Reducing Intergenerational Poverty (2024)

most micro level of housing quality to the more global issue of homelessness, which is the product of many factors related to housing.

Housing quality is the factor most closely linked with children’s outcomes, representing the physical character and maintenance of the housing unit. Housing crowding refers to the number of people within a housing unit, which is relevant for things such as noise levels and study spaces. Housing stability and tenure are tightly related, since homeowners tend to move less frequently than renters, and frequent moves can be disruptive. Housing affordability affects the availability of household financial resources to pay for children’s development. Finally, a lack of housing, or homelessness, reflects critical housing needs that make other investments in children more difficult. While there are consistent findings about the importance of these factors for children’s long-term outcomes, much of the research on housing is correlational in nature.

Housing Quality

Low-income households are more likely to experience “inadequate housing,” defined by the federal government as housing “with severe or moderate physical problems, including plumbing and heating deficiencies; rodent and cockroach infestations; and structural issues such as cracks and holes in walls and ceilings, water leaks, broken windows, and crumbling foundations”¹ (Lew, 2016; U.S. Department of Housing and Urban Development, 2014). Figure 8-1 shows the distribution of those experiencing inadequate housing by poverty status and race/ethnicity. Households with cash incomes below the official poverty line are more than twice as likely as nonpoor households to experience inadequate housing, and Native American, Black, and Latino households are disadvantaged relative to White households. Homeownership, which we discuss in greater detail below, is related to better housing quality (Haurin et al., 2002). Exposure to household and environmental toxins (e.g., mold, asbestos, dust mites, lead, smoke, heavy metals, nitrogen dioxide, fine particulate matter) is also greater for low-income and racial/ethnic minority groups (Adamkiewicz et al., 2011; Hauptman et al., 2021; Vivier et al., 2011). Further evidence on lead exposure and its effects is provided in the neighborhood section below and in Chapter 5, which covers children’s health.

Housing quality is positively correlated both with children’s health and with cognitive and social development (e.g., Barros et al., 2018; Coley

___________________

¹ For full typology, see American Housing Survey, 2019, Appendix A: Subject Definitions, and Table Index, A-15-16 (U.S. Department of Housing and Urban Development and U.S. Census Bureau). https://www2.census.gov/programs-surveys/ahs/2019/2019%20AHS%20Definitions.pdf

et al., 2013; Dunn, 2020; Leventhal & Newman, 2010). Exposure to household lead, pollutants, and allergens harms children’s health and development (Huang et al., 2021; Maciag et al., 2022; and see Chapter 5). The “Healthy homes” policy and grant initiatives at the Department of Housing and Urban Development (HUD), the Centers for Disease Control and Prevention, and the U.S. Department of Agriculture all recognize the relationship between household quality indicators and children’s well-being. Researchers and healthy housing advocates have identified a number of interventions—such as air filters, education on pollutants, remediation, and smoking bans—that can improve children’s health (Butz et al., 2011; Morgan et al., 2004; National Center for Healthy Housing, 2009; Reddy et

al., 2017). However, there are no studies that establish a direct causal link from housing quality during childhood to adult socioeconomic outcomes, and the indirect pathways are based on correlational studies.

Housing Crowding

Moving from the unit itself to the density of people who reside within it, research on household crowding hypothesizes that in a given household, more people per room increases the household’s stress and interrupts children’s study, sleep, and play. Crowding is frequently defined as more than one person per room or per bedroom (Blake et al., 2007). Household crowding is more common among low-income and non-White households (Cross, 2018; Kunesh, 2021; Mateyka, 2015). Longitudinal studies have found negative short-term (Solari & Mare, 2012) and long-term (Conley, 2001; Lopoo & London, 2016) impacts of household crowding on educational attainment and on youth criminal convictions (Blau et al., 2019). Similarly, living with extended kin and non-kin adults during childhood is associated with lower high school graduation and college attendance rates (Harvey, 2020). However, sensitivity to ethno-racial diversity in household composition is important. Native American, Asian, Black, and Latino households are more likely than White households to be multigenerational and to include extended kin, and these arrangements may be beneficial for other outcomes (Cross, 2018).

