Page 4 of 5   <       >

THE CRASH -- What Went Wrong

The firms all say they did not bend their rules. "We have rigorous policies and procedures in place to maintain our analytic independence and shield our analysts from encroachment of commercial interests," S&P said in a written statement. The company also said its criteria are "consistently applied."

In the trenches, the overworked former analyst said he never worked with investment bankers to structure any of these complex deals. Instead, as S&P's policy dictated, he engaged in back-and-forth discussions. The most contentious conversations, he said, involved bankers trying to copy some new twist in a competitor's CDO. Because CDO documents were private, the analyst could not tell the copycats how a competitor had won a top rating. He watched as bankers tried to reverse-engineer the new structure from scant public information, often leaving out important risk-mitigating elements.

The CDOs piled up at a staggering rate, and so did the complexity, along with the pressure to get them rated. "There was such a compressed time frame," the analyst said. "As soon as some new development emerged, everyone took advantage of it. There was such an explosion, so many deals with the same features, that when something went wrong, it wasn't just limited to a few hundred million, but to several billions."

The S&P statement said analysts were not compensated on the number of deals rated or on the percentage of high ratings granted. Although the firm acknowledges that its assumptions about defaults "did not serve us well on certain structured finance ratings," it also maintains that those assumptions "seemed reasonable given the projected market environment at the time they were made."

Gugliada, who praises structured finance for lowering borrowing costs and making credit more widely available, says the ratings firms misjudged the impact that a relatively few mortgage foreclosures would have on all the other mortgages in a pool. "In hindsight, we now know all mortgage-backed securities perform fine or they all default at the same time," he said.

The S&P statement called that view "overly simplistic and inaccurate," saying securities have different levels of credit enhancement, different mortgage types and different borrowers that can cause each to perform differently.

Gugliada used to believe that. But he now says that CDOs made up of BBB-rated pools of mortgage-backed securities should never have been rated any higher than BBB themselves. If the ratings agencies had understood that, "there wouldn't have been any CDOs" of mortgage-backed securities, he said.

But Wall Street didn't just create CDOs based on mortgages. As the frenzy went on, it cooked up CDOs based on CDOs.

Deluge: Morphing the CDO

The Mantoloking CDO, offered to investors in late 2006, owed its existence to an innovation that became typical of late-stage CDOs.

In Wall Street parlance, it was a "CDO squared" -- a CDO concocted from the unwanted parts of 126 other CDOs. In plain English, Mantoloking was a dumping ground.

BBB and BB bonds that had drawn few buyers in earlier incarnations were repackaged into a new set of tranches. At the top, $580 million of the bonds took on AAA designations. The remaining $185 million went into the lower, riskier tranches.

The CDO-squared carried an extra level of danger, however: If something went wrong, investors in Mantoloking's AAA tier had to stand behind the investors in the top layers of the original CDOs. The securities in the bottom tranches were a distant last in the lengthening queue.

<             4        >

© 2008 The Washington Post Company