Wall Street's money machine breaks down

Merrill Lynch CEO Stanley O'Neal lost his job.

More from Fortune

Will Mmmhops be a hit? NBA confirms L.A. Clippers sale to ex-Microsoft CEO Steve Ballmer FBI and SEC probe into Carl Icahn and golfer Phil Mickelson

FORTUNE 500 Current Issue Subscribe to Fortune

Big Names, Big Losses

The toll keeps rising Here are the losses that top banks and brokers have booked (or announced) on CDOs, asset-backed securities, and other structured products.

Write-downs on structured products
Citigroup	$9.8 billion	This is the low estimate; Citi says the figure could be $3 billion higher.
Merrill Lynch	$7.9 billion	Analysts project that the broker will have to write down billions more this quarter.
UBS	$4.4 billion	UBS still has nearly $40 billion in CDOs and mortgage-backed securities on its books.
Morgan Stanley	$3.7 billion	Morgan's total subprime exposure after write-downs stands at $6 billion.
Wachovia	$2.1 billion	Wachovia was among the top issuers of subprime mortgage CDO debt this year.
Credit Suisse	$948 million	Credit Suisse lost nearly another $1 billion on leverage loans.
Lehman Brothers	$700 million	Total includes leveraged loans; Lehman does not provide more detail.
Bank of America	$3.527 billion	CEO Ken Lewis is cutting back the company's investment-banking operations.
Bear Stearns*	$1.65 billion	Big hedge fund losses in June kicked off the subprime follies.
J.P. Morgan Chase	$339 million	CEO Jamie Dimon is credited with losing less than his peers.

Structure products include collateralized debt obligations, collateralized loan obligations, asset-backed securities, and mortgage-backed securities. They do no include leveraged loans. *Estimate.
Source: Company earnings releases

(Fortune Magazine) -- Two things stand out about the credit crisis cascading through Wall Street: It is both totally shocking and utterly predictable.

Shocking, because a pack of the highest-paid executives on the planet, lauded as the best minds in business and backed by cadres of math whizzes and computer geeks, managed to lose tens of billions of dollars on exotic instruments built on the shaky foundation of subprime mortgages.

Predictable because whether it's junk bonds or tech stocks or emerging-market debt, Wall Street always rides a wave until it crashes. As the fees roll in, one firm after another abandons itself to the lure of easy money, then hands back, in a sudden, unforeseen spasm, a big chunk of the profits it booked in good times.

"The fee engine becomes so huge that these products take on a life of their own," says Tiger Williams, CEO of Williams Trading, a leading financial services firm for hedge funds. "Everyone rationalizes that it's safe because they're making so much money. But it's far from safe."

In pure destructive power, the subprime mess has become Wall Street's version of Hurricane Katrina. It has wreaked havoc on the nation's iconic brokerage firm, Merrill Lynch (Charts, Fortune 500), and biggest bank, Citigroup (Charts, Fortune 500), which have announced billions of dollars in losses and parted ways with their celebrated CEOs, E. Stanley O'Neal and Charles Prince. Banks, brokerages, and lenders have announced thousands of layoffs, and more are sure to come.

The blow to shareholder wealth is staggering. Since June 29, Citi's share price has dropped 35%, from $51 to $33, while Merrill's stock has slid from $84 to $54, a 36% swoon. In the same period, the dozen biggest Wall Street firms and the commercial banks with the largest investment arms - a list that includes Bank of America (Charts, Fortune 500), J.P. Morgan Chase (Charts, Fortune 500), and Credit Suisse (Charts) - have lost more than $240 billion in market value. Dozens of smaller companies in the mortgage business have suffered huge losses or folded completely.

The crisis of confidence has exploded beyond Wall Street, driving the dollar to record lows - and helping send the prices of commodities, especially oil, soaring to historic highs.

The results could be devastating for the U.S. economy. "The subprime crisis and its fallout on commodity and foreign-exchange markets significantly raises the odds of a recession early next year," says Mark Zandi, chief economist at Moody's Economy.com.

And it's far from over. As stunning as today's losses are, more carnage lies ahead. Wall Street banks are holding tens of billions in risky securities on their books, and no one seems to have any idea what they're worth. In conference calls and press releases, banks have been changing their estimates of the value of these assets.

Merrill Lynch, for example, predicted a $4.5 billion subprime loss for the third quarter, then jolted investors and analysts three weeks later by announcing that its real deficit was $7.9 billion - or 76% more than the initial estimate. (Oops!)

