Mobility Lessons from Gautreaux
and Moving to
Greg
J. Duncan & Anita Zuberi
Research on the impacts of the Gautreaux
residential mobility program was first conducted nearly two decades ago.
[1]
The purpose of this article is to provide an
update on Gautreaux lessons based on a new wave of Gautreaux research. The recent research provides a much
longer-run picture of residential and personal outcomes, and draws its data
from administrative records rather than surveys. We also provide some comparison between
results from Gautreaux and those from its sister program, Moving to Opportunity
(MTO). We conclude with some thoughts
about policy implications based not only on residential mobility research but
also from evaluations of more general work-support programs that have been
conducted in the past decade.
Gautreaux One is the name we give to the
original Gautreaux program that moved thousands of
The Moving to
One important way in which the Gautreaux program
designs differ from MTO’s is that the Gautreaux programs were part of a legal
settlement involving racial discrimination and designed to provide families
living in highly segregated neighborhoods of concentrated poverty in Chicago
the opportunity to move to much better neighborhoods, where “much better” was
defined as more racially integrated.
[2] In contrast, Moving to
Program
evaluation. In
evaluating the three programs, it is crucial to understand the nature of the
comparisons that are being made. The
potential impact of a housing mobility program can be defined by the difference
between how a family provided with the program’s mobility opportunities fared relative to what would have happened to that
family had it not been given those opportunities. Since MTO’s evaluation design is based on
random assignment, it is well-suited for estimating program impacts. When families living in public housing in
MTO’s five cities were recruited for the program, they were told that they
would be in a lottery in which they had a one-in-three chance of not being able
to move in conjunction with the program.
[6] Essentially, a coin was then flipped,
determining which families became eligible for the program and which did
not. As a result, the fortunes of MTO
control group families can be tracked alongside those of families in the
program group.
There is no control group in the Gautreaux One
program; research studies have only been able to study subgroups of families,
all of whom moved in conjunction with the program. While the Gautreaux Two study follows families
who both did and did not take up the program’s offer of mobility assistance,
the two groups were formed by self-selection rather than random
assignment. The study designs of
Gautreaux One and Two thus limit their capacity to answer the important policy
impact question, although both are useful in addressing questions regarding
program design and the role of the characteristics of destination
neighborhoods.
Although control-group comparisons are not
possible with Gautreaux One, a great deal of research has come out of the
program.
[7] The best-known research studies compare the
roughly half of program families placed in neighborhoods in the city of
Long-run
Gautreaux results.
What have we learned from the more recent look at Gautreaux One? Its most stunning success is revealed by
tracking families’ addresses in the late 1990s, up to 22 years after their
original moves. With two-thirds of
Gautreaux One families placed in the suburbs still residing in the suburbs,
there is great persistence in the residential successes of the families that
moved as part of the program.
[11] Even more impressive is the program’s intergenerational residential successes:
the children placed in the suburbs with their mothers, but old enough to be
living on their own by the late 1990s, continue to reside in neighborhoods that
have lower poverty rates, have higher rates of educational attainment and are
more integrated than the ones they originally lived in.
[12] Figures 2 and 3 provide a more detailed look
at the residential fortunes of Gautreaux One mothers and children. On average, these families came from very
poor neighborhoods, with census tract poverty rates averaging 40%, the lower
boundary commonly used to define concentrated-poverty or “ghetto”
neighborhoods.
[13]
On average, Gautreaux One cut these neighborhood
poverty rates by more than half – to 17%.
[14] As of the late 1990s, both the mothers and
the children were living in communities with poverty rates that were similar to
their placement addresses – 16% for mothers
[15]
and 18% for their grown children.
[16] The persistence of the residential success of
families moving in conjunction with the Gautreaux One program is indeed
striking. Looking to the racial
composition of the neighborhoods (Figure 3), Gautreaux One families began in
highly segregated settings: on average 87% of their neighbors in their origin
communities were African American. The
program placed its families in communities that reduced this percentage by
two-thirds, to 30% African American.
[17] In contrast to rates of neighborhood poverty,
the subsequent moves of Gautreaux One families were to somewhat different
neighborhoods – on average, they contained a fairly even balance of African
Americans and individuals from other races.
[18] Almost none of the Gautreaux One mothers and
children moved anywhere close to their original neighborhoods.
[19]
Gautreaux Two began in 2002, so it is much too
early to tell what long-run residential successes and failures these families
will experience. But the early returns
are not as promising as the long-run results from Gautreaux One (Figure 4).[20] At 13%, the average poverty rates in
Gautreaux Two families’ placement neighborhoods were lower than those for
families moving in conjunction with Gautreaux One, but roughly half of the
Gautreaux Two families have already moved on, with the movers moving to
neighborhoods with poverty rates that averaged 27%.
Thus, it appears that the subsequent moves are
undoing some of the initial advantages of the original moves. This is true with respect to race as well
(Figure 5).
[21] On average, Gautreaux Two families came from
communities in which 80% of their neighbors were African American and were
placed in neighborhoods in which 11% of their neighbors were African
American. But those moving again have
moved to neighborhoods averaging 61% African American. All in all, it appears that the persistence
of the residential success enjoyed by Gautreaux One families may not hold for
Gautreaux Two families.
