Consumer Law

Costs and Benefits of Subprime Lending

By Alan White

By Alan White

The new paper by Todd Zywicki and Joseph Adamson reasserts the received wisdom that subprime mortgage lending had significant consumer welfare benefits, specifically that it increased homeownership and provided access to credit to previously excluded groups. I was disappointed to find that the sources cited do not provide any more empirical support for this folk wisdom than has been previously offered elsewhere. I have argued that subprime mortgage lending did no such thing, i.e. was in no way responsible for the slight increase in homeownership in the 1995 to 2005 period (homeownership is now going down, thanks to subprime foreclosures), in a paper I posted a while back. In my view, subprime mortgage lending had the net impact of reducing homeownership, and displacing cheaper and better home financing, especially for low-income and minority groups. I tried to survey the available research and empirical data and to look at other variables that have affected homeownership and the volume of consumer credit in the relevant period, including data from the Survey of Consumer Finances, the Census Housing Survey and other available sources. If it were true that subprime mortgages had produced significant welfare benefits, then I would agree with the conclusion that regulation needs to proceed carefully. I don't believe the evidence is there, if the question is examined carefully, although I am very interested in seeing empirical evidence to support the hypothesis that subprime mortgages improved aggregate welfare. A note on the specific sources in the continuation.

The new empirical source on which Zywicki and Adamson rely to refute the evidence that subprime mortgages did not increase homeownership is an analysis of data from Loan Performance done by the Milken Institute, Download , repeating the Center for Responsible Lending's calculations but with one significant change. CRL compared the number of first-time homebuyers using subprime mortgages (positive benefit) with the number of subprime foreclosures. The Milken Institute points out, correctly, that some subprime foreclosures were not home losses, but foreclosures on investor properties. However, they overcorrect. They tabulate foreclosures only on mortgages made to first-time buyers. Subprime mortgages caused homeownership to decrease not only for first-time buyers, but also for owner-occupants who refinanced or took out a home equity loan and got foreclosed. In my view, the calculation should be redone subtracting from subprime foreclosures only the foreclosures of non-owner-occupied (investor and second home) properties. CRL arrives at a total loss of 855,000 homes. Milken Institute arrives at a net gain of 435,000 (in any event a relatively insignificant contribution to the millions of new homeowners reflected in the homeownership rate increase from 2000 to 2005.) The true number is probably in between, but is more likely to remain a negative one. Readers are welcome to examine the original sources to reach their own conclusions.

My argument that subprime lending did not increase homeownership is not based solely on the CRL's foreclosures vs. new buyers analysis. I also consider the impact that subprime mortgages had in displacing lower-cost FHA mortgages that the same buyers might otherwise have obtained, as well as the impact of prime mortgages on the homeownership rate. The simplistic analysis asserting that subprime lending increased homeownership tends to suffer from the classic missing variables problem familiar to social scientists.

Full post as published by CL&P Blog on March 20, 2008 (boomark / email).