Gautreaux mothers and their children: an update

Abstract

The Gautreaux program was one of the first major residential mobility programs in the United States, providing low-income black families from public housing with opportunities to relocate to more affluent white neighborhoods in the Chicago suburbs and in other city neighborhoods. This paper reviews the most recent research on the Gautreaux families, which uses long-term administrative data to examine the effects of placement neighborhoods on the economic and social outcomes of mothers and children. We find that both Gautreaux mothers and their now-grown children were remarkably successful at maintaining the affluence and safety of their placement neighborhoods. As to the long-run economic independence of the mothers themselves, however, the new research fails to confirm the suburban advantages found in past Gautreaux research, although it does show that these outcomes were worst in the most racially segregated placement neighborhoods. With regard to the criminal records of Gautreaux children, it is found that suburban placement helped boys but not girls. Based on these results, we review possible new directions for successful mobility programs.

Keywords:

• neighborhoods
• mobility programs
• families
• policy

Acknowledgments

The authors would like to thank the following organizations for their generous support of the recent Gautreaux research summarized in this article: the Foundation for Child Development, the National Institute for Child Health and Human Development (U01 HD30947-06), the Spencer Foundation, the National Academy of Education, the MacArthur Foundation, the Institute for Policy Research and Graduate School at Northwestern University, and the National Opinion Research Center. James Rosenbaum and Phil Tegeler provided helpful comments on an earlier draft.

Notes

¹See Rosenbaum and Rubinowitz (2000) for a comprehensive review of the history of the Gautreaux lawsuit, the implementation of the program, and early research results.

²The collection and preparation of the data used for the studies executed from 2000– 2005 were done as a collaborative effort involving: Stefanie DeLuca, Greg J. Duncan, James E. Rosenbaum, Ruby Mendenhall, and Micere Keels. Various authors collaborated on each study, as cited in the references.

³A new round of the Gautreaux program began in 2002. Thus far, it has moved hundreds, rather than thousands, of families. Not enough time has elapsed for us to be able to assess long-term outcomes in Gautreaux Two.

⁴This practice began in 1981. While the program did not intend for these families to move to very segregated, high poverty neighborhoods, our Census-based analysis shows that some did. This unintended consequence may have resulted from the use of data from earlier census periods to characterize improving neighborhoods. For example, many Southside Chicago neighborhoods might have met the criterion of demonstrating economic revival according to the 1980 Census figures. However, if a family was placed in the late 1980s or early 1990s, these neighborhoods might have undergone demographic changes that were not represented by the previous Census figures.

⁵With 95% of Aid to Families with Dependent Children (AFDC) families having four or fewer children, Gautreaux's overcrowding restriction would eliminate only a few of the many Gautreaux families who were eligible for AFDC (US Department of Health and Human Services 1986). All three criteria reduced the eligible pool by less than 30%. Popkin (1988) found that the Gautreaux participants and the sample of AFDC recipients were similar in the length of time (seven years on average) spent on AFDC; however, the welfare group had more women who were second-generation recipients. In terms of marital status, the groups were again similar: 45% never married and 10% were currently married. The two groups also differed with respect to levels of education and age. Thirty-nine percent of the Gautreaux women dropped out of high school compared to 50% of the AFDC sample. The Gautreaux participants were slightly older (median age of 34 vs. 31).

⁶Although only about 20% of the eligible applicants ended up moving through the program, self-selection does not appear to have played a big role in program take-up (Peterson and Williams 1995). Rather than opting out of the program, most non-moving families were not offered a housing unit and thus not given the chance to participate. Housing counselors were forbidden by the consent decree from making offers selectively among eligible families, and there is no evidence that they did so.

⁷Families placed in the suburbs came from slightly more advantaged neighborhoods than the city movers, with the former originating from neighborhoods with higher census-tract family incomes, rates of employment, and education (see Keels et al. 2005, Table 1 for details).

⁸Our original sample consists of 1506 female-headed households. Through our various geocoding and address matching attempts, we were able to obtain geocoded addresses for all but 57 households (reducing the sample to 1449). For 145 cases the address was from 1994 or earlier. We judged these addresses to be too old and deleted these cases from further analyses (reducing the sample to 1304). Another 129 addresses were out-of-state and therefore precluded comparisons within the same metropolitan area, so we omitted these (reducing the sample to 1175).

⁹The crime data were obtained from two sources, the Chicago Police Department's records and the FBI uniform crime reporting data (UCR) for the suburban addresses. For addresses in the city of Chicago, we used data from the Chicago Police Department's annual reports. These data are compiled for the 25 Chicago police districts. We used census tract boundaries of each district to match city addresses to the police district via census and police maps. The UCR data are collected by a “reporting agency,” which can be a town or a city. We matched the suburban addresses to these towns and cities by zip code.

¹⁰Neighborhood poverty rates averaged 61% among Gautreaux participants that resided in public housing at the time of enrollment. Half of Gautreaux families were on the waiting list for public housing and lived in somewhat lower poverty neighborhoods.

¹¹Unless otherwise noted, all of the current versus origin neighborhood differences are statistically significant at the 5% level or less.

¹²This average masks substantial changes among participants placed in the lowest and highest fifth of neighborhood affluence. Participants placed in the lowest income neighborhoods experienced an increase of $24,700 in moving from placement to current residence, while subsequent moves for participants placed in the highest income communities resulted in an $11,100 decrease in neighborhood income.

