The Quants: How a New Breed of Math Whizzes Conquered Wall Street and Nearly Destroyed It

by Scott Patterson

  Then, in 2007, the music stopped. The carry trade blew up. The securitization machine collapsed as homeowners started defaulting on their loans in record numbers.

  As a risk manager, Brown watched it all go down at Morgan Stanley, one of the biggest players in the CDO casino, which he’d joined in 2004 after leaving Citigroup.

  Brown joined “Mother Morgan,” as it was known, under the reign of Phil Purcell, who’d taken over several years earlier in a vicious power struggle with another Morgan kingpin, John Mack. Purcell had joined Morgan as part of a $10 billion merger in 1997 between the elite storied bank and Dean Witter Discover & Co., a brokerage that largely catered to middle-income customers. White-shoe Morganites were aghast. But Purcell, named CEO in the deal, proved a savvy rival to Mack, who’d been at Morgan since 1972, starting off as a bond trader. In 2001, Mack left the bank, realizing he couldn’t unseat Purcell; he first worked at Credit Suisse First Boston before joining a hedge fund.

  After Mack left, however, earnings at Morgan hadn’t kept pace with those of its rivals, especially Goldman Sachs. Between Mack’s departure and early 2005, the market value of the company had fallen nearly 40 percent, to $57 billion. While the value of competitors had also taken a hit, Morgan’s drop was among the steepest on Wall Street. Underlings fumed at Purcell. They said he was too cautious. That he wasn’t a real risk taker. That he didn’t have the balls to make real money—like John Mack.

  Brown thrived, however. He’d been hired to help the bank’s credit system get up to speed with an arcane set of regulations known as the Basel Accord, an international standard designating how much capital banks needed to hold to guard against losses. Morgan’s chief financial officer, Steve Crawford, a Purcell protégé, had hired Brown. He wanted him to accomplish the task—which had taken commercial banks such as Citigroup several years—in eighteen months or less.

  “If you can do that, you can have any job at the bank you want,” Crawford promised.

  Brown was impressed by Morgan’s higher-ups, who seemed to appreciate the oft-ignored quants and had encouraged a number of programs to gear up the firm’s risk management capabilities. But in a palace coup, Purcell and his favorites, including Brown’s benefactor, Crawford, were forced out in June 2005 by a band of high-powered shareholders. His replacement: John Mack.

  Mack, a native of North Carolina and son of Lebanese immigrants, promised to bring back the old aggressive culture of Morgan. He found the performance of his beloved Morgan under Purcell unacceptable. Over all his years at the bank, he’d overseen the first stat arb operation under Nunzio Tartaglia in the 1980s, and had also helped manage Peter Muller’s group. He had a taste for risk. Morgan, he believed, had lost it. Upon his return, Mack marched onto the Morgan trading floor like a triumphant general on the streets of ancient Rome. The financial news channel CNBC broadcast the event live. The bank’s traders peeled their eyes away from their ubiquitous Bloomberg terminals to loudly cheer the second coming of “Mack the Knife,” a nickname earned by his willingness to slash payrolls and cut costs.

  Morgan had been left behind by fast movers such as Goldman Sachs and Lehman Brothers, Mack said, and its profits were suffering. The new paradigm for investment banks on Wall Street was risk taking. The ideal model was Henry Paulson’s Goldman Sachs, with its hugely successful Global Alpha Fund and outsized profits in private equity.

  Paulson himself had outlined the new paradigm in the bank’s 2005 annual report. “Another key trend is the increasing demand from clients for investment banks to combine capital and advice,” he wrote. “In other words, investment banks are expected to commit more of their own capital when executing transactions. … Investment banks are increasingly using their own balance sheets to extend credit to clients, to assume market risk on their behalf and sometimes co-invest alongside them.”

  Goldman’s strategy reflected the shifts that had been going on at investment banks for more than a decade. Banks were in a life-and-death struggle to keep talented traders from jumping ship and starting hedge funds—as Cliff Asness had done in 1998. They were going head-to-head with the Greenwich gunslingers and losing. No bank saw this more clearly than Goldman. Others, such as Boaz Weinstein’s Deutsche Bank and Peter Muller’s Morgan Stanley, were close behind. The only way to compete would be to offer their best and brightest massive paychecks, and open the gates wide on risk taking and leverage. In short order, Wall Street’s banks morphed into massive, risk-hungry hedge funds, Goldman leading the way and Morgan close behind.

