The Fatal Flaw in the Johnson-Crapo GSE Reform Bill

APR 22, 2014 6:57pm ET

If you saw "Titanic" you probably understand subordination, and the reason why there was a meltdown in September 2008. And you might also understand why Johnson-Crapo invites a similar disaster going forward.

More than a century ago on the Titanic, the captain set access to lifeboats according to a hierarchy of classes: First class women and children first. As a result, the mortality rate among women in first class was not catastrophic, whereas it was for men traveling in steerage, and for those working in the boiler room.

Private-label residential mortgage securitizations are designed around a strict hierarchy of classes, which determine, not loss of life, but loss of principal. These modern marvels of financial engineering had purportedly found a way to contain any potential damage, by making sure that all suffering was borne first by the lower classes.

For tranches that absorbed up to the first 10% loss in any given mortgage pool, the results might be catastrophic, but the senior tranches would feel no pain.

Private-label deals represented a small subsector, about 20%, of the $11 trillion residential mortgage market as of yearend 2007. And the deeply subordinate tranches represented a small sliver of that subsector. And in 2007 if you hadnít paid attention, it might seem like catastrophic losses, contained within a small sliver of a subsector, would not have much of a ripple effect.

Had the lower classes been spread and disbursed throughout the vast universe of the institutional debt markets, the catastrophic losses borne by a sliver of a mortgage subsector might not have had much of a ripple effect.

Unfortunately, these risk exposures were concentrated among a relative handful of major institutions that didnít think about the risks of subordination.

For the most part, those deeply subordinated tranches were never actively traded in the secondary markets. Rather, they were warehoused, bundled and sold off as portfolio investments known as collateralized debt obligations. CDO structures were designed and sanctioned by the ratings agencies, who never gave much thought about the impact of subordination in a mortgage bond portfolio.

Within any private-label deal, the geographic diversification is a lot less than meets the eye. It inevitably follows that the senior tranches get paid down from the homeowners that prepay earliest, those individuals who refinance whenever they find it advantageous to do so. The subordinate tranches are more reliant on homeowners who are unable to refinance, and are more likely to be underwater because of prices in the local market. Think of it this way, if you sold travel insurance on the Titanic, you didn't really diversify your risk by selling more policies to passengers traveling in steerage.

Top executives at large firms were ignorant of the risks from deeply subordinated mortgage debt embedded within CDOs. In case you forgot, AIG, Citigroup, UBS, Merrill Lynch, MBIA and AMBAC all faced the specter of insolvency because their collective risk exposure to subprime CDOs exceeded $300 billion.

The losses from those particular CDOs, about 3% of all mortgage debt, exceeded all the credit losses from $4.8 trillion worth of mortgage debt financed by Fannie Mae and Freddie Mac as of yearend 2007.

Because CDO losses had impaired their own capital, and because they faced with a crisis of confidence among their counterparties, AIG, Citigroup, UBS, B of A/Merrill, MBIA and AMBAC and others were forced to cut back on credit they extended to customers. So their actions drained liquidity from the marketplace, creating a downward spiral and, well, you know the rest.

All this was traceable to a tiny percentage of the nationís overall mortgage debt, where a very high concentration of risk and loss was found among tranches in the bottom 10% of their dealsí capital structure.

Johnson-Crapo reflects a similar kind of obliviousness to the risks of subordination. In a nutshell, the bill would demolish Fannie Mae and Freddie Mac, and the types of housing finance they offer, and replace them with a system offering a new type of residential mortgage securitization, one which is insured by combination of private parties and the federal government.

Private parties would insure the first 10% loss, and any losses above that threshold would be insured by the federal government. In effect, these would be two-tranche securitizations, with a risk-free senior tranche representing 90% of the total, and a second tranche assuming private credit risk, which is subordinate to 90% of the deal's capitalization.

To be clear, Johnson-Crapo securities are not like those toxic CDOs. Structurally, they are far riskier. CDOs were packed with investment-grade tranches that, though they were subordinate to 90% of a deal's capitalization, remained structurally senior to the sub-investment grade or equity tranches, which provided a 4% cushion. Johnson-Crapo securities have no equity cushion.

These deeply subordinated tranches are the only reference point we have for estimating whether the Johnson-Crapo system will be successful. Private mortgage insurers may be adept at taking on subordinate credit risk, but almost all of their business is on a loan-level basis, not on a portfolio basis. And history shows that the private mortgage insurers' collective risk appetite had its ebb and flow. By contrast, Fannie and Freddie have served as a buffer to market cyclicality.

Hereís an inconvenient truth: It is simply impossible for private-label deal to attain anything close to the level of risk diversification by a GSE securitization. Forget about any kind of implicit or explicit government support, just look at the dealsí structure and their performance record. GSE securitizations benefit from corporate guarantees, so they are, by definition, exponentially safer than private-label deals.

Comments (0)

Be the first to comment on this post using the section below.

Add Your Comments:
Not Registered?
You must be registered to post a comment. Click here to register.
Already registered? Log in here
Please note you must now log in with your email address and password.
MultimediaSee All »