Recent research on the usefulness of financial analysts’ forecasts of earnings indicates that they are significantly more accurate than predictions made by naive models that merely extrapolate past earnings trends. Accuracy, however, is only one aspect of a wider attribute that useful forecasts must possess—that is, rationality, or the ability to reflect all available information.
A rational forecast will adjust for any systematic errors discernible from past performance. Analysts’ forecasts were found to incorporate the past history of realizations and predictions in an unbiased manner. This property was not exhibited by economists’ predictions of variables such as inflation, GNP and unemployment.
The evidence on the usefulness of analysts’ forecasts indicates that they are used by investors and, in fact, can serve as a reasonable proxy for the unobservable market expectation of earnings. Moreover, the dispersion of analysts’ forecasts—differences between earnings predictions—seems to be an important measure of risk, shadowing such traditional measures as beta and return variability.