Housing Stability and Tenure

The evidence suggests that housing stability—or low residential mobility—improves children’s long-term outcomes. According to the Joint Center for Housing Studies, “Low-income Americans are more likely to move…, with 14% of people in the bottom income quartile moving between 2017 and 2018, compared with 11% of those in the top income quartile” (Frost, 2020, p. 1). Frequent moves harm children’s educational attainment, health, and delinquency in short- and long-term studies (Metzger et al., 2015; Schmidt et al., 2018; Simsek et al., 2021). A large-scale population study in Norway using sibling fixed-effects models found weak but significant effects of residential moves on dropping out of high school, lower adult incomes, and early parenthood (Tønnessen et al., 2016).

Evictions are one driver of housing instability that has been shown to harm children and their families. Each year, there are six eviction filings nationally for every 100 renting households (Eviction Lab, 2018), and rates are higher among female-headed households, families with children, and Black and Latino families (Desmond, 2014; Desmond & Gershenson, 2017; Desmond et al., 2013). Families experiencing an eviction order are

at higher risk of homelessness, and have reduced earnings and consumption (Collinson et al., 2019; 2022). Studies have shown that evictions increase the odds of losing Medicaid coverage, with subsequent disruptions in health care access (Schwartz et al., 2022). They have also found that exposure to evictions in utero is associated with higher rates of preterm birth and low birthweight among infants (Khadka et al., 2020), and that greater neighborhood-level eviction rates are associated with higher rates of child maltreatment, particularly among older children (Bullinger & Fong, 2021). Families that are already vulnerable—for example, because of a child with special health needs—are more likely to be evicted (Schwartz et al., 2021). Conversely, during the COVID-19 pandemic state policies that protected renters from eviction were associated with improvements in renters’ mental health (Leifheit et al., 2021). While these studies do not establish causation, they suggest an important relationship between housing instability and negative impacts on children that have been shown to worsen later adult outcomes.

Home ownership, in contrast, increases residential stability, including for low-income families (Kull et al., 2016). This may be one explanation for the correlational research that finds positive short-term (Cordes et al., 2023) and long-term effects on children of living in an owned home (Aarland et al., 2021; Blau et al., 2019; Galster et al., 2007; Harkness & Newman, 2003; Rostad et al., 2019), although some research finds that differences may be due to selection into home ownership and are thus weaker than previously estimated (Cordes et al., 2023; Holupka & Newman, 2012). For low-income White families, home ownership seems to be beneficial for children’s outcomes, but this does not hold true for Black families (Holupka & Newman, 2012). It should also be noted that low-income homeowners are more likely to exit from home ownership than higher-income homeowners (Herbert & Belsky, 2008), thereby losing whatever benefits home ownership may have conferred on their children.

There are sizable disparities in home ownership rates by income and race/ethnicity (Figure 8-2). Households of all races with incomes below 50% of area median income (AMI) have lower home ownership rates than their counterparts further up the income distribution. Notably, White households with incomes below 50% of AMI are just as likely to be homeowners as Black households earning between 80% and 120% of AMI. Racial disparities in asset ownership and housing wealth may explain racial gaps in intergenerational mobility (Fox, 2016; Toney & Robertson, 2022), a topic explored further in Chapter 6. While there is some evidence to support home ownership as a housing intervention (as opposed to an asset intervention) to improve low-income children’s outcomes, there is no strong causal evidence. Indeed, the relevant operating mechanisms may be residential stability and housing and neighborhood quality, rather than ownership per

se. Furthermore, there are significant financial risks of homeownership for low-income households that may lead to greater stress and more residential moves for families (Bostic & Lee, 2008; Tyuse & Birkenmaier, 2006; Wainer & Zabel, 2020).

