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

by Scott Patterson

  Asness just laughed. He knew Goldman was the place for him. In 1994, soon after finishing his dissertation, Clifford Asness, Ph.D., launched the Quantitative Research Group at Goldman Sachs. He was twenty-eight years old.

  WEINSTEIN

  One day in the early 1980s, Boaz Weinstein stared intently at an array of knights, pawns, kings, and queens scattered before him. He was nervous and on the defensive. Across the chessboard sat his stone-faced opponent: Joshua Waitzkin, the boy-genius chess master eventually profiled in the 1993 film Searching for Bobby Fischer.

  Weinstein lost the match against Waitzkin, played at the famed Manhattan Chess Club, but that didn’t dampen his enthusiasm for chess. He was soon beating his older sister so consistently that she quit playing him. Desperate to keep playing at home, he bugged his father into buying him a computerized chess game. By the time he was sixteen, Weinstein was a national “life master,” a few steps away from grandmaster, and No. 3 in the United States in his age group.

  Chess wasn’t everything to young Weinstein. There was also the tricky game of investing. A weekly ritual at the Weinstein household was watching the Friday night show Wall Street Week with Louis Rukeyser. He started dabbling in the stock market with his spare change, with some success. In his junior year as a student at New York’s elite Stuyvesant High School, he won a stock-picking contest sponsored by Newsday, beating out five thousand other contestants. Weinstein realized that in order to come out on top, he’d have to make picks with the potential for massive gains. His winning strategy was a primitive form of arbitrage: he shorted big gainers while picking beaten-down stocks he thought could rise sharply. The strategy showed that Weinstein could size up a situation and see what it would take to win—even if it was a massive gamble.

  Indeed, growing up in the privileged Manhattan neighborhood of the Upper East Side, he seemed to have money all around him. While Griffin, Muller, and Asness were all raised relatively far from the cacophonous din of Wall Street, Weinstein practically grew up on a trading floor. When he was fifteen, he took a part-time job doing clerical work for Merrill Lynch, the prestigious firm known for its “thundering herd” of brokers. During his downtime, he would scan research reports scattered around the office, looking for investing tips.

  Meanwhile, his sister had taken a job at Goldman Sachs. Weinstein would visit after hours and roam the warrens of the storied bank, dreaming of future glory. One day, while visiting her office, he made a pit stop at the men’s room. He ran into David DeLucia, a junk bond trader he recognized from a chess club they both played at. DeLucia gave Weinstein a quick tour of Goldman’s trading floor, which the starry-eyed Weinstein parlayed into a series of interviews. He eventually landed a part-time job working on Goldman’s high-yield bond trading desk when he was just nineteen years old.

  In 1991, he started taking classes at the University of Michigan, majoring in philosophy, drawn to the hard logic of Aristotle and the Scottish skeptic David Hume. He also became interested in blackjack, and in 1993 picked up Ed Thorp’s Beat the Dealer. He loved how card counting gave him a statistical ability to predict the future. It made him think of Mark Twain’s book A Connecticut Yankee in King Arthur’s Court, in which the main character, Hank Morgan, travels back in time and saves his neck by predicting a solar eclipse, having memorized all eclipses up to his own time.

  But Weinstein’s true passion was trading. He knew that once he left school, his first stop would be Wall Street. After graduating in 1995, he took a spot on the global debt trading desk at Merrill Lynch, where he’d had his first taste of Wall Street. Two years later he moved to a smaller bank, Donaldson Lufkin Jenrette, lured by DeLucia, who’d left Goldman. Weinstein thought it would be a good idea to work at a smaller firm, where he would have a better opportunity to run his own trading desk. At DLJ, he learned the nuts and bolts of credit trading by trafficking in floating-rate notes, bonds that trade with variable interest rates.

  For an up-and-coming trader with a nose for gambling such as Weinstein, it was an ideal time to launch a career on Wall Street. A boom in exotic credit derivatives was about to take off. Derivatives on stocks, interest rates, and commodities had been around for years. But not until the mid-1990s did the financial engineers concoct ways to trade derivatives linked to credit.

