Fool's Gold: How the Bold Dream of a Small Tribe at J.P. Morgan Was Corrupted by Wall Street Greed and Unleashed a Catastrophe (16 page)

In 2006, though, in San Francisco, Las Vegas, Miami, and other once-hot areas, house prices stalled. Though the slowdown wasn’t dramatic or spread around the whole country, it nonetheless triggered a wave of subprime defaults. Faced with the prospect that their homes might not be such reliable cash cows, a number of subprime borrowers who had been
depending on their home equity continuing to rise robustly decided that they were better off abandoning their mortgages. From early 2002 to early 2005, while the proportion of subprime borrowers who were delinquent on payments had fallen from 15 percent to 9 percent, by early 2006, the number had risen above 12 percent.

Dimon and the other managers debated what to do. They were deeply reluctant to abandon the idea of creating a mortgage-repackaging pipeline, but Dimon was also very uneasy. “I
know
about this stuff [of consumer lending]!” he would growl at Bill King, harking back to his days in consumer credit. Eventually, a cautious compromise was reached. The bank maintained its mortgage “production line,” but the lending unit was told to raise its underwriting standards and reduce the level of unsold mortgages it held on its books. The J.P. Morgan managers were determined not to get caught short if defaults rose any higher.

The bank also turned to the derivatives market to reduce its risk, by purchasing credit default swaps from other parties, which promised to redeem any default losses on the mortgage bonds it would begin selling. Such mortgage derivatives had barely existed a few years earlier, but they were among the products that had become so hot in the last years. Increasingly, banks were using them as insurance against losses from their mortgage-repackaging business. Then, in January 2006, an index for tracking these offerings and their values was launched, much like the Dow Jones, called the ABX. This allowed all those playing in the mortgage-based investing terrain to track prices and get a continuous read on investor sentiment. King and his colleagues dived in, thrilled that they could buy protection against the risk of a possible cascading of defaults. Somewhat to their surprise, they found that even though defaults had started to rise, there was absolutely no shortage of players willing to take the other side of their bets and insure J.P. Morgan against the danger that the market might turn sour.

At almost all other banks and investment funds, financiers were still hunting for ways to
increase
their exposure to the subprime market and boost their returns. By early 2006, in fact, CDS contracts looked to many like a particularly attractive way to place bullish bets. The big, dirty secret of the securitization world was that there was such a frenetic appetite for
more and more subprime loans to repackage into CDOs that the supply of loans had started lagging behind demand. Mortgage derivatives were an easy substitute, since there was no constraint on how many of
those
could be created, or at least not as long as some market players, such as JPMorgan Chase, were eager to hedge their risk by buying them. Ironically, JPMorgan’s Chase’s conservative stance on subprime loans thereby enabled others to continue rolling the dice on the mortgage boom, to a far greater degree than would have been possible if their bets had been limited to the world of tangible bonds. The JPMorgan Chase management thought they had reached a reasonable compromise, but before long they would be forced to carefully revisit their mortgage repackaging business.

In another corner of the bank, Bill Winters had been becoming increasingly uneasy about the mortgage market for his own set of reasons. Winters knew that the 1990s derivatives team had rejected the notion of putting mortgage loans into BISTRO deals, and he respected that judgment call. But with others making so much money with such a range of mortgage-repackaging products, the pressure was on. Dimon had made it clear that he wanted a mortgage production line, so Winters had duly asked his staff to reexamine how to create a profitable business selling mortgage-based CDOs.

When they crunched the numbers, though, they encountered a problem. “There doesn’t seem to be a way to make money on these structures,” Brian Zeitlin, one of the bankers who worked in the CDO division, reported. Baffled, Winters told Zeitlin to look again: so many banks were doing a booming business in the field, why couldn’t JPMorgan Chase do the same?

Zeitlin and his colleagues conducted more analysis but kept encountering the same sticking point. When the business was viewed through the prism J.P. Morgan employed to test whether a business was profitable, it just did not seem to make much sense. The issue was the “spread,” or the level of return, on the CDOs. Back in the days when J.P. Morgan had arranged its first BISTRO deal, investors holding junior tranches of
debt were paid around 375 basis points over LIBOR, nice compensation for the risk they were assuming. Even the AAA tranches had a spread of 60 basis points, which was a high enough return to enable the bank to sell the notes. The bank had made juicy fees on these sales. The return on the super-senior debt had been dramatically smaller, but the bank had decided to pass that, at a slight loss, to AIG, Swiss Re, and others and continue cranking out deals because the profits from the higher-risk notes were so good.

