(FORTUNE Magazine) – HEDGE FUND MANAGERS are a secretive breed, especially when they're shorting a stock. That's partly because many people still seem to feel that there's something almost unpatriotic about making money when a stock declines in value. So it raised some eyebrows, back in 2002, when a guy named Bill Ackman took on a company called MBIA in a very public way. First he shorted the company's securities, and then he issued a press release with a link to the website for the fund he ran, Gotham Partners, where he posted a 66-page analysis questioning almost every aspect of MBIA's business.

MBIA is not a household name, and by some measures it is small. It employs just 623 people and its market value is less than $8 billion. Yet MBIA is a subterranean powerhouse. It guarantees interest and principal payments totaling $890 billion on bonds that fund everything from the Eurotunnel to the city of Minneapolis to trusts backed by royalties from DreamWorks movies. As MBIA's chairman, Jay Brown, says, MBIA "touches a lot of things." Certainly, it touches a lot of investors. Major banks, pension funds, and moms and pops all own bonds that are seen as some of the safest investments in the world simply because they are insured by MBIA--carrying what it calls its "Good Housekeeping seal of approval."

The key to MBIA is that the credit-rating agencies--such as Standard & Poor's--give its insurance arm the rare and coveted rating of triple A. That means that MBIA is judged to be one of a handful of the most financially sound companies in America, up there with Exxon Mobil and GE. When MBIA insures, or in industry lingo "wraps," a bond, it gives that bond the attributes--mainly a lower interest rate--of a triple-A-rated bond. In other words, what MBIA sells is its triple-A rating. As CEO Gary Dunton wrote in his 2004 letter to shareholders, "MBIA's true constant--our North Star, if you will--is our commitment to protecting our triple-A ratings." And it was precisely MBIA's triple-A rating that Ackman was publicly calling into question.

The initial response to his report was a swift and brutal backfire. At MBIA's request, Eliot Spitzer's office launched an investigation of Gotham over possible market manipulation. That alone led many on Wall Street to dismiss Ackman's allegations--when a short-seller is questioned by the state attorney general's staff, whatever credibility he has tends to vanish.

But today the tables have turned. In March, MBIA announced that it would restate its earnings for the past six years to correct the accounting for a transaction that Ackman had identified as a "loss- deferral, earnings-smoothing device" in his initial report. Then, later that month, MBIA announced that it had received subpoenas covering a range of issues from both the SEC and--yes--Spitzer's office.

Ackman, a persistent 39-year-old who has accumulated 40 boxes of MBIA-related documents in his office, is still convinced that MBIA engages in risky practices that endanger not only its own investors but also the global capital markets--if the company were to lose its triple-A rating, what would happen to the price of the billions in bonds it insures? MBIA, of course, has a different story. To hear its executives tell the tale, it is a good company that has been unfairly targeted by short-sellers. They also say that MBIA is being swept along by the wave of regulatory attention that has struck the insurance industry lately. And Brown observes that the subpoenas "could be nothing. A lot of times you send a lot of documents off and you never hear back from anyone."

The questions surrounding MBIA are critical ones not just for the millions of investors who own MBIA-guaranteed bonds but also for everyone who wonders whether the watchdogs of our financial system--the analysts, the accountants, and most critically in this case, the credit-rating agencies--can be trusted. Most of those people are on MBIA's side. The key is that few believe that MBIA's rating is at risk, because the agencies are--so far--affirming their triple A's. To believe Ackman is right, you have to believe all those people are wrong. It's almost impossible. Almost.

Yet Ackman has already been proven right on one count. And because the standard is supposed to be so high for MBIA, Ackman doesn't have to be right about everything to be right that the company doesn't deserve its triple A. On that front, some agree. "A triple A implies a bulletproof credit, that come hell or high water a company can easily satisfy its obligations on a timely basis," says analyst Sean Egan, a managing director of the independent credit-rating agency Egan-Jones. He, like Ackman, does not think MBIA meets that standard. Others have had their faith shaken by MBIA's earnings restatement. "We believe that a breach of trust has occurred to such a degree that it may be difficult for shareholders to rely on management's statements going forward," proxy advisory firm Glass Lewis wrote in a recent report. It recommended that investors withhold their votes for Brown and Dunton as board members. For a company whose business model is built on trust, votes of no confidence are an ominous development.

