In defending itself against A.S.B.’s accusations, PaineWebber made an interesting claim: It said that Fitch had been intimately involved in the structuring of the deal and that it had relied on Fitch’s representations for assumptions about the performance of the underlying assets. The U.S. Court of Appeals for the Second Circuit agreed, writing that A.S.B. had discovered that “PaineWebber and Fitch had extensive communications about the structure of the transactions [that] concerned what PaineWebber needed to do to earn an investment-grade rating from Fitch.” The ruling also said the claim that “Fitch plays an active role in structuring the transaction is extremely credible.”

The case is notable in part because ratings agencies are rarely sued or even ensnared in other parties’ lawsuits. In the A.S.B. case, Fitch refused to turn over documents, claiming protection under the New York State shield laws that allow journalists to guard their sources and methods—a claim the court didn’t buy. Credit-ratings agencies still maintain that their ratings are simply published opinions, which investors are free to heed or ignore.

But as a result of the subprime-mortgage mess, pressure is building to rein in the agencies. Mason and Rosner, for instance, are convinced that the agencies are hopelessly conflicted. They argue that there are “fundamental flaws” in the rating process for mortgage-backed securities, suggesting that the entire world of structured finance could be suspect.

Mason estimates that direct losses from mortgage securities and other complex structures called collateralized debt obligations are already between $70 billion and $100 billion. And the damage could spread to other markets, such as the high-flying private equity world, which depends on the agencies to stamp dependable ratings on the bonds of companies that private equity firms want to acquire. “The reason this works is because the ratings agencies have said it works,” said Bill Ackman, a hedge fund manager who has about $6 billion under management, in a speech at a charity-investment conference in May. “The big point here is that everyone in the chain gets paid up front. The rating agencies get their fee . . . if they say the deal works. If they say the deal doesn’t work, well, you just go across the street” to another agency to get the rating you want.

The 2006 vintage of subprime mortgages was troubled from the start, coming as it did when real estate prices began their ­descent. Consumers were offered loans that, at times, exceeded the entire value of the homes they were about to buy. Some borrowers didn’t have to verify their income before ­receiving mortgages. These are denigrated as “liar loans” in the industry, and not surprisingly, they are going bad at a rapid pace.

While the agencies say they have tightened up their standards in recent years, the data suggest otherwise. The ABX index, which tracks the ­subprime business, shows that, beginning in the last half of 2005—long before the scope of the crisis became widely known—subprime securities were already starting to get shaky. The amount of protection for the riskiest investment-grade tranches was going down. Yet the agencies continued to assign high ratings to a big percentage of subprime deals, collecting fees along the way.

The recent crisis has led the agencies to make a series of embarrassing tweaks. In April, Moody’s said it would start doing what it should have done long ago: more aggressively scrutinizing new mortgage loans. The company acknowledged that its models, created in 2002, were out-of-date. “Since then, the mortgage market has evolved considerably, with the introduction of many new products and an expansion of risks associated with them,” a Moody’s report said. In hindsight, it seems astounding that the most influential rater of mortgage bonds wouldn’t be upgrading its models regularly to account for the growth in exotic mortgages.

The changes may be too little, too late. Last year, President Bush signed a law to have the S.E.C. monitor and regulate credit-ratings agencies, taking what has been a free-market free-for-all and putting it under the microscope. The S.E.C. formalized its rules this summer.

Other ideas for reform are flowing in. Rosner suggests that ratings for structured securities use a different scale—say, numbers instead of letters—to differentiate them from ratings for corporate and municipal bonds. He believes the agencies need to step up the training for analysts and should be compelled to re-rate transactions regularly rather than monitor them haphazardly. Furthermore, he thinks efforts should be made to distance the agencies from Wall Street. He proposes that any ­ratings-agency employee involved with a structured-finance deal for a Wall Street firm should have to wait a year before being able to join that firm. Such a waiting period already exists for auditors.

Murphy, the ex-Moody’s executive, doesn’t blame the ratings agencies alone. “But in the end,” she says, “it’s supposed to be the ratings agencies that are the purest of them all. They should be held to the highest standard. Maybe we should fundamentally rethink their position in the markets.”