The Some Loss and Stop Loss Standards

The times have been a-changin' at an exhausting rate for years now. Bond insurance as a viable business has been on trial during all that time, but only recently an alternative business model gained preliminary attention. That may or may not be all it ever gains.

Several developments have encouraged the insurance of crappier, smaller credits. S&P's leverage limits handicap insuring the senior-most tranches of anything. Earlier in 2010, Viral Acharya, Professor of Finance at NYU's Stern School of Business, proposed GSE reform in a guest article in The Economist (available here) that suggests monolines insure portions of the juniormost 20% of securitizations of Qualifying Residential Mortgages. The potential for such a regime has survived Treasury's recent proposals on GSE reform issued this week.

Even on the municipal side, the new entrant BondFactor has proposed writing business in a way essentially the same as guarantying a non-senior-most tranche of a CDO. 

To the credit of these ideas, there is no reason why it couldn't be a viable business given adequate market demand. Mortgage insurance is a form a credit enhancement in which the junior-most portion of a loan is guaranteed. While the industry has been troubled, it survives despite being at the vortex of a so-called great recession.

Furthermore, increased investor demand to hedge guarantor counterparty risk has created a large, upward pressure on the cost of such protection. This in turn results in downward pressure on a guarantors pricing power. The less insurace a guarantor issues, the less is potential demand for such hedges.

For a little more insight, lets consider three hypothetical underwritings of the same theoretical securitization trust.

1) In a traditional underwriting, lets say the financial guarantor issues a policy covering the 80% of the securitization constituting only the senior-most tranche. The guarantor considers this consistent with an almost zero loss underwriting standard. Lets say the guarantor receives a healthy 25 bps of the par insured.

2) The financial guarantor insures the par constituted by the next senior-most 10% of par. In exchange for more likely losses, the guarantor receives a higher fee in absolute terms and as a percent of par insured. This higher fee pays for itself through interest cost savings on both the insured par and the uninsured senior-most tranche. The uninsured senior-most tranche interest savings arise from the increased protection of the monoline's due diligence, surveillance, and recovery skills; the interest savings would be less than actually guarantying that tranche.

The main concern with this strategy is the exposure to tail-loss scenarios. The ability of economic catastophe to turn some loss exposure into a devastating loss experience would be difficult to overcome. Acharya realizes this in his idea for government risk-sharing to replace GSE-only mortgage pool guarantees. (In his example, a monoline could guarantee 10% of exposure and the GSE would be responsible for the other 90% with both receiving the same percentage fee determined by monoline pricing.)

3) As in most threesome's, the best place may be in the middle. In that vein, the senior-most 70%  could be uninsured and the next 10% insured. The fee could be calculated as 20 bps of the senior-most 80% or 160 bps of the par insured. The insurer would thereby maintain the same underwriting standards in terms of default expectations. In terms of loss given default, it would be higher as a percent of the insured but smaller as a percent of the senior-most 80%.

If done correctly, any three of these business models could work. But we feel that some (or more or lots of) loss model (#2) would have difficulty managing economic catastophe risk. That risk is enough of a problem for the traditional model. And while the conservative stop-loss model (#3) seems to offer a nice solution, market acceptance of viable pricing remains to be seen.

Given increased investor demand to hedge monoline credit exposure, it may make sense for guarantors to invest in testing the waters of alternative underwriting models. They should not however, feel inclined to write new forms of insurance just because of the pressure building to change their ways and a desire to show the product is still relevant.

Unlike La Roux who is doing it for a thrill, the guarantors should go in for the kill slowly, and conservatively. And as always, they should consider risk in terms of catastrophe.