  It proved to be a revolution that changed the way Wall Street worked forever. Young bucks who hadn’t been trained in the old ways of credit trading—when all you needed to worry about was whether the borrower would pay the loan back and where interest rates might go—could dance circles around dinosaur rivals who couldn’t compete in the strange new world of derivatives. What’s more, banks were increasingly encouraging traders to push the envelope and generate fat returns. And a golden age of hedge funds, once largely the domain of freewheeling eccentrics such as George Soros or math geeks such as Ed Thorp, was taking off. Banks would compete against hedge funds for profits, eventually morphing into giant, lumbering hedge funds themselves.

  In 1998, Weinstein learned that a job had opened up at the German firm Deutsche Bank, where a number of the traders and researchers he’d worked with at Merrill had popped up. Deutsche was making a big push to transform itself from a stuffy, traditional commercial bank into a derivatives powerhouse. It was laying plans to purchase Bankers Trust, a cowboy New York investment bank packed with quants who thrived on designing complex securities. The deal, announced soon after Weinstein joined the firm, would make Deutsche the world’s largest bank, with more than $800 billion in assets at its fingertips.

  Weinstein thought it might be a good fit—the job was for a small desk, with little competition, at a firm making a big push into a field he was certain had plenty of room to grow. Soon after joining Deutsche, he was learning how to trade a relatively new derivative known as credit-linked notes. Eventually they would become more commonly called credit default swaps.

  Credit default swaps are derivatives because their value is linked to an underlying security—a loan. They were created in the early 1990s by Bankers Trust, but it wasn’t until the math wizards at J. P. Morgan got their mitts on them that credit derivatives really took off. When Weinstein arrived at Deutsche, only a few notes or swaps traded every day—light-years from the megatrillion-dollar trading in swaps that went on in cyberspace a decade later.

  Weinstein was taught how the notes worked by Deutsche Bank’s global head of credit trading, Ronald Tanemura, a trailblazer of the credit derivatives world who’d cut his teeth juggling complex securities in Japan and Europe for Salomon Brothers in the 1980s.

  Credit derivatives were, in a way, like insurance contracts for a loan, Tanemura explained to Weinstein at Deutsche’s New York headquarters, which sat in the shadow of the World Trade Center. Investors who buy the insurance on the loan pay a premium for the right to collect if the borrower goes belly up. The buyer and seller of the insurance basically swap their exposure to the risk that the bond will default.

  Weinstein quickly grasped the concept. Tanemura could tell he was a quick learner and a hard worker. One colleague thought he was also a bit on the nervous side, jittery and self-conscious.

  The swaps were commonly priced according to how much a trader would pay to insure several million dollars’ worth of bonds over a certain period of time, often five years, Tanemura explained. For instance, it may cost about $1 million to purchase insurance for $10 million worth of General Motors debt over five years, implying a 10 percent chance that the automaker would default during that time period. If GM does default, the party that provided insurance would need to cough up $10 million, or some percentage of the amount determined after the bankruptcy.

  Most of the trades were “bespoke,” custom-designed between two trading parties like a tailored London suit. “Credit derivatives basically give our customers exactly what they need,” Tanemura added. “And we supply it.”

  Weinstein soaked it up like a sponge with his photographic memory—and soon realized that the CDS market wasn’t about buy
and hold until the bond matured, it was about the perception of default. Traders didn’t need to wait around for a company to blow up. A trader who purchased a swap on GM for $1 million could potentially sell the swap to another trader later on for, say, $2 million, simply on the perception that GM’s fate had worsened.

  At the end of the day, it was all very simple: traders were betting on a level, just like a stock. If the company looked shaky, the CDS cost would rise.

  In theory, hundreds of swaps, or more, can be written on a single bond. More commonly, swaps are written on baskets of hundreds or thousands of bonds and on other kinds of loans. They could metastasize without end—and did—reaching a value of more than $60 trillion a decade after Weinstein arrived on the scene.

  What’s more, since the trades were commonly done on a case-by-case basis on the so-called over-the-counter market, with no central clearinghouse to track the action, CDS trading was done in the shadow world of Wall Street, with virtually no regulatory oversight and zero transparency. And that was just the way the industry wanted it.