  Regulators lent a helping hand. On a spring afternoon in late April 2004, five members of the Securities and Exchange Commission gathered in a basement hearing room to meet a contingent of representatives from Wall Street’s big investment banks to talk about risk. The banks had asked for an exemption for their brokerage units from a regulation that limited the amount of debt they could hold on their balance sheets. The rule required banks to hold a large reserve of cash as a cushion against big losses on those holdings. By loosening up these so-called capital reserve requirements, the banks could become more aggressive and deploy the extra cash in other, more lucrative areas—such as mortgage-backed securities and derivatives.

  The SEC complied. It also decided to rely on the banks’ own quantitative models to determine how risky their investments were. In essence, in a move that would come to haunt not just the agency but the entire economy, the SEC outsourced oversight of the nation’s largest financial firms to the banks’ quants.

  “I’m happy to support it,” said Roel Campos, an SEC commissioner. “And I keep my fingers crossed for the future.”

  Initially, Morgan wasn’t eager to join the party. A mantra at Morgan before John Mack returned was that the bank “wouldn’t be another Goldman,” according to a person who worked at the bank. Morgan would exercise caution during boom times to be prepared for the inevitable bust when the music stopped.

  Mack’s return changed that. His solution was that the firm should take bigger, bolder gambles, and more of them, just like Goldman.

  Looking on, Brown grew concerned as Morgan’s risk appetite surged. The new regime seemed to act as if risk management were simply a matter of filling out the forms, dotting the i’s and crossing the t’s, but not a central part of the firm’s ethos, which was to ring the cash register.

  Brown also raised his eyebrows at one of Mack’s themes. In meeting after meeting in conference rooms near the bank’s executive offices, Mack said he wanted to double Morgan’s revenues in five years and keep costs flat. Nice idea, Brown thought. But how exactly are we going to do that?

  The answer, he feared, was simply to take more risk.

  Among the ideas Mack’s staff cooked up for reaching his goal: increase investments in the financial derivatives business, plow headlong into the booming field of residential mortgages, and take more risk with the firm’s own capital on its proprietary trading desks such as Peter Muller’s PDT.

  Morgan quickly found a way to combine all three of those goals in one area: subprime mortgages. In August 2006, Morgan rolled out a plan to purchase subprime mortgage lender Saxon Capital for $706 million. The bank’s perpetual-motion subprime machine was cranking up.

  Brown could see it all happening before his eyes. His job as a risk manager for the bank’s credit division gave him unique visibility into the positions on Morgan’s fixed-income balance sheet. For the most part, it seemed under control. But there was one area that troubled him: securitization and all those subprime mortgages.

  Subprime mortgages had become the new darlings of Wall Street. The more high-risk borrowers who could be enticed to take high-interest mortgages, the more high-risk/high-reward CDOs—and synthetic CDOs—could be created by and for Wall Street investors. As long as the carousel kept spinning, everyone would get a brass ring.

  Brown, however, was growing increasingly concerned about Morgan’s securitization merry-go-round. Just as it had at Citigroup, one of his biggest worries centered on the massive �
��warehouses” of subprime mortgages Morgan used to store the loans. Most banks, inspired by Salomon Brothers, had created off-balance-sheet vehicles that would temporarily house the loans as they were bundled, packaged, sliced, diced, and sold around the world. Such vehicles funded themselves using the commercial paper markets, short-term loans that constantly needed to be rolled over. Any hitch in the chain, Brown realized, could result in disaster. Still, he didn’t think it represented a risk that could substantially damage the bank. Profits were through the roof. He also took consolation in Morgan Stanley’s sky-high share price. If the bank suffered a huge hit, it could always raise cash on the open market with a share offering. He hadn’t factored in the possibility that Morgan’s share price would collapse in an industrywide meltdown.

