"I was deeply impressed by this book"--John le Carré
The computer revolution is speeding from Silicon Valley to Wall Street, and if David Shaw gets his wish, many of the fat-cat employees at America's major stock exchanges will soon be out of work. "Finance is a pure information processing game," says the Stanford-educated PhD and expert in artificial intelligence and computer design. "A lot of people in the business do things that should be done by computers."
Cold-blooded? Not to the 45-year-old Shaw, who launched the D. E. Shaw & Co. investment bank into the financial stratosphere by using advanced mathematical and computational techniques to play the markets. Almost everywhere Shaw looks into the gray canyons of American finance he sees a mess in need of straightening out. "If computer scientists designed the world, things would be different," he says.
If Shaw sounds smug, consider his remarkable success at turning bytes into bucks. Eight years after it was founded, his investment bank has 400 employees in offices scattered around the globe and US$800 million in capital, with annual returns averaging more than 20 percent. The firm wields a $100 million technological toolbox of secret algorithms, and on a busy day it accounts for 5 percent of the total shares traded on the New York Stock Exchange. Shaw is an adviser to Bill Clinton on computer technology, and he has wired a lot of people into cyberspace through Juno Online Services, which provides free email for about 700,000 subscribers.
That's just the beginning. Over the next year, Shaw will launch FarSight, an audacious online financial service. FarSight will bundle automated stock trading, online checking accounts, financial data, portfolio management tools, credit card accounts, and access to ATM machines into a one-stop financial supermarket. Potential customers run the gamut from savvy investors who need point-and-click stock trading to grandparents fine-tuning their retirement accounts. "Retail customers will get the same services as institutional investors," says Harold Rhodes, FarSight's marketing director. "Why should the big guys get all the goodies?"
Shaw ultimately plans to bundle into FarSight everything having to do with money - from offering insurance to mortgaging your house. It is a package meant to redesign the financial world around a D. E. Shaw interface. Sound familiar? But if we're talking about the Bill Gates of computer finance, why have you never heard of David Shaw?
You will. FarSight, along with Shaw's other financial ventures, which have already grabbed a lot of business from Wall Street's traditional players, is beginning to scare the hell out of the financial services industry. "Technology isn't just changing the tools we work with; it is changing the nature of the work itself," said Morgan Stanley managing director Elaine La Roche in a speech at the Public Securities Association's annual convention in April. "It allows end users to bypass the middleman. It allows them to bypass us."
The market for online investing and banking potentially represents a large chunk of change. In the next five years, the number of online investing accounts will grow from 1.5 million to 10 million, according to Forrester Research Inc. in Cambridge, Massachusetts. Online banking accounts are estimated to grow from 1.1 million today to 19.5 million by 2001. "The potential shift is huge," says Julio Gňmez, a senior analyst at Forrester Research. "The Internet investor will be the cream of the crop, the high net worth individual you have to get if you want to succeed in this business."
Shaw not only wants to succeed, he wants to redesign the financial world from the ground up. "We're trying to find areas that are going to be fundamentally transformed by computers," he says. "Not just made slightly more efficient, but where the whole industry is going to be turned on its head."
Money as fluid flow
In my first telephone conversation with the notoriously secretive Shaw, he sounds more friendly than sinister. He apologizes for having a tight schedule. The night before, he met with Clinton, for whom he has promised a report on computers in the schools. The president wants to spend $2 billion to wire every American classroom to the Internet by 2000. Shaw's report - a blueprint for getting the job done - is late, and he has promised Clinton he'll finish it before leaving on vacation the following week.
"Buckle down and do your assignment," I tell him. "Deadlines are deadlines." Then I casually suggest that if we can't meet before then, perhaps he should take me with him on vacation. To my surprise, Shaw jumps at the opportunity. "I'm not good at vacations," he says.
Shaw is so bad at vacations that he doesn't know where he's going. So I double back to his assistant. She gives me the phone number of a health spa in the Green Mountains of Vermont. Shaw, his wife, and two other couples are driving up from New York the following Monday. I'll be there for lunch.
