An implication of ongoing research on the economics of mortgage contracts is that loan underwriting quality may not be a very important factor in causing losses in the absence of a steep decline in asset prices or employment.
Conventional wisdom about the losses suffered by holders of subprime mortgages points to poor underwriting standards as a main culprit. Had it not been for mortgages with high loan-to-value ratios, poor documentation, or those extended to non-creditworthy borrowers, so goes the thinking, much of the subsequent defaults and losses to the securities that pooled them would not have occurred.
However, a 2009 piece by Yuliya Demyanyk questioned the relevance of poor underwriting standards to subprime mortgage losses by simply noting that default rates doubled for mortgages originated in 2006 and 2007 despite having roughly the same underwriting standards as prior years. Papers by Kristopher Gerardi, Paul Willen, and their co-authors have likewise questioned the role of poor underwriting in causing mortgage defaults. They instead argue that the collapse in housing prices was a more fundamental reason behind mortgage foreclosures. Without the housing price decline, poorly underwritten mortgages would not have foreclosed in such great numbers, and (expectations of) rising housing prices likely caused the decline in underwriting standards in the first place (rather than vice versa).
In line with the Gerardi and Willen research is a recent article in the Summer 2011 issue of the Journal of Structured Finance by David Lehman II, Tyler Simpson, and Florin Nedelciuc (submitted version freely available here). The authors analyze mortgages in what they consider to be a representative sample of four subprime mortgage securitizations from 2006. They look at the relationship between underwriting standards and loan default and find little or no relationship between most of them. In particular, the authors found that, except for loan-to-value ratios, subprime mortgage loan performance was not affected by underwriting characteristics such as documentation quality, the borrower's FICO score, or whether the property was owner occupied or the purpose of the loan was for purchase (as opposed to refinance). They did find a strong correlation between changes in employment and housing price depreciation, which adds further gloss to the Gerardi and Willen line of research to the extent employment may be a more fundamental cause than housing prices.
An important implication of the studies questioning the relationship between underwriting quality and loan performance is that in the absence of broader negative macroeconomic developments (such as price declines or increases in unemployment), underwriting quality may be relatively unimportant. Accordingly, to the extent current levels of real estate prices and employment are near bottom, then low or declining underwriting standards may not be as large a concern as would be the case otherwise. Declining underwriting standards in the post-crisis wave of loans in commercial mortgage-backed securities may therefore not present a very large problem if broader macroeconomic conditions are not poised to worsen much further. In addition, credit risk transfer governance may need to be better structured to deal with macro risks rather than asset-level problems.
The role of mortgage underwriting standards in explaining losses and defaults is nonetheless still a subject of debate. A review of relevant studies by Christoper Mayer (pages 14-15), and a recent paper by Patrick Bajari et al., present evidence on the importance of underwriting to explaining mortgage losses.
I don't get this. Of course lax underwriting standards were an important cause of the crash. A very obvious explanation that is ignored in this post is that even non sub-prime mortgages did not have satisfactory underwriting standards on average to begin with. And even if those mortgages saw an increase in their default rates, those default rates are still less pronounced than those for the sub-prime mortgages. If mortgages were required to have a max LTV of 80% (bye-bye FHA, VA, and USDA mortgages), along with a minimum credit score (say, 650) and max PITI and DTI ratios of say, 30/40, then any recession, if any, would be far less pronounced. If you had those UW guidelines, then a large housing bubble would not have developed. If a less pronounced bubble did develop, you wouldn't see as sharp a decline in home values. Any decline in this scenario would have a modest effect on default rates since (1.) people's homes suffer less depreciation ,(2.) homeowner's have more equity to begin with (think that 80% LTV requirement), and (3.) minimum credit scores and lower max ratios bolster the credit quality of homeowners.
Posted by: AJ | September 22, 2011 at 08:04 PM
The authors do not deny the importance of underwriting standards. The more important questions are how much they matter relative to other factors such as housing prices, under what conditions, and whether underwriting standards were a cause or effect of rising house prices (and expectations about them).
As noted at the end of the post, there is abundant research finding that underwriting standards have a prime explanatory role, contrary to the research highlighted.
Posted by: Houman Shadab | September 29, 2011 at 05:33 PM