 2018: (1) program entry; (2) program exit; and (3) program entry followed by program exit.

Program entry was associated with small increases in neighborhood poverty overall and for white households, but small decreases in poverty for Black and Hispanic households. These minimal changes are despite households moving an average of several miles upon entering the voucher program, which suggests that receiving a voucher does have a definite impact on residential mobility but not necessarily on the characteristics of the neighborhoods in which voucher recipients live. The overall finding runs somewhat counter to prior findings that receiving a voucher enables households to move to lower poverty neighborhoods (Fenelon et al., 2023; Gubits et al., 2009; Varady and Walker, 1999). It should be noted that the sample examined here is more heavily white (45%) than the average voucher population (30% in 2019). This may reflect the particular demographic profile of those that entered housing assistance after 2000, and particularly those that only received assistance for a limited period of time. Indeed, the patterns observed for non-white households, while not statistically significant, are more consistent with these prior findings. Nevertheless, these results suggest that while longitudinal benefits may accrue to some households at the individual level, the average trajectory of those entering the voucher program does not indicate substantial changes in neighborhood context after controlling for individual and temporal factors.

Exiting housing assistance appears to lead to significant changes in neighborhood context for most households. As with program entry, program exit also results in relatively long-distance moves, with white households that exited the voucher program moving the greatest distances. Program exits led to moderate decreases in neighborhood poverty compared with voucher neighborhoods for all but Black households, suggesting that most (though not all) groups of those exiting the voucher program do transition into lower poverty neighborhoods (Table 3, columns 6–10). This aligns with other recent findings, which suggest that exiting housing assistance leads to a modest but meaningful decrease in exposure to neighborhood poverty for HCV recipients—albeit far smaller than the decrease experienced by households exiting public housing (Ramiller and Reid, 2023).

Finally, long-term trajectories for households that both entered and exited the voucher program reveal starker differences by race. White households that entered the voucher program consistently experienced increases in neighborhood poverty rates, and their exposure to neighborhood poverty remained elevated in 2018 relative to 2000. However, white households that exited the voucher program experienced decreases in poverty exposure compared to those that entered and remained through 2018, indicating that exiting the program did enable white households to reduce their relative poverty exposure (Table 3, column 2). By contrast, non-white households that exited the voucher program ended up living in neighborhoods with very similar poverty rates to where they had resided prior to entering the voucher program. Furthermore, model results show no statistically significant difference in these trajectories associated with program exit (Table 3, columns 3–5). This indicates that voucher program participation had no discernible impact on neighborhood outcomes for non-white households whatsoever. These findings point to the persistence of racial discrimination and segregation in shaping the neighborhood outcomes of voucher recipients: white voucher recipients are exposed to higher poverty neighborhoods due to constraints that limit neighborhood choice for voucher recipients but are largely able to return to lower poverty neighborhoods upon exiting the voucher program. By contrast, non-white recipients already face a wider array of constraints and are not impacted by participation in the voucher program.

While prior research has pointed to the importance of neighborhood context in shaping life outcomes and household well-being (e.g. Chetty et al., 2016; Sampson, 2012), this analysis suggests that the voucher program does relatively little to support long-term improvements in household socioeconomic position or neighborhood context, particularly for people of color. This does not discount the value of voucher-based approaches to reducing the impacts of concentrated poverty altogether. Policy demonstrations such as MTO point to real benefits for voucher recipients provided with additional resources to move to lower poverty neighborhoods (Chetty et al., 2016). However, mobility programs are costly and have only ever been implemented at a small scale, leaving serious questions about the efficacy of such an approach under a status quo of limited funding—not to mention the inevitable barriers to large-scale implementation of low-income housing among more affluent communities. Rather, these findings imply that the HCV program, despite its nominal ability to combine housing assistance with residential mobility, does not point to long-term effectiveness in moving people to radically different neighborhood contexts. While moves within the voucher program potentially also reflect programmatic constraints, including a cap on the Fair Market Rents and source of income discrimination by landlords wary of leasing to voucher recipients, the fact that exiting households experience only modest changes in neighborhood context—and no significant change when compared to pre-voucher locations for non-white households—suggests that any aggregate impact of the voucher program on household neighborhood attainment is minimal.

