Blog

From the International Herald Tribune circa February 1, 1992 comes this article “The New High-Tech Investing: Computer as Fund Manager” By Philip Crawford:

The specter of a computer-dominated society has long been the stuff of science fiction, conjuring chilling scenarios of a world that values only efficiency. Welcome to the cutting edge of investment management. The use of computers to automatically trigger moves in a variety of markets is growing, portfolio managers and traders say. Although they acknowledge that the true stars of the industry outperform computers today, they contend that the days of homo sapiens as the highest form of investment strategist are numbered. “Put it this way,” said David Harding, research director of Adam, Harding, & Lueck, a London-based commodity trading advisor, “I don’t expect the world chess champion to be a human being in the year 2000, and I don’t expect the world’s top investment manager to be one either. We’re already at the point where the best computer systems are better than the average manager.” Why the increase in computer-managed portfolios? Investment company executives cite the seemingly age-old reasoning that first struck fear into automobile workers a generation ago when assembly-line robotics initially appeared: that a well-programmed machine performs more consistently, and less expensively in the long run, than a human being. “The computer is not subject to human frailties, other than its design and how it’s programmed,” said Michael Quenington, European manager of E.D. & F. Man International, a London investment firm. “A person can have a huge row with his wife one morning, crash his car, or have any number of things happen that can affect his emotions and judgment. And all the information a person has in his head might rest only with him. A person dies, but a computer just keeps churning out numbers.” Mr. Quenington cited a program recently devised by his firm that, given a revolving input of 100 highly liquid equities chosen by staff researchers, electronically monitors the stock market and produces profit-maximizing decisions on how much of each stock to buy or sell, and when. He declined to elaborate on which market characteristics prompted the program to make its choices, saying only that it employs a mathematical model, and is performing well. Using computers to signal an opportune moment to trade large blocks of equities is, of course, nothing new. Program trading, which involves taking advantage of discrepancies in prices between stock-index futures and the underlying stocks, came into vogue during the 1980s bull market and is sharply on the rise internationally, according to a recent study by Greenwich Associates, a U.S. research group. The practice, also known as stock index arbitrage, can cause huge swings in the market as electronic “buy-sell” programs are tripped off when prices reach a certain level. Many securities industry professionals have maintained that program trading contributed to the crash of ’87. But the search for greater returns has led to the evolution of ever more sophisticated software that monitors developments in numerous markets simultaneously, taking trading action when a list of programmed criteria are met. “We devised a set of hypothetical rules, such as ‘buy this product if the price goes over the 60-day moving average,’ or something like that,” said Peter Matthews, chief portfolio strategist for New Jersey-based Mint Investment Management Co., in describing the development of his firm’s in-house software. “Then we took the rules and applied them to the behavior of markets going 20 to 30 years back. Once you find out what works historically, then you can write a program that goes forward.” Mr. Matthews said Mint’s programs determine when the firm, which deals primarily in commodities and foreign exchange markets, should sell soybeans, for example, or buy gold. “Particularly in futures,” he said, “things move so fast that when things start to go badly, you tend to panic. The computer is totally unemotional – it never panics. It really improves consistency. You can sleep more soundly at night knowing it’s there.” Mr. Harding, acknowledging that total reliance on computers might be unsettling to some, maintained that the advantages of human perceptional abilities are largely outweighed by the computer’s capacity to process staggering amounts of information instantaneously. The use of programs based on historical trading patterns, he added, “actually helps mesh the world of the computer with those of the researcher and trader, rather than polarize them. If someone has an idea or theory he wants to test, it takes only 20 minutes, not 20 years, to simulate it.” But despite apparently impressive performances by 100 percent-computerized firms such as Mint and AHL – both claim compound annual rates of return in excess of 23 percent over the past eight years, compared with 12.4 percent for the S&P 500 – not everybody is jumping on the bandwagon. “We don’t do this type of computer program-driven trading,” said Barry Holman, a quantitative analyst at Legal & General Investment Management in London, “because we haven’t found (software) which we feel is reliable enough. But there’s no doubt that some computers are a real challenge to fund managers. There’s a lot of market knowledge out there that never gets acted on, because people don’t always do things they intend to.” Computers cannot, of course, be programmed to predict market reaction to random events, or to feel fear when perhaps fear is called for. Some say that a harbinger of a doomed race might be the creation of a computer that could get angry or fall in love. And machines obviously need humans to feed them data in the first place. But in the investment world, well-programmed computers that figure out what to do, and then do it, are the wave of the future, say many market pundits. The next generation of machines, moreover, may not attempt to emulate the human mind better, but try to incorporate it. “We’re seeing research on machines which would theoretically interface with a human mind, in which a person can be totally immersed, by being placed inside an electronic body-suit, or some similar concept,” Mr. Harding said. “It’s very sci-fi, I know, but maybe it really is possible to get the best of both worlds.”

17 years ago this was printed? Yes. 17 years ago. Now 26 years!

Related Content

Technicals Matter

Trend Following with Seykota and Harding

An Agenda

Don’t Worship the Computer

Tim Larkin Interview