What do Analysts’ Forecasts Reveal about Analysts’ Private Information? † Bjorn N. Jorgensen * and Julian Yeo ** Columbia Business School, 3022 Broadway, Uris Hall, New York, NY 10027 October 7, 2005 Abstract Analysts often provide forecasts of one-year ahead earnings, earnings two-year ahead, as well as long-term earnings growth rates. In our attempts to understand the properties of these contemporaneous multi-period earnings forecasts, we begin by examining whether analysts’ earnings forecasts can be described using the linear information dynamics (LID). We find that LID is an appropriate description of analysts’ information processing for 77% of our firm year observations. For firm-year observations that exhibit LID, we develop a theoretical procedure to infer the persistence of abnormal earnings, analysts’ other information about future abnormal earnings (not in current abnormal earnings), and the persistence of other information at individual firm year level. Demonstrating that the two symmetrical combinations of the persistence parameters derived from our estimation procedure lead to the same price estimate, we then calculate the predicted firm value (incorporating analysts’ private information). Our results show that the predicted firm value is statistically associated with the observed market value. In addition, when forecasted abnormal earnings are expressed as an ARMA (2,1) process, the median persistence in current forecasted abnormal earnings into one period ahead forecasted abnormal earnings is 0.95 and the median persistence in lagged forecasted abnormal earnings into one period ahead forecasted abnormal earnings is 0.16. Since our procedure allows us to infer analysts’ other/private information, we find that the main information incorporated in returns originate mostly from innovations in other information rather than innovations in abnormal earnings. Disclaimer: The Securities and Exchange Commission, as a matter of policy, disclaims responsibility for any private publication or statement by any of its employees. The views expressed herein are those of the author and do not necessarily reflect the views of the Commission or of the author’s colleagues upon the staff of the Commission. † VERY PRELIMINARY AND INCOMPLETE. * Currently visiting the Securities and Exchange Commission. E-mail: bnj2101@columbia.edu. ** E-mail: jy2176@columbia.edu.