RESEARCH Business Economy

WALL STREET
Companites manage earnings forecasts so actual earnings look better

Mark Reutter, Business Editor
(217) 333-0568; mreutter@uiuc.edu

12/1/2000

CHAMPAIGN, Ill. -- With the name of the game being not to fall below Wall Street's expectations, executives of public companies pay increasingly close attention to "managing" the forecasts of analysts. How does "earnings management" work, and does it bias the way that analysts estimate the prospective earnings of a company?

Dan Bernhardt, an economist at the University of Illinois, sampled 13,800 forecasts of publicly owned companies across 40 quarters between 1989 and 1998. He and colleague Murillo Campello, an economist at Arizona State University, zeroed in on "earnings forecast errors," or cases where the earnings per share varied from the average forecast made by the analysts.

"We found systematic evidence that companies try to talk down an analyst's forecast by a few cents per share," Bernhardt said in an interview. "That's perhaps not surprising. But what was surprising was that when forecasts went down as the announcement date approached, it was more likely that the earnings would exceed the consensus forecast. In other words, the latest forecast was not the most accurate."

Bernhardt said there was evidence that less experienced analysts were more likely to make late forecasts and were more likely to revise their forecasts than more experienced analysts.

"Analysts understand a company's incentive to deflate earnings forecasts, but they seem unable to completely adjust their predictions for earnings management," the UI economist said. Interestingly, earnings "surprises" were less likely to occur in firms that were followed by a small number of analysts, while the consensus forecast for firms followed by many analysts tended to be less accurate.

"Little is known about the dynamics of earnings management," according to Bernhardt. "While a number of academic papers seek to identify a correlation between discretionary components of earnings and a firm's ability to generate a positive earnings surprise, very few have examined how earnings forecasts are managed in order to convey good news and to create the greatest amount of financial slack for the company."

Lowering expectations, or creating "slack," is especially important for fast-growing companies where management may be unsure of the results of a quarter, or where earnings may change in a matter of weeks or even days.

Bernhardt said that the "greatest management sin" today is not to report a loss, but to report a loss not expected by Wall Street. Venerable companies such as Procter & Gamble and AT&T Corp. have lost billions of dollars of share-price value on the same trading day when they have sprung "surprises" on investors.

Even companies that undershoot analysts' expectations by a small amount are awarded with sharp sell-offs, making "earnings management" all the more important to executives wishing to keep their jobs.