Sunday, November 13, 2005

From the archives: Does it matter?

Where Great Ideas Come From
By Matthew Heimer
2911 words
1 February 2004
(c) 2004 SmartMoney. All rights reserved.

In Nerd Nirvana, it's time for show-and-tell.

On a crisp November Friday morning in Santa Fe, about 150 Ph.D.s, executives and entrepreneurs have successfully resisted the urge to hike the high desert or shop for pueblo pottery. Instead, they've packed themselves into a darkened conference room to catch every word of a two-day marathon of presentations by researchers from across the scientific spectrum, sponsoredby the Santa Fe Institute.

The agenda reads like the academic equivalent of a Las Vegas casino buffet -- there's plenty of everything, and everything looks filling. The evolution of human languages. The containment of SARS and other infectious diseases. The cultural factors behind a civil war in El Salvador. Supply-and-demand dynamics in the beer industry. Computer networks that can spontaneously evolve to fight hackers -- a hot topic that wins its presenter, the splendidly named Englishman Robert Ghanea-Hercock, a grant to dig deeper. At breaks, the guests grill the researchers, and each other. At one table, a credit card company executive and a Honda Motors engineer gawk appreciatively as Ed, a computer jock from the Los Alamos nuclear labs, describes the software he's designed to help the Army track Saddam. "I think they started using it before it was ready," he sheepishly admits. Sure enough, the dictator eventually gets captured without help from Ed's code, but for the moment Ed's the poster boy for Santa Fe style cool.

It seems there's only one topic that doesn't get talked about in this intellectual free-for-all: the stock market.

Well, of all the nerve. Considering that a half-dozen major financial firms -- including mutual fund powerhouse Legg Mason and Credit Suisse First Boston -- are footing much of the bill for the research and keeping a sharp eye on the results, you'd think they'd at least pick up a timely stock tip for their trouble. But that's not the game in Santa Fe. "When you spend a weekend at the institute, it's a nonstop exchange of energy and intellectual stimulation," says Robert Hagstrom, a frequent visitor and the manager of Legg Mason Focus Trust. "But if you aren't a scientist, you'll come out scratching your head and asking, 'What does this tell me to do on Monday morning?'"

Still, a tiny cadre of investors have found that the head-scratching can pay off. You'll never see a book called Investing the Santa Fe Way, but SFI thinkers have challenged the assumptions of classical economics and introduced innovative ideas about how markets behave. And for the handful of managers willing to wrestle with those ideas -- most notably Bill Miller, whose Legg Mason Value Trust was on track in late December to beat the Standard & Poor's 500 for the 13th consecutive year -- the Santa Fe mind-set has become a key to success.

When it isn't hosting hectic conferences, SFI's campus has the buzz of a student union hall at a hip liberal-arts college. The institute's new headquarters sits atop a ridge just above Santa Fe's adobe downtown. There's plenty of open space for spontaneous teamwork and brainstorming: Bearded and ponytailed graduate fellows huddle around extra-wide Apple laptops at nursery-school-height coffee tables, while others throw on fleece sweatshirts and slip outside to conspire over chai and admire the view from the patios.

At any given time, about a dozen senior researchers and a dozen more grad students work at SFI, but the institute also has more than 50 "external faculty" academics doing SFI projects on their own campuses -- scattered around the globe from New Haven to New Delhi. The Business Network, the folks at the November show-and-tell, meets once a year. Fifty member companies each pay $30,000 a year to underwrite SFI research: Sponsors include everyone from Microsoft and Cisco to Procter & Gamble and defense contractor Northrop Grumman.

SFI researchers face no pressure to invent a hot gadget or meet anyone's bottom line. So what, exactly, are these people studying on corporate America's dime? A far-reaching topic called complexity theory. Writers have filled several books trying to define and describe that concept, but the magazine-friendly version goes like this: Complexity examines the ways in which individual elements in a system -- grains of sand, geese, investors -- organize themselves into more elaborate structures -- sand dunes, goose flocks, the Nasdaq. It goes on to study how the continuous process of "self-organizing" affects the ways in which the larger structures evolve. Complexity thinking also revolves around a key insight: While complex systems tend to evolve toward long-term stability, the constant interactions of their individual elements create frequent periods of volatility and disruption: avalanches, goose famines, the tech boom and bust.

By the early 1980s, a few math-savvy academics had begun to think that complexity theory could be used to understand systems in all the sciences -- biology, chemistry, physics, economics, political science, behavioral psychology. Led by George Cowan, a physicist from Los Alamos, a coalition of complexity-curious scientists launched a research lab where they could pursue this Grand Unified Theory of Everything -- and in 1984, in a few overcrowded rooms in a former Roman Catholic convent, SFI was born.

