(FORTUNE Magazine) – Give the stock market credit: It has been saying for years that the company with by far the largest market valuation in the world was losing its luster. The evidence is in that ski-slope line on the chart. As late as 1978, the price-earnings multiple on IBM's stock (as figured on earnings for the most recent four quarters) was 80% higher than the multiple on the Standard & Poor's 500-stock index. Then the difference began to narrow, erratically but persistently. And in 1986 IBM's multiple did the unthinkable. It sank below the S&P's -- nearly 25% below. Move over, President Reagan: You are not the only U.S. symbol in deep trouble with the public. In the 1986 leg of this slide, IBM fell from an April peak of nearly $162 a share to $125 in late December. At the top the company's market value was $99.7 billion, an amount so close to a colossal milestone that headline writers must have been hyperventilating. At $125, it is IBM's stockholders who are gasping. The fall cut IBM's market value by nearly $24 billion. To put that number in perspective, it is more than the market values of Coca-Cola and PepsiCo combined. Among the people clobbered by the drop were IBM analysts on Wall Street, who in general have been egregiously optimistic. In early 1985 many were forecasting 1986 earnings of about $14 a share. By early 1986 the analysts were down to $12. The reality? Probably $8 or so. One prescient analyst was Bob Djurdjevic of Annex Research in Phoenix. In April 1983, when nothing but good news was coming out of IBM, his always opinionated monthly letter asked, in boldface type, ''Is IBM Mortgaging Its Future?'' What worried him was IBM's drive to sell rather than lease its computers, a strategy that boosted immediate revenues at the possible expense of future revenues. Djurdjevic also questioned whether the computer industry was going to grow enough to support the mammoth capital expenditures that IBM was making on new plant and equipment. Such farsightedness makes Djurdjevic's current opinions worth attention. He thinks that by 1990 IBM will have reshaped itself into a healthier company and that software in particular will be giving a lift to earnings. In the meantime, though, he expects things to get worse, with 1987 looking like the trough year. One reason the market has dealt poorly with IBM is a fundamental change in the company's character. IBM used to be a Rip Van Winkle stock: Buy it, go to sleep, wake up years later a richer burgher by far. Today the stock is quite cyclical, mainly because IBM has lost that large, steady flow of revenues it once got from leases. Says Richard Gochman, an analyst at money manager J.&W. Seligman and a longtime follower of IBM: ''Consistency of earnings is important to institutional buyers, and predictability is important. Today this company is neither consistent nor predictable.'' That has made Seligman a seller of IBM stock in the recent past. Gochman says he also dislikes owning IBM when ''it's tangled up -- and, boy, it is really that way now.'' IBM may be going through another wrenching change: a substantive, noncyclical decline in profitability. In 1986 the company's return on equity was probably around 14%, a sharp comedown from the returns of more than 20% it had recently been earning. IBM's chairman, John Akers, expects a rebound. But Ulric Weil, a veteran IBM analyst who is now a Washington consultant, expects less. ''They may occasionally go over 20%, but they won't do it consistently,'' he says. ''This company is never going to look like the old Big Blue again. Those days are gone.'' Even so, Weil believes IBM is a reasonable buy at prices close to $120, in part because its $4.40 dividend provides a large measure of support. Many Wall Streeters think IBM is a bargain now. One is Barton Biggs, chief investment strategist at Morgan Stanley, who likes IBM a lot at $125. Unfortunately he also liked it at $155. ''I'm sort of chastened,'' Biggs says. ''The stock's hurt me.'' Nevertheless, he's kept the faith. He thinks that IBM's ''sustainable, normalized'' return on equity is perhaps 18% to 20%, vs. 12% or so for the S&P 500. IBM's profits, he believes, should grow at an average 11% to 12% a year, against 7% for the S&P. These numbers, he says, suggest that IBM should be selling at a price-earnings multiple 50% higher than the S&P's, not below it. Says Biggs unhappily, ''The price doesn't make any sense.'' In agreement is Rick J. Martin, a Sanford J. Bernstein & Co. analyst and one of Wall Street's biggest bulls on Big Blue. Martin, who used to work for IBM, believes its fortunes rise and fall with product cycles, and he expects the new line of midrange computers that IBM has announced for 1987 to be a smash. They have also been priced, he says, to produce ''spectacular'' margins. He concludes that in 1987 IBM could earn $10.75 a share, which he says is high enough to put him on ''the lunatic fringe.'' Another bull seems to be roaming around Armonk. Between May and late December, IBM spent nearly $1.5 billion to buy in 11 million shares of its own stock, paying on average $135 a share. The board has okayed an additional $550 million in purchases. Troubles or no, IBM appears to be a believer in itself.

CHART: NOT AVAILABLE CREDIT: ILLUSTRATION BY MARC ROSENTHAL CAPTION: NO CAPTION DESCRIPTION: Chart shows IBM price earnings ratio relative to that of Standard & Poor 500 multiple 1976-1986.