ADVERTISEMENTS

Some industries securitize primarily to gain off-balance sheet and immediate revenue-booking accounting benefits and because it is a cheaper funding source than other methods. Industries like these often have significant skin in the game (e.g., the credit card industry, where a 7% vertical slice is the typical minimum requirement and it's usually much higher). Other industries, like non-GSE mortgages securitize primarily to shift credit risk. The whole point of securitization is not to have skin in the game.

The skin in the game requirement is being driven by the experience in mortgage securitization, not other types of securitization, and imposing a skin in the game requirement probably won't do much to non-mortgage securitization, where there might already be more than 5% retained interest. But for housing finance, skin in the game is really counter productive.

Most mortgages are originated either by thinly capitalized mortgage banks and brokers or by depositary institutions. Mortgage banks and brokers simply cannot afford to have skin in the game--their business model is not based on holding credit and interest rate risk beyond what product they have warehoused before securitizing it. And depositary institutions are ill-suited for holding the interest rate risk on long-term fixed-rate mortgages. This is the lesson we ought to have learned from the S&L crisis: if interest rates go up, depositary institutions that hold long-term fixed-rate obligations take a beating, as their cost of funds can exceed the yield on the mortgages.

If we want to do long-term fixed-rate mortgages (and we should, as consumers are not well equipped to handle interest rate risk), we need to match long-term funders (insurance companies, pension plans, etc.) with long-term borrowers. The long-term funders don't do mortgage origination (insurance companies used to, but they got out of the market for reasons I don't understand), so we need the secondary market (securitization) to match them with the long-term borrowers. Making depositary institutions sit on a chunk of the loan is just counterproductive--it places interest rate risk on institutions that we know can't handle it--rather than placing it with institutions that are happy to take the risk in exchange for a locked-in, long-term fixed-rate of return.

The problems that arose with private-label securitization were primarily from information asymmetries between MBS purchasers and originators. The ratings agencies were supposed to bridge that divide, but the nature of private-label MBS was that they were heterogenous products, and the whole ratings system is based on past performance, which requires homogenous products with known track records. Not possible for new niche products. (fwiw, the niche products were only able to gain market share because of information asymmetries between originators and consumers. If the full cost of the niche products was clear up-front, they could not compete against the 30-year fixed, which benefitted from the implicit government guarantee of the GSEs; the only way private-label products could compete was through cost obfuscation via complexity.) Private-label MBS can work, but product standardization both of the mortgages and the MBS are essential for correcting the information asymmetries that originators exploiting during the bubble. Skin in the game is just an awkward and counterproductive way of dealing with the information asymmetries that could be more easily corrected through policies that favor the 30-year fixed as the standard product for securitization.

Where skin in the game actually make sense is not with originators, but with servicers. This crisis has taught us about the importance of servicers in loan performance, and servicer compensation needs to be revised so that servicers have incentives that match all investors' (something like a vertical slice). Watch for more on this later this summer.

Bloggers, promote your law blog by nominating your blog for inclusion in USLaw.com's Law Blog Directory and RSS Reader. Benefits described.