• Embed Doc
  • Copy Link
  • Readcast
  • Collections
  • CommentGo Back
Download
 
 
WORKING PAPER SERIES
   F   E   D   E   R   A   L   R   E   S   E   R   V   E   B   A   N   K 
      o
      f
   A   T   L   A   N   T   A
Decomposing the Foreclosure Crisis:House Price Depreciation versus Bad Underwriting
Kristopher Gerardi, Adam Hale Shapiro, and Paul S. Willen Working Paper 2009-25September 2009
 
 
Parts of this paper appeared earlier in Boston Fed Working Paper 07-15, titled “Subprime Outcomes: Risky Mortgages, Homeownershipand Foreclosure.” The authors thank Chris Foote, who figured out how to identify subprime purchase mortgages in the Warren Groupdata using the HUD subprime lender list, which greatly simplified this project. They also thank Paul Calem, John Campbell, Jeff Fuhrer, Simon Gilchrist, Lorenz Goette, Robert King, Andy Haughwout, Zach Kimball, David Laibson, Andreas Lehnert, Atif Mian,Paolo Pellegrini, Karen Pence, Anthony Pennington-Cross, Erwan Quintin, Julio Rotemberg, Amir Sufi, and Geoff Tootell; audiences atBoston University, the University of California-Berkeley, Harvard University, MIT, Fordham University, Moody’s KMV, Freddie Mac,George Washington University; the Federal Reserve Banks of Atlanta, Boston, New York, Philadelphia, Kansas City, Richmond, and SanFrancisco; the Board of Governors of the Federal Reserve System; the Chicago Fed Bank Structure Conference; the Homer HoytInstitute; the Rodney White Conference at the Wharton School; and participants in the Stanford Institute for Theoretical Economics,the Society for Economic Dynamics, the National Bureau of Economic Research Summer Institute, and the American Real Estate andUrban Economics Association midyear meetings for helpful comments and suggestions. They are grateful to the hospitality of Studienzentrum Gerzensee, where this paper was completed. They also thank Tim Warren and Alan Pasnik of the Warren Group andDick Howe Jr., the registrar of deeds of North Middlesex County, Mass., for providing them with data, advice, and insight, and theythank Anthony Pennington-Cross for providing computer programs. Finally, they thank Elizabeth Murry for helpful comments andedits. The views expressed here are the authors’ and not necessarily those of the Federal Reserve Banks of Atlanta or Boston or theFederal Reserve System. Any remaining errors are the authors’ responsibility.Please address questions regarding content to Kristopher Gerardi, Research Department, Federal Reserve Bank of Atlanta, 1000Peachtree Street, N.E., Atlanta, GA 30309-4470, kristopher.gerardi@atl.frb.org; Adam Hale Shapiro, U.S. Department of Commerce,Office of the Chief Economist, Bureau of Economic Analysis, 1441 L Street, N.W., Washington, DC 20230, 202-606-9562,adam.shapiro@bea.gov; or Paul S. Willen, National Bureau of Economic Research and Research Department, Federal Reserve Bank of Boston, P.O. Box 55882, Boston, MA 02205, 617-973-3149, paul.willen@bos.frb.org.Federal Reserve Bank of Atlanta working papers, including revised versions, are available on the Atlanta Fed’s Web site at www.frbatlanta.org. Click “Publications” and then “Working Papers.” Use the WebScriber Service (at www.frbatlanta.org) to receive e-mail notifications about new papers.
FEDERAL RESERVE BANK
o 
  f 
ATLANTA 
 WORKING PAPER SERIES 
Decomposing the Foreclosure Crisis:House Price Depreciation versus Bad Underwriting
Kristopher Gerardi, Adam Hale Shapiro, and Paul S. Willen Working Paper 2009-25September 2009
Abstract:
We estimate a model of foreclosure using a data set that includes every residential mortgage,purchase-and-sale, and foreclosure transaction in Massachusetts from 1989 to 2008. We address theidentification issues related to the estimation of the effects of house prices on residential foreclosures. Wethen use the model to study the dramatic increase in foreclosures that occurred in Massachusetts between2005 and 2008 and conclude that the foreclosure crisis was primarily driven by the severe decline inhousing prices that began in the latter part of 2005, not by a relaxation of underwriting standards on which much of the prevailing literature has focused. We argue that relaxed underwriting standards didseverely aggravate the crisis by creating a class of homeowners who were particularly vulnerable to thedecline in prices. But, as we show in our counterfactual analysis, that emergence alone, in the absence of a price collapse, would not have resulted in the substantial foreclosure boom that was experienced.JEL classification: D11, D12, G21Key words: foreclosure, mortgage, house prices
 