In fact, Wall Street banks are sitting on rotting piles of highly suspect, thinly traded securities no one wants to touch. "Whenever the market turns against you, you take the biggest losses in illiquid securities," says Richard Bookstaber, former head of risk management at Salomon Bros. "Because there are so few buyers, you're forced to sell at a discount that is both huge and highly unpredictable."

What really spooks investors is the fog surrounding the future. One problem is that they can't trust management's estimates of future losses. Citi, for example, says it will take additional write-downs of $8 billion to $11 billion in the fourth quarter.

But it's impossible to know whether those numbers have any relation to reality. Presumably, they are based on a theoretical model, but such models have proved highly unreliable. When Citi actually brings the securities to market, it may have to slash their prices to unload them, forcing it to take a much bigger write-down.

The banks are also far from forthcoming with detailed information on their positions, making it difficult for analysts to assess what the future holds. "The risk to investors is far greater because we're getting so little information," says Michael Mayo, an analyst at Deutsche Bank.

Backed by Treasury Secretary Henry Paulson, Bank of America, Citi, and J.P. Morgan are trying to establish a giant fund that would buy distressed debt so that investors who own it don't have to unload it at fire-sale prices. The hope is that the market will rebound before too long and that the bonds will regain much of their value. But there's no guarantee that the bonds will ever bounce back, and the bailout fund may simply delay the day of reckoning, pushing losses further into the future.

Just how big could those losses be? Both Mayo and analyst Meredith Whitney of CIBC project that write-downs could total $50 billion or more by the end of the year. Longer term, Mayo sees losses climbing to $70 to $100 billion. The wide range simply underscores the uncertainty surrounding subprime. "This will take two to three years to play out," says Mayo, explaining that it will take that long for lenders to foreclose on troubled mortgages and sell the collateral - in this case, hundreds of thousands of homes - to recoup part of their loans.

In this special report, we take a broad tour of the financial wreckage, featuring a close look at Citigroup and comments on the plunging dollar, the echoes of Enron, and CEO accountability. Here, we'll try to dispel the biggest mystery of the subprime crisis: How did the Wall Street firms manage to pile mountains of high-risk mortgage debt, bonds that most investors and analysts thought the firms were selling to their customers, onto their own books?

Their gambit amounted to massive speculation in subprime mortgages. Investors are justly aghast that Wall Street ignored the obvious pitfalls. We'll explain how it happened by examining one of the firms deepest in the subprime swamp, Merrill Lynch. Presumably, many other banks followed similar paths -although Citigroup added a few twists of its own, as we explain in the Robert Rubin story. Hence, Merrill's story may shed light on the misadventures of its rivals, from Morgan Stanley (Charts, Fortune 500) to UBS.

Naturally we don't know the thinking behind Merrill's disastrous foray into subprime. The firm declined to speak to Fortune in detail for this story. But by talking to former employees, experts on risk management, and Wall Street analysts, Fortune has put together the most likely scenario for how Merrill's subprime business swung from a relatively steady channel for booking fees into a wild gamble that devastated the firm.

To understand how Merrill came to grief, we have to take a mind-bending trip inside the complex creation on which it nearly wagered the franchise: the collateralized debt obligation, or CDO, a type of investment vehicle that buys and sell bonds.

Merrill and other banks typically don't operate CDOs. True to their role as middlemen, they help clients create CDOs, take a fee, and then exit the deal. It was Merrill's decision to alter this template - by becoming a huge investor in the funds it assisted - that got the firm in trouble.

Here's how a typical CDO backed by subprime mortgages might work. The game starts when a client - known as the collateral manager - approaches Merrill Lynch and asks it to provide financing for a CDO that will hold, for example, $1 billion worth of bonds backed by subprime mortgages. The clients are mostly big money-management firms like Pimco, Trust Co. of the West, and Cohen & Co.

To get things rolling, Merrill makes $1 billion available to the collateral manager, taking a fee of 1.5% to 2%, or $15 million to $20 million. The collateral manager uses the balance to purchase bonds backed by pools of subprime mortgages (known as "subprime mortgage ABS," for asset-backed securities) issued by Wall Street firms, including Merrill.

At the same time, Merrill's structured-finance team gets to work creating a variety of bonds that will be backed by the interest and principal payments the CDO collects on the asset-backed securities it owns. (To make things more confusing, the bonds the CDO issues are also called CDOs. In this story, we'll use CDO to refer to the investment vehicle; we'll call the securities it issues CDO bonds.) Using complex mathematical models that predict default and payoff rates, among other things, the bankers create several tranches of securities with different interest rates and levels of risk.

