Wall Street analysts may have had the last laugh now that their bete noir, former New York Attorney General and Gov. Eliot Spitzer, got his comeuppance over assignations with gilt-edged call girls.

Mr. Spitzer's performance as a crusading reformer was arguably responsible for major Wall Street firms agreeing to pay $1.5 billion in 2003 to settle allegations that they strong-armed their analysts into touting questionable stocks in order to win investment banking business from the companies the analysts were supposed to be analyzing objectively.

The uncaped crusader may have lost his credibility, but Mr. Spitzer's claims about the shortcomings of analysts have not. They are cemented in a new study by J. Randall Woolridge, a finance professor at Penn State's Smeal College of Business.

Mr. Woolridge's previous contribution to a more informed understanding of analyst behavior was research that concluded that investors who followed analyst recommendations would have slightly underperformed the Standard & Poor's 500, even though investing in stocks touted by analysts involved slightly more risk than investing in the broad market index.

This time around, Mr. Woolridge, aided by Penn State Harrisburg assistant finance professor Patrick Cusatis, compared analyst earnings growth forecasts for the companies they covered with what actually happened. After all, expectations of earnings growth are what drives the stock market. The better analysts forecasts are, the more investors can profit by acting on them quickly.

You won't be surprised by what the Penn State profs discovered.

They examined analyst forecasts at more than 1,200 companies from 1984 through 2006. They found that although analysts predicted long-term earnings per share growth of 14.7 percent at the companies they followed, the actual earnings growth that occurred was only 9.1 percent. By comparison, earnings of the S&P 500 over five-year periods grew an average of 7 percent from 1960 through 2006.

As should be expected, analysts fell closer to the mark when they looked only one year out, but their forecasts were unjustifiably cheerful nonetheless. They predicted average earnings per share growth of 13.8 percent vs. the 9.8 percent that actually occurred.

"Analysts' earnings growth rate estimates are consistently overly optimistic," Mr. Woolridge said. "These are very bright people. They have M.B.A.s from the best schools. They get paid very well. But they only see the upside."

And they seldom see the downside. While an average of about 30 percent of the companies studied had negative earnings growth in any given year, analysts predicted shrinking profits for only less than 1 percent of the companies.

"Their models are always forecasting positive growth," Mr. Woolridge said. "They never see the downturns. History tells us things go up, things go down."

The study indicates the positive bias of analysts has persisted even after their $1.5 billion settlement with the Securities and Exchange Commission. The agreement required Wall Street firms to separate their investment banking arms from their research departments in an effort to help analysts produce less biased and more realistic reports on the companies they follow.

Mr. Woolridge and Mr. Cusatis found that the gap between the growth analysts predict and the growth that actually happens has narrowed since the settlement, but remains significant.

There are several explanations for the persistent optimism of analysts. Some of their behavior stems from career concerns or conflicts of interest. Mr. Woolridge believes that one of the reasons why analysts are seldom gloomy is that they are rewarded financially to the extent that their optimistic assessments generate brokerage and underwriting business for their firms.

Anyone who has ever listened to a quarterly earnings conference call can attest to the fact that analysts are more likely to congratulate a CEO despite a miserable performance than they are to ask tough questions.

"People who are doomsayers don't last very long in this business," Mr. Woolridge said. "That's not what people want to hear."

Secondly, analysts only follow stocks they recommend and do not generate forecasts for companies they are not fond of, he says.

"If analysts systematically believe that they follow companies that are superior to others, they will be reluctant to issue negative earnings forecasts," Mr. Woolridge said.

Finally, analysts lose their objectivity because they get too close to the companies they follow, Mr. Woolridge says. They realize that if their forecasts are negative, "companies won't talk to them," he said.

Given what his research reveals about the accuracy of analyst forecasts and the value of their recommendations, Mr. Woolridge remains somewhat puzzled that so many continue to put great weight in what they have to say.

You could say the same about meteorologists, only, unlike analysts, they are more likely to forecast the storm of the century than warm and sunny weather.