The key finding of Chetty's work with economists Nathaniel Hendren, Patrick Kline and Emmanuel Saez is that a child's odds of escaping poverty as an adult vary widely across U.S. cities and towns.
The CAP paper suggests that about half of the variations in regional mobility can be explained by middle class size.
For example, Scranton -- Scranton! -- has a significantly larger middle class than average for the largest 100 regions in the country. Its mobility is well above average, too. The Washington region has a much smaller middle class, and its mobility is lower.
None of the work shows causality -- it's not determinable from the analysis, in other words, if a large middle class creates more upward mobility, or if more mobility creates a larger middle class. But the CAP analysis suggests that mobility shares a stronger relationship with middle-class size than it does with almost any of the more than two dozen variables Chetty and his coauthors tested for hints on what makes the economic ladder easier to climb in some places than in others.
CAP casts the findings as supporting evidence for its focus on a "middle-out" economic strategy, which holds that boosting the middle class is a key driver of broad economic prosperity; President Obama has embraced that strategy in speeches.
“There’s something going on here about the size of the middle class that’s really potent,” says Ben Olinsky, who co-wrote the CAP study with colleague Sasha Post.
Chetty and his team found a variety of social and economic factors that seem to be correlated with mobility. As David Leonhardt put it in a New York Times piece summarizing the original study this summer: "All else being equal, upward mobility tended to be higher in metropolitan areas where poor families were more dispersed among mixed-income neighborhoods. Income mobility was also higher in areas with more two-parent households, better elementary schools and high schools, and more civic engagement, including membership in religious and community groups."
Olinsky and Post found that almost none of those variables were as strongly associated with mobility as the size of a region's middle class -- defined as the percentage of area residents who earn incomes between the 25th and 75th percentiles nationally. That's related to, but not the same as, income inequality between high and low earners, which the original research team found to be more modestly correlated with mobility.
The original study shows only two variables more strongly associated with mobility than middle-class size, and both were linked to decreased mobility: the region's share of single mothers and its divorce rate.
What does the middle-class link mean for a child hoping to escape poverty? Let Olinsky and Post explain:
Imagine a city in which 40 percent of the population is in the middle class. According to the data, a child who begins in the 25th income percentile could expect to reach the 37th percentile when he or she turns 30. But if the city’s middle class were larger, say, 50 percent instead of 40 percent, then a low-income child could expect to end up in the 42nd percentile, making around $26,000 a year instead of $22,000 a year. That’s almost $4,000 in additional income — a 17 percent increase.
Post and Olinsky sent their findings to Chetty. In a phone interview, he said the CAP study was "consistent with what we've shown ourselves -- there's a strong correlation."
He also offered a few caveats for interpreting the finding. Principally, he noted that for the 50 largest urban areas in the country, his group's research shows the middle-class/mobility relationship breaks down. What's more important, he said, is income segregation -- whether poor people are isolated in where they live from middle-class and wealthier people. (Reihan Salam has written extensively on that subject.)
Chetty also said there's reason to think that middle-class size is a driver of mobility, and not the other way around: In many regions where children are more upwardly mobile, young people move up the income ladder by moving to other cities. In those cases, strong mobility would not by itself produce a larger middle class.