While booking myself into Shaw's party, I learn from the spa's receptionist that I'm headed to a "camp for adults," a "wellness experience that's guaranteed to destress me and jump-start my fitness program." Not knowing whether I'm in for a powder-puff vacation or a rugged wilderness experience, I pack my car with everything from hiking boots and pitons to yoga mats and beach towels.
On a gorgeous late-summer day, I drive from my house on the upper reaches of the Mohawk River across the Adirondack Mountains into Vermont. I am sitting in the spa dining room when in walks Shaw, a large man with a high-domed forehead. He is accompanied by his wife, a smooth-faced young woman who works as a journalist, and two other couples, a chemistry professor married to a doctor and a lawyer married to a law professor. Shaw is dressed in a black- and gray-checked collarless shirt, black jeans, black socks and shoes. Nursing a bad back, he has spent the morning in his room.
"This is the first time I've ever hiked on the ground, instead of a treadmill," says Shaw's wife when recalling the morning's activities. She gazes onto a forested mountainside. "How do you ski without running into the trees?" she asks. Someone points to a nearby ski trail. I realize I won't be needing my hiking boots and pitons.
The spa director, a roly-poly elf in tennis shoes, comes around to schedule our afternoon massages. Then the lunchtime conversation turns to ideas, which is what Shaw and his friends really care about.
"I never understood the point of vacations," he later confides. "I do all the things I normally do, except with fewer resources." Shaw's friends are meant to help with the resource problem.
His chemist friend launches into a minilecture on the difference between classical mechanics and quantum mechanics and how the latter is so counterintuitive that not even Einstein believed it. The argument essentially boils down to Schrödinger's cat, which simultaneously exists and does not exist - and we don't mean on average, or at one point in time or another, but at this instant right now the cat is both dead and alive - which Einstein had a hard time swallowing. But not Schrödinger, who developed the idea while he and his mistress were summering in the Alps.
"That must have been nice for his mistress," quips Shaw's wife. Everyone laughs.
Shaw and I take this as our cue and head up to my room. I hope the spa director doesn't catch me "re-stressing" his client.
Born in 1951, Shaw grew up in Los Angeles. His stepfather, Irving Pfeffer, was professor of finance at UCLA. Pfeffer published papers supporting the efficient market hypothesis,a keystone to modern financial theory, which claims that the stock market is so well equilibrated that no one can beat it. Shaw would later make a lot of money by discovering the "inefficiencies," the little pockets of predictability, whose existence Pfeffer questioned.
Shaw's natural father (his parents divorced when he was 12) was a theoretical physicist who studied plasma and fluid flows. His mother is an artist and educator. Shaw is a perfect mix of these three inherited influences. His career exemplifies the melding of physics and finance - a new way of looking at the world that might be called "phynance." Phynance studies money as a kind of fluid flow. It searches for little eddies of predictability in the great waves of speculative investment that wash daily around the globe. Like many of its best practitioners, Shaw practices phynance with the intuition of an artist, and he prefers to live in environments that he has designed.
Planning to study marine biology, Shaw went to UC San Diego, where he was surprised to learn that the subject was offered only to graduate students. Instead, he signed up for cognitive psychology, which turned him on to all the big questions that still intrigue him: How do people work? Will we become obsolete? Will we become immortal?
Shaw became fascinated with information theory and mathematical models for pattern recognition. He also discovered neural networks - computer programs designed to mimic the parallel processing capabilities of the human mind.
"I was just amazed," says Shaw of these very simple systems that can perform exceedingly complex computations. "I was intrigued by neural networks that could automatically learn a pattern. I was very excited about finding these magical devices that, once you discovered the right design, you could turn on and they would get smarter." D. E. Shaw, the investment bank, can itself be conceived as a neural network, a $100 million learning device that just keeps getting smarter.
Disproving the traditionalists
Shaw continued exploring the ramifications of artificial intelligence after he began graduate school in computer science at Stanford in 1973. Among Shaw's colleagues in this famous and talented class were future company founders Len Bosack (Cisco Systems Inc.), Andy Bechtolsheim (Sun Microsystems Inc.),
Jerry Kaplan (Go), and Jim Clark (Silicon Graphics, Netscape Communications Corp.).
