Why Dog Stocks Had Their Day

Many analysts forgot that shares in weak companies can soar during a recovery

It seems like the perfect solution for investors fed up with conflict-ridden Wall Street stock-pickers: Charles Schwab & Co. (SCH ) feeds thousands of stocks through a computer, assigning each one an objective grade free of any human taint. Customers get lists of "A" stocks to buy and "F" stocks to shun. "Understanding our ratings is as easy as A, B, C, D, F," say ads for Schwab Equity Ratings.

Well, nothing's that easy. In the year through Jan. 12, Schwab's A stocks soared 66%. But the Fs -- the equivalent of a sell recommendation -- did even better: They jumped 70%. Says Jeremiah H. Chafkin, executive vice-president of Schwab's investment-advice business: "I do appreciate that the inversion looks strange."

Things are even more topsy-turvy at venerable Value Line (VALU ), another company that uses a computer model to pick stocks. Last year, its top-rated stocks rose 40%, but its bottom-rated ones jumped 90%. "To expect [a system] to be right each year is a bit much -- that's perfection," says Samuel Eisenstadt, Value Line's research chairman.

Raters that use human stock-pickers aren't doing any better. At Standard & Poor's -- like BusinessWeek, a unit of The McGraw-Hill Companies (MHP ) -- the one-star stocks climbed 57% for the year ending Jan. 31, while the five-star stocks mustered only 43%. The big Wall Street brokerage firms also got their picks and pans confused. For example, Merrill Lynch & Co.'s (MER ) sells jumped 46% last year, but its buys rose only 30%, according to research firm Integrated Data Consulting Services Inc.'s Marketperform.com. And UBS (UBS ) saw its buys go up 32%, while the stocks in its reduce category returned a much healthier 47%.

Should investors start using ratings as a contrarian indicator, loading up on the Fs and skipping the As? Not so fast. The upside-down performance of stocks over the past 12 months is typical of the first year of an economic recovery.

ACHILLES' HEEL. That's because short-term investors jump back into the market first, looking for big gains on the stocks that got clobbered the most during the bear run. These risk-takers know that low-quality stocks with volatile earnings often see dramatic rebounds in profits as the economy improves and the market rallies. And it doesn't take very many $1 stocks doubling or tripling in price to beat a portfolio of higher-quality stocks that didn't fall as far and don't have as much potential to rise. In turnaround periods such as the past year (the market touched bottom on Mar. 11, 2003), many investors "throw fundamentals out the window," says Sam Stovall, chief investment strategist at S&P.

In addition, the switch from a bear to a bull market suddenly forces short-sellers to buy low-quality shares they were shorting before. Chafkin says this helps account for the rise in Schwab's D- and F-rated stocks. "When shorts see a recovery coming, they do a lot of buying to cover their short positions in junky companies," he says.

The Achilles' heel of all stock-pickers -- relying too much on historical data that inevitably make steadily performing stocks look better than volatile ones -- left them especially vulnerable last year. Schwab's computer model analyzes 3,500 companies each week using dozens of measures, such as earnings growth and investment returns. The system gives an A to the top 10% of the stocks, an F to the bottom 10%, and middle rankings to the rest.

Schwab unveiled the system in May, 2002, amid a blitz of irreverent ads mocking Wall Street and pushing Schwab's independent investment advice. Despite the inverted results for last year, Schwab is confident its system will work over the long haul. Testing it on stock prices going back 15 years, "the As beat the Fs 80% of the time," says Chafkin. "The high-risk stocks happened to pay off this time, but we don't expect that to continue."

ESTABLISHED PATTERNS. Value line uses a similar method to rate 1,700 stocks. It plugs in the average price of the stock over the past year, its 10-year earnings growth, and its price momentum, among other factors. Value Line believes that evaluating stocks on established facts and patterns is a solid method. Says Eisenstadt: "We're right in 70% to 80% of years. Last year, the market went for stocks that had little to recommend them. They went from $1 to $2 or $2 to $4, for 100% gains. Other stocks couldn't compete."

Eventually, the established big-cap stocks favored by computer models and Wall Street analysts will begin leading the charge. Investors might have made a bundle on "$2 stocks with negative earnings or high p-e's, but there's no way to defend recommending stocks like that," says Schwab's Chafkin. He and others point out that as the bull matures, the market leadership will shift back to the higher-rated stocks. At that point, baffled investors can quit scratching their heads.

By Louise Lee in San Mateo, Calif.

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