¹³Unsurprisingly, there was much less change in the racial composition of the current neighborhoods of Gautreaux families who were placed in the city of Chicago. The current neighborhood percent black for these movers is 79%, which is close to the 83% black of their origin neighborhoods.

¹⁴After regression adjusting for background factors and city versus suburban placement, the probability of ending up in a relatively integrated neighborhood (30% African American or lower) is higher for those families who had been placed in low black (2.1– 5%) neighborhoods than it is for all placement neighborhoods. Holding city location at placement constant, families who moved to the highest black neighborhoods (50%) are significantly less likely to reside in integrated neighborhoods years later. These differences are statistically significant. (see DeLuca and Rosenbaum (2003) for details).

¹⁵The data covered the period from 1990 to 1992. Demographic eligibility for the program was determined by projecting the ages of children present in families at the time of their original Gautreaux moves. Estimations employed Tobit regression models and included a host of controls for family and origin neighborhood characteristics.

¹⁶Previous research by Rosenbaum and DeLuca (2000), found that both tract racial composition and the education level of tract residents in Gautreaux placement neighborhoods was related to later welfare receipt, but that the education level retained significance after both were included in the models. These results differ slightly from the more recent work presented here, largely due to differences in the way the dependent variable was defined. For example, the data for this previous study focused on AFDC receipt at one point in time, did not include neighborhood crime rates, and did not factor AFDC eligibility into spells of receipt. The independent variables also differ across these studies, since Rosenbaum and DeLuca (2000) use single Census tract variables to characterize neighborhoods, and Mendenhall, DeLuca, and Duncan (2006) use race and a neighborhood resource indicator (see footnote 16).

¹⁷We used tract-level measures of residents' education, median family income, crime rates, and male unemployment as proxies for the neighborhood level processes and resources commonly associated with child and family outcomes. The “neighborhood resource” variable was created by subtracting standardized measures of crime level and male unemployment from mean family income and the percent of individuals with a college degree or more in a tract. We used the average of the composite value.

¹⁸Keels (2006b) and Mendenhall (2004) randomly selected Gautreaux participants to interview based on the following characteristics/criteria: racial composition of placement communities (75% or higher black and 30% or lower black), residence in the city or suburbs, age of their children at the time of move, and birth cohort (born after 1951). They were thus able to interview women who represent the full range of Gautreaux participants: city movers to high black areas, city movers to low black areas, and suburban movers to low black areas.

¹⁹Research analyzing interviews conducted with mothers closer to the time of move (in the late 1980s) suggests additional ways that new suburban environments improved employment prospects, including an increased sense of self-efficacy and assistance from neighbors with child care and transportation (Rosenbaum, DeLuca, and Tuck 2005; Rosenbaum, Reynolds, and DeLuca 2002).

²⁰This is a sample of 273 children and draws residential information an average of 14 years after children moved to their placement community. Eighty-four percent of these children continued to reside within Illinois. None of these now-adult children resided in the same apartments or houses into which they were initially placed and only nine continued to reside at the same address as their mother.

²¹Based on logistic regressions (not presented here), year of move was the only factor significantly associated with whether children had complete social security number information. Families moving later in the program's operation were more likely to have provided children's social security number information, a result of the program's imposition of strict verification requirements in response to increasing demand for the program. Importantly, none of the origin or placement neighborhood Census characteristics was significantly associated with whether social security information was missing.

²²As with our discussion of Figure 1, all noted differences at statistically significant at the 5% level or below.

²³It is important to bear in mind that Gautreaux participants were placed in thriving middle-class suburban communities, not suburban areas characterized by high rates of poverty and segregation (see Harris 1999b).

²⁴Despite demographic similarities, Gautreaux families do differ from other public housing families because they volunteered for the program. Thus, our findings generalize most readily to families voluntarily choosing to participate in residential mobility programs in which the decision to move to a new, more integrated, higher income community is left up to the family. As the Housing Opportunities for People Everywhere (HOPE VI) program and the transformation of distressed public housing projects continues, families are often involuntarily forced to move, and then often move to new segregated communities; involuntarily relocating families may not result in the same outcomes as those we found for Gautreaux participants.

²⁵Early results from the Thompson program in Baltimore suggest that receiving vouchers that target the income and racial composition of neighborhoods has helped hundreds of families lease-up in less segregated, low-poverty neighborhoods (DeLuca and Rosenblatt 2006). These current and former public housing residents are moving as part of the partial consent decree in the Thompson v HUD case (#95-309-D. MD.). They receive Housing Choice vouchers and must use them in census tracts in the Baltimore metropolitan area that have less than 30% African-American residents, less than 10% poverty, and less than 5% public housing residents.

²⁶Recent qualitative research on the MTO program has shed light on interesting mechanisms behind the differences in results for girls and boys. Popkin, Leventhal, and Weismann (2006) suggest that some of the benefits for girls might be the result of less sexual harassment and predation in low-poverty areas. Clampet-Lundquist et al. (2006) show that MTO parents supervise girls more than boys, and that boys are not as socially flexible as girls when facing new environments; they are more likely to demonstrate old behaviors that will get them into trouble in lower poverty neighborhoods.

²⁷Boyd, Edin, Duncan, and Clampet-Lundquist (2006) conducted a qualitative study of a random sample of families participating in a new round of Gautreaux moves which began in 2002. They also find many reports of problems with landlords in placement neighborhoods.