Ed DeMarco, the Acting Director of the Federal Housing Finance Agency for the past three years, will deliver a speech today at the Brookings Institution that leans in the direction of allowing principal reduction on Fannie and Freddie-backed loans in certain cases. This is definitely a crack in the otherwise rigid objection to principal reduction on these loans.

DeMarco states flatly that “I will not be announcing any conclusions today,” because the analysis of the efficacy of principal reductions being carried out by his office have not been completed. However, he does provide preliminary findings from this analysis, which we can look at to assess the policy going forward. DeMarco then segued into a description of the human element of the crisis, something largely missed by most discussions of housing at the federal level.

Throughout this crisis each of us know of, or have heard about, many individual stories of homes lost through foreclosure. One cannot help but have sympathy for those who have suffered such misfortune. And surely no one can look at the dislocations in the housing market and not feel frustration at how so many people and institutions failed us, whether through incompetence, indifference, or outright greed or fraud. Yet we are also blessed in this country with people and institutions who care, who are strongly motivated to provide assistance and find solutions [...]

Whichever of these categories any particular homeowner falls into, the decline in house prices over the last few years has reduced the housing wealth of all homeowners. The Federal Reserve has estimated that from the end of 2005 through 2011, the decline in housing wealth to be $7.0 trillion.

Six years into this housing downturn, the losses persist. The debate continues about how we as a society are going to allocate the losses that remain. Asking hard questions in this debate does not make one unfeeling about the personal plight this situation has created for so many. Indeed, the majority of those most hurt by this housing crisis did nothing wrong – they were playing by the rules but they have been the victims of timing or circumstance or poor judgment.

That part about “people who care” is debatable. And the whole thing sounds like a defense mechanism against being called, well, a person who doesn’t care. But let’s see what DeMarco’s got.

First, he rationalizes the steps Fannie and Freddie have already taken to assist troubled borrowers. This includes reducing interest rates, extending the term of the loan, or forbearance (setting aside principal to reduce interest payments; the principal gets paid off at the end of the loan). DeMarco is very jazzed on forbearance, because he says that the GSEs still retain a stake in the upside equity of a loan if the value appreciates, as opposed to in a principal reduction scenario. Of course, when the borrower has trouble paying the balloon payment when it comes due, these fantasies about the efficacy of forbearance will crash.

Yet, DeMarco comes with some interesting statistics. He notes that the GSEs hold 29% of all seriously delinquent loans, yet they have performed 52% of all the active HAMP modifications. Fannie and Freddie modifications also have lower re-default rates than their countrerparts (although still high, at around 27% within 12 months of the mod). This is more an indictment of the banks and what they’ve done with their portfolios than anything else.

In another chart, DeMarco claims that loan-to-value, the percentage of underwater-ness on loans, has little bearing on whether one of the Fannie and Freddie modifications is successful. Of course, none of the GSE mods include principal reduction, so it’s hard to draw any conclusions from that. DeMarco’s trying to make the point that more underwater GSE borrowers aren’t experiencing a high re-default rate after their mods. Rather, the percentage of the payment decrease correlates with re-defaulting.

Then DeMarco gets to principal reduction, and in particular the new incentives through HAMP, which triple the financial benefit for a principal write-down by the GSEs, mainly to encourage them to perform them. First DeMarco acknowledges some realities:

In fact, historical data has shown that the probability of default correlates with the borrower’s current LTV ratio; the higher the ratio, the greater the likelihood of default. So, in theory, by forgiving principal and reducing a borrower’s current LTV ratio, the probability of default is reduced and losses are reduced. This type of relationship between default and current LTV, supported by previous analytic work, is embedded in the HAMP NPV model, and thus has been explicitly factored into FHFA’s repeated analyses of principal forgiveness.

DeMarco summarizes FHFA’s previous analysis of principal reduction, saying basically that forbearance allows the GSEs to share in the upside while still reducing payments a similar amount that you would see from principal reduction. He shows an analysis from January that made some assumptions about borrowers with over an 115 LTV, and how much the GSEs would be impacted by various strategies.

• If borrowers are offered no modifications, Enterprise losses would be $101.8 billion;
• With forbearance-only modifications, losses would be $24.3billion less, or $77.5billion;
• With forgiveness-only modifications losses would be $21.0 billion less, or $80.8 billion.

That’s not a heck of a lot of difference on a $4.5 trillion portfolio. And it acknowledges lower re-default rates from principal reductions.

Then DeMarco addresses the second lien issue, the elephant in the room:

Before moving on, a few words about credit enhancements are appropriate. While FHFA did not analyze the impact of credit enhancements, it should be clear that principal forgiveness confers benefits on parties in a first loss position, such as mortgage insurers and subordinate lien holders. In fact, any modification of a first lien that reduces the probability of the homeowner suffering foreclosure makes these third parties better off since they are in the first loss position in the event of foreclosure. In the case of principal forgiveness the Enterprises bear all the losses of the write down and share the benefit of the lower probability of default with the third party. In the event of a subsequent foreclosure, the Enterprises bear all the losses of the write down and the third party realizes all the benefits of it before incurring losses, if any.

In plain English, a write-down without mandatory write-offs of second liens enriches the banks. That’s just a fairness issue. DeMarco reveals that one of its enterprises (he doesn’t specify, Fannie or Freddie) found that “almost 50 percent of its underwater,
seriously delinquent loans have at least one subordinate lien on the property.” When you combine that with mortgage insurance or recourse agreements, DeMarco believes that over half of all the seriously delinquent underwater mortgages in Fannie and Freddie’s portfolio have some third party credit interest. So it’s not a small issue. And he makes the obvious point about the Second Lien Modification Program (2MP), which only forces write-downs of second liens on equal terms as the firsts:

While HAMP and 2MP provide for similar treatment of first and second liens, these modifications are more favorable for second lien holders because first lien holders share in the second lien holders overall losses, which is inconsistent with lien priority. Without modifications, second lien holders would absorb all of their own losses, instead of sharing them with the first lien holder.

Finally, DeMarco shows a preliminary finding based on the tripled principal reduction incentives. And this one does show an improvement for principal reduction relative to forbearance. For the group of loans studied, principal reduction with the tripled incentives reduces losses to the GSEs by $1.7 billion. The taxpayer would pay for those tripled incentives, however, with a net cost to the taxpayer of $2.1 billion. However, those incentives have already been authorized long ago through the HAMP portion of TARP and were meant to go toward reducing foreclosures, which have their own economic effects not counted in this analysis.

Again, these are pathetically small numbers relative to the total GSE portfolio. I can’t believe we’re spending this much time on it.

DeMarco then goes on about the mythical possibility of strategic defaults in order to get a principal reduction, which is a dog that just hasn’t barked throughout the foreclosure crisis. Nonetheless, DeMarco goes through a long sequence on supposed “strategic modifiers” and how many of them would wipe out the benefit from principal reduction. He closes by highlighting potential operational costs to a new program.

DeMarco kind of invalidates his initial claim about human costs with this analysis. The idea that every delinquent borrower is just waiting to take advantage of the government is unsupported by reality. But it’s driving some decision-making inside the GSEs. Furthermore, DeMarco only looks at principal reduction as affecting a small universe of seriously delinquent borrowers. Current borrowers would not get such a reduction, though they could be eligible for refinancing. DeMarco sets potential eligibility at around 1 million borrowers. “This is not about some huge difference-making program that will rescue the housing market,” he says. Again, you wonder why it has become such a central focus of everyone, in that case.