Can the inclusion of expectational macroeconomic data improve the predictive power of earnings per share forecasting models? The author used a model that included forecasts of industry and economic performance to predict annual earnings per share of 31 firms over a four-year time period. The results were compared with those of a naïve model that incorporated only historical earnings per share. The proposed model proved to be more accurate.
The proposed model was also used to predict earnings per share of 24 firms for a one-year time period. In this case, the forecast accuracy of the proposed model was examined relative to the accuracy of management forecasts. The management forecasts proved to be more accurate than the model forecasts. On the other hand, the model’s results did point up cases in which management had significantly overestimated future earnings per share.