During the buildup to the financial crisis, mortgage underwriting standards weakened dramatically.  Although there have long been loans made with low or no down payments, or to people with FICO scores below 660 or even 600, or to people who did not wish to provide full documentation, it was rare to find loans made that had every one of the these features.  But that was the case in the mid-2000s.  Even as underwriting standards weakened, mortgage credit performance (measured by rates of delinquency, foreclosure and loss) were historically low and mortgage investors were not only content but eager to buy more of these types of loans (generally, in securitized form).  Undoubtedly, the greatest driver of the excellent credit performance through 2005 was the strength in housing price appreciation during this period.  Even if a borrower was facing difficulty with debt service, it was typically possible to either refinance or sell the underlying property.

However, once housing prices peaked out in 2006 and began to recede, the dynamic changed dramatically.  Many lenders, securities underwriters and mortgage investors found themselves in precarious straights with highly levered investments in complex MBS funded with volatile short-term borrowings.  As housing prices softened, many mortgagors, particularly those who were speculating on a second or third property, chose to walk away from their mortgages and there was a surge in so-called early payment defaults.

Ultimately, the positive feedback loop that had driven home prices and mortgage debt higher during the boom turned around and fueled an accelerating housing bust.  Worsening credit performance caused lenders to demand greater haircuts on loans to mortgage lenders and investors.  The latter responded by raising capital or dumping positions.  Lenders tightened underwriting standards and the effective demand for housing declined.

After the dust settled, MBS investors and guarantors claimed that the losses were primarily the responsibility of mortgage lenders.  Many large lawsuits were brought by plaintiffs arguing that the quality of the loans that they purchased was misrepresented by the sellers or underwriters.  Even while arguing that the vast majority of losses were due to the combination of major declines in housing prices and weak underwriting standards that were widely known at the time, most lenders have acquiesced to settle these lawsuits, sometimes with very large damages.

But at least one defendant refused to settle.  Nomura Securities and Nomura Credit (collectively, Nomura) was sued by the Federal Housing Finance Administration (FHFA), the regulatory agency that oversees the giant housing agencies Fannie Mae and Freddie Mac (the GSEs), for losses suffered on mortgage securities sold to Fannie and Freddie by Nomura.  The FHFA’s position is that there were material errors in the information provided by Nomura.  In particular, the percentage of loans where the owner was the occupant was overstated and the aggregate loan to value ratio (LTV) was understated.  Nomura’s position appears to be that data errors, if any, were not the key factor driving the losses suffered by the GSEs.

Earlier this month the judge in the case ruled in favor of the FHFA on the grounds that “The offering documents did not correctly describe the mortgage loans.”  Regarding the magnitude of the damages, the judge ordered the FHFA to submit a revised estimate.  It did not appear that the judge bought the Nomura argument data errors were not the primary cause of the losses.  On that issue, she commented that “the banks have not quantified the loss that they say is due to macroeconomic factors.”

While I have zero insight or knowledge about the legal issues involved, the question of how one might reasonably allocate the losses is interesting.  Surely Nomura is correct that a key factor driving mortgage losses over the period 2007-2012 was the dramatic drop in housing prices in the period 2006-2010.  Also, surely the FHFA is correct that higher incidence of non-owner occupancy and higher LTV led to greater losses.

One approach would be to argue that the lender is responsible for only that portion of the loss attributable to incorrect information.  For example, suppose the data show that losses on a pool of non-owner loans were on average 10% of the pool balance while losses on an otherwise identical pool of owner-occupied loans were on average 8% of the pool balance.  Further suppose the lender had stated the loans were owner-occupied but that turned out not to be correct.  Then it seems to me that it would be reasonable for the lender to absorb 20% of the losses (2% of the pool balance).  In this case the investor is on the hook for 80% of the loss (that part attributable to the housing price decline) and the lender is on the hook for 20% of the loss (that part attributable to incorrect data).

That is probably not how it is going to turn out.