As interest rates had continued to fall and so many more banks and investment funds began offering CDOs, though, the profits to be made from each deal declined. The returns on the underlying loans repackaged into the CDOs had fallen substantially. Whereas loans might have once produced a cash flow of 10 percent a year, say, it was now just 5 percent. Worse, banks were competing so fiercely to purchase the raw material needed to create CDOs that the price of that raw material was rising. If a mortgage had a face value of $100,000, say, banks might now pay $102,000 as opposed to $100,500 to buy that loan. They kept making more and more of the purchases anyway because they could still turn a profit from turning around and selling the CDO notes. But the margins were being squeezed. The business had fallen victim to a vicious cycle: banks and other lenders were issuing more and more mortgages, which were riskier and riskier, so that those loans could be repackaged into more and more CDOs in order to make up for the declining profit margins.

In the process, those selling these products were creating huge masses of super-senior risk, and Winters and his team wondered how the others were coping with the expense of either selling that off or insuring it. The cost of that cut even further into profits. One solution was to simply park the super-senior on J.P. Morgan’s books, but Winters didn’t like that idea at all. Like Bill Demchak eight years earlier, Winters thought that it was just a bad idea to let assets pile up on the balance sheet. And while Winters was not opposed to taking calculated risks, he always wanted to be properly compensated for doing so. To him, the returns on super-senior were far too low to justify holding it in any appreciable volume.

Some of Winters’s staff floated the idea of buying insurance against the super-senior risks, as Demchak’s group had arranged with AIG. By 2005, a host of insurance companies were offering that service. Some of the most active were MBIA and Ambac, which were two industry leaders in the so-called monolines niche, companies that were specialized bond insurers. These monolines had first sprung up three decades earlier to offer insurance for bonds issued by government municipalities, but they had started to insure CDO notes in the 1990s.

Back in the 1990s, the Demchak team had always been cynical about the monolines. They were very thinly capitalized, for one thing. At some, the volume of assets they were insuring was 150 times larger than their underlying equity. Demchak had reasoned that with so little in reserves, the monolines wouldn’t really be able to pay on any of that insurance in the event of the kind of credit shock it would take to produce losses on the senior risk. If that financial disaster ever
did
hit, it probably would be bad enough to simply wipe out the monolines. Buying insurance from them was therefore pointless. Demchak said: “To us, that business model seemed nuts!” Winters shared that view, preferring to keep any dealings with monolines to a modest level. “We did not really use monolines [much] because we reasoned that they would not be there whenever you really needed them.” Winters told his team, “You have got to get rid of the senior debt in a clean sale—until you do that, you cannot do the deals.”

Frustrated, the team began hunting for ways to sell the irritating super-senior risk, but there were few takers. By 2005, the spread on super-senior was around 15 basis points, and few investors wanted assets yielding so little. Some banks were trying to solve that problem by repackaging super-senior liabilities into new products; in mid-2005, Deutsche Bank launched one, for example, known as “leveraged super-senior.” Those were generally sold to Canadian and German investors. The banks were also selling super-senior to investment groups, like hedge funds and the proliferation of SIVs closely affiliated to the banks. But JPMorgan Chase had never created a network of SIVs attached to it. Back in the late 1990s, when Citigroup and other banks started creating SIVs, J.P. Morgan executives debated whether they should do so too, but in the end opted out. Winters and others calculated that these vehicles were
quite risky, because they raised their funding in the short-term commercial paper market. Meanwhile, the assets they were selling would pay the bank over a much longer time frame. This mismatch alarmed Winters. Such was his wariness that when he became deputy CEO of the investment bank, he slashed the $12 billion of credit lines that J.P. Morgan had already extended to other banks’ SIVs to a mere $500 million. He also campaigned to get rid of a SIV that had come along with the merger with Bank One. “I could never work out why anyone thought that SIVs were a good idea,” Winters later remarked.

Without an SIV network to take on the super-senior risk it would accrue by cranking up its mortgage-repackaging pipeline, Winters and Zeitlin just couldn’t see a way to make the business pay. Reluctantly, Winters told the J.P. Morgan management that the bank should not open the spigots on its pipeline after all. The decision was greatly frustrating, though. The other banks were pushing JPMorgan Chase further and further down the league tables largely due to the bonanza from their mortgage pipelines. So were they just ignoring the risks? Or had they found some alchemy that made the economics of their machines work? Winters simply could not work it out.

