Why the US Mortgage Market Will Remain Heavily Dependent on Government Support

In Senate Banking Committee testimony today, Georgetown law professor Adam Levitin explains why the private label (non-government guaranteed) mortgage market is a textbook case of what Nobel prize winning economist George Akerloff called a “market for lemons”. While that conclusion is not news to those who’ve been following the financial crisis and its aftermath, Levitin reaches a second conclusion that many policymakers have been keen to finesse: that the US mortgage market, even if private mortgage securitizations were reformed (which they have not been), is bound to remain heavily dependent on Federal guarantees. Even under generous assumptions, Levitin concludes:

Screen shot 2013-09-28 at 2.52.00 PM

The implication is that the pet dream of Republicans, to kill the GSEs, isn’t realistic unless they want to kill the housing market as it goes through a brutal transition period to more on-balance sheet bank lending. The reason is not simply securitization allowed for mortgages to be issued much more cheaply due to eliminating the cost bank equity and FDIC insurance. A bigger reason is that the 30 year fixed interest rate mortgage with an unrestricted borrower right of prepayment would never exist absent government support.

So it should be no surprise that the various mortgage insurance “reform” programs are simply to have supposedly better GSEs replace Fannie and Freddie but still wield a government guarantee (remember that the perverse structure of having private sector entities run a government guarantee program is to avoid consolidating the debt related to this mortgage insurance activity).

The private label mortgage market has barely made a contribution to housing finance post the crisis. Since 2008, there has been a grand total of 33 private label mortgage securitizations, financing a grand total of under 17,000 mortgages during this entire period versus a total of 8.6 million mortgages in 2012. And these mortgages weren’t just “jumbos,” meaning larger than “conforming” or Fannie/Freddie mortgages. They were also super duper high credit quality (click to enlarge):

Screen shot 2013-09-28 at 3.32.01 PM

“Fully documented” means that, among other things, the lender verified the borrower’s income.

And why did these mortgages need to be of such impeccable quality? Mortgages represent two types of risk: interest rate risk and credit risk. Interest rate risk is more complicated than for normal bonds, since consumer mortgages in the US have an option that works against the investor, namely, an unrestricted right of the borrower to refinance if interest rates fall. So the time when you really like being a bondholder, when interest rates have dropped but you are holding fixed income investments made when interest rates were higher, is exactly when mortgage investors take your nice high-yielding instrument away by prepaying. That produces a related unattractive feature, that you don’t know very well what the maturity of your investment will be. So investors demand, and get a premium over simple coupon bonds (such as corporate bonds) to compensate for these unattractive characteristics.

So investors can be compensated enough to get them to accept the peculiar interest rate features of mortgage-backed securities. By contrast, Levitin explains that there never was much investor willingness to take on the credit risk of consumer mortgages. Remember, private label securities are tranched to create different levels of credit risk. The AAA tranches were supposedly insulated from losses due to overcollateralization and excess spread (finance-speak for different loss cushions built into the deal) and the fact that there were other tranches sitting below them in payment priority who would take losses before them if something really bad happened.

But as we know now, that proved to be a fantasy. The AAA investors typically bought blindly, relying on the rating and the assumption that the lower-tranche investors, particularly the ones in the worst rated tranche, the BBB investors, would do the homework. A deal could not be funded unless the BBB tranche was sold, since the investment banks who were structuring and selling these securitizations didn’t want to wind up holding the bag. And indeed, for some of the life of the private label securitization market, that sort of due diligence took place. Up through 2003 or so, there was a group of “B” investors who would acquire loans tapes and analyze the deals before investing. But the pool of investors both willing to do the work and with the risk appetite constrained the size of the private label market.

But then, enter the Ponzi scheme knows as the CDO. CDOs displaced the traditional “B” investors as the buyers of BBB tranches. But CDOs were not able to sell all of their BBB tranches (even after hawking them to every stuffee around the world they could find, enticed by liberal use of expensive wine, drugs, and hookers*) to investors. So those BBB pieces were rolled into other CDOs (even first generation CDOs could contain a specified percent of low-rated CDO tranches, and some were sold into the more obviously toxic CDOs known as CDO squared and CDO cubed**).Thus Levitin concludes that the maximum size of a sustainable PLS market it is level before the CDO bid overwhelmed the market (and common sense). At its peak, it was $300 billion versus a US mortgage market whose annual financing needs ranged from $2.5 to $4 trillion and thus able to supply at most a quarter, and most likely more like one eight, of America’s housing finance demand.

Even that level of activity assumes either investor amnesia (which typically takes 10 to 20 years as new managers replace the ones that were burned) or reforms. But the sell-side (the large securities firms and the bank servicers) have fought sensible notions, like a short list of improvements put forth by the FDIC in 2010, tooth and nail. Further complicating the situation is that reform would have to address multiple aspects of how private label mortgage securitizations work. For instance: servicers need to be compensated more so that they can service delinquent loans properly (right now, it is more profitable for them to foreclose, so that’s what they do absent crude government cudgels and bribes). Trustees have simply refused to do their job of policing servicers and representing investors; they need to be held to account. And anyone who reads the business press even casually knows that rating agency reform (or approaches that reduce the importance of ratings for certain types of investors) seems much more remote than in 2009. So there is no reason to think the PLS market will return to anything dimly representing even its pre-CDO size any time soon.

The issues discussed above mean that the “let’s kill Fannie and Freddie and let private capital fill the void” is a fantasy. The issue is not whether we continue to have government backstopping of a lot of mortgage credit risk, it’s how we go about doing it (yes, it might have been better not to be in this position, but life is path-dependent, and there is no easy and low-damage way to undo a housing finance system so dependent on Federal guarantees). And as we expect to discuss in future posts, some of the supposed “private sector” replacements for Fannie and Freddie give a central role to the government-sponsored enterprises know as too big to fail banks. Sadly, the debate over an issue this important to consumers and taxpayers is almost certain to remain shrouded in deliberately misleading messaging and arcane but critically important technicalities. But why should we expect GSE reform to be any different than financial reform generally?

Update: Neglected to embed Levitin’s very readable testimony. Apologies.

Levitin Senate Banking Testimony 10-1-13

This piece is cross-posted from Naked Capitalism with permission.