A Less Than Random Walk

Why some stocks don't shift course

According to efficient-market theory, stock market movements are essentially random. That is, stocks tend to show no predictable pattern from day to day, month to month, or even moment to moment because new information is instantly reflected in share prices.

Notwithstanding the general validity of this theory, however, analysts have found a number of market anomalies that suggest that the market is less than perfectly efficient. One of the most notable is the so-called momentum effect--the fact that high performers over a six-month period, for example, tend to continue to outperform the market in the subsequent 6-to-12 months while losers continue to lose ground.

What's responsible for the momentum effect? Some observers claim that the notion that the market responds immediately and fully to new information is highly exaggerated. They argue that prices tend initially to underreact to changes in corporate fortunes because new information may actually be diffused slowly to the investing public.

In a study of U.S. stocks from 1980 to 1996, economists Harrison Hong of Stanford University, Terence Lim of Dartmouth College, and Jeremy C. Stein of the Massachusetts Institute of Technology offer evidence backing this view.

Figuring that information about smaller companies gets out more slowly than news about large companies, the researchers predicted that momentum effects would be far larger among smaller-cap stocks. And that's exactly what they found. For stocks with a market cap averaging $100 million, for example, they found that buying recent winners (the top 30% in performance) and selling losers short (the bottom 30%) produced a cumulative return of 8.3% in six months. For stocks in the $500 million range, the six-month gain was 5.3%. But for stocks with a market cap over $5 billion, it approached zero (chart).

Since analyst coverage directly affects information flow, the authors also predicted that momentum strategies would pay off more for stocks followed by relatively few analysts. Indeed, taking company size into consideration, they found that such strategies for low-coverage stocks had a cumulative return of 15% after 12 months, and 20% after two years--far larger than the return for high-coverage stocks, which peaked at less than 10% after 12 months.

Finally, the study found that poorly covered past losers tended to decline further than poorly covered past winners tended to rise. "Companies are a lot less eager to let bad news out than good news," theorizes Stein, "and lack of analyst coverage probably helps them keep it under wraps."

Should one pursue momentum strategies? Given high transaction costs and the complexities of selling short, Stein doesn't recommend it for individual investors. But it's clear that if you're intent on buying recent past winners, your chances of making a big score are greater if you avoid large-cap stocks.