Can computer programs invest better than humans?

BOSTON (AP) -- Computers make managing money easier. They can remind you when a bill is due, and help you stick to a budget. With investing, however, emotion often trumps automation. Many exited the stock market in late 2008 when it seemed the Dow might never stop falling. Well, the market roared back, and many anxious mutual fund investors missed out. Now they may be realizing they need a more disciplined approach. One possibility: quantitative or "quant" funds, the closest thing to emotion-free investing. There's no manager with human failings choosing the stocks. Instead, a computer program unique to that fund is in charge. Experts devise investing formulas and translate them into computer code. Yet there's a key weakness. Those formulas are largely based on how stocks have performed in the past, which won't necessarily guide what they do in the future. Indeed, most quant funds performed poorly when the market meltdown upended what many experts thought they knew. This is raising questions about whether rigid, numbers-driven approaches can adapt to a market where human stock-pickers have recently fared better. The disappointing results led Starmont Asset Management, a San Francisco-based money manager, to cut the amount of its clients' assets in quant funds from about 12 percent in 2007 to about 5 percent now. "The world changed dramatically in 2008, and the quants have to adjust," says CEO Harvey Rowen, whose firm manages more than $100 million for wealthy individuals. "They may be able to successfully change their programs to reflect the world as it is now, but it will take a while." Quant fund managers try to outwit the market by feeding data into their computer models about such stock-picking factors as shifts in a company's debt, earnings and revenue, or the stock's recent movement. Managers closely guard program details, giving investors only the basics about the designs and investing goals. "It tends to be a disciplined strategy, and it takes emotion out of the equation," says Greg Carlson, an analyst with fund tracker Morningstar. It's not always easy to determine if you have money in a quant strategy. Just a handful of funds use "quantitative" or "quant" in their names. The best place to go is a fund's prospectus, and review the disclosure of the investment process. It will specify how much of the fund's portfolio is managed using quant strategies. The approach isn't quite mainstream, but it's not far off. The largest fund company, Vanguard, offers eight funds entirely run using quant strategies, including funds found in 401(k) plans. Another seven Vanguard funds use quant strategies to manage pieces of their portfolios. All told, Vanguard has about $20 billion in quant strategies, or 2 percent of its total stock assets, according to a spokesman. Charles Schwab offers nine funds using quant strategies to some degree. Fidelity Investments doesn't offer any quant funds. However, many fund managers there sometimes use quantitative tools to guide their decisions. A look at three challenges quant funds face now: -- Poor performance: Over the past 3-year period, more than two-thirds posted below-average returns for their peer groups -- some 46 of the 68 quant funds that Morningstar tracks. Over 10 years, the quant record is average. About half of the 37 quant funds with 10-year records have beaten their peers. -- Losing investors: While quant funds went through that recent performance skid, assets in quant funds geared toward institutional investors, such as, pension funds and foundations, slid from $627 billion in 2007 to about $262 billion last fall. That's according to eVestment Alliance, an Atlanta-based investment technology firm. Over that period, investors withdrew a net $200 billion. -- Error message: Complex computer codes run quant programs, and errors can happen. Last week, quant fund manager AXA Rosenberg agreed to a $242 million settlement with the Securities and Exchange Commission over its alleged mishandling of a coding error. Starting in 2007, the error disabled a feature designed to reduce investing risks. SEC investigators concluded the company's top managers didn't learn of the problem from lower-level staffers until more than two years later. Instead of disclosing it immediately, top officials directed others to keep quiet, the SEC said. AXA Rosenberg disclosed the error to clients last April. Institutional investors and fund companies that had hired the firm to oversee some of their quant fund management -- including Vanguard and Schwab -- eventually dropped AXA Rosenberg. The California-based company didn't admit or deny the SEC findings. Robert Khuzami, head of SEC's enforcement division, noted that quant fund managers often protect trade secrets by leaving oversight of their computer models to a few programmers. "The secretive structure and lack of oversight of quantitative investment models, as this case demonstrates, cannot be used to conceal errors and betray investors," he said in announcing the settlement. The case illustrates the need for greater accountability at some quant funds, Morningstar's Carlson says. And the poor recent returns, he adds, are spurring some managers to adapt their computer models so they're more nimble. That means building more flexibility into the programs so they can protect portfolios in a broader array of market scenarios. It could even mean giving managers the capability to override programs under special circumstances such as the days following the Lehman Brothers bankruptcy when the market tanked. Carlson says a few quant funds appear to be successfully adapting. His three current favorites: Bogle Small Cap Growth (BOGLX), Bridgeway Aggressive Investors 1 (BRAGX), and Vanguard Strategic Equity (VSEQX). As for selecting one to pick, he says the criteria isn't much different than for other funds. "In the grand scheme of things," he says, "look at the manager, and the fund's costs, and the overall strategy -- whether it's quant or not." ---------- Questions? E-mail investorinsight@ap.org. Published: Fri, Feb 25, 2011