THIS week's print edition includes a look at the changing trajectory of the American recovery. From 2007 to 2011 many of the extreme points in America's metropolitan distribution, in employment terms, could be found in the Sunbelt: cities in Texas and Oklahoma were among the few metropolitan areas to manage net employment gains over the period while those in the Southwest and Atlantic Southeast performed miserably, notching some of the highest unemployment rates of the downturn. Since 2011, however, the relatively rapid job growth has spread across the Sunbelt, which now seems to be outperforming most of the country's other regions. California's large metro areas have joined cities like Phoenix and Atlanta alongside Dallas and Houston, which job growth in cities like Boston and Minneapolis has fallen down the league tables a bit.
As the piece notes, the initial divide largely reflected differing experiences in housing, and the reconvergence correspondingly owes something to the nationwide improvement in housing. But in the Sunbelt that improvement, and the broader resurgence in Sunbelt economic activity, is very much associated with the continuation of a great sunward migration that has helped reshape America for a generation. Migration into the South and West slowed during the crisis and recession but it didn't stop. From 2011 to 2012, according to the latest Census estimates, even Clark County, Nevada cracked the list of counties with the biggest increase in population, despite a double-digit unemployment rate in Las Vegas.
Below, I've put up a chart that didn't make the print piece, which shows the contribution to overall GDP growth from population and from the change in output per person. This is a bit of a rough calculation. The metropolitan GDP data from the Bureau of Economic Analysis aren't perfect (choices have to be made concerning how to allocate output from firms operating in multiple cities) and only run through 2011. And population and per capita output obviously aren't independent of each other; other things equal a rise in one will generate a decline in the other. But the chart (which selects from the 50 largest metropolitan areas the ten fastest and slowest growing over 2007-11) gives a decent sense of where there has been a lot of population growth relative to GDP growth and where there has been surprisingly little:
One striking fact is that even among the best performing metropolitan areas, overall increases in output per capita have been hard to come by. They have been limited to a handful of very brainy cities, especially West Coast tech centres. In general, growth has been a product of population increase large enough to offset falling output per person. Through 2011 lots of Sunbelt bubble metros were on the wrong side of that calculation. Increasingly they are back on the right side.
But there is a qualitative gap between the top performers and bottom performers on this list; the level of per capita GDP (which isn't shown) is generally quite different at the top and the bottom. Average real per capita GDP in the top cities is $58,426, to $38,724 in the bottom. (The median for the top cities is $58,058 to $40,804 in the bottom; it isn't one or two outliers driving the gap.) And so we find ourselves confronting a very interesting question: why were Americans moving in droves to metropolitan areas with low income levels and high unemployment rates?
In a 2007 paper Edward Glaeser and Kristina Tobio pinned growth in the Sunbelt from 1980 on the relative affordability of its housing, which was itself due to rapid construction in Sunbelt metros. Some readers may question that hypothesis. What about coastal California, for instance, where NIMBYism runs rampant? A good point: coastal metro areas in California became very restrictive in the 1980s and 1990s, a change which sent housing and office costs there soaring. That brings us to another potential hackle: how could Sunbelt cities with soaring prices, like Las Vegas and Phoenix, have been attractive on the basis of cost? Well, despite rapid housing price growth in those metros they remained vastly cheaper relative to coastal California, which was the source of many migrants to those areas.
The argument, it's worth clarifying, is not simply that cheap housing makes a metro area attractive; if that were true the Midwest would be booming. It is that a very elastic housing supply in the face of steady demand will allow for massive population growth.
When recession struck, the bubbly Sunbelt metropolitan areas fared horribly. But there was an upside to the crash; it restored their cost competitiveness:
Foreclosures came fast and furious in California and Florida. This contributed to a deep economic swoon, but it also helped borrowers escape crushing debts. A return to affordability has also helped. By 2011 house values in Washington, DC, were still 79% above their level in 2000, and prices in New York and Boston were 63% and 49% higher. In San Francisco prices were just 29% higher, however. Prices in Phoenix fell to 2000 levels, and Atlanta prices sank 11% below them. America’s southward migration has duly resumed, providing new buyers for local houses and new demand for local businesses.
Industry trends have also helped. In the immediate wake of the crisis the fastest rebounds occurred in manufacturing and finance oriented areas like the Midwest and Northeast. Those industries have since faced headwinds—a global economic slowdown in the case of the former and a secular decline in financial employment in the latter—while others have boomed. Energy, technology, health care, and construction have bouyed many southern cities in recent years.
But one important takeaway from these trends is that for many Americans the best way to get a real wage increase is to move to the Sunbelt. Compensation is generally lower there. But housing costs are much, much lower.
Importantly, this does not seem to be the case for one subset of the American population: the rich. The supply and quality of amenities prized by better-off households is unmatched in rich coastly metro areas like New York and San Francisco (see this and this). If you have a lot of money to spend, you get more for it in expensive metros than cheap ones, enough so that real incomes for the rich may actually be higher in the more expensive cities than the cheap ones. This dynamic is a powerful force for inequality across the country's geography.
The meaning of that separation isn't entirely clear. It could be that the rich are simply segregating themselves into gilded consumption centres. But it's also possible that the concentration of high incomes corresponds to a concentration of high levels of productivity and innovation. If that's the case, then America is filtering a very large share of its citizenry out of its most opportunity-rich places.