Until then, Shaw had little direct experience with computers. But in a year or two he had caught up to his classmates and begun pushing into software consulting and computer design. A computer basically consists of a central processing unit attached to memory. You can jazz up the CPU or expand the memory, but the connection between the two will always be limited by something called the von Neumann bottleneck. This bottleneck is caused by the fact that a standard computer can transfer only one chunk of data at a time between its CPU and memory.
What if we threw out the ground rules and designed a completely new type of computer? Shaw wondered.
To overcome the limitations of conventional computer design, Shaw created a new, non-von Neumann computer (called NON-VON, for short), which was loosely based on a different model - the human brain. The blueprint of the NON-VON computer looked like a tree leafed out with thousands of tiny computer chips. Each of these leaf-chips contained several processors,and more processors were attached where the branches met the trunk.
The computer had an ingenious design, similar in many ways to what Danny Hillis later built at Thinking Machines. The machine was incredibly fast. It was also incredibly expensive. "We needed $30 million to get going and $100 million to break even," Shaw said. Planning to start his own supercomputer company, he began shopping the idea to venture capitalists, some of whom would laugh out loud when shown the price tag. "That's not a good sign," Shaw remarks wryly.
In 1980, he snagged a job as an assistant professor of computer science at Columbia University in New York, where he began building a NON-VON prototype. In the course of looking for investors, Shaw became known as a local computer whiz good at running big projects. These skills were in short supply on Wall Street in the mid-1980s, where the big banks and investment firms were desperate to junk their IBM mainframes and boost their computer power with desktop workstations.
Head hunters started calling, and Shaw used their introductions to keep pitching his $100 million idea for a new computer. He never intended to leave academia, until Morgan Stanley & Co. put an extra zero on his annual salary and created an offer too good to resist.
Morgan Stanley wanted Shaw to be part of an independent, top secret group that was being formed to exploit anomalies in stock market prices. These little pockets of predictability disproved the efficient market hypothesis and opened the way to beating the market. Shaw would put together the group's computer facilities.
Shaw - then a longhaired, bearded, ex-surf musician from Southern California - had some qualms about working for a white-shoe Wall Street firm, but curiosity got the better of him. "What appealed to me was the challenge of trying to beat the market," he says. "I was raised to believe it was impossible, and here they were telling me they knew how to do it."
Morgan Stanley was exploiting a technique called pairs trading, which is based on the idea that prices of related stocks should be correlated. Ford and General Motors, for instance, tend to fluctuate in price around the same news events. But what if an unusual gap - in which Ford lags in price and GM pushes ahead - develops?
A hedge fund might rush in to sell GM and buy Ford. If the gap persists and the entire stock market goes up or down, the hedge fund will neither win nor lose. Such strategies, in Wall Street lingo, are called "market neutral." But if the gap between GM and Ford narrows, as predicted by statistical models, then the fund will make money.
The trick is to get your strategy truly market neutral, not only for fluctuations in the stock market, but also for fluctuations in interest rates, foreign exchange movements, and global economic risks that can come at you faster than a shark lunging forred meat. Wall Street is filled with former employees of failed hedge funds.
"This is a high-risk game at the frontier of mathematics, physics, and finance," says Andrew Lo, director of the financial engineering program at MIT. "Many correlations exist among securities at different points in time, and solving the optimization problem that implements a successful market strategy involves the same tools that we use for guided missiles. It's stochastic control theory. You're trying to hit a moving target."
Shaw treated Morgan Stanley like a college campus with a terrific smorgasbord of courses to be sampled. He was acting the part of a New York banker with such gusto that he got cast for the role in a movie. He was walking home one night at 2 a.m., after a long day at Morgan Stanley, when he found his street roped off by a film crew and one of the buildings on his block miraculously transformed into a high-class nightclub. Shaw joined what turned out to be a crowd of extras standing near his apartment. Then the casting director came over and singled him out for a walk-on role in the Tom Hanks film Punchline.
