Confidence Game - Christine Richard [30]
The highest rating category is triple-A, an indication that Moody’s sees “minimal risk” of an issuer defaulting on its debts. The U.S. government is rated triple-A based on the revenue-generating capability of the world’s largest economy and the Treasury’s ability to increase taxes and even print money if it faces a shortfall. The lowest rating is C, which indicates “the bonds are in default, with little prospect for recovery of principal or interest.” Typically, the lower a company’s credit rating, the more it has to pay investors to borrow money.
Though criticized for missing risks at Enron and WorldCom, credit-rating companies make it possible for investors to sift through the billions of dollars of debt sold every week around the world. At a glance, investors can compare whether it’s riskier to lend to the government of Indonesia or to Mattress Discounters.
By 2003, the credit-rating business was less about confidence in men or in the companies they ran than it was about models that attempted to predict the probability of loan defaults. Moody’s and its only two serious competitors, Standard & Poor’s and Fitch Ratings, were earning an ever-larger share of their profits from rating asset-backed securities (ABS). To create these securities, investment banks bundle hundreds and even thousands of mortgages, credit-card receivables, or other types of loans into special-purpose vehicles (SPVs), which sell bonds to finance the purchase of these assets. The key to assessing the credit of these securities is in modeling how the underlying loans will perform under various economic scenarios. The Holy Grail for Wall Street is finding ways to manufacture triple-A-rated securities out of higher-risk assets.
In many ways, the bond insurers were the template for the entire structured finance market. Bond insurers, like SPVs that sell asset-backed bonds, are essentially diverse pools of credit risk. The companies, like the lowest-risk piece of a securitization, were structured in such a way that they would receive triple-A ratings from Moody’s and Standard & Poor’s.
To get that top rating, bond insurers had to show the credit-rating companies that they could pay all claims on defaulted bonds in 99.99 percent of the scenarios generated by rating-company models. Gary Dunton, MBIA’s president, described it this way in the company’s 2000 annual report: “Our no-loss underwriting standard—no losses under the worst probable scenario—is the most important discipline that we have as a triple-A-rated credit enhancement company.”
Read that again. “No losses in the worst probable scenario.” At first glance, the statement seems to inspire confidence. It’s also easy to misread it as: “No losses in the worst possible scenario.” But there’s a big difference. In the worst possible scenario, all the bonds MBIA guaranteed would default. Defining the worst probable outcome is a matter for statisticians.
The bond insurers and the credit-rating companies used so-called Monte Carlo simulations to determine how likely it was that a bond insurer would be unable to meet its claims. This modeling technique was named after the Mediterranean gambling city in Monaco because it relies on randomly generated numbers to create the different scenarios.
Monte Carlo simulations are like computerized crystal balls that generate thousands of versions of what the future might look like. In the case of the bond insurers, the model analyzed what would happen to the insurer’s claims-paying resources under various economic scenarios. What if unemployment was 1 percent and home prices rose by 5 percent? Or what if unemployment was