Having spent all his career at J.P. Morgan, Bill Winters assumed that the logic he employed was roughly similar to that prevailing in the other banks. He knew perfectly well that banking cultures varied, but it was fiendishly difficult for anyone to ever know for sure just
how big
the variations might be and just how crucial decisions were made at other banks. The internal corporate dynamics and incentives schemes of banks were closely guarded secrets, and when staff hopped from one bank to another, they were interviewed by the senior management to pick their brains in the way that the CIA would have debriefed a defecting KGB spy. It was years before that mystery of what the others were doing was solved, and when the story of what they were up to was revealed, Winters was shocked.

O
n April 13, 2005, some of the former J.P. Morgan bankers from the old derivatives team assembled at the Acropolis Center in Nice, southern France. The occasion was an industry conference to discuss the credit markets. A whole decade had passed since they had helped build the CDS world. Many were both stunned and thrilled by the sheer pace of growth. By mid-2005, there were $12 trillion in CDS contracts alone in the market, a sum equivalent to the size of the entire US economy. The CDO market was also exploding. “The speed of growth is just astonishing, even for those involved,” observed Terri Duhon, who had worked on J.P. Morgan’s team in New York. By 2005, she had left ABN AMRO and was working as a consultant in London, giving advice to investors trying to jump into the frenzied business.

Yet Duhon was also becoming uneasy. “There is a type of euphoria about getting into structured credit products right now, and so you have to ask, Could we be getting into a state of irrational exuberance?” she observed. Cynthia McNulty, another former J.P. Morgan banker, agreed. “There is such a buzz about credit derivative products now that there are hedge funds getting into it without the requisite abilities.”

Ian Sideris, a London-based lawyer who was handling some of the structured products that banks were producing, was also becoming worried. “You have to really wonder about the capacity of investors to handle this,” he said at the time. With a grim half smile, Sideris revealed that he had recently seen deals that were being sold not only to retail investors but to government bodies, pension funds—and even a charity serving
disabled people in Australia. All of those investors, it seemed, were chasing higher returns. Yet few of them, Sideris pointed out, had the ability to fully comprehend these new products. Even some of the bankers seemed overwhelmed.

Those concerns were prescient. A few weeks after the conference in Nice, ratings agencies downgraded the debt of General Motors, taking the markets entirely by surprise. That triggered panic among some investors, and many rushed to sell CDSs and CDOs, causing their prices to drop, an eventuality not predicted by the models. JPMorgan Chase, Deutsche Bank, and many other banks and funds suffered substantial losses.

For a few weeks after the turmoil, the banking community engaged in soul-searching. At J.P. Morgan the traders stuck bananas on their desks as a jibe at the so-called F9 model monkeys, the mathematical wizards who had created such havoc. (The “monkeys” who wrote the statistical models tended to use the “F9” key on the computer when they performed their calculations, giving rise to the tag.) J.P. Morgan, Deutsche, and others conducted internal reviews that led them to introduce slight changes in their statistical systems. GLG Ltd., one large hedge fund, told its investors that it would use a wider set of data to analyze CDOs in the future.

Within a couple of months, though, the markets rebounded, and the furor died down. Most banks and ratings agencies continued to use the Gaussian copula approach to risk assessment. Having built their structured finance machines, they could not suddenly turn them off, and there was no obvious alternative model on offer. As the former J.P. Morgan “quant,” David Li, pointed out, “It’s not the perfect model…[but] there’s not a better one yet.” Meanwhile, some bankers comforted themselves with the observation that the industry had been tested and had quickly recovered. “We have had a shake-up, but that is good in a way,” Tim Frost observed a few months after the storm. “People have learnt lessons.” Frost liked to quote Friedrich Nietzsche’s observation that “What does not kill us makes us stronger.” He passionately believed in free-market principles, to such an extent that he even stood for election for the British Parliament, for the right-wing Conservative Party (he
lost). As part of that philosophy, he believed that a shakeout would make the credit derivatives markets more vibrant and efficient.

On the other side of the Atlantic, though, one of Frost’s former colleagues didn’t quite share Frost’s optimism. That man was Andrew Feldstein. Back in the days when the J.P. Morgan team had concocted its derivatives dream, Feldstein had believed deeply in the
intellectual
arguments behind financial innovation. He was utterly convinced that if tools such as derivatives were implemented in a rational, efficient manner, they would vastly improve the financial system and economy. It was the dream that drove them all.