Housing Affordability

Low-income households are less likely than high-income households to be able to afford their housing. The standard measure of affordability is “housing cost burden,” or the condition of paying more than 30% (or 50%, defined as severe housing cost burden) of gross income toward housing. Figure 8-3 shows housing cost burdens for 2020 by income, home ownership

status, and race/ethnicity. Renters are more burdened than owners, low-income households are more burdened than higher-income households, and Black, Latino, and Asian households are more burdened than White households. Native American households experience slightly higher rates of housing cost burden than White households, but this fact is accompanied by significantly higher rates of overcrowding and lower housing quality in tribal areas (U.S. Department of Housing and Urban Development, 2017). Greater than 70% of households earning less than $30,000 a year pay more than 30% of their income toward rent or mortgages, and roughly half pay more than 50% of their income.

Limited research has been done on the direct impacts of housing affordability (without housing assistance) on short-run child outcomes (Holme,

2022), and to our knowledge no research has been done on later-life outcomes. Using propensity score matching and instrumental variable techniques, Newman and Holupka (2014) found an inverted U-shaped pattern for the effects of housing cost burden on children’s cognitive outcomes. That is, as housing becomes more expensive relative to families’ resources, children’s math and reading scores improve, but then they decline as housing cost burden exceeds 30% to 40%. Newman and Holupka (2015) find a similar inverted-U relationship between housing cost burden and parental investments in child enrichment. In other words, very low-cost and very high-cost housing relative to income seems to be associated with poor short-term child educational performance. The negative effect of low-cost housing may be due to the correlation of such housing with poor housing quality. Housing unaffordability may also contribute to other housing-related factors that affect children’s long-term outcomes, such as housing quality and stability (Holme, 2022). We found no studies on the direct relationship between housing costs and intergenerational poverty.

Homelessness

Being without housing entirely is the most severe form of housing deprivation. The 2020 Annual Homeless Assessment Report (U.S. Department of Housing and Urban Development, 2021) estimates the population living in unhoused families with children to be 171,575, or 30% of the total population without housing. This represents a decline from a peak of over 240,000 people in unhoused families with children in 2010. All nonWhite groups, with the exception of Asian Americans, are overrepresented among unhoused families relative to their proportion of the total U.S. population (see Table 8-1). Native American and Pacific Islander families

are overrepresented by a factor of more than seven. Point-in-time counts of people without any form of shelter underestimate the total number of families in precarious housing situations or who may experience homelessness over the course of the year. For example, the U.S. Department of Education reports that over the course of the 2019–2020 school year, 52,307 enrolled school children were unsheltered, 146,769 lived in shelters or other transitional housing, and 991,300 lived doubled-up with other families (National Center for Homeless Education, 2021, Table 3).

Correlational studies find negative effects of being unhoused during childhood on children’s education and health (Perlman & Fantuzzo, 2010; Sandel et al., 2018) and on high school graduation, adult employment, and the likelihood of being stably housed as an adult (Bassuk et al., 2014; Cobb-Clark & Zhu, 2017; Parpouchi et al., 2021). Most studies cannot distinguish between the unique effect of being unhoused from the general effect of growing up in a poor household (Buckner, 2008). However, an experimental policy treatment showed that giving unhoused families permanent housing reduced homelessness, food insecurity, the number of schools that children attended, school absences, and child behavior problems (Gubits et al., 2018). Many of these measures are correlated with later adult outcomes. For example, childhood food insecurity is related to poorer adult outcomes (see Chapter 5 for more detail), and school mobility correlates with lower educational attainment, lower occupational prestige, and greater likelihood of arrest in adulthood (Herbers et al., 2013). Hence, there is indirect evidence that reducing homelessness in childhood would increase the likelihood of upward mobility. Permanent housing subsidies have been shown in a randomized controlled trial (RCT) study to reduce family homelessness (Gubits et al., 2018); accordingly, we discuss housing choice vouchers later as an intervention to interrupt this pathway of intergenerational poverty.

Conclusion 8-1: The evidence on the effects of housing on intergenerational poverty is nearly all correlational or drawn from longitudinal panel surveys. The most consistent correlational evidence is on the effects of housing quality on children’s short-term outcomes, with the strongest evidence on the long-term effects of lead exposure. There is also correlational evidence on the negative effects of homelessness, overcrowding, residential mobility, and very low or high housing costs on children’s short and long-term outcomes.