  Soon after Weinstein took the job, his boss (not Tanemura) jumped ship. Suddenly he was the only trader at Deutsche in New York juggling the new derivatives. It was no big deal, or at least it seemed that way. It was a sleepy business, and few traders even knew what they were or how to use the exotic swaps—or had any idea that they represented a new front in the quants’ ascendancy over Wall Street. Indeed, they would prove to be one of the most powerful weapons in their arsenal. The quants were steadily growing, moving ever higher into the upper echelons of the financial universe.

  What could go wrong?

  As it turned out, a great deal—a four-letter word: LTCM.

  In 1994, John Meriwether, a former star bond trader at Salomon Brothers, launched a massive hedge fund known as Long-Term Capital Management. LTCM was manned by an all-star staff of quants from Salomon as well as future Nobel Prize winners Myron Scholes and Robert Merton. On February 24 of that year, the fund started trading with $1 billion in investor capital.

  LTCM, at bottom, was a thought experiment, a laboratory test conducted by academics trained in mathematics and economics—quants. The very structure of the fund was based on the breakthroughs in modern portfolio theory that started in 1952 with Harry Markowitz and even stretched as far back as Robert Brown in the nineteenth century.

  LTCM specialized in relative-value trades, looking for relationships between securities that were out of whack. It made money by placing bets on pairs of securities that drifted out of their natural relationship, ringing the cash register when the natural order—the Truth—was restored.

  One of LTCM’s favorite bets was to purchase old, “off the run” Treasuries—bonds that had been issued previously but had been supplanted by a fresh batch—while selling short the new “on the run” bonds. It was a trade that dated back to Meriwether’s days at Salomon Brothers. Meriwether had noticed that the newest batch of bonds of equal maturity—ten years, thirty years, five years, whatever—almost always traded at a higher price than bonds that had gone into retirement. That made no sense. They were essentially the same bond. The reason for the higher price was that certain investors—mutual funds, banks, foreign governments—placed a premium on the fact that newer bonds were easy to trade. They were liquid. That made them more expensive than more seasoned bonds. Okay, thought Meriwether, I’ll take the liquidity risk, and the premium, betting that the bonds eventually converge in price.

  One problem with this trade is that it doesn’t pay much. The spread between new and old bonds is fairly small, perhaps a few basis points (a basis point is one-hundredth of a percentage point). The solution: leverage. Just borrow as much cash as possible, amp up the trade, and you basically have a printing press for money.

  Meriwether spent $20 million on a state-of-the-art computer system and hired a crack team of financial engineers to run the show at LTCM, which set up shop in Greenwich, Connecticut. It was risk management on an industrial level.

  The principal risk management tool used by LTCM had been created by a team of quants at J. P. Morgan. In the early 1990s, Wall Street’s banks were desperate for a methodology to capture the entire risk faced by the bank on any given day. It was a monumental task, since positions could fluctuate dramatically on a daily basis. What was required was a sophisticated radar system that could monitor risk on a global level and spit out a number printed on a single sheet of paper that would let the firm’s CEO sleep at night.

  Getting the daily positions was hard but not impossible. Advances in computer technology enabled rapid calculations that could aggregate all the bank’s holdings. The trouble was determining the global risk. The model the J. P. Morgan quants created measured the daily volatility of the firm’s positions and then translated that volatility into a dollar amount. It was a statistical distribution of average volatility based on Brownian motion. Plotted on a graph, that volatility looked like a bell curve.

  The result was a model they called value-at-risk, or VAR. It was a metric showing the amount of money the bank could lose over a twenty-four-hour period within a 95 percent probability.

  The powerful VAR radar system had a dangerous allure. If risk could be quantified, it also could be controlled through sophisticated hedging strategies. This belief can be seen in LTCM’s October 1993 prospectus: “The reduction in the Portfolio Company’s volatility through hedging could permit the leveraging up of the resulting position to the same expected level of volatility as an unhedged position, but with a larger expected return.”

  If you can make risk disappear—poof!—in a quantitative sleight of hand, you can layer on even more leverage without looking like a reckless gambler.