AT FIRST GLANCE, MBIA's business is both elegant and simple. The core idea is that a municipality--say, Kansas City--will be willing to pay a small premium to have its bonds insured, because the triple-A rating will lower its interest cost. As long as the premium MBIA charges is less than what the client is saving in interest costs, it can market its product as a way for issuers to save money. MBIA is often paid its premium up front, which it then recognizes as revenue over the life of the insurance. That means that a substantial portion of future years' earnings are already in the bank. Analysts often cite MBIA's smooth, predictable earnings as a reason to own the stock.

Bond insurance has proved to be wildly popular--and profitable. MBIA was founded by a group of insurers in 1974, and today about half the municipal bond market is insured; MBIA is the market leader with about 18%. In addition, over the past 15 years, MBIA has begun to insure what are known as structured finance deals--such as trusts backed by manufactured-housing loans and airplanes. Last year MBIA's insurance operations reported just under $2 billion in revenues and $804 million in net income, meaning that its net margin was a stunning 40%. (Like all insurance companies, MBIA is structured as a publicly traded holding company with an insurance company subsidiary. MBIA Insurance Corp., which has that all-important triple-A rating, provides almost all of MBIA Inc.'s earnings.)

Look beneath the surface of this business, though, and you find the potential for important conflicts of interest. The credit-rating agencies--Standard & Poors, Moody's, and Fitch--are often referred to as the industry's "de facto regulators" because they control that triple-A rating. But agencies are paid for the ratings they issue. That means they are paid by the bond insurers and also by the issuers of debt--which might not bother to get a rating at all if it weren't a requirement for obtaining insurance. These two streams of money are sizable. A former S&P executive says that when he was involved in that agency's municipal finance division, fees from bond insurance accounted for roughly 50% of revenues. Furthermore, both MBIA's former CFO, Neil Budnick, and its current CFO, Nick Ferreri, are agency veterans. Debra Perry, who became an MBIA board member last spring, was a senior Moody's executive.

The rating agencies point out that they are always paid by issuers, and they say they can manage any conflicts inherent in that arrangement. S&P also says most issuers would pay to get a rating whether or not the insurer required it.

To outsiders, there is another aspect of the business that is truly through the looking glass. Unlike other insurers, which expect to suffer some level of losses, MBIA prides itself on "no-loss underwriting." In other words, the company never intends to actually have to pay a claim. MBIA says that, with rare exceptions, it insures only debt that is deemed relatively safe, or "investment grade," by the rating agencies, though some issues get downgraded after the initial guarantee. Today the company says that 98% of the debt it guarantees is investment grade. MBIA has had occasional losses, but if they were to become a regular, sizable occurrence, that could threaten its triple-A rating and blow a hole in its business model.

At MBIA's conference for its investors in March, Gary Dunton presented what he calls "my favorite slide." It shows that over MBIA's three decades in business, the company has guaranteed more than $1.8 trillion of debt service and suffered only $586 million in losses, or about 0.03%. To many investors, this is all the proof they need that MBIA can come very close to its promise of zero losses. But it also raises a question. Given how important that track record is, what measures has the company taken to preserve it?

BILL ACKMAN--who declined to be photographed, but who is otherwise anything buy shy--first became interested in MBIA back in the late spring of 2002. He was talking to a dealer of credit-default swaps, which enable buyers and sellers to place a bet on the likelihood that a company will default on its bonds. He asked if the guy could think of any companies that were rated triple A but shouldn't be. The dealer told him that he had never understood the bond guarantors like MBIA. How, the dealer asked, could a company that was leveraged 139 to one--at the time, MBIA had $764 billion of outstanding guarantees and just $5.5 billion of equity--be a triple A company?