NEIGHBORHOODS AS A DRIVER OF INTERGENERATIONAL POVERTY

As discussed in Chapters 2 and 3, the neighborhood environments of low-income children and non-White children differ significantly from those

of high-income and White children and are an important correlate of intergenerational mobility. Residential segregation by income has increased since at least the 1980s (Logan et al., 2020; Reardon et al., 2018), particularly among families with children (Owens, 2016). This means that struggling and affluent families are more likely to live in separate neighborhoods and more likely to live close to families of a similar socioeconomic status now than in the past.

The racial residential segregation of Black and White people has declined steadily since about the 1970s, but it is still high in many large cities. Residential segregation between Latino and White households has remained steady or increased slightly in some areas (Elbers, 2021; Logan, 2013). Because of residential segregation by both race and class, poverty is disproportionately concentrated in the neighborhoods in which nonWhite families and children live. Black and Native American children are more than seven times as likely and Latino children more than four times as likely as White children to live in neighborhoods with poverty rates of 30% or more (see Figure 8-4). The child poverty rate on Native American reservations was 42% in 2010 (Akee & Taylor, 2014). Children who grow up in high-poverty neighborhoods have worse adult outcomes than children living in low-poverty neighborhoods (Chetty et al., 2016).

Children in families with incomes below poverty and children living in low-income neighborhoods are disproportionately exposed to lead through air pollution, soil-based lead, leaded water pipes, and lead paint and dust in older housing structures (Hauptman et al., 2021; Vivier et al., 2011). Children in families below poverty experience greater negative cognitive and physiological effects from living in census tracts where the risk of lead is high (Marshall et al., 2020). As discussed in more detail in the chapters on health (Chapter 5) and crime and criminal justice (Chapter 9), childhood lead exposure is linked with greater delinquency in adolescence and young adulthood (Aizer & Currie, 2019; Manduca & Sampson, 2019; Wright et al., 2008) and with lower adult IQ, higher rates of teenage pregnancy, and declines in occupational status and income relative to parents (Reuben et al., 2017; Zhang et al., 2013; also see Boyle et al., 2021; Manduca & Sampson, 2021). While the magnitude of the impact seems to be the same for Black, White, and Latino children, the levels of exposure are much higher for Black and Latino children (Manduca & Sampson, 2021). More research on household and environmental exposures in Native American communities is needed (Barros et al., 2018).

Concentrated poverty also increases young people’s exposure to violence, which is negatively correlated with children’s educational, labor market, and delinquency outcomes (Beland & Kim, 2016; Burdick-Will

et al., 2021; Eitle & Turner, 2002; Nader et al., 1990; Sharkey, 2010) and reduces intergenerational mobility (Chetty et al., 2014; Sharkey & Torrats-Espinosa, 2017). We review this literature more fully in Chapter 9.

Experimental and quasi-experimental evidence demonstrates that exposure to “better neighborhoods” during childhood has a significant association with children’s outcomes in adulthood. But what is a better neighborhood? Experimental interventions to improve the physical quality of neighborhoods show that beautifying vacant lots and fixing abandoned homes significantly reduce crime rates (see Chapter 9). Just as housing quality seems to matter, so does the physical quality of the neighborhood. However, understanding which aspects of neighborhoods matter most—e.g., poverty levels, crime rates, labor markets, or residential segregation that

decreases resources for racial/ethnic minorities in particular—is challenging, as these factors are often highly correlated. As discussed in Chapter 2, a literature has emerged using geographic variation and following families that move across neighborhoods to understand what factors explain differences in intergenerational mobility across areas (Chetty, 2014; Chetty et al., 2020, 2022). Studies have found several systematic predictors of differences in intergenerational mobility, such as poverty rates, school quality, the degree of inequality, the fraction of children living with single parents, connectedness to high-income people, and historical factors such as redlining and Jim Crow laws. Because many of these factors can be interrelated, most studies therefore consider a single dimension of neighborhood disadvantage—often neighborhood poverty—or examine a composite measure of a few facets of disadvantage.

