Last week's plunge in the computer giant's stock price raises questions about how much research Wall Street analysts actually do.

Gee, you would have thought that a meteor had smashed into IBM's headquarters in scenic Armonk, N.Y., last Thursday and wiped out the place. Or that Big Blue Chairman Lou Gerstner, whom Wall Street considers IBM's savior, had renounced the profit motive, taken a vow of poverty and announced he was quitting to devote the balance of his life to good works. But neither thing happened.

So why did IBM's shares drop almost 20 percent in a single day? That's the kind of percentage move you're used to seeing in Internet stocks and tacky, thinly traded issues, but not in a giant stock like International Business Machines Corp.

The answer: On Wednesday, after the New York Stock Exchange closed, Big Blue announced that sales of its mainframe computers would be soft for the next six months. Mainframe customers, IBM said, are cutting back on purchases of big machines because they're preoccupied with Y2K problems.

Duh. You would think that the extravagantly paid analysts and money managers who follow IBM would have known about this problem well before the company dropped its Big Blue bombshell, and warned investors what lay ahead. After all, analysts--who, as the term implies, are supposed to analyze companies' businesses and prospects--are supposed to talk with big customers of the companies they cover rather than relying on the firms to spoon-feed them.

And one would think that even an occasional call to a mainframe buyer or two would have uncovered IBM's mainframe problems. That way, IBM shareholders would have known what to expect, and wouldn't have freaked out and dumped the stock. The big investors who dumped IBM en masse didn't exactly cover themselves with glory, either. But that's a whole other story.

Chuck Hill, research director of First Call/Thomson Financial, which compiles analysts' recommendations, says that four of the 22 IBM analysts in his database downgraded IBM after its announcement. This is what rural types call locking the barn door after the horse has been stolen. How many analysts had downgraded IBM this year before Wednesday's announcement? One, who in May cut IBM to "buy" from "strong buy."

"We've said for a while that there are too many analysts getting a free ride in a market like this one, and they don't bother to go out and kick the tires," says Hill. "If more than one analyst [Kevin McCarthy of Donaldson, Lufkin & Jenrette Inc., who cut his earnings estimate on June 7 but left his "hold" recommendation intact] had done their homework," Hill adds, "the stock wouldn't have gotten as high as it was, and it wouldn't have fallen as much as it did."

The fact that one of the most widely followed stocks on the planet can move so greatly in a single day without a fundamental change having taken place shows how volatile the stock market has gotten.

David Braverman, a senior investment officer at Standard & Poor's Corp., points out that IBM's stock has had two big price rises this year as well as Thursday's big meltdown. On April 22, IBM rose 13 percent because earnings were higher than expected, and on May 13 it rose 9 percent after Gerstner told analysts that a quarter of IBM's sales came from Internet-related businesses. "IBM is no less volatile than any tech stock these days," Braverman says.

One of the stranger aspects of IBM's Thursday drop is that you may have seen it calculated two different ways.

Here's why. Most newspaper stock tables and financial sites on the World Wide Web showed IBM dropping 15 percent Thursday, to $91 from $107. But if you calculate the way S&P and Dow Jones & Co. do--and the way we're doing here--the drop was 19.3 percent, to $91 from $112.75. The reason? Dow Jones and S&P use the New York Stock Exchange's 4 p.m. (New York City time) closing figures, but stock tables and Web sites generally include NYSE after-hours trading, which runs until 5:15. IBM made its Wednesday announcement during after-hours trading, and the stock dropped to $107 from its 4 p.m. closing price of $112.75.

The $21.75 decline whacked $39 billion from IBM's stock market value and reduced the value of Gerstner's stock and options by more than $175 million, poor guy. IBM's drop cost the 30-stock Dow Jones industrial average 110.19 points, which is more than the Dow's Thursday drop of 94.67 points (0.9 percent).

Thus, all by itself, IBM turned a positive day for the Dow into a reasonably negative one. In addition, IBM seriously distorted the S&P 500, which fell 5.82 points, or 0.45 percent. S&P says IBM accounted for more than 80 percent--4.76 points--of the loss.

The point of this exercise? To show you not to take Wall Street "research" too seriously. And to remember to laugh the next time someone tells you that the stock market is a rational place and that big-money investors know what they're doing.

Sloan is Newsweek's Wall Street editor. His e-mail address is sloan@panix.com.