But after living through the mess of the Chase–J.P. Morgan merger, Feldstein became cynical. He still believed derivatives had the theoretical potential to make markets function better, but in practice, dysfunctional management and warped incentives for traders and the ratings agencies were badly distorting the CDO market. He understood the ways in which the banks were playing around to garner good ratings and make end runs around the regulatory system, and the situation troubled him. But it also presented a trading opportunity.

When he left J.P. Morgan, he created a hedge fund, together with two partners, Gery Sampere, a J.P. Morgan analyst, and Stephen Siderow, a McKinsey consultant who had been Feldstein’s classmate at Harvard Law School in the class of 1991. (Another member of that year was Barack Obama, whom Feldstein later supported.) They rented a small, windowless room in a skyscraper on Madison Avenue and set up shop under the name BlueMountain Capital Management. Key to their strategy was placing bets that CDOs were being mispriced.

Feldstein was convinced that the models banks and funds used were miscalculating the true degree of default correlation in loan bundles, and he was convinced that eventually economic reality would hit. The true risk of the CDOs would then become clear, and prices would drop. “The models are great, but at the end of the day, in credit, anyone relying solely on them will lose,” Feldstein explained soon after he created his fund.
“Those who are successful know this and overlay their own idiosyncratic views about risk, actual correlation among and between groups of credits, and other factors on any model they use.”

Ironically, however, such hedging on the market only further fueled the boom. After all, for every buyer in a market there must be a seller; a market in any commodity—be it equities, art, or synthetic CDOs—can operate only if there are parties on both sides of the trade. Feldstein and some others putting their money on the contrarian view helped to make a lively new business of trading in default swaps take off.

By 2005 there were more tools available to conduct such trading, too. In the early years, bankers who wanted to trade credit default swaps generally used only contracts that related to single names. From 2004 onward, though, indices of credit default swaps sprang up, known as “CDX” in the US and “iTraxx” in Europe. They tracked the cost of insuring against default on a basket of companies, offering a handy way for investors to evaluate trends in pricing, in the same way that the S&P 500 shows how the whole equity market is moving. They could also be traded as contracts in their own right.

The ABX index, which tracked the price of derivatives written against risky home equity loans, provided another way for investors to trade, after its launch in January 2006. Other mutations proliferated, too. The LCDX was an index of loan derivatives; the TABX tracked derivatives on mortgage tranches; CMBX was an index of derivatives on commercial mortgages. It was a veritable alphabet soup.

By 2006, bankers and investors were using all of these indices to trade as if their flashy computing models were infallible. Feldstein and his colleagues at BlueMountain just did not believe that they had assessed the true risks. To make the point, he and his colleagues placed an old-fashioned abacus in their office and labeled it “correlation calculator” as a black joke.

They also constructed an investment strategy to take advantage of the shortcomings of the banks’ models. To do this, they quietly constructed CDS portfolios that they traded with different investment banks and then used detective work to guess how each bank was modeling CDO risk. Then, like hackers tapping into a computer, they hunted for
the flaws in those models, which they could exploit. It was not hard to find such weaknesses, since the models varied, sometimes in a haphazard manner. Sometimes the banks’ traders guessed what the geeks at BlueMountain were doing. They rarely complained, though. Individual bank traders did not usually have any personal incentive to worry about whether Feldstein was exploiting their models or not. They got paid according to their trading results,
as measured by the internal models at the banks
—and as long as those internal models produced flattering prices, those traders got fat bonuses.

In any case, few managers sitting at the top of the investment banks had much idea what their traders were doing, let alone whether or not the models were accurate. By 2005, very few men running investment banks had extensive experience in structuring and trading derivatives. The field was just too young to have produced many high-level leaders, and many derivatives experts were too cerebral to play the type of internal corporate political games needed to rise to the top at most banks. Bankers who had started their careers as corporate advisers or salesmen tended to be better at charming their superiors. One exception was Lloyd Blankfein, the CEO of Goldman Sachs; Anshu Jain, the co-CEO of Deutsche’s investment bank, was another. But Citigroup, Merrill Lynch, UBS, and numerous others were run by former bond and equity salesmen, lawyers, wealth managers, and commercial bankers. Such men had little instinctive interest in the technical details of managing risk. Moreover, the wider competitive climate provided an overwhelming incentive for them to focus on revenues above all else. If they didn’t deliver higher revenues, their company’s stock prices would be pummeled. “Banks now face the challenging task of sustaining their success in creating value,” Boston Consulting Group opined in a report on the industry’s performance. “Continuing to increase profitability remains important but becomes more and more difficult as profitability has already reached a new level.”

