Red Faced Over Big Blue: Why The Analysts Blew It

Some 50 securities analysts--who collectively earn eight figures in salaries and bonuses, hang dozens of MBA sheepskins on their office walls, and tally up several hundred years of on-the-job experience--spend most of their working hours doing nothing but tracking the travails of IBM. They quiz IBM executives about new products and shipments, sound out the company's customers to see if they're satisfied with products and service, ask them about buying plans, and then cross-check their findings by running the same drill on Big Blue's competitors. When it's all done, they distill their findings into earnings forecasts.

Yet on Mar. 19, first in a brief press release and later in a conference call with those fearless 50 analysts and dozens of institutional investors, IBM executives told them something that not one analyst had turned up: Business is bad. So bad, in fact, that the analysts' average-earnings forecast--what Wall Street calls the "consensus estimate"--was about 100% too high. Instead of first-quarter earnings weighing in at $1.82 a share, said the IBM execs, the number would be only about half that.

How could so many analysts be so wrong? The question reverberated through Wall Street brokerage firms, trading rooms, and the offices of mutual-fund and pension-fund managers. In the short run, at least, the so-called negative earnings surprise cost IBM's 800,000 shareholders--about half of whom are individual investors--dearly. The stock lost 12 3/4 points, or 10%, on Mar. 19 and fell an additional 4 points by week's end, to 111 5/8. All told, Big Blue lost some $9.6 billion in market value that week, and by Mar. 27, the stock had recovered to only 112 7/8.

FUZZY VISTA. Earnings surprises are common enough, yet IBM's was a stunner. Smaller, lesser-known companies sometimes turn up big surprises. But IBM, with $69 billion in revenues, is hardly a stealth operation. Big Blue told the analysts that in the face of the gulf war and a softening world economy, demand for most of its hardware--from PCs to minicomputers and mainframes--is poor, and only workstations and software are doing well.

Where did the analysts fall short? Several suggest that they misjudged the macroeconomic picture, especially the impact that the gulf war had on already-slowing economies abroad. With 61% of IBM revenues coming from overseas, that was, no doubt, a major factor in the projected earnings shortfall. "We didn't gauge what has happening on the economic front as well as we should have," says Bear, Stearns & Co. analyst Clifford Friedman. Adds David P. Soetebier of A. G. Edwards & Sons Inc.: "A lot of us were asleep on the strength of the dollar and, especially, how that would play on overseas earnings."

A small error in estimating revenue can balloon on the bottom line. Forecasting technology-companies' profits is tricky because of their high gross margins and high fixed costs. Says Soetebier: "If you're 1% too low on revenue and 1% too low on expenses, your forecast can be off by a third."

The analysts--as well as some institutional money managers--say IBM could have spoken up sooner to provide them with more information on the sluggish pace of sales worldwide. While analysts must do their own legwork, the most important insights often come from the company itself in the form of "guidance." IBM is so big, diverse, and far-flung that no one but top management could possibly know what's really going on. "There was no earnings guidance this quarter," says William J. Milton Jr. of Brown Brothers Harriman & Co. In the absence of help, he says, "most analysts keep their estimates close to the consensus." If they're wrong, they're all wrong together--and that's what happened.

COCKROACH THEORY. Even the bearish IBM watchers, such as Brown Brothers' Milton, were caught off guard. In a Mar. 12 memo to clients, Milton warned of "unfavorable developments that continue to unfold in foreign markets" and raised the strong possibility of a negative earnings surprise. Still, Milton was forecasting $1.65 per share for the first quarter--which he says was the lowest estimate on the Street.

What was ironic about Big Blue's negative surprise is that in the fourth quarter, the company had a modest positive surprise, with profits 8% better than expected. That led analysts to the conclusion that IBM, struggling for the past six years to slim down and become more competitive, had finally begun to show the fruits of the trying transformation. And what reinforced that view for many was the tendency for positive-earnings surprises to repeat themselves for several quarters. That's what is known as the "cockroach theory." Earnings surprises--both positive and negative ones--for a company, say the pros, are like roaches: You don't see just one.

Little wonder, then, that through January, the analysts' first-quarter consensus forecast stood steadfast at $1.99, slightly ahead of the 1990 figure. But in February, the consensus started to edge down, a few pennies at a time. Initially, analysts lowered estimates because of a surge in the dollar, which would lower the value of foreign earnings. In addition, the analysts started to sense the economic slowdown abroad and mused that it would lower profits. On Mar. 4, Rick J. Martin of Prudential Securities Inc. cut his 1991 projection by 10~, but left his $1.80 first-quarter estimate intact. He also changed his IBM "buy" recommendation to a "hold." (On Mar. 19, his advice changed to "sell.") And on Mar. 11, Merrill Lynch & Co.'s Dan Mandresh, one of the most-followed IBM analysts on the Street, knocked down his forecast from $1.81 to $1.75 per share. Although he lowered his rating a notch, it wasn't until Mar. 19 that he cut his rating down to "neutral."

Moreover, while lowering the first-quarter outlook, most analysts were still upbeat on 1991's second half, when shipments of new, highly profitable mainframes are expected to kick in. Indeed, on Mar. 8, Bear Stearns' Friedman, referring to his meeting with senior IBM management, said that while a dreary economic picture clouded the near-term prospects, the company's long-term plan of better cost control and faster deployment of products was on track. He rated the stock a "good long-term investment"--and still does.

WHO KNEW? If IBM made the analysts look foolish, their clients don't seem to blame them. David A. Polak, president of NWQ Investment Management Co., lays the blame at IBM's doorstep: "When the fourth-quarter results came out, IBM was telling analysts and institutional investors that things looked pretty good going into the new year." Vincent Bajakian, who runs the $2.7 billion Wellington Fund, thinks IBM management was probably as surprised as everyone else. "I'm sure they're trying to figure out how business came to a screeching halt," he says. An IBM spokesman said that the bulk of first-quarter computer orders don't come until March anyway, and top management did not know the magnitude of the slump until a few days before the announcement.

The flap over the first-quarter earnings raises the perennial question of the value of analysts' earnings estimates, especially for huge companies such as IBM. "No one outside the company really has the numbers," says Tony Estep of New Amsterdam Partners, a money-management firm that tracks earnings forecasts, "so the analysts end up extrapolating from past results." And the analyst who might break away from the pack with a provocative view is discouraged from doing so. "If I had put out a first-quarter estimate of 90~ a share on IBM, I'd have been laughed out of town," says Steve Cohen of Soundview Financial Group Inc. His estimate was $1.80.

Still, those who put little faith in earnings estimates want the analysts' evaluation of the company. "We're interested in what they have to say about industries and companies," says W. Anthony Hitschler, chief investment officer at Brandywine Asset Management. Hitschler is bullish on IBM for the long term. When the stock price tumbled, he bought more shares. Indeed, despite the licking the analysts have taken, many are telling their clients to do the same.

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