Strategic default: how big an issue?
March 12, 2014 12:00 PM
by Maia Woluchem
[Reprinted with permission. View the original article on Urban Institute's MetroTrends blog.]
An interesting element of the Great Recession’s foreclosure crisis has been the increase in strategic default - when borrowers who can still make their monthly mortgage payments instead choose default and possible foreclosure. A popular explanation for this counterintuitive choice is that many homeowners whose mortgages are bigger than their home’s value simply “give up”.
But is that the whole story? At Monday’s Monthly Data Series, an event hosted by the Urban Institute’s Housing Finance Policy Center, researchers Michael Bradley and Paul Willen shed new light on the issue through their recent research.
Neighborhood Effects Matter
In a paper co-written by Amy Crews Cutts and Wei Liu, Bradley used data from CoreLogic’s databases, as well as credit bureau and income data from Equifax, to explore the factors determining strategic default, both at the macro-level and at the ZIP-code level.
They discovered that tallying how many borrowers strategically default on properties depends on exactly what “strategic default” means. When it’s defined narrowly to mean borrowers who did not make the next four payments after already having missed two payments on their home, were not delinquent on any other trade line, had negative equity, and did not lose a big share of their income, less than 10 percent of all defaults were strategic.
That is a much lower number than was found in earlier studies that used a broader definition of strategic default. Among the factors that may increase one’s propensity to strategically default: higher credit scores, higher combined loan-to-value ratio (CLTV, the value of all liens on the property as a percentage of the property value), and living in a non-recourse state where the borrower cannot be pursued for payment after a foreclosure.
Geography provides yet another clue about who might strategically default. There is clear a relationship between strategically defaulting and having neighbors who do the same. It’s possible that defaulting neighbors reduce the stigma of strategic default, and may even offer an example of the consequences, or (lack thereof) from such actions.
Unemployment Changes the Picture
Researchers Kris Gerardi and Paul Willen find evidence that the “negative equity” explanation for strategic default is over stated. Co-writing with Kyle Herkenhoff and Lee Ohanian, they argued that the effect of unemployment is just as important as the effect of negative equity, in terms of what pushes borrowers to default.
Willen and his co-authors used Panel Study of Income Dynamics data to test the effect of certain life events like unemployment, divorce, and low wealth on homeowners with an already high CLTV. They find that at lower CLTVs, these events raise default rates by a huge order of magnitude, while at higher CLTVs the effect is smaller. For example, for borrowers with CLTVs between 75 and 100 percent, 10 percent of the unemployed defaulted, versus 2 percent of the employed. At over 125 percent LTV, 28.8 percent of the unemployed defaulted versus 17.6 percent of the employed.
Alternatively, there is a high fraction of borrowers who remain current on their loans despite multiple triggers. In his paper, Willen shows that nearly 90 percent of borrowers with no job and no wealth are still current on their mortgages (a figure that drops to 73.3 percent for borrowers with an LTV above 125).
Essentially, the model should look more like the choices a buyer faces in a call option, says Willen. “Should I make one more payment to preserve the option on this house?” Ultimately, it’s a highly individual decision.
The Audience Speaks
Bob Avery from FHFA was the discussant for both papers. He complimented the authors on the fact that they had used better data than previous work, noting that the Willen paper used “ideal data”, despite the limited sample. The Bradley paper used the best possible data for a larger sample pool.
Both speakers, as well as their co-authors Amy Crews Cutts and Kris Gerardi, emphasized that the predictive nature of their research can be applied to loan origination, as well as to loss mitigation and effective modifications.
During Monday’s session, an audience member asked, “Why would you say we’re talking about all this?” Bradley replied, “The first sentence in the [mortgage] contract is ‘I promise to pay’”, and if we can’t help to avoid defaults, we might be facing a cost structure for home mortgages with interest rates as high as credit cards.