Housing / Employment Correlation Underscores the Chaos of This Economy

by: Ryan Avent

In comments to Tuesday’s post, my friend David Schleicher noted:

Ryan — you should pair this with the lists on which cities have seen year-on-year increases in property values. The in-flow data is interesting all on its own (thanks for blogging it!), but it doesn’t paint a full demand-side picture without price.

If supply was determined everywhere simply by the cost of building, inflow would provide that picture, but with zoning and other supply restrictions, it paints a funny picture. Demand for moving to boston, for instance, pushes up both prices and people, whereas demand for phoenix — where supply is, well, available and cheap — only effects inflows.

He’s right. This is the point that Ed Glaeser has made — that recent migration to the Sunbelt is primarily about housing supply and housing costs. There is high demand for housing in places, largely on the coasts, where housing supply is relatively tightly constrained. This leads to large increases in price in those places, which in turn drives migration from those cities. So you wind up with a world in which there are some markets with high prices, low construction levels, and low population growth, and others with low prices, high construction levels, and high population growth. In this world, migration data alone doesn’t give us the complete picture of demand.

So here’s a little more data for the top 20 gainers in terms of domestic migration. These are percentage changes from July of 2008 to July 2009 (for house prices, third quarter on third quarter, FHFA data):

Migration Home Prices Employment
Raleigh-Cary, NC 1.8% -2.4 -4.7%
Austin-Round Rock, TX 1.6% -0.4 -2.4%
New Orleans-Metairie-Kenner, LA 1.3% -1.7 -1.1%
Charlotte-Gastonia-Concord, NC-SC 1.1% -2.5 -6.5%
Charleston-North Charleston-Summerville, SC 1.0% -5.6 -5.4%
Columbia, SC 1.0% -0.9 -5.0%
San Antonio, TX 0.9% 0.6 -2.6%
Houston-Sugar Land-Baytown, TX 0.9% 0.9 -3.3%
Denver-Aurora-Broomfield, CO /1 0.8% -0.7 -4.9%
Portland-Vancouver-Beaverton, OR-WA 0.7% -8.5 -6.9%
Oklahoma City, OK 0.7% 0.6 -3.2%
Dallas-Fort Worth-Arlington, TX 0.7% -0.1 -4.5%
Nashville-Davidson–Murfreesboro–Franklin, TN 0.7% -2.0 -5.7%
Seattle-Tacoma-Bellevue, WA 0.5% -9.8 -5.9%
Indianapolis-Carmel, IN 0.4% -1.0 -5.5%
Washington-Arlington-Alexandria, DC-VA-MD-WV 0.3% -5.6 -1.6%
Atlanta-Sandy Springs-Marietta, GA 0.3% -5.2 -6.2%
Phoenix-Mesa-Scottsdale, AZ 0.3% -17.2 -8.3%
San Francisco-Oakland-Fremont, CA 0.2% -6.9 -6.9%
Boston-Cambridge-Quincy, MA-NH 0.1% -3.4 -4.1%
Click to enlarge

Spot any patterns? The closest correlation is actually between change in home prices and change in employment — no surprise, really. While Glaeser’s relationship holds up over longer periods, there’s just too much going on with the recession and housing bust for us to see that much in the data.

I actually suspect that the relationship is flipped for bubblier markets at the moment — employment-oriented migration is likely to provide a firmer floor for falling housing prices. You see this in Washington, for instance, where a strong labor market has driven migration, which has in turn supported home prices. Put another way: Employment prospects may currently be trumping issues of home affordability.

But it’s difficult to say. This is a chaotic period for labor markets and housing markets.