Confidence Game - Christine Richard [31]
Fitch’s model for assessing the bond insurers’ claims-paying resources tests the companies’ portfolios under 500,000 different scenarios. The model uses so much computing power and takes so long to run that the analysts input all of the required data and push the start button as they are headed out the door for the evening. By the time they arrive back at work the next morning, they have their result.
If the portfolio was successfully constructed to a triple-A level of safety, then the Monte Carlo simulation would find that the insurer had enough capital to cover claims 99.99 percent of the time. In other words, a triple-A-rated bond insurer would fail the test only once in every 10,000 scenarios. It was a very reassuring-sounding outcome.
But was it? Models often rely on historical data to predict what is possible. Unfortunately, the past is not always a reliable guide to the future.
By 2003, CDOs were the hottest securities in the structured finance market, although not everyone was sure why. Frank Raiter, a former director at Standard & Poor’s, testified at a congressional hearing in 2008 that he and many of his colleagues were baffled by the success of CDOs. “To a lot of us analysts that were outside the CDO area but were looking at it through the glass, intuitively it didn’t make a whole lot of sense,” Raiter said.
The business apparently made sense to the bond insurers, though. MBIA defended the business in a report it posted on its Web site shortly after Gotham released its report. The securities were a positive development because they allowed financial institutions and investors to transfer risk, the report said. For that reason, “CDOs are extremely beneficial to the global capital markets,” MBIA asserted. “I feel strongly that if you don’t like CDOs, you don’t like MBIA,” Jay Brown told attendees of the company’s investor meeting that spring.
Most of the equity analysts covering MBIA liked the CDO business, too. Though sometimes they seemed hard-pressed to explain exactly why. “Potential losses are carefully monitored, new entrants are entering the business, and the rating agencies are confident that the business is sound,” wrote Joshua Shanker, an analyst with Blaylock & Partners, in a February 2003 report. MBIA’s expansion into backing CDOs “carries the aura of a good business,” Shanker concluded.
When MBIA talked about CDOs, the statements were reminiscent of how it talked about its own business. “MBIA has structured its synthetic CDO transactions to withstand a very large number of investment-grade corporate defaults—typically about 7 times the average historical default rates and more than 2.5 times the historical maximum default rates—before losing a single dollar,” MBIA said in its December report on CDOs. In the company’s opinion, this level of defaults would be “an extremely unlikely catastrophic event, approaching a global meltdown of corporate credit.”
That’s why the Gotham report’s estimate of MBIA’s CDO losses was so shocking. The securities were supposed to be extraordinarily stable—just like the bond insurers themselves. When the New York attorney general’s office launched an investigation of Gotham, it wanted to know if Bill Ackman’s estimates were reasonable. MBIA had insisted that any estimate by an outsider who lacked knowledge of specific CDOs backed by MBIA was meaningless.
The attorneys investigating Gotham called in Boaz Weinstein, one of the credit-default-swap (CDS) traders Ackman spoke with to make his estimates. Weinstein, a University of Michigan philosophy major and Life Master chess player, who headed up credit trading for North America at Deutsche Bank in 2002, told investigators that Ackman’s approach was reasonable: “I said that at best it would be a rough approximation, but that certainly you could get it in the ballpark.”
Weinstein argued that CDOs that originated around the same time were similar. “These deals by nature