The Quants: How a New Breed of Math Whizzes Conquered Wall Street and Nearly Destroyed It (52 page)
Authors: Scott Patterson
The models that gauged the risk:
“Behind AIG’s Fall, Risk Models Failed to Pass Real-World Test,” by Carrick Mollenkamp, Serena Ng, Liam Pleven, and Randall Smith,
Wall Street Journal
, October 31, 2008.
“That patient has had a heart attack”:
The account is based on an interview with New York senator Chuck Schumer.
A FLAWIn a May 2005 speech:
“Risk Transfer and Financial Stability,” by Alan Greenspan, remarks made to the Federal Reserve Bank of Chicago’s Forty-first Annual Conference on Bank Structure, Chicago, Illinois, May 5, 2005.
“Ken, you guys are getting killed”:
Several details of Citadel’s late-2008 turmoil, and the conference, are based on an interview with Ken Griffin and interviews with numerous people familiar with the fund who requested anonymity.
Others, including the James Forese quote, are based on “Citadel Under Siege: Ken Griffin’s $15 billion Firm Was Flirting with Disaster This Fall,” by Marcia Vickers and Roddy Boyd,
Fortune
, December 9, 2009; and “Hedge Fund Selling Puts New Stress on Market,” by Jenny Strasburg and Gregory Zuckerman,
Wall Street Journal
, November 7, 2008; and “A Hedge Fund King Comes Under Siege,” by Jenny Strasburg and Scott Patterson,
Wall Street Journal
, November 20, 2009.
It was a rallying cry:
Columbus did not write “Sail on” in his 1492 journal.
By outward appearances, Boaz Weinstein:
A number of details of Saba’s final days were first reported in “Deutsche Bank Fallen Trader Left Behind $1.8 Billion Hole,” by Scott Patterson and Serena Ng,
Wall Street
Journal
, February 6, 2008.
Cliff Asness was furious:
Several details of AQR’s struggles in 2008 were first reported in “A Hedge-Fund King Is Forced to Regroup,” by Scott Patterson,
Wall Street Journal
, May 23, 2009.
a fund with ties to Nassim Taleb:
Universa’s gains were first reported in “October Pain Was ‘Black Swan’ Gain,” by Scott Patterson,
Wall Street Journal
, November 3, 2008.
THE DEVIL’S WORKPaul Wilmott stood before a crowded room:
The account is based on firsthand reporting and interviews with Paul Wilmott.
Together that January, they wrote:
The full “manifesto” can be found on Wilmott’s website,
http://www.wilmott.com/blogs/eman/index.cfm/2009/1/8/The-Financial-Modelers-Manifesto
.
“They’re usually doing the devil’s work”:
Interview with Charlie Munger.
Even before the fury of the meltdown hit:
The account is based on a series of interviews with Mandelbrot in his Cambridge apartment.
In February 2008, Ed Thorp gazed:
The account is based on a meeting with Ed Thorp in his office, and a subsequent meeting with Thorp and Bill Gross in Pimco’s office. The Q&A appeared in “Old Pros Size Up the Game—Thorp and Pimco’s Gross Open Up on Dangers of Over-Betting, How to Play the Bond Market,” by Scott Patterson,
Wall Street Journal
, March 22, 2008.
DARK POOLSOn a sultry Tuesday evening:
The account of the poker night is firsthand.
Muller had been working on a new business model:
“Morgan Stanley
Eyes Big Trading Change,” by Aaron Lucchetti and Scott Patterson,
Wall Street Journal
, April 24, 2009.
There were other big changes in Simons’s life:
“Renaissance’s Simons Delays Retirement Plans,” by Jenny Strasburg and Scott Patterson,
Wall Street Journal
, June 11, 2009.
“They’re starting to sin again”:
“After Off Year, Wall Street Pay Is Bouncing Back,” by Louise Story,
New York Times
, April 26, 2009.
a new breed of stock exchange:
In part derived from “Boom in ‘Dark Pool’ Trading Networks Is Causing Headaches on Wall Street,” by Scott Patterson and Aaron Lucchetti,
Wall Street Journal
, May 8, 2008.
Arbitrage:
The act of buying and selling two related securities that are priced differently with the expectation that the prices will converge. If gold costs $1,000 an ounce in New York and $1,050 in London, an arbitrageur will buy the New York gold and sell it in London. Quants use formulas to detect historical relationships between assets such as stocks, currencies, and commodities, and place bets that any disruption in the relationships will revert back to normal in time
(see
statistical arbitrage)
. Such bets are placed under the assumption that past performance in the market is predictive of future performance—an assumption that isn’t always true.
Black-Scholes option-pricing formula:
A mathematical formula that describes the price of a stock option, which is a contract that gives its owner the right to buy a stock (a call option) or sell a stock (a put option) at a certain price within a certain time. The formula has many components, one of which is the assumption that the future movement of a stock—its volatility—is random and leaves out the likelihood of large swings (see
fat tails)
.
Brownian motion:
First described by Scottish botanist Robert Brown in 1827 when observing pollen particles suspended in water, Brownian motion is the seemingly random vibration of molecules. Mathematically, the motion is a random walk in which the future direction of the movement—left, right, up, down—is unpredictable. The average of the motion, however, can be predicted using the law of large numbers, and is visually captured by the bell curve or normal distribution. Quants use Brownian motion mathematics to predict the volatility of everything from the stock market to the risk of a multinational bank’s balance sheet.
