How to Reduce the Government's Trillions In Housing Credit Risk
Over $5.8 trillion in home mortgage debt in the United States is now either owned or guaranteed by a federal entity - be it the Federal Housing Administration (FHA), Ginnie Mae, the Veterans Housing Administration, or one of the two government-sponsored enterprises (GSEs) under "conservatorship" since 2008. This mother lode of guaranteed debt constitutes a large contingent liability facing American taxpayers. If a future economic crisis triggers widespread defaults within this mortgage pool, hundreds of billions of dollars could be added to the nation's already unsustainable deficits.
The solution to this crisis in the making is the wholesale transfer of housing credit risk back to the private sector. While an asset sale akin to the post savings and loan crisis clean up effort may not be possible, the U.S. can work down these guarantees by reawakening the private mortgage finance market from its post-crisis slumber. The objective would be to replace government mortgages with private loans as owners move or refinance (on average, mortgages are refinanced every seven to eight years). To accomplish this, however, and remove this liability from the taxpayers, there are four large roadblocks that have to be removed before private capital can take on residential mortgage risk.
Three of these roadblocks have widely discussed remedies. The first roadblock is the high conforming loan limits for Fannie Mae and Freddie Mac (266 percent of the current Case-Shiller median mortgage price), and an even higher conforming loan limit for the rapidly-increasing-in-market-share-duo FHA and Ginnie Mae (311 percent of median mortgage price). These high caps on the mortgages eligible for their subsidized guarantees mean that the private sector can't compete on nearly all mortgages in the U.S.
The answer here is simply to lower conforming loan limits gradually to give the private sector more room to compete.
The second roadblock is a combination of mortgage provisions in the Dodd-Frank Act, the most restrictive of which is the risk retention requirement. While well intended, there are a host of unintended consequences in this requirement that residential mortgage-backed securities (RMBS) underwriters hold five percent of risk on all mortgages except those determined to be super safe - the yet to be finalized "qualified residential mortgage" (QRM). The worst is that with Fannie, Freddie, and FHA exempted from these requirements they will have a cost advantage against which the private sector can't compete.
The remedy to this and related problems is to repeal the QRM and risk retention provisions of Dodd-Frank and instead have strict resolution and bankruptcy procedures established for those that take on too much mortgage-backed security risk.
The third roadblock is the complex legal framework governing RMBS. The high level of defaults following the subprime meltdown have exposed an industry that never prepared (legally or emotionally) for losses on this scale, resulting in years of "tranche warfare" - the fight among various classes of investors as to who shoulders these losses.
The solution to this process is evolving, with a combination of legislative clarity, new regulatory frameworks, and judicial review. One private issuer - Redwood Trust - has already demonstrated the ability to create RMBS that investors can trust, having successfully launched five jumbo deals since 2010.
But to move beyond small scale RMBS issuance, the U.S. must clear the fourth roadblock to private capital returning to housing finance: a profound lack of confidence in the models used by credit rating agencies to assess residential mortgage-backed securities and in the rating agencies themselves.
Two tools to overcome this roadblock can be found in our newly published Reason Foundation study, "Restoring Trust in Mortgage Backed Securities," which outlines a series of policy initiatives in regards to ratings problem for mortgage assets.
To start, we suggest that Congress authorize underwriters to include property-level address data in RMBS disclosures so that investors or independent analytic firms can perform more detailed and accurate risk assessments at lower cost. Currently, it is both illegal and taboo for mortgage investors to receive the addresses for the properties backing their bonds. Without address level data it is impossible to get a precise fix on a mortgage's value in real time.
Prevailing wisdom has suggested that the zip code is enough to perform loan level analysis, but the build up of toxic debt and subprime crash demonstrated this view to be mistaken. Consider that zip codes can contain several thousand properties and may also embrace large, often heterogeneous areas. For example, in northeast D.C.'s 20002 zip code, Zillow.com recently listed a price range of $250,000 to $2.8 million, depending on whether you want to live in Little Trinidad or in an upscale townhouse on Capitol Hill.
We also suggest that the mortgage-finance industry enforce common formatting of RMBS collateral data and the inclusion of cashflow-waterfall models with prospectuses to make investor due diligence easier, more competitive, and less costly. This would enable more third party analysts and research firms to more readily analyze the risk of an RMBS offering, providing more competition for the major rating agencies.
This, combined with the availability of address level data, would enable investors to perform their own in-house risk assessment and track the value of their investment over time without having to rely on an outside ratings assessor.
In the spirit of finding alternatives to letter grades from ratings agencies, we further propose the mortgage finance industry should create a Mortgage Underwriting Standards Board loosely based on the Financial Accounting Standards Board (FASB) model. This organization would provide self-regulation against misrepresentation and could even replace the QRM standards with industry established categories of mortgages that are transparently defined by risk appetite - as compared to the status quo binary standard of AAA, BB-, CCC+ that conveys little information about the true nature of a product.
We recognize our proposals are just the beginning of addressing the confidence problem that plagues the private sector. Our study has more proposals including standardizing the formatting of loan level data, and authorizing third parties to challenge ratings provided by the ratings agencies when they are used for capital adequacy purposes. But we do not suppose this is the perfect solution (we are fallible economic analysts like many others after all). Our fear, though, is that without a debate about the best ways to tear down this roadblock, the pace of private capital's return to mortgage finance will be sluggish at best. We hope this contributes to the conversation.
Last year, over 90 percent of mortgage originations received some form of federal backing - either from the dominant giants Fannie Mae and Freddie Mac, or from the new rising-star, FHA, recently granted an increased conforming loan limit. Taxpayers will remain at substantial risk as long as these entities remain the bedrock of American housing finance.