It’s unusual to flag an economics article as a “must-read” for general audiences: but if you care about cities and place, and about the prospects for the American Dream in the 21st Century, you owe it to yourself to read this new article by Raj Chetty, Nathaniel Hendren, “The Impacts of Neighborhoods on Intergenerational Mobility: Childhood Exposure Effects and County-Level Estimates.”  (The Executive Summary is just six pages long—you can download it here.)

This work strongly confirms the growing belief that the kind of community you grow up in has a huge impact on your lifetime economic opportunities.  Specifically, the Chetty-Hendren study shows that some communities do a much better job of helping kids from low income families achieve economic success than do others.  And these communities tend to be ones that have low levels of economic and racial segregation, better schools, less violent crime, and fewer single-parent families.  An important part of how we assure opportunity to all hinges on how we build communities.

Seattle ranks as one of the most mobility-friendly metropolitan areas. Credit: Jonathan Miske, Flickr
Seattle ranks as one of the most mobility-friendly metropolitan areas. Credit: Jonathan Miske, Flickr

This study is remarkable for a number of reasons:  it’s clearly and simply presented, based on an extraordinarily large and powerful database, provides detailed findings (down to the county level), and provides strong evidence that its findings are cause-and-effect, not mere correlation.

One of the big bugaboos of economic research is that, unlike other scientific inquiries, economists are not generally allowed to run random selection controlled experiments on human beings—which we would all probably agree is a good idea.  Economic research, like most social science, typically must rely on statistical inference from sample data often gathered for other purposes, with its attendant margins of error.  And secondary statistical data make it especially difficult to make definitive cause-and-effect statements: For example, did a community’s environment cause children to have particularly high levels of economic mobility, or did the unseen choices of some parents to move in and out of particular neighborhoods lead “natural” high achievers to locate in some places and “natural” low achievers to locate elsewhere.

Using the combination of a massive, long-term longitudinal data set (created from anonymized tax return data), and data from the Federal Government’s quasi-experimental “Moving to Opportunity” program which gave low income families vouchers to enable them to move to non-poor neighborhoods.

It’s highly unusual in the world of economics to use the word “causal” to describe one’s reported findings, but in this new report you’ll see this term used early and often to describe the findings.  The use of data on siblings and exploiting the differential effects for boys and girls is clever and impressive.  One of the criticisms levied of other work is that it can’t control for the fact that intergenerational mobility for some families may represent selection effects: the most energetic, ambitious families are more likely to move away from worse environments and to better ones.  It is very rare in social science to be able to make this kind of strong claims about causality.

And New Orleans ranks as one of the worst. Credit: Chuck Coker, Flickr
And New Orleans ranks as one of the worst. Credit: Chuck Coker, Flickr


The great thing about the Chetty-Hendren research is that you can drill down to the county level to see what impact the local community has on economic outcomes for kids.  And the measure of success couldn’t be clearer: they show how much each additional year spent growing up in a particular neighborhood is likely to influence a child’s income as an adult.  As they explain in their report:

Every extra year spent in the city of Baltimore reduces a child’s earnings by 0.86% per year of exposure, generating a total earnings penalty of approximately 17% for children who grow up there from birth.

The differences among metropolitan areas are substantial: a poor child growing up in Seattle would be expected to earn about $29,000 (about $3,000 or 12 percent more than the national average for children in the bottom quintile of the population), while a poor child growing up in New Orleans would be expected to earn a little more than $22,000 at the same age, ($3,800 or almost 15 percent less than the national average.)  You can see data for individual counties and for commuting zones (metropolitan areas and their surrounding hinterlands) at the New York Times website.

To provide a quick snapshot for large metropolitan areas, we’ve created a graphic showing the Chetty-Hendren estimates for central counties (the county that includes the first-named city in a metropolitan area) and for the surrounding commuting zone.  These data show how much more (or less) than the national average a child in a family in the lowest quartile of the income distribution growing up in the central county or commuting area would make at age 26.  (Orange dots represent commuting zones; blue dots represent central counties.)

A couple of patterns are apparent: in general, central counties have lower rates of economic mobility for poor children than in commuting zones.  Central counties tend, on average, to have more concentrated poverty, lower-performing schools, and higher rates of single-head households—all of which are correlates of low economic mobility.

In a companion paper, Chetty and Hendren and Harvard Economist Larry Katz re-examine an important, and previously discouraging set of findings from the Moving to Opportunity (MTO) project.  MTO was a federal project that gave poor families vouchers to move from poor neighborhoods to middle income neighborhoods.  The previously reported results found that the moves produced little economic improvement for adults, and modest results for children.  In their re-analysis, Chetty, Hendren and Katz, show that when children moved made a huge difference:  those who moved as very young children (under age five) showed significant gains, while those who moved at an older age showed few if any gains.  Consistent with their larger analysis of inter-neighborhood moves, the gains to moving to better neighborhoods were directly correlated to how long children were exposed to better conditions.  (For a more detailed review of these studies and their import, it’s worth reading Justin Wolfers’ commentary.)

At City Observatory, we think these findings are the strongest evidence yet that addressing neighborhood and urban development is critical to promoting equal opportunity for all.  As Chetty and Hendren conclude that while the evidence shows that some children can gain opportunity by moving to a new, better neighborhood that this isn’t a scaleable solution for everyone; as a result:

. . . one must also find methods of improving neighborhood environments in areas that currently generate low levels of mobility. . . our findings provide support for policies that reduce segregation and concentrated poverty in cities (e.g., affordable housing subsidies or changes in zoning laws) as well as efforts to improve public schools.