Limitations and future research

Although propensity score matching and fixed-effects modeling are intended to address demographic differences between the groups, there are likely structural differences that remain. Households that were already participating in the program prior to 2000 and remained through 2018 likely represent individuals that face additional structural impediments to exiting housing assistance, particularly compared with households that had relatively short tenures only spanning between 2000 and 2018. Additionally, this analysis lacks a comparison to households that were eligible for the voucher program but never participated, whether because they never received the opportunity to use a voucher or because they were unwilling or unable once the opportunity arose. Data limitations prevent the construction of a reliable control group: not all low-income households are likely to resemble those households that participate in the voucher program, making it difficult to isolate those households that would have used a voucher if given the opportunity. Prior studies have used waiting list data to assess the impacts of voucher participation, but rarely at the geographic or temporal scale of this analysis. Nevertheless, this study provides important context regarding the neighborhood trajectories of those that do enter and exit the voucher program over time.

Notwithstanding the scope and granularity of the data in question, there remain several directions for future research. Whereas this analysis focuses on longer-term trends in neighborhood trajectories across several broad time periods, the longitudinal nature of program data could be leveraged to detect shorter-term changes in neighborhood conditions, residential location, or program participation. Furthermore, identifying the distinction between “positive” and “negative” exits would likely reveal important divergences in neighborhood trajectories based on individual circumstances. Finally, while neighborhood poverty rate is the primary focus of this analysis, there are many other neighborhood measures that could be considered in future analyses of long-term outcomes. These include concentrated poverty, the presence of affluent households, school quality, crime, proximity to employment opportunities, and exposure to environmental hazards.⁴

Conclusion

Increasing access to high-opportunity neighborhoods and reducing exposure to concentrated poverty and under-resourced neighborhoods are regarded as important objectives of US federal housing subsidy programs. The HCV program has long been viewed as a potential tool for allowing low-income households to move away from concentrated poverty and segregated neighborhoods, by providing a source of housing support not tied to a specific location. Research has consistently indicated, however, that the voucher program largely fails to precipitate significant changes in neighborhood outcomes for recipients moving within the program. Voucher recipients face ongoing discrimination by housing providers, and programmatic limitations prevent voucher recipients from accessing many higher-cost neighborhoods. While the limitations on neighborhood attainment within the voucher program are well-established, the question remains whether participation in the program might yield long-term benefits in terms of neighborhood attainment after households exit the program. The voucher program might, for example, provide households with the residential stability and financial assistance necessary to subsequently attain a higher socioeconomic status. The analysis presented here suggests that this is only true for certain voucher participants—namely, white households that temporarily participate in the voucher program. People of color do not experience the same long-term benefits from program participation, and Black households do not experience poverty decreases after exiting even relative to their neighborhoods while in the voucher program.

These results suggest that focusing on the voucher program as a means of upward neighborhood mobility may not be viable without significant changes to the structure of the program and to the social safety net more generally. While some voucher recipients may experience socioeconomic ascent during their participation in the program—by lightening the financial burden in the face of already limited household resources—the voucher program cannot itself solve the challenges of trenchant economic and spatial inequality. Households often spend years on waiting lists to receive a voucher precisely because their economic precarity is so entrenched, and the voucher provides a crucial resource for getting by—not necessarily as a resource for neighborhood attainment per se. It is unreasonable to expect that simply providing housing support in the absence of other support such as employment opportunities will single-handedly enable positive long-term trajectories. Such expectations must face not only the many well-documented barriers to the use of housing vouchers on the private housing market but also the difficulty of incentivizing moves to lower poverty neighborhoods without significant additional resources (as in the MTO demonstration).

