CLEARWATER, Fla. - LBS Capital Management Inc., a firm well respected in the world of artificial intelligence but unknown to the average investor, is taking on mutual fund giant Fidelity Investments.
LBS has been named subadviser of a new mutual fund sponsored by the Quantitative Group of Funds, Lincoln, Mass., that will invest in stocks through neural networks and other techniques based on artificial intelligence. The only other mutual funds believed to use such techniques are funds sponsored by Fidelity and managed by Brad Lewis and Greg Fraser.
LBS' first mutual fund assignment comes as the firm's institutional business is growing rapidly. Since Jan. 1, assets under management have doubled to $600 million from 13 institutional clients as well as individuals with a minimum investment of $100,000. Half the money came from new accounts from the pension fund of John Deere & Co., Moline, Ill., and an international bank.
Dean Barr, chief investment officer of the firm, expects assets under management to reach $900 million by year end, based on a number of potential new accounts.
LBS and the handful of other investors using artificial intelligence believe price behavior of securities and markets is predictable and non-linear, meaning the same size cause can have different-sized effects. While the average human being can only analyze a handful of non-linear "events" at once, AI-based computerized systems - incorporating neural networks, expert systems and/or genetic algorithms - can analyze a virtually unlimited number of events in a matter of minutes.
"Managers have so much information coming at them at once. How do you harness it? We're sticking a nozzle on a fire hydrant and letting a huge amount flow through," Mr. Barr said.
Artificial intelligence techniques are not a black box. For instance, a neural network, a mathematically-based software model, can recognize patterns and statistical aberrations in financial data, in an effort to find excess returns. It is able to process new information and adapt it to new input. Thus, it "learns" from its mistakes and readjusts to new parameters of the market.
"Its computational power overwhelms any traditional research method," said Don Waechter, director of marketing.
A technology that has its roots in biology applications and was used by the Defense Department in the 1940s, AI is working its way into the world of high finance. Academic research on artificial intelligence, some of it sponsored by LBS, is widespread. Big brokerage firms are using it for forecasting volatility of their security selections; and in the derivatives area, institutions are using it to analyze risk of counterparties in transactions.
Mr. Barr predicts that eventually the cutting-edge technology will become ingrained in the investment management world. Managers will be divided between gifted intuitive stock pickers - like former Fidelity Magellan Fund manager Peter Lynch - and users of artificial intelligence technology.
"Going forward there will be two schools of money management. The middle ground will disappear," Mr. Barr said.
"The technology is rapidly gaining acceptance with consultants and plan sponsors," he added.
LBS' performance on middle-capitalization stock portfolios, which represent the bulk of its assets, has handily beat the Standard & Poor's 400 Stock Index. In 1992, the firm returned 21.3% vs. 5.71% for the index; in 1993, the firm returned 16.6% vs. 9.05% for the S&P 400; in 1994, its year-to-date performance through Aug. 31 was 4% vs. 0.85% for the index.
Most of LBS' assets are in midcap stocks, but it also runs TAA and global asset allocation portfolios and plans to develop an international stock-picking capability. The firm's existing global asset allocation portfolios invest in index funds or country funds.
Despite LBS' lack of experience in small cap, the Quantitative Group decided to launch a small-cap fund because it soon plans to close an $80 million small-cap fund subadvised by John Bogle Jr.'s firm, Numeric Investors L.P., Cambridge, Mass., when assets reach $100 million.
"Artificial intelligence is definitely the wave of the future in financial services," said Lynn Wickwire, senior vice president of the Quantitative Group.
The new fund will have two classes of shares, including a no-load share for institutions. The retail shares will have a 1% redemption fee and a 12(b)1 fee of 0.50%.
Brad Lewis' three funds at Fidelity are: the $1.039 billion Disciplined Equity, a no-load fund, rated five stars by Morningstar Inc., the firm's highest rating; the $825.6 million Stock Selector, a four star-rated no-load fund; and the $685.8 million Small Cap Stock fund, the newest addition to the group, launched in June 1993., a load fund. The fund has no rating as yet. Also using the techniques is the $365 million Diversified International fund, a no-load fund until next June, managed by Greg Fraser.
The LBS fund will be named - perhaps "a little tongue in cheek" - Disciplined Growth, Mr. Wickwire said.
While Mr. Lewis' strategy incorporates many top-down elements, LBS' approach is bottom up, looking at fundamentals as well as price-derived information. The firm has built a neural network for each one of the 4,000 stocks in its universe.
"It's like a little analyst for each stock," Mr. Barr said.
Where do the human beings at LBS factor into the picture? With 4,000 forecasted alphas each week, they create portfolios, be they small-cap, midcap, long-short or enhanced index. They attempt to optimize returns and control risk by adjusting such factors as the number of stocks, weightings, style biases and industry or sector biases. Humans also handle the execution of trades, a formidable task with annual turnover of 170% to 300%.
Mr. Barr proudly pointed out that for three months in a row - December 1993 through February 1994 - the firm owned at least three of the top 10 stocks in the S&P 400. The odds of picking the top stocks three months in a row: "three in 1 million," he declared.
"People should not get scared by the techno-babble associated with this. It's just a tool just as regression is," Mr. Barr said.
