Whispers, a hidden force in stock trading
By Jack Reerink
NEW YORK, Nov 26 (Reuters) - A company posts quarterly profits that solidly beat Wall Street analysts' estimates but its stock gets hammered. Investors licking their wounds are left wondering what happened.
Chances are the company failed to beat the ``whisper number,' or the unofficial profit figure circulating in Internet chat rooms and on Wall Street trading desks. These numbers, which represent earnings per share, are often different from the average analysts' forecasts collected by market research firms, or consensus estimates.
And in this age of rapid-fire stock trading, a company that misses profit forecasts -- whispers or consensus -- by a penny a share can lose billions of dollars of market value in a matter of hours.
``If you miss the whisper number or meet it, generally that is seen as a disappointing factor, and (traders) sell the stock and whack it,' said Paul Hauck, co-founder of the Web site Whispernumber.com, which publishes whisper numbers. ``It's a short-term indicator but it can be a catalyst for a long-term move.'
The buzz numbers have long floated on trading desks, but in the past year they have become available to anybody with a computer and Internet connection. About five Web sites, which aim to make their money on advertising, now publish free whisper numbers on thousands of U.S. companies.
Industry experts agree whisper numbers can affect stock prices, at least in the short run. The buzz numbers gained credibility earlier this year after professors from two U.S. Midwest universities found whispers are more accurate in predicting profits than analysts' consensus estimates gathered by market research firms such as First Call/Thomson Financial.
That study, however, derived its whisper numbers from Internet message boards of Web sites such as Silicon Investor and Yahoo, and did not gauge the accuracy of numbers published by the new whisper Web sites. In fact, some say these newcomers use either historical data or -- worse -- just add a penny or two a share to the First Call consensus estimate.
The whisper Web sites' founders, by contrast, dispute those charges and say they obtain their data from a variety of sources, including analysts, company insiders, investors and brokers. But none of them has kept score and compared the accuracy of their numbers with consensus estimates.
TRACING WHISPERS' ROOTS
Whisper numbers exist because most companies exceed Wall Street expectations, undermining the credibility of analysts' forecasts. For example, almost two-thirds of U.S. companies beat analysts' projections in the third quarter and just 12 percent missed forecasts.
The reason is that companies often give analysts a conservative outlook so they can show an upside surprise when the actual earnings come out.
``The companies either by default or by design allow the pattern to develop,' said Chuck Hill, First Call's director of research. ``By default I mean you don't give enough guidance whereas by design you purposefully try to low-ball the numbers so you can beat them.'
Many analysts play ball and help companies tone down expectations, Hill said.
``Part of the problem is that there are too many rookie analysts who aren't going out, kicking the tires, talking to competitors, to come up with their own independent judgment,' Hill said. ``They are too much parroting what the companies are telling them. You get by with putting out in some fancy looking report what the company told you.'
Hill, a former technology analyst, said he dislikes whisper numbers because he thinks companies and analysts should give investors their most accurate forecast.
``Our first objective is to try and make the whisper numbers go away,' Hill said. ``Ideally, the analysts and the companies will wise up and start telling it like it is rather than playing games. Then the whisper numbers will disappear.'
SOURCES OF WHISPERS
But Hill realizes whisper numbers are here to stay and has started to publish ``hisper numbers' -- the ``h' stands for historical -- on companies that consistently beat the First Call consensus estimate, mostly in the technology and securities industries. To arrive at a hisper number, First Call takes the percentage by which a company has exceeded expectations in the past two years and accordingly adjusts the next quarter's consensus estimate for that firm.
Hill said his hisper numbers are identical to most whisper numbers, because both rely on crunching numbers from the First Call's historical data base.
``Our whole premise is that the great majority of genuine whisper numbers out there -- forget the rumor mills -- are coming from historical analysis,' Hill said.
Others, including the university professors who published the study on whisper numbers, disagree.
``If Mr. Hill's conjecture is right, (whisper Web sites) are lousy at doing the calculation,' said Mark Bagnoli, an associate professor at Purdue University. ``We have some evidence that says that, while some of what he says may be true, it is certainly not the whole picture.'
Whispernumber.com, for one, runs more than 100,000 Internet searches each day to locate whisper numbers on some 300 companies, Hauck said. The site also feeds off Web comments by employees of securities firms and has five data miners who throw out ``the garbage,' wild estimates based on wishful thinking or attempted stock manipulation, he added.
In other cases, Hauck's firm bases its whisper numbers on guesses by individual investors posted on the Web.
``We're trying to see the behavior of the individual investor. We act as a polling indicator,' Hauck said. ``The goal is to get the expectations from the public, which is a pretty powerful force in the market these days.' |