While residential stability is essential, enabling access to opportunity remains important in efforts to reduce concentrated poverty and improve the economic prospects of low-income households. However, the limitations of existing programmatic frameworks as illustrated here suggest that the prospects of other policy strategies—such as varying the subsidy formula for housing vouchers based on local housing market conditions via Small Area Fair Market Rents (Dastrup et al., 2019; Reina et al., 2019) or distributing subsidized housing to higher-resource neighborhoods via LIHTC allocation schemes (Ellen and Horn, 2018; Owens and Smith, 2023)—deserve significant attention moving forward. Nevertheless, enthusiasm for the existing voucher program as a spatial strategy for poverty deconcentration remains robust. The key finding of this study—that the impact of program participation on locational outcomes is limited both in the short and long term—suggests that it may be more worthwhile to focus on the role of the voucher program in promoting housing stability and affordability. The voucher program serves an essential role in providing greater economic and residential stability for precarious households, and the persistent focus on neighborhood outcomes for voucher recipients (to which this article admittedly contributes) should not overshadow the primary role of housing assistance programs in delivering stable housing.

Footnotes

Appendix

Acknowledgements

The author would like to express profound gratitude to Rebecca Walter, Arthur Acolin, and Vince Wang for the many ways they have supported this project, as well as to Carolina Reid and three anonymous reviewers for their helpful comments and suggestions. The author would also like to thank Carlos Becerra, Survey Statistician and Administrator of the Northwest Federal Statistical Research Data Center.

ORCID iD

Alex Ramiller

Funding

The author disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Partial support for this research came from a Eunice Kennedy Shriver National Institute of Child Health and Human Development research infrastructure grant, P2C HD042828, to the Center for Studies in Demography and Ecology at the University of Washington. The content is solely the responsibility of the author and does not necessarily represent the official views of the National Institutes of Health. Additional support for this research came from support to CSDE from the College of Arts and Sciences, the UW Provost, eSciences Institute, the Evans School of Public Policy and Governance, College of Built Environment, School of Public Health, the Foster School of Business, and the School of Social Work.

Declaration of conflicting interests

The author declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Disclaimer

Any views expressed are those of the author and not those of the US Census Bureau. The Census Bureau has reviewed this data product to ensure appropriate access, use, and disclosure avoidance protection of the confidential source data used to produce this product. This research was performed at a Federal Statistical Research Data Center under FSRDC Project Number 2137 (Disclosure Review Board Approval Number CBDRB-FY25-P2137-R11842).

Notes

References

Ambrose

(2005) A hazard rate analysis of leavers and stayers in assisted housing programs. Cityscape 8(2): 69–93.

Basolo

Yerena

(2017) Residential mobility of low-income, subsidized households: A synthesis of explanatory frameworks. Housing Studies 32(6): 841–862.

Bergman

Chetty

DeLuca

, et al. (2024) Creating moves to opportunity: Experimental evidence on barriers to neighborhood choice. American Economic Review 114(5): 1281–1337.

Carlson

Haveman

Kaplan

, et al. (2012) Long-term earnings and employment effects of housing voucher receipt. Journal of Urban Economics 71(1): 128–150.

Chetty

Hendren

Katz

(2016) The effects of exposure to better neighborhoods on children: New evidence from the moving to opportunity experiment. American Economic Review 106(4): 855–902.

Collinson

Ganong

(2018) How do changes in housing voucher design affect rent and neighborhood quality? American Economic Journal: Economic Policy 10(2): 62–89.

Colombara

Quinn

Pennucci

, et al. (2024) The relationship between exits from federally subsidized housing and wages, King County, WA. Housing Policy Debate 34(3): 286–304.

Dastrup

Finkel

Ellen

(2019) The effects of small area fair market rents on the neighborhood choices of families with children. Cityscape 21(3): 19–48.

DeLuca

Garboden

PME

Rosenblatt

(2013) Segregating shelter: How housing policies shape the residential locations of low-income minority families. The Annals of the American Academy of Political and Social Science 647(1): 268–299.

