When President Bush last month called Saddam Hussein's performance with U.N. inspectors "a rerun of a bad movie," he could just easily have been talking about the year-end statements investors just got from their mutual funds.

Over the five years ended Dec. 31, the 20 largest stock mutual funds -- repository of more than half a trillion dollars in savings -- returned, on average, a grand total of 10 percent, including dividends. That's the good news. The bad news is that the Standard & Poor's 500-stock index lost 23 percent last year, its third decline in a row.

Most of the biggest funds -- the ones that attract conscientious long-term investors -- have been flat since 1998, overall either losing a little or making a little. That's no surprise, since large funds tend to perform roughly the same as the S&P benchmark, which is down a total of 4 percent during those five years.

But meanwhile, amid all this stagnation and devastation, a group of stock portfolios has been thriving. With few exceptions, you can't invest in them, but, with some effort, you might be able to mimic them. And, at the very least, their success provides important lessons on how to pick stocks -- but it also, I have to admit, undermines my once-sturdy faith in blindly investing in index funds or large, diversified funds managed by actual human beings.

I ran across the portfolios by accident, flipping through the pages of a magazine published by the American Association of Individual Investors (www.aaii.com), a no-nonsense organization, based in Chicago, that helps educate its members, about 150,000 small investors.

Over the past five years, John Bajkowski, the group's vice president for financial analysis, has been introducing readers to the joys and intricacies of stock screens -- mathematical filters, or rules, that are applied to a huge universe of stocks. The object of applying screens is to find a few dozen gems that meet stringent criteria.

Bajkowski's aim is to teach, not to discover hot stocks. But, as sometimes happens when serious people put on their green eyeshades and simply do their work quietly and well, he has struck gold. Until I called last week to ask questions about his remarkable discoveries, he didn't really know what he had.

Here it is. Of the 54 stock portfolios (note: no bonds or cash) that Bajkowski has been following for the past five years:

* Forty-four beat the S&P in 2002.

* Forty-six beat the S&P for the entire five years, from Jan. 1, 1998, to Dec. 31, 2002.

* Sixteen doubled (not even including dividends) in value over the past five years, compared with a loss of 9 percent (again, without dividends) for the S&P.

* Ten produced positive returns in each of the five years.

* The best of the AAII portfolios returned a total of 388 percent over five years; the second best, 350 percent.

The portfolios use a wide range of screens. Some attempt to follow the strategies of famous investors of the past (such as Benjamin Graham and Philip Fisher) and the present (William O'Neil, Ralph Wanger, Martin Zweig). Others focus on particular indicators, such as the ratio of a stock's price to its sales or its book value.

Bajkowski points out that, in general, they choose companies smaller than the average S&P 500 business but larger than the average for the market as a whole. The portfolios have a bias toward value stocks -- that is, companies with low valuations. But four out of five growth-stock portfolios also beat the S&P, and many of the best strategies combine both value and growth screens.

As an example, let's examine a portfolio that uses a strategy Bajkowski calls "Low Price-to-Free-Cash-Flow." Cash flow, rather than reported earnings, is the concept on which classic stock valuation is built, as I wrote here two weeks ago. It's a figure that is less easily manipulated than earnings.

Since 1987, all listed firms have had to file a cash-flow statement, along with a conventional earnings statement, with the Securities and Exchange Commission. The AAII has a computer program called Stock Investor that calculates "free cash flow" for each stock. The program takes "operating cash flow" from the company's financial statement and subtracts capital expenditures and dividends. In other words, these are actual dollars of profit that the company's management can use as it wishes -- to invest in new factories and machines, to give to the shareholders as dividends, or to keep in the bank.

Now, here are the screens that Bajkowski uses to build the portfolio:

1. Begin with the universe of about 8,000 listed companies.

2. Eliminate all stocks with a market capitalization (that is, value according to investors) of less than $50 million. Stocks smaller than that are not very liquid -- that is, have little trading activity -- and so are very volatile in price.

3. Use Stock Investor, or another computer program, to find companies that have had positive cash flow in each of the past five fiscal years and for the most recent 12 months.

4. Take those remaining companies and for each, determine the ratio of the price to the free cash flow over the past 12 months. Eliminate all companies whose P/FCF ratio is higher than the norm for their industries as well as any company whose ratio is below its own five-year average. The idea is to find bargains.

5. Now, rank the remaining stocks by P/FCF ratio and compose a portfolio of the 50 with the lowest figures.


The surviving stocks include companies of various sizes and sectors, from Ford Motor (F), with a market cap of $19 billion, to Valley National Gases (VLG), which sells gas to welders and has a market cap of $53 million. Others that passed the screen test include Banta (BN), which sells printing and logistics services; Centex Construction Products (CXP); Tommy Hilfiger (TOM), clothing; Digi International (DGII), computer hardware; and Finlay Enterprises (FNLY), retail jewelry.

What I love about screens is that they turn up stocks in obscure businesses. This one found FinishMaster (FMST), whose niche is selling touch-up paints to auto-body shops. Besides generating gouts of cash, the company has been increasing its sales at an average annual rate of 18 percent annually and carries a price-to-earnings (P/E) ratio of less than 8. Another obscure find is Prime Medical Services (PMSI), which owns more than 60 mobile lithotripters, machines that travel from hospital to hospital, using shock waves to turn kidney stones into harmless fragments.

As you can see, the screening process that identifies these companies is not easy to describe, but with a software program you should be able to apply it fairly quickly. The AAII charges $247 for Stock Investor Pro, its top-of-the-line software. Morningstar (www.morningstar.com) sells screening software called Premium Stock Selector; Standard & Poor's (www.standardandpoors.com) has software for its well-known Compustat database; and other research firms have choices as well.

