Sociology’s revenge: Moving to Opportunity (MTO) revisited


Neighborhoods remain the crucible of social life, even in the internet age. Children do not stream lectures — they go to school. They play together in parks and homes, not over Skype. Crime and fear of crime are experienced locally, as is the police response to it.

But wide income gaps and America’s legacy of racial segregation result in wide differences between neighborhoods on a range of measures. Two major new studies from Harvard economists Raj Chetty and Nathaniel Hendren show that neighborhoods matter not only for daily life, but for the life chances of the children raised there.

Drawing on a unique data set based on the tax records of 44 million households, the first study shows that locality matters a great deal for the future income of children. The second study of roughly 13,000 children is smaller but packs a big policy punch, since it directly contradicts recent evaluations of a major policy initiative — Moving to Opportunity (MTO) — carried out by leading social scientists.

In short: MTO seems to work, after all.

MTO, Act I: Ideals

MTO was started in 1994 by the Department of Housing and Urban Development. In a handful of large cities, a few thousand public-housing residents were randomly assigned to one of three programs:

  1. An experimental group that received a rental subsidy (voucher) but had to move to a low-poverty neighborhood for at least one year
  2. A Section-8 group that received a voucher but no restriction on movement
  3. A control group that received no voucher.

Economists love random assignment, because it helps overcome one of the biggest problems in social science: isolating causation from correlation. It’s like a social science pharmaceutical trial. Around half the experimental group “took their medicine” and moved to less poor areas; that is, the proportion living in neighborhood poverty fell by roughly half, from 40 percent to 20 percent. There were limits, however: the voucher didn’t allow them to buy their way into affluent neighborhoods with great schools, so much as somewhat less poor neighborhoods with schools that were slightly better.

Teams of economists and other social scientists published analyses of MTO data in leading academic journals (see the table at the end of the piece).

MTO, Act II: Disappointment

These studies were disappointing, especially for those, like myself, who believe that harmful neighborhood qualities have impeded the social progress of the poor. The key findings were that:

  1. Neighborhood poverty has no effect on adult earnings or employment.
  2. Neighborhood poverty has no consistent positive effects on the behavior of children or their academic performance.
  3. Neighborhood poverty improves some aspects of adult mental and physical health.

MTO Act III: Reflection

These findings, especially the second, contradicted a large body of social science theory and non-experimental evidence, mostly from sociologists, like William Julius Wilson and Douglas Massey. They and other sociologists suggested that the limitations of the study explained the absence of observable effects. In a very recent example, Massey and I found that neighborhood income during childhood strongly predicts adult incomes. Our evidence suggested a causal effect, too, since neighborhood income even explained differences in sibling incomes.

Less noticed, the first wave of MTO findings also contradicted a powerful body of experimental evidence from school lottery and voucher programs. These studies had consistently found that attending better schools—measured in various ways—boosted the academic performance (and eventual earnings) of children, especially those from poor families.

MTO Act IV: Vindication

With their new study, Chetty and Hendren (along with Lawrence Katz, an author of many of the previous studies) provide very strong evidence for the positive impact of MTO. Specifically, moving to a less poor neighborhood in childhood (i.e., before the age of 13):

  1. Increased future annual income by the mid-twenties by roughly $3,500 (31%)
  2. Boosted marriage rates (by 2 percentage points)
  3. Raised both college attendance rates (by 2.5 percentage points) and quality of college attended

The age of the child moved was a critical factor: moving to a less poor neighborhood in the teenage years had no significant impact on later earnings or other adult outcomes.

Lessons from MTO evaluation

There was nothing wrong with the earlier round of MTO evaluations in themselves: the main problem was that the positive effects of leaving poor neighborhoods as a child could not be observed until the children were old enough to finish college and enter the adult labor market. In measuring adult outcomes, scholars of MTO thought creatively about capturing alternative outcomes, like mental health, that had not previously been studied in this context.

Still, some of the scholars who conducted earlier studies were too quick to write off MTO specifically and, more importantly, neighborhood effects more generally. As Chetty and his colleagues show, even a few extra years of data can make all the difference. Now we can be even more confident that when it comes to equality and opportunity, place matters.

List of influential MTO studies, in chronological order





Characterization of finding by authors

Lawrence Katz, Jeffrey Kling, Jeffrey Liebman

Quarterly Journal of Economics (2001)

Two years after program entry

Some positive health effects for experimental group; no economic benefits for household heads; some positive behavioral effects for boys, not girls.

