They're playing a familiar song with different words in the stock market these days.

The tune is "Regression to the Mean," a big hit among bears in the soaring 1990s. The skeptics warned, quite rightly as it turned out, that stock prices had climbed beyond the limits of mathematical reason, that the market couldn't go on indefinitely at a pace exceeding its long-term average rate of advance.

Now, after more than five years of retrenchment, the beat has been picked up by bulls, who say stocks may have gone too far in the other direction.

"Whatever people's fears tell them to expect, history argues strenuously against today's popular pessimism," says Milton Ezrati, senior economic strategist at Lord Abbett & Co. in Jersey City, N.J., which manages $94 billion in mutual funds and other accounts.

Among all the 10-year periods from 1925 through 2004 -- starting with 1925-35, then 1926-36, and so forth -- Ezrati says stocks posted a return of less than 3 percent a year only five out of 70 times. The most recent was 1964-74 when, according to Bloomberg data, the Standard & Poor's 500-stock index rose at a 1.2 percent annual pace.

For the entire seven decades, Ezrati said in a recently published update of a study he first did in 2002, the index's average annual return was 10.4 percent. That's right in line with the 9 percent to 10 percent range that kept showing up in 20th-century studies as an average market return.

There's no telling, of course, whether the 21st century will see anything like the same average value. A big problem with using mean reversion to forecast stock prices is that nobody can be sure what the mean is.

Still, 10 percent can serve as a handy reference point to gauge how the market fared over a period that included both strong, sustained economic growth and spells of big trouble.

In the 1990s, the S&P 500 averaged an 18 percent annual gain. By the end of 2002, its 10-year annualized return was down to 9.2 percent, putting on a pretty convincing show of mean reversion. Checking for an up-to-the-month reading, we find 10.2 percent for the 10 years ended May 31.

A market index that can run above its historic average can also sink below it for significant periods. So just because the S&P 500 is back down around its mean gives us no assurance that we have hit any sort of solid bottom.

Even so, says Ezrati, a few further calculations serve as "a powerful argument against any lasting future declines." I did the math myself: In the 1964-74 stretch, the S&P 500's total return -- not annualized -- was 12.7 percent. In the five years from the end of 1999 through the end of 2004, it declined 11.1 percent.

According to the personal savings plan analyzer on Bloomberg, even if the index is going to emerge from the 1999-2009 period with nothing better than its 1964-74 result, it will average a 4.8 percent gain for the last five years of the '00s.

Suppose the '00s are going to duplicate the '70s, when the S&P 500's annual gain of 5.8 percent produced a total return of 75.3 percent. In that case, the index will need to climb more than 14 percent a year in the last five years of the '00s to get where it is going.

Ezrati says the economic setting is far healthier now than it was in the worst decades of the last century.

"Real economic growth is averaging 3.5 percent to 4 percent at an annual rate," he says. "Capital spending is on the rise, as is employment. There is every reason to believe that the market should do better, not worse."

Yes, I know, none of this proves anything. The stock market can do whatever it wants, and isn't the least bit bound by precedent. But Ezrati's study highlights nicely what extraordinarily bad results stocks would have to produce to keep faring as poorly through the second half of this decade as they did in the first.

Investors who paid attention to the mean-reversion crowd in the '90s gained some valuable perspective. Maybe the same will prove true for those who listen now.