10.

de Souza Briggs

Popkin

Goering

(2010) Moving to Opportunity: The Story of an American Experiment to Fight Ghetto Poverty. Oxford University Press.

11.

Desmond

(2012) Eviction and the reproduction of urban poverty. American Journal of Sociology 118(1): 88–133.

12.

Devine

Gray

Rubin

, et al. (2003) Housing Choice Voucher Location Patterns: Implications for Participants and Neighborhood Welfare. Office of Policy Development and Research, US Department of Housing and Urban Development.

13.

Dwyer

(2012) Contained dispersal: The deconcentration of poverty in US metropolitan areas in the 1990s. City & Community 11(3): 309–331.

14.

Ellen

Horn

(2018) Points for place: Can state governments shape siting patterns of low-income housing tax credit developments? Housing Policy Debate 28(5): 727–745.

15.

Ellen

Turner

(1997) Does neighborhood matter? Assessing recent evidence. Housing Policy Debate 8(4): 833–866.

16.

Ellen

O’Regan

Harwood

KWH

(2023) Advancing choice in the Housing Choice Voucher Program: Source of income protections and locational outcomes. Housing Policy Debate 33(4): 941–962.

17.

Feins

Patterson

(2005) Geographic mobility in the Housing Choice Voucher Program: A study of families entering the program, 1995–2002. Cityscape 8(2): 21–47.

18.

Fenelon

Slopen

Newman

(2023) The effects of rental assistance programs on neighborhood outcomes for U.S. children: Nationwide evidence by program and race/ethnicity. Urban Affairs Review 59(3): 832–865.

19.

Finkel

Buron

(2001) Study on Section 8 Voucher Success Rates. Office of Policy Development and Research, US Department of Housing and Urban Development.

20.

Galster

(2013) U.S. assisted housing programs and poverty deconcentration: A critical geographic review. In: Manley

van Ham

Bailey

, et al. (eds) Neighbourhood Effects or Neighbourhood Based Problems? A Policy Context. Springer Netherlands, pp. 215–249.

21.

Gubits

Khadduri

Turnham

(2009) Housing Patterns of Low Income Families with Children: Further Analysis of Data from the Study of the Effects of Housing Vouchers on Welfare Families. Joint Center for Housing Studies. Available at: https://www.jchs.harvard.edu/research-areas/working-papers/housing-patterns-low-income-families-children-further-analysis-data (accessed 22 December 2022).

22.

Hangen

O’Brien

(2023) The choice to discriminate: How source of income discrimination constrains opportunity for Housing Choice Voucher holders. Urban Affairs Review 59(5): 1601–1625.

23.

Hess

Walter

Kennedy

, et al. (2023) Segmented information, segregated outcomes: Housing affordability and neighborhood representation on a voucher-focused online housing platform and three mainstream alternatives. Housing Policy Debate 33(6): 1511–1535.

24.

Hungerford

(1996) The dynamics of housing assistance spells. Journal of Urban Economics 39(2): 193–208.

25.

McClure

(2018) Length of stay in assisted housing. Cityscape 20(1): 11–38.

26.

McClure

Schwartz

(2019) Small area fair market rents, race, and neighborhood opportunity. Cityscape 21(3): 49–70.

27.

McClure

Schwartz

Taghavi

(2015) Housing choice voucher location patterns a decade later. Housing Policy Debate 25(2): 215–233.

28.

Massey

Kanaiaupuni

(1993) Public housing and the concentration of poverty. Social Science Quarterly 74(1): 109–122.

29.

Matheson

Colombara

Pennucci

, et al. (2023) A good farewell? Positive exits from federal housing assistance and lower acute healthcare utilization. Journal of Urban Health 100(6): 1202–1211.

30.

Metzger

(2014) The reconcentration of poverty: Patterns of housing voucher use, 2000 to 2008. Housing Policy Debate 24(3): 544–567.

31.

