In just a few months, companies will start to report their annual earnings. And that means observant investors have a chance to witness what one finance professor calls the annual walk down to earnings the gradual ratcheting down of analysts estimates. Projections start optimistic and deflate, says Penn State s J. Randall Woolridge, getting more accurate as the announcement gets closer: If the company s announcing earnings at four this afternoon, they re usually pretty close.

What s a long-term investor to do? It s hard enough to focus on the long term under the best of circumstances, but long-term forecasts can help keep attention on the goal that is, outperformance over the next year or three or five. Now, a new study suggests that those long-term forecasts especially very rosy ones are the most unreliable. According to the work by researchers from Notre Dame and Singapore Management University, investors looking for reliable or even roughly realistic predictions of what that long term will look like for the companies in their portfolios probably shouldn t rely on analysts forecasts.

It appears that while analysts work quickly to incorporate new information into their short-term earnings forecasts, they don t adjust their long-term outlook as quickly, says Mitch Warachka, an associate professor of finance at Singapore Management University and co-author of the study. Plus, the average stock analyst s career is only about four years long, so they may have little incentive to continually fine-tune five-year projections. Eventually, the information influencing short-term forecasts makes its way into long-term expectations, too, but it takes about six months for the stock price to catch up, he says.

For investors, this could mean there s opportunity in identifying stocks where there s a big difference between short- and long-term forecasts, Warachka says. According to this theory, you d want to buy companies with strong short-term forecasts and weak long-term projections, or short companies where forecasts indicate they ll suddenly take off a few years down the road. A couple of words of caution: The typical pace of earnings growth is different in different industries. Tech companies might be reasonably expected to grow earnings faster than utility companies, for example, Warachka says. So you d want to evaluate any projected spike or dip in future earnings against other companies in the same sector. And, of course, there could be compelling reasons for an individual company to dramatically out- or under-perform its sector in the next few years you d just want to know what that reason was before relying on a long-term forecast, Warachka says.

Warachka and co-author Zhi Da, an assistant professor of Notre Dame, aren t the first to find fault with analysts estimates. Consensus estimates for earnings of companies in the S&P 500 have consistently overestimated earnings growth for the past 25 years, according to research by McKinsey & Company, a management consulting firm. Over the same time period, analysts have forecast 15% annual earnings growth and companies have delivered 6%, says J. Randall Woolridge, a professor of finance at Pennsylvania State University, who has done similar research. I don t know that it s dishonesty, it s just looking at things through rose-colored glasses, Woolridge says. The post-boom Global Settlement of 2003 aimed to reduce the incentives for analysts to inflate estimates to help their firms draw investment-banking businesses by separating analysts from other parts of a firm s business. But analysts predictions are still overly optimistic, Woolridge says.

ARE THESE COMPANIES LONG-TERM FORECASTS OVERLY OPTIMISTIC?

According to data from Capital IQ, coal giant Massey Energy is expected to earn just a penny per share in fiscal 2010, but analysts project it will earn $5.39 per share in 2014, representing a growth rate that s far above other energy companies. Over the next three to five years, analysts expect the company to grow earnings at a rate of 112% a year, according to Bloomberg.

Coeur d Alene Mines Corporation, a silver and gold miner, is projected to lose 13 cents a share in fiscal 2010, but earn 57 cents a share in 2014, according to Capital IQ. Analysts consensus estimates have it growing earnings at a rate of 184.5% a year in the next three to five years, according to Bloomberg.

Salix Pharmaceuticals, which makes drugs that treat gastrointestinal disorders, is expected to earn 30 cents a share in 2010 and $6.17 a share in 2014, according to Capital IQ, with earnings growing at a rate of 42.7% a year in the next three to five years, according to Bloomberg. Warachka s research suggests that investors who buy these long-term forecasts without a clear rationale for a growth surge may need strong stomachs.