MTO
results. The MTO evaluation gathered far more extensive
information about the personal characteristics of family members than Gautreaux
One researchers have been able to measure about families participating in its
program. MTO experimental and control-group families were tracked for five
years and then questioned extensively about their personal, family and
neighborhood situation. Although control families were not given assistance in
moving to low-poverty neighborhoods, many of them did move. Given their
interest in this program, it is hardly surprising that most control families
moved to somewhat better neighborhoods, but their new neighborhoods had far
higher poverty rates than the placement neighborhoods of the MTO families that
moved in conjunction with the program.
Neighborhood safety improved as well for the experimental group.
[22]
The most striking successes for MTO were in its
inventory of mental health measures.
[23] Considerably fewer MTO mothers were depressed
than control mothers.
[24] The reduction in depression was similar in
magnitude to what has been observed in clinical studies testing best-practice
depression treatment regimens.
[25] This substantial improvement in mental health
is not so surprising given what these families sought with the MTO program in
the first place. Almost all families
enrolling in MTO reported that their primary motivation in signing up for the
program was moving away from violent, gang-ridden neighborhoods to safer ones.
[26]
Moving to neighborhoods with less crime and violence, which satisfies families’
major program desires, is likely linked to the striking improvements in mental
health.
[27]
Thus, by the criteria of what mattered most
to the participants themselves, MTO was very successful.
MTO was much less successful in promoting
self-sufficiency. Program designers
hoped that MTO moves would improve mothers’ employment prospects, reduce their
welfare dependence, and boost the school achievement of children.
[28] On these counts, the results have been
disappointing.
[29] MTO families were no more likely to be
employed, earned no more and received welfare no less often than families
assigned to the control group.
[30] Here it is important to remember that MTO’s
evaluation took place in the late 1990s, in the midst of welfare reform and a
booming economy.[31] MTO families did indeed boost their
employment and reduced their reliance on welfare,
[32]
but these changes were no different, on average, from those experienced by the
comparison group.
Early evidence from Gautreaux One conducted by
James Rosenbaum and his colleagues compared maternal employment rates for city
and suburban movers and found substantial differences.
[33] But our longer run look at employment and
welfare receipt, drawing on data from the late 1990s, showed no employment
advantages to being offered housing in a suburban as opposed to a city
neighborhood.
[34] Mendenhall et al. (forthcoming) demonstrate
the importance of a finer-grain distinction among types of placement
neighborhoods, with the lowest maternal earnings associated with placement
neighborhoods with fewer whites and less favorable employment and welfare
outcomes for the one-fifth of families placed in high poverty, largely African
American neighborhoods.
[35] Overall, the most recent evidence from MTO
and Gautreaux One does not replicate the dramatic employment differences
observed in the original Gautreaux research.
Also disappointing, and this is also better
established in MTO than Gautreaux, is that moves to better neighborhoods did
not boost children’s school achievement.
[36] Part of the reason is that MTO-related moves
improved neighborhood quality much more than school quality. Families that moved in conjunction with the
program did not always attend a local school; all five MTO cities had citywide
school districts with school choice programs.
Despite the greater distance, some MTO families stuck with their
familiar original schools.
[37] Others appeared to pick schools that were
close to relatives who might provide after-school care.
[38] Whatever the reason, it is clear from the
experiment that there was essentially no impact of MTO moves on school
achievement. Even younger children, who
were preschoolers when they moved in conjunction with the program, did not
improve their school achievement relative to their control-group counterparts.
[39] In the case of delinquent behavior, MTO
appeared to benefit girls modestly.
[40]
In contrast, it increased delinquent behavior on the part of boys in families
that moved as part of MTO compared with boys in the control group.
[41]
Summary
and conclusion.
Drawing policy implications from Gautreaux One and MTO requires us to
weigh the considerable successes with the disappointments. In terms of why participants were interested
in the program, both programs succeeded admirably. Families successfully fled their violent,
gang-ridden neighborhoods, and their mental health improved. But in terms of what many policy makers were
hoping for – more work and earnings, greater independence from welfare,
second-generation successes – the programs were not as successful. Merely changing neighborhoods, even changing
to much better neighborhoods, does not produce the kind of achievement-oriented
successes in either generation that many had hoped for, although Gautreaux One
children were able to sustain the residential successes of their mothers.
Perhaps mobility programs need to go beyond
merely placing families in better neighborhoods and provide them with needed
family and personal services and supports.
But here it makes sense to think more broadly, since low-income families
involved in mobility programs are not the only ones in need of these kinds of
family-based supports.
An example of such a broader program is a
Milwaukee-based work support program called
An important policy issue addressed in neither
Gautreaux nor MTO research is their impact on receiving neighborhoods. Designers of both programs worried that
individuals living in the receiving neighborhoods might be subjected to higher
crime rates and/or lower property values.
On the other hand, there might be potential spillover benefits to these
moves if, for example, moving minority families into white neighborhoods helps
promotes racial tolerance and understanding.