Ackman flipped open MBIA's 2001 annual report and read the letter that then-CEO Brown had written to shareholders. The first thing that caught his attention was a comment Brown made about how MBIA had "expanded the disclosures" on $8.6 billion of debt in off-balance-sheet special-purpose vehicles--the kinds of financial instruments that had become notorious in the Enron debacle. Ackman couldn't find any previous disclosure of MBIA's SPVs.

SPVs are complex creatures, but here's a crude model of how one might work: The SPV raises money by issuing debt. It takes those funds and lends them out to, say, a credit card company, which in turn gives the SPV its receivables. If cardholders continue to pay their bills, the bondholders will get their money. In MBIA's case, the insurance company also issues a guarantee, promising that if, say, the cardholders stop paying, it will step in. Separately, it guarantees bondholders that they will be paid. In a commonsense way the second guarantee seems duplicative, but it contains the potential for risk: What if the holders of a particular bond can demand payment right away, but the credit-card holders--or MBIA in their place--are scheduled to pay their bills over time?

Because the owners of the debt are guaranteed payment, they don't have to care what assets the SPV owns. MBIA itself deems the contents "confidential." In the post-Enron climate, Ackman had two concerns: First, he didn't see how MBIA would be able to avoid putting the SPV debt on its balance sheet. And second, he wondered whether MBIA was using the SPVs to lend money to troubled entities that had issued bonds it had guaranteed--thereby helping them avoid defaulting and preserving its crucial "no-loss" promise.

By picking apart other companies' filings, Ackman was able to identify some $4 billion of assets that were or had been in the SPVs. Many were, indeed, lines of credit to troubled companies that also had other debt that was guaranteed by MBIA. An example was American Business Financial, which declared bankruptcy in early 2005. That company had deals like securitized home equity loans that were guaranteed by MBIA. To Ackman, that was a red flag. But he couldn't prove that MBIA had loaned American Business Financial--or any other company--funds to avoid a default. MBIA denies that it did. It also says that each deal is structured so that bondholders can't demand payment early and trigger a large payment for which MBIA is on the hook.

As Ackman dug further, he saw more signs that worried him, including a strange deal MBIA had done back in 1998 after a hospital it insured had gone bankrupt. On July 17, 2002, he placed a bet against MBIA using credit-default swaps that would pay off in the event of MBIA's bankruptcy. On July 24 he shorted the stock. Then he released his report. The first page noted his short position.

The rating agencies responded by reiterating their triple-A ratings. Other Wall Street analysts also voiced their support. But Alice Schroeder, a well-respected analyst then at Morgan Stanley, downgraded the stock to "underweight." Morgan Stanley soon dropped coverage of MBIA. Egan-Jones, which does not accept payments from companies it rates, and which has a good track record--it issued warnings about Enron and WorldCom long before the major rating agencies did--released a report that rated MBIA several notches below triple A.

MBIA, for its part, issued a press release just over half an hour after Ackman posted his report in which Brown called his analysis "patently false" and warned investors to "carefully consider the credibility of any negative report issued by a hedge fund that would directly benefit from a decline in MBIA's share price."

And after that ... nothing really happened for nearly two years. By late 2003, MBIA's stock, which hit a low of $35 earlier that year, had reached $60. Both Dunton and Brown had provisions in their employment contracts that gave them big paydays if the stock stayed above $60 for ten consecutive trading days. On Jan. 26, 2004, over $30 million of options and restricted stock for the two men vested. In December the New York State Insurance Department allowed MBIA's insurance arm to pay a $375 million dividend to the holding company. Those dollars belong to policyholders, so by allowing the money to leave, the NYSID is also showing its faith in MBIA. All this is a big part of why many observers don't even feel as though they need to read or understand Ackman's report to know that he was wrong.