Credit default swap:
Created in the early 1990s, these contracts essentially provide insurance on a bond or a bundle of bonds. The price of the insurance fluctuates depending on the riskiness of the bonds. In the late 1990s and 2000s, more and more traders used the contracts to make bets on whether a bond would default or not. At Deutsche Bank, Boaz Weinstein was a pioneer in the use of CDS as a betting instrument.
Collateralized debt obligation:
Bundles of securities, such as credit-card debt or mortgages, that are sliced up into various levels of risk, from AAA, which is deemed relatively safe, to BBB (and lower), which is highly risky. In the late 1990s, a team of quants at J. P. Morgan created “synthetic” CDOs by bundling credit default swaps linked to bonds and slicing them up into various portions of risk. In the credit meltdown of 2007 and 2008, billions in high-rated CDO and synthetic CDO slices plunged in value as borrowers defaulted on their mortgages in record numbers.
Convertible bonds:
Securities issued by companies that typically contain a fixed-income component that yields interest (the fixed part), as well as a “warrant,” an option to convert the security into shares at some point in the future. In the 1960s, Ed Thorp devised a mathematical method to price warrants that anticipated that Black-Scholes option-pricing formula.
Efficient-market hypothesis:
Based on the notion that the future movement of the market is random, the EMH claims that all information is immediately priced into the market, making it “efficient.” As a result, the hypothesis states, it’s not possible for investors to beat the market on a consistent basis. The chief proponent of the theory is University of Chicago finance professor Eugene Fama, who taught Cliff Asness and an army of quants who, ironically, went to Wall Street to try to beat the market in the 1990s and 2000s. Many quants used similar Fama-derived strategies that blew up in August 2007.
Fat tail:
The volatility of the market is typically measured using a bell curve, which represents the normal distribution of market movements captured by Brownian motion. The tails of the distribution—the left and right sides of the curve—slope downward. A fat tail represents a
highly unlikely “black swan” event not captured by the bell curve, and visually is captured by a bulge on either side of the curve. Benoit Mandelbrot first devised methods to describe such extreme market events in the 1960s, but he was largely ignored.
Gaussian copula:
A model developed by financial engineer David X. Li that predicted the price correlations between various slices of collateralized debt obligations. Copulas are mathematical functions that calculate the connections between two variables—in other words, how they “copulate.” When X happens (such as a homeowner defaulting), there’s a Y chance that Z happens (a neighboring homeowner defaults). The specific copulas Li used were named after Carl Friedrich Gauss, the nineteenth-century German mathematician known for devising a method to measure the motion of stars through the bell curve. The connections among the default risks of the slices in a CDO were, therefore, based on the bell curve (a copula is essentially a multidimensional bell curve). In the credit crisis that began in August 2007, the model failed as the correlations between CDO slices became far tighter than expected.
Hedge fund:
Investment vehicle that is open only to wealthy individuals or institutions such as pension funds and endowments. Hedge funds tend to use copious amounts of leverage, or borrowed money, and charge high fees, typically 2 percent of assets under management and 20 percent or more of profits. One of the first hedge funds was launched in 1949 by Alfred Winslow Jones, a reporter, who “hedged” positions by taking offsetting long and short positions in various stocks. Ed Thorp started a hedge fund in 1969 named Convertible Hedge Associates, later changed to Princeton/Newport Partners.
Law of large numbers:
The law states that the more observations one makes, the greater the certainty of prediction. Ten coin flips could produce 70 percent heads and 30 percent tails. Ten thousand coin flips are far more likely to approach 50 percent heads and 50 percent tails. Thorp used the LLN to win at blackjack and went on to use it on Wall Street. Many quant formulas are based on it.
Statistical arbitrage:
A trading strategy in which computers track the relationships between hundreds or thousands of stocks and implement trades based on those relationships. The computers look for periods when the long-term relationship breaks down and makes bets that the relationship will revert back. The strategy was first deployed by a computer programmer, Gerry Bamberger, at Morgan Stanley in the 1980s. It quickly became one of the most powerful and popular trading methods ever devised, helping launch the giant New York hedge fund D. E. Shaw and others. Peter Muller at Morgan Stanley’s Process Driven Trading was one of the most adept stat arb traders. The strategy imploded in the quant crisis of August 2007.
A cast of thousands
, it seems, helped me with this book, including a multitude of unnamed sources behind the scenes who explained the inner workings of these highly secretive investors. My agent, Shawn Coyne, helped bring the idea to life and deserves enormous credit for helping develop it. My editor, Rick Horgan, and his gifted associate editor, Julian Pavia, had a wealth of ideas that gave a healthy kickstart to the book when it needed it. Mitch Zuckoff was an ideal sounding board and provided fantastic insights into how to put the book together and make the ideas understandable. Thanks to my editors at
The Wall Street Journal
, particularly Jon Hilsenrath and Nik Deogun, who encouraged my interest in writing about this strange group of traders; and Anita Raghavan, who helped me crack open the quant group at Morgan Stanley. A virtual army of traders and professors helped me better understand the world of the quants, including Mark Spitznagel, Nassim Taleb, Paul Wilmott, Emanuel Derman, Aaron Brown, Benoit Mandelbrot, and so many others. Ed Thorp devoted far too much time to help me understand the true nature of trading and risk management, as well as his own amazing career. As promised, I’d like to thank ANONYMOUS. Mostly, I thank my wife, Eleanor, whose understanding, patience, and constant encouragement over the past few years made this book possible.