“Early analysis” of the program

Lisa Sanbonmatsu, Kling, Duncan, Jeanne Brooks-Gunn

Journal of Human Resources (2006)

Four to seven years after random assignment

“We did not find evidence of improvements in reading scores, math scores, behavior or school problems, or school engagement, overall or for any age group.”

They explain these results by showing that, “Participation in MTO did not result in attendance at high-performing schools,” and resulted in only small changes in school quality . . . “It appears that interventions focused exclusively on neighborhoods rather than on factors directly related to the child, family, and school are unable to solve the myriad problems of children growing up in poverty.”

Jeffrey Kling, Jeffrey Liebman, Lawrence Katz

Econometrica (2007)

Five years after random assignment, on average

We find no significant overall effects of this intervention on adult economic self-sufficiency or physical health. Mental health benefits of the voucher offers for adults and for female youth were substantial. Beneficial effects for female youth on education, risky behavior, and physical health were offset by adverse effects for male youth.

“Our analysis provides direct evidence on the existence, direction, and magnitude of neighborhood effects for important socioeconomic and health outcomes in both adult and youth populations … housing mobility by itself does not appear to be an effective antipoverty strategy—at least over a 5-year horizon.”

Susan Clampet-Lundquist and Douglas Massey

American Journal of Sociology (2008)

Four to seven years after random assignment

We show that living in low-poverty neighborhoods is positively associated over the long term with higher levels of employment, greater earnings, and lower levels of public service dependency, as expected under the neighborhood effects hypothesis.”

Re-analysis needed to consider time spent in low-poverty neighborhoods: “As a result of widespread and selective out-migration from neighborhoods after initial relocation, experimental subjects accumulated relatively little time in low poverty settings, and the time spent within low-poverty neighborhoods.”

Jens Ludwig, Liebman, Kling, Katz, Greg Duncan, Ronald Kessler, Lisa Sanbonmatsu

American Journal of Sociology (reply to Clampet-Lundquist and Massey) 2008

Four to seven years after random assignment

“[T]here is no evidence that extra time spent in low-poverty integrated neighborhoods improves economic outcomes, while the estimated effects of time in low-poverty segregated neighborhoods are quite small. We also find no evidence that living in low-poverty neighborhoods in general (pooling segregated and integrated areas together) improves economic outcomes.”

“[MTO] shows us that moving out of a disadvantaged, dangerous neighborhood into more affluent and safer areas does not have detectable impacts on economic outcomes four to seven years out.” But there are some important health effects and effects on violent behavior.

Sanbonmatsu, Ludwig, Katz and others

Department of Housing and Urban Development (2011)

10-15 years after random assignment

Better health outcomes, no adult economic benefits (income and earnings); for youth, little to no effect on access to higher quality schools (based on test scores); no effect on actual test scores or educational attainment between groups.

Long-term effects: “A majority of MTO children were still attending majority-minority, overwhelmingly low-income public schools located in the districts serving the five original MTO cities… MTO was more successful in improving mental and physical health in poor families that signed up to participate in the program than in bringing about the improvements in education and labor market outcomes needed to boost family incomes.”

Jens Ludwig, Greg Duncan, Lisa Gennerian, Katz, Kessler, Kling, Sanbonmatsu

Science (2012)

10-15 years after random assignment

“We found that moving from a high-poverty to lower-poverty neighborhood leads to long-term (10- to 15-year) improvements in adult physical and mental health and subjective well-being, despite not affecting economic self-sufficiency.”

“MTO-type policy efforts to improve the
neighborhood conditions of poor families would not be part of an effective antipoverty strategy.”

Jens Ludwig

Third Way (2014)

10-15 years after random assignment

“[Living in] poverty does have extremely important impacts, but not on the outcomes emphasized by [William Julius] Wilson. Rather the impacts are seen in areas such as physical and mental health, and the overall sense of well-being or happiness of poor families.”

“The MTO findings raise the possibility that very distressed neighborhood environments may be less important for outcomes like children’s schooling and adult earnings than hypothesized in William Julius Wilson’s landmark book The Truly Disadvantaged.”

Raj Chetty, Nathaniel Hendren, and Lawrence Katz

National Bureau of Economic Research working paper (2015)

15-18 years after random assignment, using federal tax records to study adults who moved as children.

Growing up in a low-poverty neighborhood raises incomes by 31% for those in their mid-twenties. Effects are evident for children who moved before age 13 but not after.

“The results of this study demonstrate that offering low-income families housing vouchers and assistance in moving to lower-poverty neighborhoods has substantial benefits for the families themselves and for taxpayers.”