Much of the experimental evidence on the importance of neighborhood factors comes from HUD’s Moving to Opportunity (MTO) experiment, which began in the 1990s and has since been revisited by numerous researchers using contemporaneous data to measure participants’ long-run outcomes along a range of dimensions. This work has revealed that some of the children who moved to lower-poverty neighborhoods as part of MTO experienced positive effects not only on their average earnings and educational prospects (Chetty et al., 2016), but also on their physical and mental health as well (Pollack et al., 2021). Importantly, it was young children in particular whose long-run outcomes improved after they moved, suggesting that the age prior to adolescence may be a key point to intervene on the children’s neighborhood environment; similar benefits were not observed for older children or adults in this study.

These results have been replicated in many quasi-experimental and observational studies of children who move across areas at different ages using larger samples. Using de-identified tax records covering more than five million children whose families moved across counties between 1996 and 2012, Chetty et al. (2016) showed that, on average, children whose families move to a better neighborhood have better outcomes, and the beneficial effect increases as the amount of time they spend growing up in the better area increases. This pattern holds for a range of outcomes including earnings, college attendance, incarceration rates, and teenage birth rates. Similar findings were documented for children who moved out of severely distressed public housing projects in Chicago (Chyn, 2018).

Analogous exposure effects on educational and economic outcomes have been shown in datasets covering movers in international settings, including Australia, Canada, and Denmark (Deutscher, 2020; Faurschou, 2018; Laliberté, 2021). Researchers have observed similar patterns for health. A series of studies examined the effects of quasi-randomly assigned neighborhoods of arrival among refugee children and adults in Denmark

and Sweden (Foverskov et al., 2022a,b; Hamad et al., 2020; White et al., 2016) and found that relatively more disadvantaged neighborhoods were associated with poorer long-run physical and mental health outcomes.

Yet despite the importance of high-opportunity neighborhoods for children’s prospects of achieving upward mobility, the vast majority of low-income families in the United States, including those that receive housing assistance, live in neighborhoods with relatively low rates of intergenerational economic mobility (see Figures 2-8 and 2-9 in Chapter 2; Mazzara & Knudsen, 2019; Metzger, 2014). A recent randomized experiment conducted in Seattle showed that families stayed in low-opportunity neighborhoods largely because of barriers that impede residential choice, such as limited time and resources to search for housing, challenges in communicating with landlords, and a lack of information about neighborhood opportunities (Bergman et al., 2019).

Some interventions aimed at increasing neighborhood opportunity and reducing residential segregation have proved effective at helping low-income families access better neighborhoods regardless of their racial background (Bergman et al., 2019). Nevertheless, Native, Black, and Latino families currently disproportionately live in high-poverty, low-opportunity neighborhoods. Recent quasi-experimental work has demonstrated the adverse effects of residential segregation on intergenerational mobility, academic achievement, and teenage birth rates (Chyn et al., 2022). Expanding access to high-opportunity neighborhoods therefore has the potential to narrow racial disparities, although this will not eliminate them entirely, because disparities persist even within higher-opportunity neighborhoods (Chetty et al., 2020).

Conclusion 8-2: Strong evidence shows improvements in low-income children’s long-term economic, educational, and health outcomes when they move to less disadvantaged neighborhoods. Less is known regarding which characteristics of neighborhoods foster upward mobility.

HOUSING AND NEIGHBORHOOD INTERVENTIONS

The main public policies that address housing problems and neighborhood characteristics involve housing assistance. Housing assistance can be in many forms, such as subsidies for the construction and maintenance of housing units, which result in lower rents for low-income households, or vouchers issued to low-income households who use them in the private rental market to pay the difference between the asking rent (up to a certain threshold) and what they can afford, which is set at roughly 30% of their gross income. Housing assistance may improve low-income children’s outcomes as adults by freeing up parental income for investments in children,

by improving housing quality and stability, and by reducing homelessness. Access to neighborhoods that promote upward mobility also requires being able to afford the often-higher costs of housing in such neighborhoods.

