Revisiting the Securities that Caused the Meltdown

Now that the country is headed for, in the best case, a pretty terrible recession it is perhaps a good time to look back and reflect on the housing bubble that really perpetrated this current mess.

The origins are easy enough to understand: cheap credit and a society whose government which was pushing home ownership, even on to those who could not afford it. What is not as well understood is how these toxic assets, mortgages to less than credit worthy borrowers, spread their way through the financial system and past safeguards the government had put in place in order to protect the solvency of the banking and insurance system.

The main culprit in this story is the vehicle known as a Collateralized Debt Obligation, or CDO. Every loan, in addition to having a certain yield, also has a certain idiosyncratic probability of default. This, in turn, translates into a credit rating assigned by a rating agency, which are then relied upon by the constituents which make up the financial system in order to properly guage the risk of the assets on their balance sheets. A CDO is a pooling of these loans, designed to take advantage of the fact (assumption is perhaps more appropriate) that the defaults of these instruments are not correlated, and therefore a portion of these assets become more valuable than if they were owned individually.

Take a simple example: if there are two $1 bonds in a pool and each has a 50% probability of default, the expected value of the pool is $1. If the defaults are uncorrelated, there is actually only a 25% chance that (50% times 50%) that pool will be worth $0. Said another way, there is a 75% chance the pool will be worth at least $1. Is the first $1 of the pool not worth more than the second dollar of the pool? Is the first $1 of the pool not worth more than each of the underlying assets? That is the ultimate logic behind the CDO. Thus, the first $1 of the CDO would be sold to an investor as a highly rated instrument (a senior tranche), and the second $1 of the CDO would be sold as something rated closer to junk status (a junior tranche).

There are ways here to rinse and repeat. One can imagine a scenario where there have been several hundred or even several thousand CDOs created in this manner, all with more assets and more tranches than the simple example above. Is it not possible to take junior tranches from a set of these uncorrelated CDOs, pool them together, and then engage in the same process as above? In fact, this is what the investment banks did when they created Collateralized Mortgage Obligations, or CMOs. Obviously, the supply did not exist in a vacuum. In an easy credit environment, yields were at all-time lows, and many institutions were clamoring to find new assets which were highly rated (to satisfy internal or external risk management) and promised better yields than those that were traditionally available.

So what went wrong? Let's revisit our assumptions. It turns out that the mortgages that made up the underlying asset pool were correlated. Most of the speculative home buying (which provided the supply of assets) were actually taking place in three major geographies: California, the Southwest, and Florida. Additionally, in the face of a severe economic event (see financial crisis, or asset bubbles bursting) most assets are correlated. One observes this phenomenon where asset prices everywhere more or less mirror the US equity market in severe negative events. Furthermore, many of the mortgages themselves may not have been priced properly due to lax origination standards and outright fraud. If one takes the simple CDO above, perfect correlation means that both tranches are actually worth exactly the same. In some sense, the "AAA" rated structured credit instruments are nowhere near investment grade status, while the lowest tranches of those same instruments may have actually been priced fairly attractively.

Thus, by the time everything was said and done, the market for purchasing these AAA rated securities had dried up, and the underwriting investment banks ended up holding these as assets on their own balance sheets. These securities were marked to market one fated day last fall, and the rest, as they say, is history.

For a more in-depth read, refer to this working paper by Joshua Coval, Jakub Jurek, and Erik Stafford. Josh and Eric are on the faculty of the Harvard Business School, and Jakub is on the faculty of Princeton University.