At the top of the heap are the super-senior bonds, which have the first claim on the cash coming into the CDO. They are designed to earn ratings of AAA from agencies like Standard & Poor's - meaning they're supposedly as sound as the best corporate bonds - and they pay the lowest rates. The bonds that have a lesser claim on the cash flow pay higher rates and get lower ratings.

Merrill turns the bonds over to its sales force to peddle to hedge funds, pension funds, and other investors. The appeal of CDO bonds is simple: They pay better rates than corporate issues with identical credit ratings. And in the low-rate environment of the past seven years, yield-hungry hedge funds were eager to buy any paper that offered extra returns. "The whole idea," says Brad Hintz of Bernstein Research, "is taking a pool of risky, illiquid bonds and, through the magic of securitization, offering higher yields than on similarly rated securities."

How are CDOs able to offer premium yields on their bonds? Most of them did it by purchasing the riskiest, lowest-rated mortgage-backed bonds - you know, the ones built on loans to borrowers with spotty credit and dubious résumés. Such bonds paid what were then super-high rates of 9% to 11% in 2006.

Didn't those loans carry a high risk of default? Well, yes, they did. But here's the key point, and where Wall Street went astray. During the early years of the housing boom, default rates on all mortgages were unusually low. That led bankers - and more important, rating agencies - to build unrealistic assumptions about future default rates into their valuation models.

And because the subprime mortgage bonds paid such lofty interest rates, the financiers figured that an unimaginably high percentage of them would have to default to cause any real problems for the top-rated CDO bonds. Of course, that's just what happened.

None of this would have been a dire problem for Merrill if it hadn't gone from simply manufacturing CDOs and reaping fees to becoming a huge investor in the CDOs it created - getting high on its own supply, you might say. Merrill was willing, even eager, to speculate with its own balance sheet because of a dramatic change in culture engineered by Stan O'Neal. Until 1997, Merrill didn't engage in a lot of speculative trading for its own account; its trading unit concentrated on making markets for clients. Merrill made its money from relatively safe, fee-generating businesses, courtesy of its army of brokers, now numbering 16,000, and a thriving underwriting operation for stocks and bonds.

After O'Neal became co-head of the institutional business in March 1997, things began to change. There was a shift from trading for customers to trading for the firm's own account. Under O'Neal's watch, Merrill expanded its relationship with Long-Term Capital Management, providing the high-powered hedge fund with lucrative financing.

By the time LTCM collapsed in September 1998, O'Neal had become CFO. Then-CEO David Komansky, a popular former broker, was struggling to cut Merrill's bloated costs, and the board judged that O'Neal, despite the LTCM debacle, had the tough, clinical approach needed to do the job. O'Neal succeeded in putting Merrill in trim.

But he also remained a strong advocate for proprietary trading, especially because Goldman Sachs was generating such huge returns making aggressive bets with its own capital. "Management should have learned from the LTCM and Asian crises that when liquidity runs dry, it happens in a day," says former Merrill Lynch investment banking chief Barry Friedberg. "When it's over, you can't get out."

At first, Merrill treated the CDO trade as a client business. The idea was to get in and out quickly - help structure the CDO, hand it to the manager, and pocket the fees. But as the fees rose, so did Merrill's hunger for market share. The driving force was CDO chief Chris Ricciardi, who constantly pushed the troops to top the league tables. Merrill rose from a bit player in mortgage CDOs in 2003, with just $3.4 billion in underwritings, to the leader from 2004 through 2006, posting $44 billion in deals backed by mortgages last year.

O'Neal primed the pump by purchasing First Franklin, one of the nation's largest originators of subprime mortgages, in December 2006 for $1.3 billion. In early 2007 one unit at Merrill was busy packaging First Franklin's loans into subprime ABS that another Merrill unit bought for the CDOs. Incredibly, in the first half of 2007, Merrill underwrote $28 billion in mortgage CDO bonds, far exceeding its pace for 2006.

The market for CDO debt, however, changed radically starting in 2006. Anyone who picked up a newspaper read constantly about the woes in housing. Foreclosures were rising, home prices were falling in dozens of major markets, and the Fed was closing the era of supercheap money.

In February international bank HSBC suffered big losses on its subprime portfolio, sending tremors through the market. Under normal circumstances, such worries would have led to higher interest rates on subprime mortgage bonds. But these were not normal times. Instead of rising, rates on subprime mortgage bonds remained abnormally low until the summer of 2007, and in some months even dropped below 2006 levels.