  Others weren’t so sure. In 1994, a financial engineering firm doing consulting work for LTCM was also working with Ed Thorp, who that year had started a new stat arb fund in Newport Beach called Ridgeline Partners. An employee of the consulting firm told Thorp about LTCM and said it would be a great investment.

  Thorp was familiar with Scholes, Merton, and Meriwether—but he hesitated. The academics didn’t have enough real-world experience, he thought. Thorp had also heard that Meriwether was something of a high roller. He decided to take a pass.

  For a while, it looked like Thorp had made the wrong call. LTCM earned 28 percent in 1994 and 43 percent the following year. In 1996, the fund earned 41 percent, followed by a 17 percent gain in 1997. Indeed, the fund’s partners grew so confident that at the end of 1997 they decided to return $3 billion in capital to investors. That meant more of the gains from LTCM’s trades would go to the partners themselves, many of whom were plowing a great deal of their personal wealth into the fund. It was the equivalent of taking all of one’s chips, shoving them into the pot, and announcing, “All in.”

  Meriwether and his merry band of quants had been so successful, first at Salomon Brothers and then at LTCM, that bond trading desks across Wall Street, from Goldman Sachs to Lehman Brothers to Bear Stearns, were doing their level best to imitate their strategies. That ultimately spelled doom for LTCM, known by many as Salomon North.

  The first blow was a mere mosquito bite that LTCM barely felt. Salomon Brothers’ fixed-income arbitrage desk had been ordered to shut down by its new masters, Travelers Group, which didn’t like the risk they were taking on. As Salomon began to unwind its positions—often the very same positions held by LTCM—Meriwether’s arbitrage trades started to sour. It set off a cascade as computer models at firms with similar positions, alerted to trouble, spat out more sell orders.

  By August 1998, the liquidation of relative-value trades across Wall Street had caused severe pain to LTCM’s positions. Still, the fund’s partners had little clue that disaster was around the corner. They believed in their models. Indeed, the models were telling them that the trades were more attractive than ever. They assumed that other arbitrageurs in the market—Fama’s piranhas—would swoop in and gobble up the free lunch. But in the late summer of 1998, the piranhas
were nowhere to be found.

  The fatal blow came on August 17, when the Russian government defaulted on its debt. It was a catastrophe for LTCM. The unthinkable move by Russia shook global markets to their core, triggering, in the parlance of Wall Street, a “flight to liquidity.”

  Investors, fearful of some kind of financial collapse, piled out of anything perceived as risky—emerging-market stocks, currencies, junk bonds, whatever didn’t pass the smell test—and snatched up the safest, most liquid assets. And the safest, most liquid assets in the world are recently issued, on-the-run U.S. Treasury bonds.

  The trouble was, LTCM had a massive short bet against those on-the-run Treasuries because of its ingenious relative-value trades.

  The off-the-run/on-the-run Treasury trade was crushed. Investors were loading up on newly minted Treasuries, the ones LTCM had shorted, and selling more seasoned bonds. They were willing to pay the extra toll for the liquidity the fresher Treasuries provided. It was a kind of market that didn’t exist in the quantitative models created by LTCM’s Nobel Prize winners.

  As Roger Lowenstein wrote in his chronicle of LTCM’s collapse, When Genius Failed: “Despite the ballyhooed growth in derivatives, there was no liquidity in credit markets. There never is when everyone wants out at the same time. This is what the models had missed. When losses mount, leveraged investors such as Long-Term are forced to sell, lest their losses overwhelm them. When a firm has to sell in a market without buyers, prices run to the extremes beyond the bell curve.”

  Prices for everything from stocks to currencies to bonds held by LTCM moved in a bizarre fashion that defied logic. LTCM had relied on complex hedging strategies, massive hairballs of derivatives, and risk management tools such as VAR to allow it to leverage up to the maximum amount possible. By carefully hedging its holdings, LTCM could reduce its capital, otherwise known as equity. That freed up cash to make other bets. As Myron Scholes explained before the disaster struck: “I like to think of equity as an all-purpose risk cushion. The more I have, the less risk I have, because I can’t get hurt. On the other hand, if I have systematic hedging—a more targeted approach—that’s interesting because there’s a trade-off: it’s costly to hedge, but it’s also costly to use equity.”