"Great suit. Perfect for the part," said the casting director, draping a woman on Shaw's arm and telling him she was his wife.
Shaw spent the rest of the night shooting one scene over and over again. Hanks is watching their taxi arrive at the club. Out jump Shaw and his wife.
"Sir, did you tip the cabbie?" inquires Hanks. "You cheap bastard, tip the cabbie!" he yells at the well-dressed, fast-moving Shaw.
When they broke at dawn for breakfast and Hanks discovered he had been playing his scene with a real investment banker instead of an actor, he cracked up laughing. The director, on the other hand, went apoplectic. Shaw's scene was cut from the movie, probably for want of a union card (which Shaw didn't get until later, when he played another bit partin Dangerous Love with Elliott Gould).
Shaw poked his nose into every aspect of Morgan Stanley's business, pestering colleagues with questions and suggestions on how they could do their jobs better.
"He is personable, but intense," says a former colleague. "He drills down on questions and won't let them go. This is great, if you're willing to die for the answer. But if all you want to do is get on with business, it can be a pain."
Shaw is more blunt: "I got on people's nerves."
RIP: tassel-loafered leeches
After a year and a half at Morgan Stanley, Shaw struck out on his own. He founded D. E. Shaw & Co. in 1988 in a loft over a communist bookstore near Greenwich Village. The company had six employees and $28 million in capital invested by Donald Sussman and three of Sussman's friends. Sussman runs Paloma Partners, a $1 billion fund-of-funds in Greenwich, Connecticut. He's a high net worth individual, as they say on Wall Street, and a tough cookie. Sussman on the telephone sounds as if he's growling out answers to my questions while simultaneously reading The Wall Street Journal, watching his Bloomberg box, and trading Euroyen futures in Zürich.
"Why did you decide to invest in David Shaw?"
"He's the smartest person I've ever met. I'm lightweight compared with him."
"But you're richer than he is."
"So what if his fund is now smaller than mine? By the end of our lifetimes his net worth will vastly exceed mine."
During Shaw's education at Morgan Stanley, he decided pairs trading was not going to be the way he would make his fortune. Instead, Shaw's secret is algorithmic trading, a complex financial juggling act that exploits tiny price differences between multiple international markets. Maybe the stock of Sun Microsystems is selling for $50 in New York and $50.50 in Hong Kong. This is called a market "inefficiency." So you buy low in New York and sell high in Hong Kong and pocket the difference.
Such discrepancies don't last long.
With computers and communication links spreading market information everywhere in nanoseconds, these price gaps tend to close almost as fast as they open. You also have to juggle currency exchange rates and transaction costs. To play this game, you have to be quick and surefooted.
Which Shaw always seems to be. He makes his bets with the aid of mathematical models whose exact nature is more closely guarded than the recipe for Coca-Cola. Until April 1996, when Juno was launched and Shaw had no choice about developing a public persona, he rarely let visitors inside his company. The few details that leaked out were a bit scary: Shaw was a control freak who kept his own traders in the dark about how the firm's computer models worked.
"We follow the same principles as the CIA or NSA," Shaw says of company security. "Information is partitioned off and distributed on a need-to-know basis." When asked if he used neural networks to design his trading strategies, Shaw said, "I could tell you, but then I'd have to kill you afterward." He was joking, I think.
Once Shaw got his hedge fund running, he realized he had enough computational models and machines to branch into other businesses. He moved into basket trading, where you buy billion-dollar stock portfolios from big investors who want to unwind their positions in a hurry. He also moved into something called the "third market." This is automated, off-exchange trading of listed stocks. Volume on the third market is already 15 percent as large as that on the New York Stock Exchange, and Shaw is well positioned to get a big chunk of this business.
Stock trading for Shaw is nothing more than a network in which computers match buyers and sellers and store unfilled orders until the next customer dials in. Order flow is key.
