Lottery winners and losers were both tracked over the ensuing years, and an important study last year by the Stanford economist Raj Chetty, with Nathaniel Hendren and Lawrence F. Katz of Harvard — a study I’ve previously written about — found that children who moved when they were young went on to enjoy substantially higher earnings than people of similar ages whose parents lost the lottery. (Another disclosure: Mr. Katz was my Ph.D. adviser.)
The random assignment of slots in this program means that we can be confident that these differences result from moving. But Mr. Chyn argues that this experiment substantially understates the importance of neighborhoods. The problem, he says, isn’t in comparing those who win the lottery with those who lose.
Rather, he argues that both the treatment and control groups had already partly inoculated their children against the effects of bad neighborhoods. Only a quarter of the families that were eligible for the lottery actually applied for it, and Mr. Chyn says the applicants were particularly motivated to protect their children from the negative effects of a bad neighborhood.
Indeed, qualitative and survey evidence from earlier research supports his conjecture. In-depth interviews by Jeffrey R. Kling, now with the Congressional Budget Office, and Jeffrey B. Liebman and Mr. Katz of Harvard revealed that these families “organized their entire lives around protecting their sons and daughters from the genuine dangers of ghetto life.” These mothers were “intensely focused on their children,” and as a result “younger children in particular were seldom allowed outside of the apartment, and never beyond the mother’s watchful gaze.”
The implication is that the housing experiment reveals the effect of moving from a bad neighborhood, for those least affected by the bad neighborhood.
In order to test his theory, Mr. Chyn would need to compare the results of a voluntary housing lottery with an experiment that essentially forces all families to enter the lottery.
Fortunately for Mr. Chyn, the demolition of many public housing projects in Chicago in the late 1990s effectively provides precisely this experiment. From 1995 to 1998, the Chicago Housing Authority demolished many high-rise public housing buildings, including the projects you might recognize from the 1970s sitcom “Good Times.”
These demolitions were effectively a lottery, because they led to the dislocation of some families — those whose buildings were demolished — but not those whose buildings were left standing. Thus those families that left did so for essentially random reasons. And just as in the Moving to Opportunity lottery, those families that “won” the lottery — that is, those families whose building was set for demolition — were offered a housing voucher to help them pay the rent if they moved out of the projects. Those families whose buildings remained standing were effectively a control group, as they continued to live in public housing, undisturbed.
Importantly, this real-world natural experiment differs from the Moving to Opportunity experiment, because all families could be pushed to move, not just those who volunteered for a government relocation program.
Mr. Chyn finds that these demolitions had very large — and very positive — effects on the children who were forced to move out of the projects. The children forced out of public housing went on to have annual earnings that were 16 percent higher than those who remained, and they were 9 percent more likely to be employed. Over all, being kicked out of public housing might add about $45,000 to each child’s lifetime earnings. The effects may be even larger for those who moved while they were young.
Around the same time, the Chicago Housing Authority also ran a small-scale lottery that gave the winners the same housing vouchers. Much like the Moving to Opportunity lottery, this one was optional, so only motivated public housing residents applied. Mr. Chyn’s analysis of this alternative experiment finds that it yielded much less impressive results, and the children whose families won the lottery went on to register roughly similar adult outcomes as those whose families lost.
The contrast is rather striking, suggesting that housing policies that also aim to help those who would not otherwise apply may yield a much larger bang for the buck.
Of course, there remain a number of open questions. Both the demolition study and the Chicago lottery study are hampered by their relatively small sample sizes. And there are many factors — including statistical chance — that might explain why the different experiments yield different effects.
But the underlying logic — that there’s good reason to think that a lottery may understate the true effects of housing programs — strikes me as sound. Moreover, Mr. Chyn has developed an elegant mathematical model to explore his logic, and it suggests that the effect of giving housing vouchers to a typical public housing resident may be many times larger than the effect on a lottery winner.
This important research also contains insights likely to extend beyond housing policy. For conservatives who are suspicious about the government’s ability to enact useful social policy, the study highlights the difficulty in targeting government programs to those who are most likely to benefit, rather than those most likely to seek them out. And for liberals, the logic that Mr. Chyn applies to housing suggests that the experiments used to evaluate other social policy interventions may understate the effectiveness that these programs could have when rolled out to a broader population.Continue reading the main story
- A core challenge of neighborhood effects research is distinguishing the role of individual and family circumstances from the effect of the neighborhood itself.
- Nevertheless, many neighborhood level indicators are linked to important outcomes for people residing in neighborhoods of concentrated poverty, including crime and delinquency, education, psychological distress, and various health problems.
