Michael Brocker and Christopher Hanes write (NBER, $)

We find that cities which had experienced the biggest house construction booms in the mid-1920s, and the highest increases in house values and homeownership rates across the 1920s, saw the greatest declines in house values and homeownership rates after 1930. They also experienced the highest rates of mortgage foreclosure in the early 1930s. These patterns look very much like those around 2006, despite the gap between the house-market peak in 1925 and the businesscycle downturn in 1929. They are consistent with a bubble. They show that the effects of the mid-1920s boom on house markets were still present as of 1929. They suggest that in the downturn of the Great Depression house values fell further, and there were more foreclosures, because the 1920s boom had taken place.

This appears to be part of a conference volume. The introduction to the volume, ungated and recommended, is by Kenneth Snowden.

The conference volume also includes Gjerstad and Smith, who view the housing cycle as important in the Great Depression. But we saw that Alexander Field, another author in the conference volume, disputes the view that housing leverage was important in the 1920s and 1930s.

So confusing! I think that all of these economic historians agree that the 1920s boom peaked in 1925. All seem to agree that the economy survived the ending of the boom quite well until 1930. Apparently, the big decline in house prices took place in the 1930s, although the authors note that good data on house prices for this period is lacking. Note also that there was general deflation, so that the decline in real house prices was much less than in our recent financial crisis. Of course, that is of little comfort to someone who borrows at a positive nominal interest rate.

Based on this, I am reluctant to assign housing a major causal role in the Depression. If the big decline in house prices took place in the 1930s, then that could be an effect of, or a part of, the overall Depression. Note, however, that the authors write,

This [the cross-sectional correlation between severity of house price declines in the 1930s and the extent of the construction boom in the 1920s] remains true when we control for measures of the local severity of the depression – changes in family income, changes in retail sales – or for changes in average rents.

The Snowden piece is filled with interesting background information, such as

The home mortgage market of the 1920s grew even more rapidly than the nonfarm housing stock, with nonfarm residential debt tripling (from $9 to $30 billion) in less than a decade, while the ratio of debt to residential wealth doubled from 14 to nearly 30 percent

He summarizes other papers in the volume, including one by Eugene White.

He argues that the double liability rule faced by bank shareholders and the restrictions on mortgage lending meant that both national and state-chartered banks were well-capitalized relative to the modest risks that they carried on real estate loans.

White argues that although there were some common factors that affected the banks during the building booms of the 1920s and the 2000s…the important difference between the two episodes is that banks were induced in the modern period to participate in risky real estate finance by a set of policies that were missing in the 1920s—deposit insurance, the “Too Big to Fail” doctrine, and federal subsidization of risky mortgage lending and securitization.