But unlike market makers on the major stock exchanges, who charge money for the finger-wagging and yelling required to fill your order, D. E. Shaw pays money, between one and two cents a share, to brokers who send them their business. (They make their money not on commissions, but on the spread, the difference in price between buy and sell orders.) You now know why the cost of stock trading at your local cut-rate brokerage firm is falling through the floor.
"I find a lot of finance highly amusing," says Shaw. "There's a lot of hocus-pocus practiced by people with fancy suits. Quite often they're selling you financial products that are terrible deals or providing very mechanical services at inflated prices."
One such group of fancy suits are the market makers, specialists, and brokers who run America's stock exchanges. "Many of them are doing something that should be done by computers," he says. "It's just not very complicated. All you have to do is match a buyer and a seller. In serving this middleman function, you collect exorbitant amounts of money for a very simple process."
In Shaw's view, once computers have replaced these tassel-loafered leeches, a new financial era will dawn. Shaw calls this the golden age of "disintermediation," because it involves "pulling intermediaries out of the loop and letting customers get closer to each other."
He envisions the possibility of an Internet exchange in which market makers no longer pocket hefty fees for matching buyers and sellers. No longer will investment bankers suck up 6 percent of the capital raised when companies sell stock during initial public offerings. "In the long run, people shouldn't have to pay us or other Wall Street firms as much as they do for many of the services we provide," says Shaw.
Shaw admits that his visionary ideas, if carried far enough, might put him out of business, at least the market-making business. But he calls this "a historical inevitability."
Shaw occasionally assembles his staff and exhorts them to think about the financial future: Internet markets, disintermediation, crossing networks that allow people to trade with each other directly. "What we are doing now is not the end product," he declares. "Let's try to imagine what things will look like 10 years from now. Is there some way we can leapfrog into the future, without losing our shirt in the process?"
In recent years, Shaw has become increasingly bold in his investments and has moved into venture capital. He has put money into two companies researching new ways to design drugs. Shaw thinks the old method of drug design, by a laborious process of trial and error, is as outdated as traditional stock trading. "It may be my computer science prejudice," he says, "but as soon as you learn the rules of the game, you realize the right way to do this is mathematically, as opposed to grinding and mixing things in test tubes."
Shaw's preference for redesigning the world from the ground up has been applied to his Internet activities as well, including Juno. Although the company has been weak on the revenue side (advertisers are just beginning to climb on board), Shaw says he is carrying on an "incredible program of experimentation. We watch over the shoulders of paid experimental subjects by trapping their inputs and looking at all the places where they make mistakes," he says when describing the process of designing the system. "Then we experiment with moving buttons around or changing their colors, which results in a dramatic drop in errors."
Good design is a critical factor for Shaw. Another environment bearing his signature is D. E. Shaw's headquarters, which moved in 1992 to the top two floors of a skyscraper near Times Square. The D. E. Shaw company logo is inspired by an electronic switch, and its office, with its shiny black floors and cutout walls bathed in reflected light, is meant to evoke the feeling of sitting inside a computer chip.
D. E. Shaw's stock trading room is a gear-filled, black hexagonal chamber that makes the Challenger space capsule look antiquated. Shaw sits at a brushed-aluminum, wing-shaped table ergonomically designed so that his neck swivels no more than 15 degrees between computer monitor and desk surface.
The icy atmosphere belies Shaw's populist sensibilities. In this cynical age, I may be faulted for believing, but I think it true, that Shaw has other motives besides money for offering the world free email. His political work for Clinton has convinced him more than ever that there is a growing gap in the United States between the digerati and the computationally illiterate - a gap that ultimately threatens to cripple the country's productivity.
While Shaw is providing free email to everyone who wants it, including inner-city residents and hard-to-wire rural customers, he is simultaneously pushing the larger political agenda that will get every classroom in the United States wired into the Internet.
"From a business viewpoint, it's not ideal to offer free email to subscribers who aren't demographically appealing to advertisers," he says. "But even if we don't make any money on these subscribers, I think for political and social reasons it has to be done.
