In a well-documented recent paper (*), Joseph R. Mason of Drexel University and Joshua Rosner of Graham Fisher & Co. ask if a chain of events has not begun unfolding that will end up in a situation where funding for the mortgage industry is massively curtailed. In their paper, Mason and Rosner draw a parallel between the role played presently by subprime loans within Collateralized Debt Obligations (CDOs) and a situation that arose in 2003 leading to the demise of the main collateral types used in those days to enhance CDO credit. These sectors were: manufactured housing, aircraft leases, franchise business loans and 12-b1 mutual fund fees. Should Mason and Rosner’s parallel be justified, then securitization of subprime loans will similarly be in for a major setback in the years to come, the effects of which would ripple through the whole mortgage capital markets and dry up their current main source of funding.
Collateralized Debt Obligations which, as the authors show, were issued in 2006 at a volume nearly equal to that of the three prior years, provide investors with access to high-yield debt, bypassing through the packaging inherent to securitization their otherwise lack of liquidity. CDOs often simply re-securitize low-grade certificates belonging to Asset-Backed Securities (ABSs). CDOs allow putting to their best possible use as credit enhancers these low-grade certificates. A Moody’s analysis in October 2006 reported collateral for outstanding CDOs comprising 28% of subprime and home equity loans and 12% of prime loans. Unlike what is the case with MBSs or ABSs, CDOs are managed, with the Collateral Manager free to populate the security dynamically during its lifetime.
The reason why subprime loans played recently in CDOs the role held in earlier days by manufactured housing et al. is easy to grasp: what makes their attraction is that until the end of 2006 they were undervalued as their high yield – supposedly reflecting their embedded default risk premium – was more conservative than the risk actually incurred, providing investors with a possible arbitrage.
The reason for this opportunity for arbitrage is obvious: in the absence of any safe method for assessing the market risk component of future losses (as opposed to the credit risk component that can be estimated with the help of an individual “credit score”), risk premiums are calculated to mitigate business cycles in their entirety. The implication is that in low default times, the risk premium, i.e. the high yield, will be overvalued, constituting for the investor an attractive proposition. Conversely, in the depressed part of the business cycle, the risk premium is undervalued and fails to cover the actual financial risk appropriately, or in any case with a too thin profit spread. Collateral types like subprime mortgages, manufactured housing et al. will out of necessity only be of interest to CDO Collateral Managers during their periods of risk overvaluation and may meet their doom or at least their periodic eclipse whenever they enter the depressed part of their business cycle as demand then essentially vanishes.
In Milken’s glory days, junk bonds’ embedded risk premiums were overvalued compared to their actual default risk, making the market an attractive one. When their high yield ceased to reflect adequately their rising default rate, the market evaporated, swallowing with it its protagonists. Manufactured housing, aircraft leases, franchise business loans and 12-b1 mutual fund fees met the same fate in 2003 as far as CDOs are concerned: “We argue that the shrinkage in those sectors arose from decreased funding by the CDO markets” (p. 33), contend Mason and Rosner. They add that the exact same phenomenon is currently taking place with subprime certificates carved out from ABSs: “We therefore maintain that the shrinkage in Residential Mortgage Backed-Securities’ sector is likely to arise from decreased funding by the CDO markets as defaults accumulate. (This) could set off a downward spiral in credit availability that can deprive individuals of home ownership and substantially hurt the U.S. economy” (p. 33).
Their argument is too complex to be faithfully reflected in a short note like this but it is in my view flawless. The only reproach I would make is that they mix two considerations about subprime, introducing an unnecessary distraction in their demonstration. The first consideration about subprime is legitimately relevant to their argument: a borrower’s capacity at making monthly payments which is now lacking with the defaults recently observed in the subprime sector. The second consideration, crucial for the mortgage industry as a whole, but extraneous when discussing CDOs, is the borrower being “underwater,” i.e. owing debt on the house larger than the potential net proceeds of selling the house. The first consideration impacts any type of mortgage-backed security as it affects directly the capacity of recovering a loan’s principal. The second consideration is but a looming threat as long as the homeowner stays put; its effects are complex and spread out necessarily over the long term.
By contrast, the discussion by the authors of the role played by the rating agencies is at the core of their demonstration. Will these be able to do any better at rating in a time of crisis than they did in November 2001 with Enron when the public became wary at their methods? The authors’ reminder that rating agencies claim they need anywhere between three and seven weeks for “notching” i.e. for reflecting in their own rating a change of grade by another agency, casts here a giant question mark. That they promptly put their act together is no luxury but a must: as Mason and Rosner state it, here lies reputational risk for the U.S. capital markets.
(*) Joseph R. Mason and Joshua Rosner, “How Resilient Are Mortgage Backed Securities to Collateralized Debt Obligation market Disruptions?” presented at the Hudson Institute on February 15th, 2007