As
we think about these three different programs – Gautreaux One, MTO and
Gautreaux Two – we need to be mindful of cohort and housing market differences.
Can the residential successes of Gautreaux One be realized today? Families moving in conjunction with Gautreaux
One faced very different circumstances than families moving in conjunction with
MTO and Gautreaux Two. Most Gautreaux
One families were first or second generation residents of
At
the same time, it is crucial to realize the formidable opportunities provided
by the residential mobility that follows on the heels of neighborhood
revitalization efforts. Gautreaux One proves that families – both adults and
their children – placed in affluent, integrated and much safer neighborhoods
are able to build new lives for themselves and maintain these residential
successes. MTO and early results from
Gaureaux Two suggest that neighborhood changes may be necessary but not
sufficient conditions for improvement; many families in these more recent
programs appear to need other supports in order to be truly successful.
[1]
For a comprehensive review of the
history of the Gautreaux lawsuit, the
implementation of the program, and early research results, see generally Leonard S. Rubinowitz & James Rosenbaum,
Crossing The Class And Color Lines: From Public Housing To White Suburbia (2000). See
also James E. Rosenbaum & Julie E. Kaufman, The Education and Employment of Low-Income African-American Youth in
White Suburbs, 14 Educ.
Evaluation & Pol’y Analysis 229 (1992) and Susan J. Popkin, James E. Rosenbaum & Pamela M. Meaden, Labor Market Experiences of Low-Income Black
Women in Middle-Class Suburbs: Evidence from a Survey of Gautreaux Program
Participants, 12 J. of Pol’y
Analysis & Mgmt. 556, 573
(1993) for specific details on early employment and educational outcomes,
respectively.
[2] The Gautreaux
One program was designed to place families in neighborhoods with less than 30%
Black residents. See Rubinowitz & Rosenbaum, supra
note 1, at 25. In 1981, program rules
changed so that families could move to neighborhoods with more than 30% Black
residents if the neighborhoods were deemed to be revitalizing.
[3] See Larry Orr et al., Dept. Hous. and
Urban Dev., Executive Summary of Moving to
Opportunity: Interim Impacts Evaluation
v (2003), http://www.huduser.org/Publications/pdf/MTOFullReport.pdf.
[4]
[5]
[9] See Micere Keels et al, Fifteen
Years Later: Can Residential Mobility Programs Provide a Long-Term Escape from
Neighborhood Segregation, Crime, and Poverty?, 42 Demography 51 (2005);
Ruby Mendenhall et al, Neighborhood
Resources, Racial Segregation, and Economic Mobility: Results from the
Gautreaux Program, Soc. Sci. Res. (forthcoming
2006).
[10] Mendenhall et
al, supra note 9 (manuscript at 6, on
file with the Northwestern Journal of Law and Social Policy).
[11] Keels et al, supra note 9, at 60.
[13] See Keels et al, supra note 9, at 58. Poverty
rates in
[14] Keels et al, supra note 9, at 62.
[15]
[16] Keels,
supra note 12, at 15.
[17] Keels et al, supra note 9, at 54.
[18]
[19]
[20] All of the
data on Gautreaux Two mobility rates and neighborhood characteristics presented
here were calculated by Micere Keels using program records and census tract
information.
[21] See supra note 20.
[22] Orr et al, supra note 3, at ix.
[24] Kling et al, supra note 23, at 20.
[25]
[26] Orr et al, supra note 3, at ix.; see also
Jeffrey R. Kling et al, Bullets Don’t Got
No Name: Consequences of Fear in the Ghetto, in Discovering Successful Pathways in Children’s Development: Mixed
Methods in the Study of Childhood and Family Life (Thomas S. Weisner ed., 2005).
[27] See Kling et al, supra note 26, at 15.
[28] Orr et al, supra note 3, at xi.
[29] Id. at xii-xiii; Kling et al, supra note 23, at 27-28; see also Jeffrey R. Kling & Jeffrey
B. Liebman, Experimental Analysis of
Neighborhood Effects on Youth (Nat’l
Bureau of Econ. Res., Working Paper No. 483, 2004).
[30] Orr et al, supra note 3, at xii-xiii; Kling et al, supra note 23, at 27-28; see
also Kling & Liebman, supra
note 29.
[32]
[34] Mendenhall et al, supra
note 9.
[35]
[36] Lisa
Sanbonmatsu et al, Neighborhoods and
Academic Achievement: Results from the Moving to Opportunity Program i
(Nat’l Bureau of Econ. Research, Working Paper No. 11909, 2006).
[37] Susan J. Popkin et al,
[38]
[39] Sanbonmatsu et
al, supra note 36, at 27.
[40] See Jeffrey R. Kling et al, Neighborhood Effects on Crime for Female and
Male Youth: Evidence from a Randomized Housing Voucher Experiment, 120 The Q. J. of
Econ.
87, 105-127 (Feb. 2004).
[43] Bos et al, supra
note 42.
[44] David P.
Varady & Carole C. Walker, Vouchering
Out Distressed Subsidized Developments: Does Moving Lead to Improvements in
Housing and Neighborhood Conditions?, 11 Housing Pol’y Debate 115, 138 n.30
(2000).