  By early 2007, Morgan Stanley was on one of the hottest streaks in its history. The bank was coming off its best quarter ever, and its best year ever, in profitability. One big success, Mack noted in the firm’s April 2007 conference call, was the institutional securities group run by Morgan’s high-flying co-president, Zoe Cruz. The group managed “a tremendous amount of risk in a very smart and disciplined way,” Mack said.

  Its leverage ratio, however—the amount of borrowed money it uses to trade every day—was a whopping 32 to 1. In other words, Morgan was borrowing $32 for every $1 it actually owned. Other investment banks, such as Bear Stearns, Lehman Brothers, and Goldman Sachs, also had sky-high leverage ratios. Internal measures at some of these banks showed the leverage was even higher than the official numbers reported to the SEC.

  One of Cruz’s desks was a proprietary credit-trading group formed in April 2006. Howard Hubler, a managing director with years of experience trading complex securities at Morgan, was in charge of it. It was a hedge-fund-like trading outfit that would wager Morgan’s own money in the credit markets, a bond-trading mirror image of PDT.

  At first Hubler’s group was a roaring success, netting the firm $1 billion by early 2007. Hubler was among Wall Street’s new breed of correlation traders using David Li’s Gaussian copula to measure the risks of defaults among various tranches of CDOs. His strategy involved shorting the lower rungs of subprime CDOs (or derivatives tied to them) while holding on to the higher-rated CDO tranches. By the quants’ calculations, those high-quality CDO slices had little chance of losing value.

  As subsequent events proved, correlation trading turned out to be a hornet’s nest of risk. Hubler thought he was shorting subprime. But in a cruel twist, Hubler ended up long subprime. He got the correlations wrong.

  Brown, meanwhile, had been growing more and more alarmed about the risks the firm was taking in the subprime mortgage market. The loans were fed into one end of Morgan’s securitization machine by subprime lenders such as Countrywide and New Century Financial and pumped out the other end to investors around the world. Indeed, though few realized it at the time, Morgan was one of the biggest players during the peak years of subprime, 2005 and 2006, underwriting $74.3 billion in subprime mortgages, according to Inside Mortgage Finance (Lehman was number one, with $106 billion in mortgages underwritten).

  Morgan was also lending aggressively elsewhere, backing massive volumes of credit card debt and corporate loans. Brown realized it was an unsustainable process bound to come crashing down. “It all only made sense if the people we were lending to could pay us back,” recalled Brown. “But it became clear that the only way they could pay us back was by borrowing more. There were all kinds of deals we were doing that only made sense if credit is good. We knew at some point that the musical chairs would stop, and we would own a lot of this stuff, and we wouldn’t have the capital to pay for it.”

  Brown believed the quants who worked on the CDO models were often narrowly focused on the Byzantine details of the deals and rarely looked at the big picture—such as the looming bubble in the housing market. “They were showing zero risk,” he recalled. It was the same mistake everyone made, from the rating agencies to the banks to the home builders to the buyers of those homes who expected to refinance their mortgages as soon as the payments shot up. On the surface there was little reason to think otherwise. Home prices had never declined on a national level since the Great Depression. Plus, everyone was happy. Some were getting filthy rich.

  Despite his concerns, Brown didn’t raise serious alarms at Morgan. He realized that the bank was going to take losses, large ones, once the credit cycle turned. But it wouldn’t be fatal. Plus there was that sterling stock price.

  Indeed, virtually no one on Wall Street had any notion of the massive implosion heading its way. The industry, fueled by greater and greater feats of financial engineering, seemed to be hitting on all cylinders. Huge profits were certainly rolling in at Morgan Stanley. Hubler’s bet on subprime, initiated in December 2006, contributed significantly to Morgan’s 70 percent increase in net income to a record $2.7 billion in its fiscal first quarter of 2007.

  But problems in the complex bet sprouted in the spring of 2007 as homeowners started defaulting in huge numbers in states that had seen massive run-ups in prices, including California, Nevada, and Florida. The higher-rated subprime CDO tranches also started to quiver.