What was going on? Instead of backing away from subprime paper, Merrill Lynch and other big players were gobbling all they could, because they needed it to feed their CDOs. Their bottomless appetite for the stuff kept prices high and yields low, against all economic logic. "What we had was a perpetual-motion machine driving mortgage prices to uneconomic levels of risk vs. reward," says Friedberg.

With the yields on the subprime paper falling, the yields on the CDO bonds sank as well. Even so, hedge funds and other investors continued to snap up the lower-rated paper, which still offered relatively generous yields. But they were rejecting the AAA-rated bonds, which paid just 30 to 50 basis points over LIBOR, the international interbank borrowing rate.

That put Merrill in a bind. It couldn't raise those yields on the AAA-rated bonds because it had paid such high prices for the subprime mortgage bonds that provided funding for them. "The hedge funds and other customers were demanding higher yields, so Merrill couldn't sell that supposedly safe AAA-rated paper," says Mark Adelson, a consultant on structured finance and a former analyst for Nomura Securities.

To keep the merry-go-round spinning, Merrill apparently made a pivotal - and reckless - decision. It bought big swaths of the AAA paper itself, loading the debt onto its own books. "Merrill took the top tranches onto its own balance sheet," says Scott Sprinzen, an analyst with S&P. "The amounts were staggering."

By the end of June, Merrill held $41 billion in subprime CDO and subprime mortgage bonds. Since the average deal is between $1 billion and $1.5 billion, and the AAA debt is around 80% of each deal, Merrill must have been buying nearly all the top-rated debt from dozens of CDOs.

The question is why Merrill would purchase bonds its customers were rejecting. Merrill hasn't given a detailed explanation of how it came to own such a large volume of subprime bonds. At first, Merrill made money on the bonds because it was able to benefit from the "carry trade" - borrowing money at low rates and using it to buy CDO bonds paying higher rates.

Also, Merrill execs apparently believed that the credit market turmoil would ease and the bonds would once again be easy to sell. That turned out to be far too optimistic, of course. But the overarching explanation is probably that Merrill became addicted to the fees that flowed from financing CDOs, which reached $700 million in 2006. "They must have had their eyes on the fees and not the risk," says Friedberg. Other big players, like UBS and Morgan Stanley, may have followed the same script.

That turned out to be one of the worst miscalculations in the annals of risk management. In late June the collapse of the Bear Stearns hedge funds helped trigger a big rise in rates on subprime mortgages. In early October, Merrill shocked the markets by predicting the $4.5 billion loss on subprime. But when it made its official earnings announcement two weeks later, investors were appalled to learn that the actual number was $7.9 billion.

Merrill's $41 billion exposure to subprime paper was more than its entire shareholders' equity of $38 billion. That this huge position went unhedged astonishes everyone on Wall Street. The $7.9 billion write-down meant that Merrill lost 19% on its bonds.

How did bonds rated AAA take the kind of hit you'd expect on junk bonds? One reason is that the rating agencies enormously underestimated the chance of default in subprime mortgages. And as subprime borrowers stopped making their payments, the cash flows weren't big enough to make full interest payments, or in some cases, any payments, to even the highest-rated tranches.

As a result, much of the AAA debt now sits at junk-bond status. "It never deserved to be AAA to begin with," says risk-management guru Bookstaber.

The high ratings are no excuse for Merrill. The spreads on the debt it was buying were far below the historic average. It should have realized that even if rates moved back to the levels that prevailed just a few years ago, it would take huge losses. "The banks were in denial," says Adelson. "They thought they were smarter than the market."

Merrill still has $21 billion of unhedged exposure to subprime bonds. It's also holding $11 billion in CDO bonds that it says are fully hedged. Analysts are worried about how effective those hedges will be as prices plunge. Analyst Mayo estimates that Merrill will have to write down another $4 billion in the fourth quarter.

The SEC is requesting information from Merrill and other firms on what they knew when they made sunny statements about subprime this summer, and how they priced their portfolios of CDO bonds.

The subprime saga is far from over. The markets remain on high alert. Each day seems to bring new rumors or announcements of losses. A golden age for banks and brokers has come to a sudden end.

The moral is clear. When Wall Street appears in genius mode, raking in huge profits on mysterious products and complex trades, the secret isn't genius at all. It's that hubris is running wild, and so is risk. And whether it's tomorrow or five years hence, risk will jump from the shadows, knife in hand, to cut genius down to size.

Reporter associates Marilyn Adamo, Lisa Bergtraum, Katie Benner, Susan Kaufman, Eugenia Levenson and Joan Levinstein contributed to this article.