- HUD’s Moving to Opportunity (MTO) demonstration found that those assigned vouchers restricted to low-poverty neighborhoods typically experienced better physical and mental health at the five- to seven-year followup than those with unrestricted vouchers or continued project-based assistance.
- Collective efficacy, as represented by measures of informal social controls, social cohesion, and trust, can help buffer communities against the negative effects of concentrated poverty.
Neighborhoods of concentrated poverty can isolate residents from the services and supports they need.
Neighborhoods of concentrated poverty isolate their residents from the resources and networks they need to reach their potential and deprive the larger community of the neighborhood’s human capital. Since the rise of inner-city poverty in the United States, researchers have sought to interpret the dynamic between neighborhood and residents in communities of concentrated poverty. Through articles and books such as The Truly Disadvantaged and When Work Disappears, sociologist William Julius Wilson has been a key figure in first popularizing the discussion of neighborhood effects. Wilson emphasizes that a “spatial mismatch” between increasingly suburban job opportunities and the primarily minority residents of poor urban neighborhoods has magnified other challenges, such as crime, the movement of middle-class residents to better neighborhoods, and a perpetual shortage of finance capital, stores, employment opportunities, and institutional resources.1 This combination of barriers creates communities with serious crime, health, and education problems that, in turn, further restrict the opportunities of those growing up and living in them. Wilson also consistently addresses the effect of family structure on the outcomes of residents in such communities, cautioning against both “culture of poverty” arguments and the assumption that individuals are helpless victims of racism.
As the study of neighborhood effects of concentrated poverty has developed, researchers have also confronted significant challenges. These hurdles include properly defining the boundaries between neighborhoods, conducting detailed longitudinal studies, and accounting for resident choice in neighborhood selection. Although technological advancements and increased research funding can address many of these challenges, distinguishing between neighborhood effects and family effects remains difficult. Researchers can control for basic family characteristics such as race, income, and education, but other, unobserved variables can result in either over- or understating neighborhood effects, which further complicates the interconnected nature of many neighborhood factors.2 As Margery Austin Turner, an expert in poverty research with the Urban Institute, tells EM:
The major question that continues to be asked is, does living in these places harm residents in and of itself? [Neighborhood effects are certainly not] the only factor; individual and family circumstances can overcome the effects of concentrated poverty but can also leave a family vulnerable. What is worrisome is that we don’t know enough about the interaction between vulnerable families and their neighborhoods. These families are the most likely to live in poverty areas but are also the most likely to have bad outcomes no matter where they reside. We need to learn more about the process by which a neighborhood transitions from low to high opportunity and, similarly, how that process influences individuals already affected by concentrated poverty.3
Working together to accomplish goals, strong neighborhood networks can lessen the effects of concentrated poverty.
Despite this limitation, researchers have found that for people residing in neighborhoods of concentrated poverty, a number of neighborhood level indicators are linked to important outcomes. Studies have illustrated that crime and delinquency, education, psychological distress, and various health problems, among many other issues, are affected by neighborhood characteristics. Thresholds, or tipping points, also prove important. In a recent review of research, Galster notes that studies suggest “that the independent impacts of neighborhood poverty rates in encouraging negative outcomes for individuals like crime, school leaving, and duration of poverty spells appear to be nil unless the neighborhood exceeds about 20 percent poverty, whereupon the externality effects grow rapidly until the neighborhood reaches approximately 40 percent poverty; subsequent increases in the poverty population appear to have no marginal effect.”4 Housing values and rents, key indicators of neighborhood decline, show a similar pattern. Using data from the 100 largest U.S. metro areas from 1990 to 2000, Galster, Cutsinger, and Malega find little relationship between neighborhood poverty rates and decline in neighborhood housing values and rents until poverty exceeds 10 percent, at which point values decline rapidly before becoming shallower at very high poverty levels.5
Several HUD studies have also contributed significantly to neighborhood effects research. One of these, the Moving to Opportunity study, has been a rare occasion to use random assignment, allowing researchers to better distinguish neighborhood effects from the selection bias that neighborhood choice creates.6 Volunteer families in five cities who lived in public or assisted housing were randomly assigned to one of the following groups:
- The treatment group: Participants received a voucher restricted to low-poverty census tracts and assistance in locating and moving to housing.
- The comparison group: Participants received a standard, geographically unrestricted voucher.
- The control group: Participants received continued project-based assistance.
After five to seven years, families who participated in the treatment group lived in better neighborhoods, and adults experienced better physical and mental health compared with the control group. Girls in these families showed significant mental health improvements, although boys may have fared worse.7 Despite these improvements, the MTO study has not shown gains in economic self-sufficiency, which was initially expected to be the primary outcome. Results of the final evaluation will be published by early 2011.”