THERE ARE a few more reasons people find it easy to dismiss Ackman's concerns. You'll often hear that he's a short-seller who has an economic interest in seeing MBIA fail. (You'll rarely hear that the company's supporters have an economic interest in seeing it succeed.) Then, Gotham Partners had its own problems with a large private investment it had made, and the fund began to liquidate right after Ackman issued the MBIA report. Negative stories about Gotham began appearing everywhere. Wasn't Gotham trying to bail itself out by issuing reports touting its positions in stocks like MBIA?

More bad press came when MBIA requested that Spitzer's office investigate the hedge fund. MBIA alleged that Gotham and a cabal of other funds were manipulating the market by spreading rumors about its business. Ackman says that in early 2003 he spent six days in depositions at the New York State attorney general's office, and more than $2 million in legal fees. But no charges were ever filed, and although the AG's office, as is its policy, won't comment on the Gotham investigation, it is a safe bet that it is closed.

After all that, most people would have thrown in the towel. Not Bill Ackman. At a charity dinner, he introduced himself to Sam Di Piazza, CEO of PricewaterhouseCoopers, MBIA's accountants, and later sent him the report. He cornered the CEO of S&P at the Harvard Club, and he hounded Moody's CEO by e-mail to read his report. He says he got no response.

Ackman has since started a new fund, Pershing Square. He says it now has more than $600 million under management and was up over 50% last year. He is again short MBIA's stock, and as before, he has bet against MBIA in the credit derivatives market.

AT MBIA'S headquarters in Armonk, N.Y., the two words you'll hear over and over again are "transparent" and "conservative." Brown describes the culture as "nerdy," and indeed, the offices are well-appointed but not at all luxurious. MBIA executives are zealous about their mission and venomous about their critics. Says Dunton: "Think about what we do. We build infrastructure around the world. We enable countries to build markets. We do good. I don't know why anyone would question the legitimacy of this business." He adds, "My mom says there are evil people out there."

MBIA's executives are particularly outraged about Ackman's allegations regarding the special-purpose vehicles. MBIA executives insist that it is categorically untrue that they paper over trouble, and say they are simply a way to provide an additional service--access to capital--to the company's clients. Dunton notes that the rating agencies review every deal that goes into the SPVs and maintains that "there is not a single non--investment grade credit" in any of these vehicles.

MBIA points out that even if a company whose assets it has guaranteed is troubled, that doesn't mean the assets themselves are in trouble. In other words, just because, say, a credit-card company goes bankrupt, that doesn't mean cardholders will stop paying their bills. MBIA says it hasn't taken any losses in its SPVs and doesn't expect to. But there's no way for an outsider to judge--and skeptics like Ackman distrust "black boxes" whose contents are known only to company executives and the rating agencies.

As Ackman expected, MBIA consolidated the SPVs on its balance sheet after the Enron scandal drew attention to such financial instruments. Today, when you look at MBIA's balance sheet, you see $7 billion in SPV assets and $6.4 billion in SPV debt (MBIA itself owns the remaining $600 million). But because MBIA guarantees payment on the assets, the rating agencies don't include this debt when they look at MBIA's finances. After all, as long as the assets are paying, the debt is covered. This is critical, because under the rating agencies' own guidelines, MBIA's holding company's debt cannot exceed roughly 15% to 20% of its capital or the insurer could lose its triple-A rating. If the SPV debt were counted, the ratio would be way over that limit.

Ackman and others say it simply defies economic common sense to say that the SPV debt--which after all is on MBIA's corporate balance sheet, and which the insurance company guarantees--doesn't belong to it. "We think it should be considered MBIA's debt," says Egan of Egan-Jones.

MBIA at times is oddly cagey about the fact that it guarantees the debt in its SPVs. On its most recent earnings conference call, a questioner asked if the SPV debt had "any recourse" to MBIA. In response, Dunton said: "That's nonrecourse debt.... It's got only the assets to look at for recovery." But because MBIA's insurance arm guarantees the debt, this is not true. MBIA says now that Dunton misspoke, and that when he said the debt was nonrecourse he was referring to the holding company.

