The importance of where you live for U.S. economic mobility

Robust intergenerational economic mobility is supposed to be one of the defining characteristics of the United States. Yet economic research conducted over the past several years—and then revisited in a study released this week—finds that the degree of mobility across generations seems to depend on where you live and who your neighbors are.  Figuring out what drives that variation has become an important academic endeavor.

The most recent contours of this investigation were set in early 2014, when a group of economists released a study looking at the variation in the level of United States intergenerational mobility in the United States by where families lived. The results were striking. Some sections of the country had mobility levels that rivaled those of high-mobility countries such as Denmark. But other parts of the United States recorded levels so low that there was no comparison among other high-income countries.

The findings of that report, like lots of good research, sparked more questions. Were these patterns a result of certain kinds of people moving to regions where high mobility was already happening? Or was there something about these areas themselves that promoted high levels of mobility? Two of those researchers returned to this question with a new paper. The authors, economists Raj Chetty and Nathaniel Hendren of Harvard University, sort out the effects of location on intergenerational economic mobility. Using about five million observations of family tax records and looking at families that moved when children were young, Chetty and Hendren study the effects of a new location on mobility.

In short, they find that there is something about specific locations that promote economic mobility across generations. More precisely, the authors find that within the United States, 50 to 70 percent of the variation in mobility is due the “causal effects of place.” In other words, if we want to boost intergenerational mobility in the United States we can’t just focus on the attributes of specific individuals. We have to look at how to improve communities and neighborhoods.

But then another question comes up: What are the characteristics of those mobility-promoting places? Chetty and Hendren don’t find a definitive explanation, but they do  suggest some answers by returning to evidence they uncovered in 2014.  In that study, Chetty and Hendren found several variables that were strongly correlated with intergenerational mobility:

• Measures of race
• Segregation
• Income inequality
• The quality of Kindergarten through 12^th grade schools
• Social capital, or the strength of civic engagement in a community
• Family structure

In their latest work, Chetty and Hendren test to see if these six correlations are due to the causal effects of location itselfor the characteristics of those that live there—what the authors call the “sorting component.”

Chetty and Hendren find that areas with more African-Americans and single-parent families have lower levels of intergenerational mobility. While about half of this relationship is due to individual characteristics, the other half is due to the negative effects of the place where these people live.  . So areas with high levels of single parents have lower mobility, for example, but only part of that relationship is due to the effect of moving into an area with high levels of single parents.

For two other variables, their initial correlations were much more influenced by the causal effects of place. The correlation of social capital with mobility appears to be entirely caused by the place itself. In other words, if you’re looking to move to an area that has a causal effect on mobility, social capital is a good indicator of where you and your family want to live.

The same goes for the quality of a K-through-12 education. At the level of commuting zone (the larger jurisdiction Chetty and Hendren use, comprised of 741 areas of the country),  correlation between mobility and education is entirely about a correlation with the causal effect of place. But then focusing on the smaller and more numerous counties within commuting zones, they find significant correlations with the sorting component.

When it comes to poverty, inequality and segregation, the results are particularly interesting. In a companion paper with Larry Katz, also of Harvard, Chetty and Hendren find that the Move to Opportunity program, which encouraged several thousand low-income families to move out of high-poverty neighborhoods into low-poverty areas, was quite successful at boosting intergenerational economic mobility. But in this paper, the three authors find that poverty rates are not strongly correlated with the causal effects of places. What are much more strongly correlated with the causal effect are measures of income segregation and income inequality.

While all these correlations are interesting and suggestive, they are still correlations. Yes, they are correlations with a causal effect but they don’t tell us if, for example, social capital is a key factor underpinning the causal impact of living in one place over another. Policymakers no doubt will find this research intriguing as they seek to promote the social and economic variables that produce more mobility-inducing areas, but further research drilling down on causation is still required.