What Leonhart suggests, and apparently the still unavailable study also suggests, is that sprawling cities that put the poor far away from job opportunities tend to have low upward social mobility. I did a bit of data-crunching here. I took my favorite measure of sprawl, population-weighted density in thousands per square mile, versus the new study’s measure of upward mobility, the probability of children born in the bottom quintile making it into the top quintile; the data points are the top ten metro areas by population, not including New York, which is so much denser than anyone else that including it would scrunch everyone into the lower left corner:



Yep, there’s a pretty strong correlation, although not perfect. What’s the matter with Chicago?(And what’s not the matter with Houston?) And as Leonhart suggests, Atlanta is the real poster child here: massive sprawl and very low social mobility.

Is the relationship causal? You can easily think of reasons for spurious correlation: sprawl is associated with being in the sunbelt, with voting Republican, with having weak social safety net programs, etc.. Still, it’s striking.

And the William Julius Wilson “spatial mismatch” theory of urban ills is looking pretty good, isn’t it? (Actually, the emergence of “underclass” problems among low-opportunity whites points in the same direction).

http://krugman.blogs.nytimes.com/2013/07/22/did-sprawl-kill-horatio-alger/