A Market Scholar Strikes Gold

As an accounting professor at the University of Michigan, Richard G. Sloan was happy to receive a call inviting him to speak to a couple of hundred top institutional investors in New York City. Trouble was, Sloan was going to be on sabbatical in Western Australia when the conference was held in March. So Bernstein Research, the conference sponsor, flew him back -- in business class.

Such is 39-year-old Sloan's star power now that a Wall Street firm will hire him to travel 30 hours to give a one-hour talk and chat with some clients. His acclaim is founded on a paper he published in 1996 in a tiny journal, The Accounting Review. Using decades of data, Sloan was the first to find that investors habitually overlook the role accounting estimates play in determining a company's earnings and hence its stock performance. He sorted through clues in the financial reports and came up with the 10% of companies using the biggest estimates and the 10% using the smallest. His discovery, now known as the accrual anomaly, revealed that companies routinely using the highest estimates had stock prices most likely to fall, while those with the lowest tended to rise.

The market's inability to detect this pattern means that big money is left on Wall Street trading floors. How much? Well, if you had sold short the basket of stocks with evaporating earnings while buying the 10% on the rise, you would have beaten the market, before trading costs, by an average of 18 percentage points a year from 1962 to 2001, Sloan figures. Now investors are clamoring to exploit this market inefficiency. "They seem to be in a bit of a frenzy about it," he says.

It wasn't always so. For years, Sloan didn't get much credit for his insight. Academics didn't believe the market would consistently miss the relationship between estimates and stock price. Sloan even doubted it himself when he first ran the numbers in 1991. Then, the reigning view was that markets efficiently use all available financial information to determine a stock's worth. "I thought I must have done something wrong," he says.

He checked and rechecked the numbers. Then he submitted his paper to an academic journal four times in five years, only to have it returned each time with questions and suggestions for investigation. He finally sent the paper to The Accounting Review, which accepted it on his second try. Even so, Sloan was almost apologetic about his results in the article's conclusion. "He was still kind of keeping the door open for efficient markets even though he had discovered this anomaly," says Bruce A. Gulliver, president of Jefferson Research & Management, a Portland (Ore.) firm that uses Sloan's analysis in its stock research and investing.

Doubters among his peers persisted. "A very large number of papers followed up to check on him," says Charles M.C. Lee, a Cornell University accounting professor now on leave to help Barclays (BCS )create portfolios using Sloan's ideas. "It is not a statistical fluke." The big accounting scandals proved Sloan was right: Investors had put too much trust in reported earnings.

Debate continues over why this anomaly occurs. Some scholars think it's less a matter of earnings manipulation and lazy investors than executives trying to grow their companies too fast and investors going along. Sloan and a colleague, Scott Richardson at the University of Pennsylvania, have concluded that growth isn't the big factor. They say the primary blame lies, if not with manipulation, with the hubris of executives and investors who want to believe that their assumptions will drive the stock higher.

Whatever the cause, Sloan and others say there's still money to be made. But as more people catch on, this trading opportunity should diminish. How long it lasts depends on the ability and determination of investors to review earnings estimates skeptically.

By David Henry in New York

    Before it's here, it's on the Bloomberg Terminal.