If you decide to buy such software and try to copy the AAII system, you may get exhausted quickly. Unfortunately, Bajkowski's amazing results were produced by running each screen on a monthly basis. I asked him if he had tried each formula annually, but, alas, he had not. My guess is that the systems don't really require a new computer run every month, but I could be wrong. To follow the strategies to the letter, a lot of buying and selling is required. That's expensive, both in sales commissions and short-term capital gains taxes.

But I am not encouraging you to rush out and play computer stock analyst. Nor is Bajkowski, who warned me not to get readers too excited. "Following these screens is not practical for most investors," he said. True, but the message is that even in this miserable market, there are strategies that can make money -- in many cases, lots of it.

Here are some of my favorite strategies from the list of AAII winners.

* High Return on Equity (five-year total return, 79 percent): The objective here is to find companies that earn the most on the investment that their shareholders have made (shareholders' equity on the balance sheet). The strategy also uses screens to eliminate firms with high debt, and it requires five-year earnings and sales growth that exceed industry averages. Stocks passing the screens at the end of December included Nautilus Group (NLS), fitness equipment; Too Inc. (TOO), clothing retailer; Fresh Del Monte Produce (FDP), pineapples and bananas; Williams-Sonoma (WSM), high-end kitchenware; and General Dynamics (GD), defense contractor. My favorite find: American Locker Group (ALGI), apartment mailboxes and coin-operated baggage carts, at a P/E of just 8, despite average annual earnings growth in double digits.

* Martin Zweig Approach (388 percent): Zweig, a popular author and money manager, uses a strategy that finds companies with what he calls "reasonable gains in sales and earnings." The screens here are complicated (all screens are detailed on the AAII Web site for members), but the result is a classic portfolio of growth stocks that have shown the strength to stay the course. No flashes in the pan for Marty. What's remarkable is that the portfolio, choosing stocks using Bajkowski's computer simulation of Zweig's style, has outperformed Zweig himself -- by a mile. Last year, for instance, the AAII's Zweig portfolio returned 17 percent while the Zweig Fund (ZF), a closed-end portfolio that trades like a single company on the New York Stock Exchange, lost 33 percent. This is like a copy of a Picasso selling for more than a real Picasso. Only 14 stocks managed to pass the latest AAII Zweig screen, many of them financials, such as Commerce Bancorp (CBH), RenaissanceRe Holdings (RNR) and Freddie Mac (FRE). Favorite find: Tuesday Morning (TUES), a felicitously named retail chain that sells closeout cookware, linens and home accessories, at a P/E of 18.

* Low Price-to-Sales (145 percent): Kenneth Fisher, son of financial genius Philip, brought this absurdly simple indicator to the attention of the public about 20 years ago, and James P. O'Shaughnessy resurrected it in his important book "What Works on Wall Street." The idea is first to calculate a company's total business revenue, or sales, for each of its shares of stock outstanding. Then divide that number into the stock price. The P/S ratio tends to vary across industries, but low numbers (especially below 1.0) often denote bargains. Bajkowski adds screens to eliminate companies with heavy debt, slow growth and languishing stock prices compared to the S&P index. The strategy did not produce a losing year in the last five, and it offered relatively low volatility (price swings). Passing stocks in December included Audiovox (VOXX), wireless phone handsets; Dave & Buster's (DAB), restaurant chain; American Dental Partners (ADPI), dental-practice management; and Norsk Hydro (NHY), Norwegian aluminum, oil, agriculture. Favorite find: OAO Tatneft (TNT), one of Russia's largest integrated energy companies, trading, like Norsk, as an American depositary receipt. Based in deepest Tatarstan, it has virtually no debt and trades at a P/S of 0.4 and a P/E of less than 3.

Those are only a few highlights. In addition, a system using screens that reflect the value-stock strategy of John Neff, legendary former manager of the Vanguard Windsor fund, returned 235 percent between 1998 and 2002; a Benjamin Graham-style "defensive" strategy returned 112 percent (with no losing years); and a system called "value on the move," which identified companies with low ratios of price to average annual earnings growth, returned 103 percent.

There remains the problem of using the screens in a practical way. You would think that some entrepreneur would make the best of the mechanical strategies the basis for several mutual funds. Remove the human element -- except for the initial choices of screens -- and let the computer do the work.

Only two funds, as far as I know, use any of the screens on the AAII's list to pick stocks. They are Hennessy Cornerstone Growth (HFCGX) and Hennessy Cornerstone Value (HFCVX). I have rhapsodized about the growth fund (which actually uses both value and growth techniques) in the past and don't want to belabor the point, but the two funds are based on the research of O'Shaughnessy. (Bajkowski's O'Shaughnessy Growth portfolio returned a total of 108 percent for the five years ending in 2002; Value, 21 percent.) It's a shame that the mutual fund establishment doesn't seem to appreciate the computer-screen approach. Brian Lund of Morningstar wrote in November, "Investing in Cornerstone Growth is largely a matter of faith, and we don't have any." This is a fund that over the past four years has beaten the S&P by an annual average of 18 percentage points. Why is an emotional human being or a dumb index better than a wisely chosen bunch of screens?

Of course, past AAII results are no guarantee of future success, but I calculated that the average screened portfolio returned a total 72 percent over the last five years, beating the most popular index by more than 80 percentage points. So, let's say that over the next five, the average is only 40 points better. That's still awfully good.

James K. Glassman's e-mail address is jglassman@aei.org.