As word about complexity and SFI spread through the academic world, a few dissatisfied economists -- among them two Stanford professors, Nobel Prize winner Kenneth Arrow and technology expert W. Brian Arthur -- took particularly eager notice. These thinkers felt that classical economic theory often didn't apply well in the real world, especially when it came to understanding stock market bubbles and crashes, or big economic changes. Textbook economics assumes that markets and economies exist in a stable state of equilibrium, run by rational individuals who always act in their enlightened self-interest. But anyone who runs a business or owns a stock portfolio knows that markets are painfully chaotic, as the Internet bubble so graphically illustrates.

Complexity helped these economists. You'll never see a book called Investing the Santa Fe Way, but SFI has challenged classical economics and introduced innovative ideas about how markets work. bridge the gap between theory and reality. Arthur, who soon joined SFI's faculty full time, could talk about ideas such as "increasing returns" and "positive feedback" -- ideas that would later prove influential -- without getting laughed out of the room by colleagues. Arthur went on to develop the El Farol theory, a formula that represents the thought processes SFI staffers use to decide whether to go drinking at a popular local bar of the same name but also offers insights about how investors in a stock market choose among competing strategies.

As SFI became a sort of Worldwide Complexity Headquarters in the late 1980s, Santa Fe programmers and mathematicians worked with the professors to develop predictive computer models that tested their ideas. (In SFI circles, the phrase "I've been doing some modeling" gets you noticed for very different reasons than it would, say, in Hollywood or SoHo.) Since then Arthur, Carnegie-Mellon University economist John Miller and physicist-turned-economist J. Doyne Farmer have been among those working with "artificial stock markets" -- writing programs that know not only how to trade hypothetical stocks, but how to react to the results of those trades and then adopt new strategies to try to get an edge.

These economic experiments drew the corporate support that helped transform SFI from a shoestring operation to one with a $7 million budget today. Still, when SFI marks its 20th anniversary this summer, Wall Street will barely notice. The institute's cyber-stock markets and feats of number-crunching gymnastics have been good at identifying big-picture trends and global economic forces, but they've never delivered -- or tried to deliver -- information that would help ordinary Joes devise an investment strategy. SFI's ivory-tower orientation drives some supporters crazy. "There's an appetite there to remain very theoretically pure," says Dean LeBaron, founder of the investing firm Batterymarch Financial Management and an SFI trustee. "You ought to be able to wear the theory hat and then take it off and say, 'Let's apply this in the real world.' The data is there, and it's a disappointment that they haven't taken that route." An often repeated quote from John Maynard Smith, an evolutionary biologist who helped get complexity theory off the ground, sums up SFI's flaws: "I can spend a whole week there . . . and not hear a single fact."

Bill Miller agrees. The institute "isn't cranking out solutions for people; it's not going to change your business overnight," he concedes. But that hasn't stopped Miller, or his colleague Robert Hagstrom, from wringing some profitable insights out of the creative chaos at Santa Fe.

Miller first heard about the Santa Fe gang in the late 1980s, when he came across a newspaper article about an SFI conference that shed light on the forces behind the 1987 stock market crash. A few years later, when Legg Mason was working with Citicorp on a financing deal, Citicorp's then-CEO John Reed sang the institute's praises -- Reed, now the New York Stock Exchange interim chairman, was one of SFI's biggest corporate backers at the time.

Reed introduced Miller to SFI insiders who invited him to the institute. For Miller, a man known for the breadth of his intellectual background -- he was an Army intelligence officer and a philosophy Ph.D. student before joining the finance world -- it was platonic love at first sight. "These were truly minds of the highest caliber, and they dealt with the most difficult and exciting scientific questions," says Miller, who today sits on SFI's board of trustees.

Miller soon discovered that SFI's thinkers, and Brian Arthur in particular, had the tools to help him grapple with the 1990s' biggest investing challenge: the tech sector. Arthur, a wry, mild-mannered native of Northern Ireland, had observed that a single format or a single company often dominated areas of technological innovation, in a phenomenon known as "increasing returns." "Say it's 1975, and people are buying video recorders in two formats, VHS and Betamax," Arthur explains. "Beta works a bit better technically. But if VHS sells a little better at the outset, then stores are more likely to carry VHS [tapes], which means the consumer's more likely to buy VHS players. This spiral grows wider, and soon one bandwagon is far ahead of the other." The same pattern can emerge in information technologies, Arthur notes: "Think of Quicken in financial software, or Microsoft in operating systems."

For Miller, a value-oriented investor, the ideas were a revelation. "The general consensus had been that technology trades rich compared with other stocks, and that it's too highly volatile," he notes. "But Brian's work opened our eyes to the fact that that isn't true." Confident that companies achieving dominance in technology could count on a stable market share, Miller bought PC maker Dell and cell phone giant Nokia early in the 1990s and held them even at prices that doctrinaire value managers said were unreasonably high.

