The Resilient Center

Ed Glaeser has written some really great stuff on the geography of the recession. He continues today with a post tracking the correlates of high unemployment. Unsurprisingly given the sectoral bias of the recession, low unemployment is strongly correlated with the stock of human capital. Glaeser notes that low unemployment in this recession is actually correlated with the stock of human capital in the 1940s, which mainly shows the persistence of concentrations of human capital. As he has noted in previous research, human capital levels have actually diverged over time, with smart cities getting smarter and vice versa.

Glaeser follows this up by focusing on the connection between manufacturing employment and high unemployment, which is related to the human capital correlation. And then he notes something interesting:

A third variable that predicts the level of unemployment is the centralization of the metropolitan area. About five years ago, the U.C.L.A. economist Matthew Kahn and I wrote a couple of papers about sprawl in America’s cities. One benchmark measure of decentralization that we used was the share of employment in ZIP codes that were within five miles of the metropolitan area’s urban center. As the next figure shows, January’s unemployment is lowest in those areas that are most centralized.

While it is true that skilled places are more centralized, and manufacturing cities are less so, this effect continues to be statistically significant even when I control for skills and manufacturing.

One thing that’s worth considering is that housing markets in more centralized places may have performed better (though it’s also likely that this effect is offset by non-bubble sunbelt metro areas that haven’t experienced high unemployment). Another possibility is that more centralized places were less exposed to high gas prices last summer, and therefore entered the depths of recession with local budgets and businesses in a stronger position.

But this may simply reflect the productivity gains and labor market advantages of dense agglomerations. These factors may simply make for a more resilient and flexible economy. Interestingly, Glaeser closes by saying:

During the recessions, thousands fled the Dust Bowl. In the 1970s and 1980s, there was a massive exodus from the Rust Belt. Today’s recession will also prompt mobility, probably toward more skilled, more centralized cities with less historical commitment to manufacturing.

Which reminded me to link to this, by Chris Nekarda, who writes:

In fact, geographic mobility is moderately countercyclical—that is, more people move during recessions than during booms (relative to trend). This may seem counter-intuitive but makes economic sense.

Geographic mobility is a means of reallocating resources, in this case labor, to more efficient uses. In the past, 70 percent of people who move indicated having moved for economic reasons and up to 50 percent of those moves occurred because of a job separation [Lansing and Morgan (1967); Bartel (1979)]. In particular, there is a significant positive relationship between unemployment and geographic mobility [Bartel (1979); Schlottmann and Herzog Jr. (1981, 1984)]. Thus, countercyclical mobility is consistent with reallocation of idle workers across space.

It’s quite healthy for workers to move to more productive areas of the country (though it’s worth noting, as Nekarda does, that housing market troubles will make the population a bit less mobile than in a typical recession). What’s also worth noting is that many of these centralized metropolitan areas are pretty green, and could be made greener, more productive, and more affordable with well targeted investments in infrastructure. When I talk about achieving climate goals while also enjoying strong and sustainable economic growth, this is what I have in mind.