IN AUTUMN 2004, very suddenly, MBIA's world changed. MBIA announced that it had received subpoenas from both Spitzer's office and the SEC over a hospital deal that had gone sour--the very one Ackman had noted in his report.

The hospital chain in question was called the Allegheny Health, Education, and Research Foundation, or AHERF. When it went bankrupt in the summer of 1998, MBIA, which insured AHERF's debt, was on the hook for over $300 million. This was the biggest loss any bond insurer had ever suffered. In the fall of 1998, MBIA took a highly unusual step: It entered into "retroactive reinsurance" agreements with three reinsurers. Converium gave MBIA $70 million, and AXA Re and Munich Re each provided $50 million. In exchange for that combined $170 million, MBIA agreed to give the reinsurers $296 million of premiums on some $45 billion of business it had guaranteed. This type of deal --reinsuring for a loss that has already taken place--is, believe it or not, permitted by accounting rules, as long as the reinsurers are taking on actual risk. Supposedly they were, and so MBIA executives wrote in the 1998 annual report to shareholders that "our loss reserves, our earnings, and our triple-A ratings were unaffected" by the AHERF bankruptcy, and that "we are fiercely proud of our near-perfect record of no losses."

Later, when questions were raised about this transaction, MBIA insisted that the deal had been fully vetted. But on March 8 the company made an abrupt about-face and acknowledged that there was "likely" an oral agreement by an unnamed MBIA executive or executives committing MBIA to ultimately assume all but a sliver of risk on the business that was supposedly transferred to Converium. In other words, even under accounting rules, that money shouldn't have been booked as earnings.

Although the company has restated its earnings for the past six years, reducing profits by $57 million because of the Converium slice of the transaction, it has not restated the other parts of the deal.

Much of what happened is still shrouded in secrecy. Ongoing probes--including a criminal investigation launched by the U.S. Attorney's Office for the Southern District of New York --may provide answers. In the wake of the AHERF transaction, then-CEO David Elliott stepped down, and Jay Brown, a board member at the time, took his place. Gary Dunton became president. Both say that they had nothing to do with the AHERF deal; Neil Budnick, who became CFO that fall, makes a point of saying that he "didn't sign the 1998 financials." MBIA says that if any oral agreement was made, it was done by an executive who has since left the company.

MBIA has another odd deal in its past. Another huge chunk--some $114 million--of the $586 million in losses MBIA has declared have come from a company called Capital Asset, which purchased tax liens. MBIA bought roughly 50% of the company in 1996 and took a majority position in late 1998. MBIA immediately tried to sell the company; then, in 1999, it wrote down the value of Capital Asset by $102 million. Then MBIA refinanced the tax liens via securitizations. MBIA guaranteed the deals, and in 2004 it was still taking losses on them.

The question is, why did MBIA increase its stake in Capital Asset in 1998? At the time, an MBIA spokesman told the press, "MBIA thinks the business ... has even more promise in the future." In an interview with FORTUNE in March, Brown said that Capital Asset was a "fraudulent shop," and that MBIA made its purchase to "take control of the problem." It's not clear what happened at Capital Asset, but one person familiar with the company says that MBIA knew it had a big problem. This person suspects that MBIA bought its second stake and refinanced the tax liens to delay taking losses. MBIA denies doing so. Its attorney says the idea that MBIA "viewed or could have viewed [Capital Asset] as a 'fraudulent shop' is simply wrong and not supported by the record." Brown now says his earlier declaration was a "emotional statement" that has been "taken out of context."

It's true that the entire bond insurance industry tries to avoid losses. And MBIA is proud of what it calls "surveillance" and "remediation." For every time MBIA does take a loss, "there are hundreds that we find a way" to avoid a hit, says Brown. That may range from MBIA's own collection agents calling nonpaying credit card holders to what the head of this division, Mitch Sonkin, calls "moral suasion." Brown even says that MBIA likes to do business with companies that need it. "When you're trying to fix something, you want to be the big guy in the room," he says. Adds Dunton: "We provide you with access to low-cost capital, so if there's a problem, fix it. I wouldn't say it, but that's an unspoken rule in the industry."

