Analysis And Comparison Of Earnings Forecast Agents

Recent policy making activities of the SEC [1978; 1979] and the AICPA [1980] have rekindled interest in the subject of management's public disclosures of forecasts. Much of the forecasting literature and related policy considerations concern the comparative accuracy of managerial forecasts relative to other sources of earnings forecasts. The basic premise is that if management forecasts are more accurate than those of other sources, then it may be socially desirable to have the FASB or the SEC govern forecast disclosures.' The economic implications of this policy issue were first discussed by Gonedes, Dopuch, and Penman [1976]. Much of the subsequent empirical work has centered on assessments of the accuracy of managements forecasts in an effort to determine if they are potentially more informative. A review of published research suggests that a number of conflicting results have been obtained. One study reports that firm-specific Box and Jenkins [1970] forecasts are superior to (more "accurate" than) those of management (Lorek, McDonald, and Patz [1976]), while another concludes that analysts' forecasts are superior to firm-specific Box-Jenkins forecasts and several other naive forecasting models (Brown and Rozeff [1978]). If one ignores differences in the samples and the time periods of these studies, the results imply that analysts' forecasts are more accurate