The securitization of mortgages and other debt obligations gives senior tranche holders less risk of individual defaults, but increases the risk to a general economic downturn. Coval, Jurek, and Stafford demonstrate this in The Economics of Structured Finance a working paper published last week by the Harvard Business School. The paper contains exceptionally lucid descriptions of how structured finance works and uses simple examples to demonstrate the sources and magnitudes of under-appreciated systemic risks. It's 36 pages including charts and graphs and well worth the 1-2 hours it might take to understand it. Here's my summary of the essence. (Except as otherwise noted, all citations are to the PDF of the working paper.)
The main reason people buy investment grade bonds is that they want a very high degree of security that the promised income stream will not be interrupted even if there is a severe economic downturn causing a deep slump in the stock market. In other words, they want investments that are as weakly "correlated" as possible with general economic conditions because the values of other classes of investments, such as corporate stocks and commodities, are highly correlated with general economic conditions.
Rating agencies (Moody's, S&P, Fitch) developed methods to compare the creditworthiness of bond issuers against one another. For example, they tell us that this AA rated General Electric bond is less likely to default than that A+ rated utility bond and both are less likely to default than a certain A- rated municipal bond. "However, credit ratings, by design, only provide an assessment of the risks of that security's expected payoff, with no information regarding whether the security is particularly likely to default at the same time that there is a large decline in the stock market or that the economy is in a recession." At 18-19 (emphasis added).
When the rating agencies began to rate collateralized debt obligations ("CDOs"), they continued to ignore correlation of risk to general economic conditions. Regrettably, while pooling of mortgages and other debt obligations does reduce the risk to senior tranches of individual defaults, it increases the correlation to general economic conditions. As a result, a AAA rated CDO is more subject to systemic risk than a single AAA rated corporate or municipal bond. "[T]ranches written against highly diversified collateral pools have payoffs essentially identical to a derivative security written against a broad economic index." At 20. Consequently, much to the consternation of most of them, holders of CDOs had bet on the overall health of the economy, not on mortgage default rates in Tallahassee.
If this had been widely understood—it was not—buyers of AAA rated CDO tranches would have demanded higher yields than they did, and the business of "structured finance" would likely have been much smaller. However, the Basel I and II international banking agreements created additional impetus by providing that banks need to hold only half the reserves against AAA rated securities that they must hold against lower rated securities. At 26.
It gets worse. There developed a practice of securitizing the mezzanine tranches of CDOs into CDO-squareds, thereby creating additional AAA rated bonds. At 17. Coval et al. show mathematically that any tiny underestimate of default rates or correlation of performance among the underlying obligations is hugely magnified in CDO-squareds. At 15. According to Moody's, CDO-squareds accounted for 55% of the notional value of all CDOs issued in 2006! At 17. It seems shocking that a AAA tranche from a CDO-squared is much more likely to default in a recession than a AAA tranche from a CDO, but that seems irrefutable.
There is some evidence that Wall Street executives realized it would end one day, but in the meantime, they had little incentive to move to the sidelines. In July 2007, the then-CEO of Citigroup, Chuck Prince, acknowledged that the cheap credit-fueled buy-out boom would eventually end, but that in the meantime, his firm would continue to participate in structured finance activities (as reported in Nakamoto and Wighton, 2007): "When the music stops, in terms of liquidity, things will get complicated. As long as the music is playing, you've got to get up and dance. We're still dancing."
At 26. See also Risk management means taking the risks your competitors take.
A nice overview of how the CDO industry developed is here.