But among its peers MBIA has a reputation for aggressiveness, and this raises questions: Is there such a thing as inappropriate pressure? When will the company admit that a deal has gone bad? And just how much do the rating agencies really know?

In Grady County, Okla., MBIA insures $12.7 million of $17.6 million in debt that was used to build a jail. In late 2004 the jail, struggling to make interest payments, said it would lay off staff. And so, in what state officials say was an apparent attempt to get Oklahoma's government to take over the jail's debt--which it is in no way obligated to do--MBIA for a period stopped doing business in Oklahoma. "It's easy to say now that the deal was too risky, but MBIA made a bad judgment, and they should not expect us to step in and bail them out," says James Joseph, Oklahoma's state bond advisor.

As yet, there is no resolution to the jail's crisis, and Grady County sheriff Kieran McMullen says that the jail is looking for a way to make the bond payment due in May. But MBIA has not yet taken a reserve.

MBIA also seemed reluctant to warn investors that it had a problem with $73 million of debt that belonged to the Fort Worth Osteopathic Hospital. As early as 2003, the hospital was unable to meet certain terms of its debt agreement, and in mid-2004 it missed a payment. But the first warning of trouble any outsider would have gotten came in late September, when Moody's, which until that point had rated the hospital's bonds investment grade, downgraded the noninsured bonds to junk. And it wasn't until Oct. 7--just one day before Fort Worth Osteopathic announced that it was shutting its doors--that MBIA said it would take a $50 million reserve. MBIA says that this was "consistent with its loss- reserving methodology." MBIA also says it can collect another $22 million from a "sale or liquidation of the hospital's assets," which is why it reserved just $50 million instead of the full $73 million. But in early 2005 the hospital was sold for less than $8 million. MBIA says it can recover the rest from sales of real estate and other assets.

Fort Worth also makes you wonder how closely the rating agencies scrutinize MBIA's existing deals. Brown, who claims that the agencies regularly review some 98% of transactions, says that this is why MBIA is not a black box. "I don't care who you mention, Citigroup, Bank of America--they don't have anyone looking over their shoulder at every single transaction," he says. But if that's the case, how did Moody's miss the problems at Fort Worth Osteopathic? A Moody's spokesperson told Dow Jones that "we know as much as Fort Worth Osteopathic Hospital and MBIA were willing to disclose. It comes down to an honor system as to how much these guys share." A former senior agency employee says the agencies simply don't have the manpower to regularly review deals. "It's a lie that the rating agencies look at every transaction," he says.

The agencies all say they devote significant resources to regular surveillance. Moody's says the idea that MBIA doesn't regularly reveal information is "inaccurate."

A problem deal like Fort Worth Osteopathic also raises an issue that has been hovering over MBIA since Ackman identified it in his initial report. Are the company's tiny reserves really an adequate reflection of the risk in its portfolio? MBIA has $693 million in net reserves, $401 million of which are already dedicated to cover known losses on specific deals like AHERF. MBIA says its policy is to book a so-called case reserve only when it knows for certain it has a loss, and when that loss "can be reasonably estimated." The remaining $292 million are "unallocated" reserves, or money set aside to cover potential losses that have not yet been identified. MBIA has arrived at this figure by setting aside 12% of the premium it is scheduled to earn. In other words, it assumes an 88% gross profit margin each year. On the face of it, 12% may sound like a large number, but remember, MBIA's premium is just a sliver, perhaps 20 basis points, of the total value of its insurance. For example, if MBIA insured a $1,000 deal, and it earned 20 basis points, or $2, as a premium, then over the life of the deal it would put away just 12% of $2, or 24 cents, as reserves.

The heart of the matter, though, may not be so much the actual level of reserves, but the fact that in bond insurance there is discretion in how companies determine both revenues and reserves. As MBIA's filings note, determining reserves is an "inherently uncertain process involving numerous estimates and subjective judgments by management." As for revenues, when the company collects an up-front premium, its policy is to recognize a greater percentage of that in the early years of a deal. Changing this methodology "would materially affect the company's financial results," notes MBIA's financial statements. Indeed, the way in which MBIA records revenues and expenses like reserves are precisely why it has that 40% net margin.