Newman

Schnare

(1997) “… And a suitable living environment”: The failure of housing programs to deliver on neighborhood quality. Housing Policy Debate 8(4): 703–741.

32.

Owens

Smith

(2023) Producing affordable housing in higher-opportunity neighborhoods: Incentives in California’s LIHTC program. Journal of Urban Affairs 47(7): 2423–2451.

33.

Park

(2013) Housing vouchers as a means of poverty deconcentration and race desegregation: Patterns and factors of voucher recipients’ spatial concentration in Cleveland. Journal of Housing and the Built Environment 28(3): 451–468.

34.

Pendall

(2000) Why voucher and certificate users live in distressed neighborhoods. Housing Policy Debate 11(4): 881–910.

35.

Petrakos

Liu

, et al. (2023) Associations between exit type from federal housing assistance and subsequent homelessness. Housing Policy Debate 8: 1–19.

36.

Ramiller

Reid

(2023) Examining the Housing and Neighborhood Trajectories for Former HUD-Assisted Households with Children. US Department of Housing and Urban Development Office of Policy Development and Research. Available at: https://www.huduser.gov/portal/publications/Examining-the-Housing-Neighborhood-Trajectories.html (accessed 20 March 2024).

37.

Reina

Acolin

Bostic

(2019) Section 8 vouchers and rent limits: Do small area fair market rent limits increase access to opportunity neighborhoods? An early evaluation. Housing Policy Debate 29(1): 44–61.

38.

Sampson

(2012) Great American City: Chicago and the Enduring Neighborhood Effect. The University of Chicago Press.

39.

Schwartz

McClure

Taghavi

(2016) Vouchers and neighborhood distress: The unrealized potential for families with housing choice vouchers to reside in neighborhoods with low levels of distress. Cityscape 18(3): 207–228.

40.

Sharkey

(2013) Stuck in Place: Urban Neighborhoods and the End of Progress toward Racial Equality. The University of Chicago Press.

41.

Smith

Popkin

George

, et al. (2015) What happens to housing assistance leavers? Cityscape 17(3): 161–192.

42.

Tighe

Hatch

Mead

(2017) Source of income discrimination and fair housing policy. Journal of Planning Literature 32(1): 3–15.

43.

Varady

Walker

(1999) Vouchering out distressed subsidized developments: Does moving lead to enhanced feelings of safety? Environment and Behavior 31(1): 3–27.

44.

Varady

Wang

, et al. (2010) The geographic concentration of housing vouchers, Blacks, and poverty over time: A study of Cincinnati, Ohio, USA. Urban Research & Practice 3(1): 39–62.

45.

Walter

Acolin

Wang

(Vince) , et al. (2024) Exploring the association between household compositional change and mobility of subsidized householders in the United States: A life course perspective. Journal of Urban Affairs. Epub ahead of print 29 July 2024. DOI: 10.1080/07352166.2024.2371380.

46.

Walter

Atherwood

(2015) Moving to opportunity? An examination of housing choice vouchers on urban poverty deconcentration in South Florida. Housing Studies 30(7): 1064–1091.

47.

Wang

(2018) Tracking “choice” in the Housing Choice Voucher Program: The relationship between neighborhood preference and locational outcome. Urban Affairs Review 54(2): 302–331.

48.

Wang

Walter

(2018) Tracking mobility in the housing choice voucher program: A household level examination in Florida, USA. Housing Studies 33(3): 455–475.

Entering and leaving housing assistance: Neighborhood trajectories of housing voucher recipients in the United States

Abstract

Keywords

Introduction

Literature review

Housing assistance and neighborhood effects

Housing vouchers and neighborhood access

Entering and exiting housing assistance

Data and methods

Results

Neighborhood trajectories

Fixed-effect model

Discussion

Key findings

Limitations and future research

Conclusion

Footnotes

Appendix

Acknowledgements

ORCID iD

Funding

Declaration of conflicting interests

Disclaimer

Notes

References