Before the crash, there was a real estate bubble in Atlanta

Kevin Erdmann argued in a Washington Post column on Thursday that the main problem with U.S. housing policy is Fannie Mae and Freddie Mac's overly restrictive lending rules. While current policies may be too restrictive, that does not explain the collapse in housing prices in 2007-2009 or the current inadequate supply of housing.

The example of Atlanta that Erdmann uses in his article is very helpful in supporting these points. Erdmann says that there was no housing bubble in Atlanta and therefore nothing to burst. He attributes the sharp decline in housing prices in 2007-2009, especially in the lower end of the housing market, to the tighter lending requirements of Fannie and Freddie.

The data Do support the thesis that a real estate bubble formed in the decade before 2007, especially in the lower segment of the real estate market. Here is the inflation-adjusted Case-Schiller index for the lower segment of the real estate market (lower third) from 1992 to today.

As you can see, the index rose sharply from 1996 to mid-2005. Then it stabilized and began to fall rapidly in 2007. During the price increase, inflation-adjusted home prices in the bottom third of the market rose 38.8 percent. In contrast, rents in Atlanta rose almost exactly in line with general inflation.

This was the general pattern of home prices in the period before the housing bubble. Home prices nationwide rose roughly in step with the rate of inflation from 1896 to 1996, with enormous differences across regions. In places like New York and San Francisco, prices far outpaced inflation, while in Detroit, St. Louis, and many small cities and towns, they lagged far behind inflation.

Erdmann points out that lower-end home prices fell much more during the Atlanta crash than higher-end home prices. That's true, but high-end home prices rose much less during the bubble. High-end prices rose 27 percent in real terms between 1996 and the peak in 2005.

While this was still a bubble in terms of rent trends, it was significantly smaller than the low-end in Atlanta. Therefore, it is not surprising that real estate prices in the low-end fell more sharply.

Another notable point in this chart is that real estate prices for the lowest tier in Atlanta had largely recovered from their bubble peaks just before the pandemic. Since the pandemic, real estate prices for the lowest tier have even exceeded their bubble peaks. This is also true for the higher tiers.

This suggests that developers have serious incentives to build lots of housing in Atlanta and elsewhere, but for some reason they aren't doing it. Fannie and Freddie's tightening of credit standards cannot explain this failure to build more housing, because that should be reflected in housing prices, but it clearly isn't.

There is another point worth mentioning about Erdmann's arguments about Fannie and Freddie's lending standards. Average credit scores have risen significantly over the past two decades, meaning that using a fixed credit score as a cutoff would mean excluding a smaller proportion of potential borrowers. It would also have been helpful if Erdmann had included data on mortgage lending in the 1990s, before lending standards were loosened and the bubble began to grow.

But that point is secondary. If excessively high credit standards were the real factor clogging up the housing market, real housing prices would not be above their bubble levels. While those prices provide plenty of incentive for builders to build, for some reason they are not building anywhere near as fast as they were before the bubble, or even as fast as they were before the bubble.

This first appeared on Dean Baker’s blog “Beat the Press.”