The SEC has asked the Financial Accounting Standards Board to rule on the reserving question for the entire industry. MBIA executives say they are not concerned. "Our reserving is as perfectly disclosed as I know how to disclose it," says Brown. MBIA's latest financial report says that the company "cannot currently assess how the ... ultimate resolution of this issue will impact its loss-reserving policy or the effect it might have on recognizing premium revenue."

Ackman and MBIA's other critics argue that the company's reserves are too small and that it recognizes revenue too quickly. In other words, they contend, MBIA's accounting makes its business look more profitable than it actually is. MBIA says it is comfortable with its reporting.

IT WASN'T UNTIL March 30 that the debate over MBIA exploded into the mainstream of business news. That was the day MBIA announced that it had received a series of subpoenas from both Spitzer's office and the SEC. This time the subpoenas weren't limited to AHERF. They requested documents relating to, among other things, MBIA's accounting for advisory fees and its reserving methodology.

There isn't a smoking gun in the subpoenas, but they raise more questions about MBIA's claim that it is "conservative." Take the advisory fees. From 2000 through 2004, MBIA has recognized $219.5 million of advisory fees, a portion of which are paid up front instead of being recognized over the life of a guarantee, as the premiums are. That means higher revenue today but lower revenue tomorrow. MBIA earns such fees when, for example, it provides advice on a complicated transaction, although according to Budnick, it is paid only if the transaction is completed. MBIA says PricewaterhouseCoopers has signed off on its accounting. "There is a very specific set of guidelines we follow on advisory fees, and it has been approved," says Brown. But MBIA's competitors don't recognize fees in a similar manner. Its largest competitor, Ambac, says it recognizes all but a minuscule amount of fees over the life of a guarantee. Another competitor, Financial Security Assurance, even said in its financial filings, "Some industry peers receive so-called structuring or advisory fees ... and earn such fees upon receipt." FSA says it recognizes all but a tiny sliver of fees over the life of the transaction.

After the subpoenas were issued, two firms came out with negative reports on MBIA. One was Egan-Jones, which downgraded MBIA's credit another notch. The other was Standard & Poor's equity division, which slapped a "strong sell" rating on MBIA's stock. "We are concerned that weakness in the company's fundamentals ... could be exacerbated by uncertainties raised by these subpoenas," wrote the analyst. The rating agencies still rate MBIA triple A, although they say they are monitoring the situation.

Back in 2002, after Gotham's original report, then--Morgan Stanley analyst Alice Schroeder wrote, "We would not expect rating agencies to do an immediate 180-degree turn as a result of a report such as Gotham's, for several reasons: their deliberate process, the potential impact on issuers and markets of downgrades, and the hurdle of reversing a long-standing point of view. In addition, the rating agencies are an explicit participant in the guarantor's business model and in effect are now in the awkward position of passing judgment on themselves." The rating agencies say they would downgrade MBIA if it deserved it.

It is clear that for MBIA, the consequences of losing its triple A would be severe. Its stock price would probably fall sharply, because after all, what MBIA sells is its triple A: Without the rating, much of its new business would dry up. As Moody's itself once noted in a report, that could be the start of a vicious circle. "It may well be that the additional negative impact on the firm's future franchise value from the downgrade itself could cause Moody's to downgrade the company by an extra rating notch." In fact, the company's defenders cite this as a reason for their confidence: Because MBIA's triple-A rating is so critical to its franchise, its executives would never endanger that. Or so goes the logic.

What is unknown is the impact a downgrade might have on the capital markets. Some of the billions that MBIA insures would be repriced to reflect the fact that it would no longer have a triple-A rating, but it's unclear whether the reverberations of that repricing would be big or small. It's worth hoping we don't have to find out.

FEEDBACK bmclean@fortunemail.com