Miller also gained insights about tech from another hot SFI topic: network effects. Several Santa Fe projects have noted that in all kinds of biological, physical and social systems, networks -- once formed -- are hard to disrupt. Later research, by sociologist Duncan Watts, showed that Internet-oriented companies -- the ultimate in business networks -- tended to settle into stable patterns with regard to market share. The Santa Fe research helped Miller ride America Online to big gains during the 1990s; today he sees Watts's ideas in play for "We believe it's going to be difficult for someone like Wal-Mart or Barnes & Noble to knock them off," Miller says. "Whatever risks they face, one of them [isn't] competition."

SFI's insights on networks have also inspired Hagstrom. Some of his tech holdings in Legg Mason Focus Trust, including eBay, Yahoo and InterActiveCorp -- the parent company of Expedia, Ticketmaster and -- are plays on the strength of networks. "You may have people yelling at you because they're [trading at] 50 times earnings," Hagstrom says. "But there's huge value in these networks, and they're not easily displaced. Had I not been to Santa Fe, I might not have appreciated that."

For investors still smarting over their losses in the last tech crash, Hagstrom's and Miller's strategies may sound a bit like déjà vu. But just as some SFI ideas have encouraged Hagstrom and Miller to take risks, other research has helped them know when to be cautious. When tech-stock prices became wildly inflated in 1999, Miller recognized a positive feedback loop -- investors buying stocks for no other reason than because everyone else was. "Bill came to a class I was teaching in New York in March of 2000 and said, 'It's over for technology,'" recalls Michael Mauboussin, chief U.S. investment strategist at Credit Suisse First Boston. "There was a pretty uniform optimism, a euphoria. When you see that, you need to head the other way." Miller's instincts led him to exit from AOL and other tech stocks before incurring major losses. Hagstrom also got out, and his network insights have helped to fuel a stunning 2003: Legg Mason Focus was up 55 percent for the year through Dec. 16.

Miller invited Mauboussin to join the Santa Fe circle in the mid-1990s, over hot dogs at a Baltimore Orioles game; today Mauboussin writes broader think pieces about complexity and the economy in a CSFB newsletter, the Consilient Observer. He's quick to point out that SFI's work has implications far beyond tech investing. "Complexity shows that the aggregate behavior of investors -- the long-term effect of how they interact -- is much more important than the behavior of individuals," Mauboussin says. As a result, he says, investors can spend less time dwelling on what he calls "short-term chatter," anything from an ugly earnings report to a mini-crisis like last summer's Northeast blackout, and focus on the big picture.

Such thinking requires flexibility. "One lesson I've drawn from Santa Fe is that there's no single, algorithmic way to outperform, because the market learns and adapts so quickly," says Hagstrom. To do well year after year, like a Miller or a Warren Buffett, managers need to continually readjust their strategies -- not unlike the programs in Santa Fe's artificial stock markets. "It's not a methodology that a lot of money managers use, but in the biological world, that's how you survive -- that's essentially what evolution is."

Above all, Hagstrom, Miller and Mauboussin follow the goings-on at SFI in search of new analogies, new metaphors to apply to the markets. The newest idea on their radar screens: biological networks, as studied by Geoff West, a Los Alamos physicist and SFI member. At a recent talk, Miller recalls, "West held up a picture of the human circulatory system, then of the Mississippi River delta, then of the Santa Fe road network, then of a corporate organization chart. They all seemed to have common features. The question is, why?" West and other thinkers believe that the organizational structure of most systems can be a major factor in determining how large that system can get -- an idea with big implications for the way investors analyze very large companies. "How big can companies grow?" Miller wonders. "And once you're at the top, can you keep growing?" And if you can't, should investors leave you behind?

If answering such questions sounds like nailing Jell-O to a wall, Miller and Hagstrom are ready to hammer. Both managers rarely hold more than 40 stocks at once, so they have the time to do more abstract analysis. And they can be pretty confident that their competitors won't be looking at things the same way. Asked whether his and Hagstrom's track records might tempt more money managers to join the SFI fold, Miller chuckles. "From the institute's point of view, that would be great. Selfishly, I hope not."

Of course, SFI could probably live without any attention from Wall Street, which is part of its charm. At an end-of-conference reception -- featuring lasagna, red-checkered tablecloths and cheap red wine -- the heroes of the weekend are Ghanea-Hercock, the computer-evolution guy, and Lauren Ancel Meyers, a young medical doctor using network theory to fight influenza and SARS. The guests are so eager to bounce ideas off each other that they pay only passing notice to the scientific marvel outside -- a total lunar eclipse in the clear desert night.

Afterward, Garrett Thornburg, CEO of a boutique mutual fund company, gushes about the weekend. His Santa Fe based firm, which manages around $8 billion in assets, is about to move into a new office, and contacts he's made at SFI have given him ideas about how to design the workspace and protect his trading network from hackers. Speaking of trading, has Thornburg followed SFI's economic research? "A little bit." Has it affected his investment strategy? "To be honest, that hasn't even occurred to me." Somewhere, Bill Miller smiles.


Post a Comment

<< Home