We first discuss the intervention of housing choice vouchers (HCVs) with supports for moving to low-poverty neighborhoods, for which promising causal evidence of direct effects on long-term outcomes is available. Then we discuss housing choice vouchers alone, which shows promising evidence for improving children’s outcomes in the short term, specifically for unhoused families, but mixed evidence for addressing intergenerational poverty. (In Appendix C: Chapter 8, we give additional information on HCVs alone and housing assistance beyond vouchers, and we discuss the limited evidence on the effects of housing production, neighborhood improvement, and targeted initiatives for Native American families. Neighborhood interventions associated with reductions in neighborhood crime are reviewed in Chapter 9.)

As discussed in Chapter 1, we characterize the evidence on some of the programs or policies as “strong” and denote them with an “*.” In the case of our direct-evidence HCV program idea, the evidence is “promising” but not “strong” because this program has not been scaled up and tested in other cities.

Policy and Program Ideas Based on Direct Evidence

Enhancing the Housing Choice Voucher Program

The HCV program allows low-income households to lease an apartment in the private market using subsidy funds allocated by the federal government to the local public housing authority. The HCV program ensures that households pay no more than 30% of their income toward housing costs. The evidence that an enhanced version of the HCV program generates positive later-life outcomes for children is robust. Specifically, pairing the HCV program with enhanced and customized mobility services is a key evidence-based intervention for increasing mobility out of poverty. The Family Stability and Opportunity Vouchers Act (FSOVA) is one example of bipartisan, evidence-based legislation aimed at this goal. Originally introduced in 2019, and reintroduced in 2021, the FSOVA bill would expand rental assistance to 500,000 families with young children. In addition to providing a 25% increase over the current 2.2 million families served annually in the HCV program, the bill also includes provisions aimed at bolstering low-income families’ access to higher-opportunity neighborhoods. The legislation would provide families with access to customized counseling and case management services designed to overcome barriers to residential choice. It would also provide housing authorities with new resources to

engage landlords in the HCV program and to partner with community-based supports to support the moving process. Additionally, the FSOVA bill includes provisions for voucher programs to reduce administrative burdens and to employ “small-area fair market rents”² to align vouchers with the cost of living in specific neighborhoods.

The FSOVA bill is built on evidence of positive outcomes that result from opportunity moves. Its design was based on research-backed practices from the Creating Moves to Opportunity (CMTO) program, a housing mobility intervention in the Seattle area (Bergman et al., 2019). Run as an RCT, CMTO showed that it was barriers, not preferences, that caused the majority of HCV participants to reside in low-opportunity areas. The CMTO program, which provided customized services to families along with flexible financial assistance, resulted in a dramatic increase in opportunity moves: 53% of families assigned to receive CMTO services moved to high-opportunity areas, as compared with 15% of families in the control group (who also received a voucher but without additional supports). The annual program cost was $2,700 per family—an amount that would be offset by anticipated lifetime earnings increases and other positive outcomes among participants who moved to high-opportunity neighborhoods. The elements included in the FSOVA bill, such as a streamlined search process, landlord mediation, and customized support from housing navigators are all based on program elements shown to be effective in CMTO.

Although expanding enhanced voucher assistance meets the committee’s criteria for being supported by direct evidence, it is not yet supported by evidence that has been replicated across several sites. As explained in Chapter 1, we therefore characterize its supporting evidence as “promising” rather than “strong.”

• Expand and enhance choice-based residential mobility assistance. Expand the housing choice voucher program’s rental assistance to an additional 500,000 families with young children (at an estimated cost of $5 billion) and couple it with customized counseling and case management services to facilitate low-income families’ access to higher-opportunity neighborhoods.

___________________

² Small Area Fair Market Rents are Fair Market Rents (FMR) calculated for ZIP Codes within metropolitan areas. The use of Small Area FMRs is expected to give HCV tenants access to areas of high opportunity and lower poverty areas by providing a subsidy that is adequate to cover rents in those areas, thereby reducing the number of voucher families that reside in areas of high poverty concentration. https://www.huduser.gov/portal/datasets/fmr/smallarea/index.html; https://www.hudexchange.info/programs/public-housing/small-area-fair-market-rents/