The antics of Goldman Sachs had a particularly significant impact on the psychology of many senior bankers. In 2002 and 2003, Goldman Sachs startled its rivals by delivering a hefty increase in profits. In 2004 that winning streak intensified when the brokerage raised its revenues by
a third, to more than $16 billion—or more than any other player. That cranked up the pressure on other banks to deliver equally impressive growth. “We all got Goldman envy—it became like an obsession,” one European bank CEO later recalled.

So, as “Goldman envy” had spread, rivals had frantically embraced ideas for boosting profits that in retrospect would look like lunacy. What was worse, many even stepped up their sales of mortgage-based products after the housing market had begun to turn. Merrill Lynch, the home of the “thundering herd” of skilled salesmen in the equities world, was one such.

Back in 2003, the bank had appointed a new chief executive officer, Stanley O’Neal. He decided to bring Merrill into the securitization business with a vengeance, hiring teams of traders who were ordered to start cranking out masses of CDOs. By 2006, Merrill topped the league table in terms of underwriting CDOs, selling a total of $52 billion that year, up from $2 billion in 2001. (J.P. Morgan was in seventh place in 2006, selling a mere $22 billion.) The herd had crashed the party in style. But behind the scenes, Merrill was facing the same problem that worried Winters at J.P. Morgan: what to do with the super-senior debt?

Initially, Merrill solved the problem by buying insurance for its super-senior debt from AIG, just as Demchak’s team had done. However, in late 2005, AIG told Merrill it would no longer offer that service. One corner of the AIG empire was involved in extending mortgages, and AIG officials were getting alarmed about subprime risk. That left Merrill Lynch’s CDO desk with a big headache. But unlike at J.P. Morgan, that stumbling block was not enough to prompt Merrill Lynch to duck out of the game.

The CDO team decided to start keeping the risk on Merrill’s books, and a senior trader named Ranodeb Roy was appointed to manage it. As the super-senior quickly piled up, some of it was insured with monolines and some was buried on the books. Traders joked that if they could not protect themselves by handing the risk to a monoline, they would use the “Ronoline” instead. A few bank officials expressed unease. Jeffrey Kronthal, one senior manager, tried to impose a $3 billion to $4 billion limit on the amount of super-senior risk that the bank could take on. Also, in
the summer of 2006, one trader protested when his colleagues created a $1.5 billion CDO deal called Octans and asked him to put almost $1 billion of that risk on the bank’s books. However, these protests were quickly squashed. The leaders of the CDO department—Harin de Silva and Osman Semerci—were determined to keep cutting deals. So was their boss, Dow Kim. In 2006, sales of the various CDO notes produced some $700 million worth of fees. Meanwhile, the retained super-senior rose by more than $5 billion each quarter.

Very few bankers inside Merrill Lynch had much idea what the CDO desk was doing. At Merrill, as with most Wall Street banks, departments competed viciously for resources and power, and the department that was most profitable usually had the most clout. As the CDO team posted more and more profits, it became increasingly difficult for other departments, or even the risk controllers, to interfere. O’Neal himself could have weighed in, but he was in no position to discuss the finer details of super-senior risk. The risk department did not even report directly to the board. O’Neal faced absolutely no regulatory pressure to manage the risk any better. Far from it. The main regulator of the brokerages was the SEC (Securities and Exchange Commission), which had recently removed some of the old constraints.

Until 2004, the SEC had imposed controls on the amount of assets a brokerage could hold on its balance sheet relative to its core equity. In April of that year, however, the SEC’s five commissioners had decided—by a unanimous vote—to lift that so-called leverage ratio control. The ruling attracted almost no attention in the press at the time, since it seemed highly technical, but it had one very considerable consequence: it raised the competitive pressure on the brokerages even further. By 2005, it had become clear that a key reason why Goldman Sachs was producing such stellar earnings was that it had raised its leverage. Merrill Lynch and the others were therefore under intense pressure to follow suit, and, that being the case, the increased leverage of super-senior risk was tolerated.