"I actually believe that technology, including email, can be harmful if it serves to drive a wedge between information haves and have-nots," says Shaw. "Even if we get parity of computer access within the schools, we're still seeing an incredible gulf in terms of home access to computers, and this gulf is now beginning to show up educationally."
Juno is free in the sense that network television is free. In exchange for the connection, a subscriber gets targeted by advertising, and soon subscribers will get targeted by D. E. Shaw, which wants to lure them into using its online financial service, FarSight.
Despite FarSight's lofty ambitions, it needs to clear some formidable hurdles. "D. E. Shaw has little experience dealing with consumers or partnering with financial behemoths," says Forrester's Gňmez. "And it also remains to be seen how a company whose stock-in-trade is secrecy will interface with online investors."
Fidelity Investments, E*TRADE, and Charles Schwab are already selling stock online, while bankers from Citibank to American Express are creating "thin branches" as fast as they can. But the World Wide Web may, in the end, prove no match for Wall Street's very human, and very hard-selling brokers. The trick to the success of FarSight, like Shaw's other ventures, will be extreme integration, efficiency, and ease of use. "For all we know, they might be able to come in and reinvent this thing from scratch," says Gómez.
Visitor from parallel universe
We have spent the day talking, and Shaw looks as if he could go all night. But the dinner bell interrupts.
We walk into the dining room to find Shaw's party already seated. Famished, I plow into my lemon-spiced, undercooked broccoli, until I am brought up short by the comments of Shaw and his friends, who find the spa cuisine ludicrously inadequate. A guy who tolerates zero human error in his stock trading wants zero fat in his dinner. The table is stripped bare of salt, sugar, butter, wine, coffee. But what's that dry hunk of sesame chicken doing under my broccoli? Do I know how much saturated fat there is in a piece of chicken? Soybean burgers and soybean hotdogs, that's what I should be eating!
"Don't forget the red wine," I remark. "You know, for keeping down your cholesterol."
My comment provokes Shaw into a lecture on cholesterol, in which we learn about the latest medical findings concerning the difference between lipoprotein, a form of LDL, or bad cholesterol - which gloms onto blood vessels and for which there is no known treatment, except, possibly, vitamin C - and HDL, which is actually a good kind of cholesterol responsible for reducing the risk of heart attacks.
"Everyone in the family consults him for medical advice," Shaw's wife tells me, and I begin to understand why some of his venture capital investments center around drug research.
As Shaw bears down on his topic, giving us the latest Journal of the American Medical Association findings on the subject, I get the sneaking suspicion I'm sitting next to someone who looks like a human being, and who plays the role quite convincingly, but is actually a visitor from a parallel universe.
Everywhere Shaw gazes he sees that we humans have made a mess of things. With wonderful good humor and efficiency, he is offering to help straighten it out with a wave of his computational wand. He wouldn't want to be president of the United States, Shaw told me. The pace of government is too slow, and it involves too many compromises. But I get the idea he wouldn't mind being president of everything else.
The following morning, after eating a bowl of wheat-berry gruel, I make my goodbyes and drive out of town, heading for the most high-fat, megacaloric, wine-drenched lunch I can find. On the way I keep thinking about money and the way it presents itself to people like Shaw. It's an amusing game. Money is really just a puzzle, a way of keeping score. It's revelatory about people's personalities and preferences. But in the end, it's just another form of information, which can be reduced to bits and bytes and zapped around the world at the speed of light. If one sits in the middle of this information, processing it, routing it rapidly from point A to point B, the view from this central vantage point is ever changing and always fascinating.
"Sometimes I think I'm going to wake up and discover there's been a clerical error - that we're really losing millions of dollars instead of making them," says Shaw. "Then I pinch myself and get back to reality."
Shaw describes the various people who have offered to buy him out. "'We'll never sell the firm,' I tell them. 'That's not in the cards. It doesn't matter what the price is.'
"'Oh, come off it,' they say. 'Everyone has a price. Would you sell for $5 billion?'
"'No,' I tell them. 'My lifestyle wouldn't change if I had more money. This is the most fun I've ever had.'"