Another key question in understanding the relationship between neighborhood and family effects is whether protective factors are family- or neighborhood-based. A major interdisciplinary study, the Project on Human Development in Chicago Neighborhoods, tested this through the concept of collective efficacy, a shared belief that a neighborhood’s residents can accomplish important tasks, such as preventing crime and delinquency, by working together in formal or informal neighborhood organizations. Communities that share expectations effectively and collectively exert social control over neighborhood conditions and behavior appear better able to counter the negative effects of concentrated poverty.
As a component of the project, which combined an intensive study of Chicago neighborhoods with coordinated longitudinal studies of randomly selected individuals, Sampson, Raudenbush, and Earls surveyed residents of neighborhood clusters on informal social controls, social cohesion, and trust.8 The researchers found that, even when accounting for factors such as personal characteristics, concentrated disadvantage, immigrant concentration, and residential stability, collective efficacy was strongly linked with decreased violence and weakened the relationship between violence and the neighborhood’s social composition. Such evidence supports the notion that, just as parents can buffer their children against the effects of violence and other negative outcomes, strong neighborhood networks can collectively lessen the effects of concentrated poverty. The project, which was jointly funded by the National Institute of Justice and the John D. and Catherine T. MacArthur Foundation, also explored neighborhood effects on health, crime, education, social processes, and other topics, yielding more than 100 publications.9
Choice Neighborhoods will coordinate with other placed-based programs to improve housing, education, communities, safety, and services in areas of concentrated poverty.
HUD recognizes the importance of creating neighborhoods of opportunity, and its Choice Neighborhoods initiative is designed to deconcentrate poverty and address the interconnected problems caused by living in neighborhoods of concentrated poverty. The initiative’s goal is to strengthen the underlying social structure of neighborhoods through competitive grants, which will encourage strong local partnerships and allow some funding flexibility to catalyze local improvement of key neighborhood assets. Choice Neighborhoods will ensure that HUD-assisted housing is financed and managed in a way that attracts a mix of uses, incomes, and stakeholders, recognizing that the program must simultaneously address housing and public safety, education, employment, well-being, and institutional resources. Choice Neighborhoods will also coordinate extensively and leverage resources with place-based programs at the Departments of Education, Justice, and Health and Human Services, among others. This partnership will help empower communities to address many of their most pressing social problems.
Because the relationship between neighborhood and family structure remains complicated, supporting mobility is also crucial. The Transforming Rental Assistance initiative, a companion to Choice Neighborhoods, will enhance tenant choice and access to a broader range of neighborhoods. As the interrelated nature of neighborhood effects shows, a comprehensive set of strategies and partnerships will be necessary to help promote opportunity in neighborhoods struggling with poverty
- William J. Wilson. 1996. When Work Disappears: The World of the New Urban Poor. New York: Knopf, xvii–xvx.
- Ingrid Gould Ellen and Margery Austin Turner. 2003. “Do Neighborhoods Matter and Why?” In John Goering and Judith D. Feins, eds., Choosing a Better Life? Evaluating the Moving to Opportunity Social Experiment. Washington, DC: Urban Institute Press, 314–8.
- Interview with Margery Austin Turner, Urban Institute, 6 April 2010.
- George C. Galster, “The Mechanism(s) of Neighborhood Effects: Theory, Evidence, and Policy Implications.” Presentation at the ESRC Seminar, St. Andrews University, Scotland, UK, 4–5 February 2010.
- Geoge C. Galster, Jackie M. Cutsinger, and Ron Malega. 2008. “The Costs of Concentrated Poverty: Neighborhood Property Markets and the Dynamics of Decline.” In Nicolas P. Retsinas and Eric S. Belsky, eds., Revisiting Rental Housing: Policies, Programs, and Priorities. Washington, DC: Brookings Institution, 116–9.
- Robert J. Sampson, Jeffrey D. Morenoff, and Thomas Gannon-Rowley. 2002. “Assessing ‘Neighborhood Effects’: Social Processes and New Directions in Research,” Annual Review of Sociology 28: 466.
- Margery Austin Turner and Lynette A. Rawlings. 2005. Overcoming Concentrated Poverty and Isolation: Lessons From Three HUD Demonstration Initiatives. Washington, DC: Urban Institute, 8.
- Robert J. Sampson, Stephen W. Raudenbush, and Felton Earls. 15 August 1997. “Neighborhoods and Violent Crime: A Multilevel Study of Collective Efficacy,” Science 277: 922.
- “About PHDCN,” ICPSR website (www.icpsr.umich.edu/ icpsrweb/PHDCN/biblio/series/00206/resources). Accessed 25 June 2010.