1. Introduction
There are two competing theories to explain the explosion of foreclosures in the United States in 2007–2009. The first focuses almost entirely on underwriting standards, and attributes the crisis to loans thatborrowers had trouble repaying, either because they had bad credit and little income to begin with, orbecause the loans were unrealistically generous at the time of origination and later became unaffordable.There is persuasive
prima facie
evidence that is consistent with this theory. For example, subprime loans,which expanded credit by offering loans to borrowers who previously could not obtain any mortgage credit,and by offering large loans to those who previously could only obtain small loans, account for a dispropor-tionate number of foreclosures. Subprime lending emerged as a major force in mortgage markets shortlybefore foreclosures began to accelerate, and were concentrated in communities that later became foreclosurehotspots. An alternative explanation focuses on house prices and points to the precipitous decline in pricesthat started in 2005 or 2006, depending on where in the country one looks, as an explanation for the crisis.There is
prima facie
evidence that is consistent with this theory as well. For example, until house prices be-gan falling, subprime mortgages performed very well. In addition, states where house prices began decliningsooner experienced foreclosure waves sooner than in states where prices began falling later.In this paper, we assess the merits of the “poor underwriting” and “falling house price” theories. Weargue that, in a sense, both explanations are true, but that prices have primacy. We find that had prices notfallen, we would simply not have had a major foreclosure crisis, regardless of whether lenders had loweredunderwriting standards in 2003 and 2004. By contrast, the observed fall in prices would have generated asubstantial increase in foreclosures, even if lenders had retained the underwriting standards that prevailedin 2002. To be sure, the increase in foreclosures would have been substantially smaller without subprimelending because as we show that subprime loans are far more sensitive to a decline in house prices thanprime loans, but the foreclosure rate would still have been very large relative to historical levels, and wouldhave been still considered a major public policy problem.The central issue that any researcher must address in such an analysis is how to identify the effect of house prices on foreclosures. One natural explanation for the observed relationship is that the increasein foreclosures caused prices to fall. If that were the case, then the observed relationship between pricesand foreclosures becomes evidence for the bad underwriting theory not the falling prices theory. The firstcontribution of this paper is to use a novel identification strategy to test this explanation of the crisis. Weexploit within-town, within-time variation in household-level measures of housing equity to refute the theorythat price falls resulted from differences in foreclosures over time and across communities. If foreclosurescaused the fall in prices, we would expect to see little effect of prices on foreclosures within communities andwithin time periods, and large differences in the effect of prices on foreclosures across communities and overtime due to simultaneity bias. In fact, we find that the observed relationship between prices and foreclosuresis virtually identical whether we use variation across communities, within communities, within a given timeperiod, or over time.We use a dataset based on deed records in the Commonwealth of Massachusetts from 1989 through2008, which has been previously used to study housing issues, most notably, in an influential paper onforced sales and foreclosures by Campbell, Giglio and Pathak (2009). The dataset was constructed by theWarren Group, and has both advantages and disadvantages over other datasets used to study foreclosure andmortgage performance. The main disadvantage is that we do not observe detailed borrower characteristicssuch as credit score, or loan characteristics such as whether the interest rate is adjustable or fixed. However,we do observe the identity of the lender and we use this information to match the data with the HUD list of subprime lenders in order to identify loans as subprime or prime. Using a subset of loans for which the dataprovider recorded more detailed loan information, we show that these loans conform to generally accepted1
of 00

Leave a Comment

You must be to leave a comment.
Submit
Characters: ...
You must be to leave a comment.
Submit
Characters: ...