  Cracks had first started to appear in February 2007, when HSBC Holdings, the third-largest bank in the world, boosted estimates of expected losses from subprime mortgages by 20 percent to $10.6 billion. Just four years earlier, HSBC had piled into the U.S. subprime market when it snapped up Household International Inc., which became HSBC Finance Corp. Household’s chief executive at the time, William Aldinger, had boasted after the deal closed that the company employed 150 quants who were whizzes at modeling credit risk. Other firms, ranging from Seattle banking giant Washington Mutual to mortgage lenders such as New Century and IndyMac Bancorp, were also warning of large losses from subprime mortgage holdings.

  Brown started to think of jumping ship, and that’s when he began talking to AQR.

  It seemed like propitious timing. This new Morgan, this subprime-fueled leverage-happy hot rod, wasn’t a place he wanted to be a part of anymore. In late 2006, he’d taken a call from Michael Mendelson, a top researcher at AQR. Brown had recently published a book called The Poker Face of Wall Street, a mix of biographical reflections and philosophical ruminations about gambling and finance. The quants at AQR loved the book and thought Brown would be a good fit.

  More important, AQR was considering an IPO and needed someone familiar with the nitty-gritty compliance details that went with becoming a public company. Disillusioned with Morgan and disappointed with a less-than-stellar bonus that spoke to the company’s lack of appreciation for his talents, Brown was intrigued. He had several interviews with AQR and met Asness, who seemed to speak his language and clearly understood quantitative risk management (though in their first meeting they primarily compared notes on a shared passion: old movies). By June 2007, Brown was making the daily commute on the Metro-North Railroad from New York to AQR’s Greenwich headquarters.

  By then, serious trouble was erupting in subprime. The same month Brown joined AQR, news emerged that a pair of Bear Stearns hedge funds that had dabbled heavily in subprime CDOs—the mind-numbingly named Bear Stearns High Grade Structured Credit Strategies Master Fund and Bear Stearns High Grade Structured Credit Strategies Enhanced Leverage Master Fund—were suffering unexpected losses. Managed by a Bear Stearns hedge fund manager named Ralph Cioffi, the funds had invested heavily in subprime CDOs.

  Broadly, Bear Stearns was optimistic that while the housing market was shaky, it wasn’t poised for serious pain. A report issued on February 12, 2007, by Bear researcher Gyan Sinha argued that weakness in certain derivatives tied to subprime mortgages represented a buying opportunity. “While the subprime sector will experience some pain as it removes some of the froth created by excesses,” he wrote, “an over-reaction to headline risk will create opportunities for nimble investors.”

  Such thinking was a recipe for a blowup. Cioffi’s Enhanced fund first started to lose money
the same month Sinha wrote his report. The more sedate High Grade fund, which had posted positive returns for more than three years in a row, slipped 4 percent in March. The leveraged fund was on the cusp of imploding. In April, an internal Bear Stearns report on the CDO market revealed that huge losses could be on the way. Even those sterling AAA bonds could be in trouble. One of the Bear fund’s managers, Matthew Tannin, wrote in an internal email that if the report was correct, “the entire subprime market is toast. … If AAA bonds are systematically downgraded then there is no way for us to make money—ever.”

  Spooked investors started to ask for their money back. Goldman Sachs, which acted as a trading partner for the Bear funds, said its own marks on the securities the funds held were much worse than Cioffi’s marks. From there, it was only a matter of time. On June 15, Merrill Lynch, a creditor to the funds, seized about $800 million of their assets. The following week, Merrill started to sell off the assets in a series of auctions, triggering shock waves throughout the CDO market. The fire sale forced holders of similar CDOs to mark down the prices of their own securities.

  Back at Morgan Stanley, Howie Hubler was beginning to sweat. The collapse of low-rated CDO tranches was exactly what he’d bet on. But weakness in the high-rated tranches, the AAAs, wasn’t in his playbook. Hubler was short $2 billion worth of low-quality CDOs. Disastrously, he held $14 billion of high-rated “supersenior” CDOs—the kind that in theory could never suffer losses.

  In July, panic set in. Credit markets began to quake as investors in subprime CDOs all tried to bail out at once. The commercial paper market, which had been used to fund the off-balance-sheet vehicles that were the engine of Wall Street’s securitization machine, started to freeze up. With all the forced selling and few buyers, the losses proved far worse than anyone could have imagined.