On The Predictability of Corporate Earnings Per Share Behavior

Many studies on the predictability of corporate annual income figures suggest that such series are generally random processes. These studies have been limited by complete reliance on extrapolatory or time-series models of annual earnings per share (EPS). The proposition that annual EPS figures may have predictable properties when considered in their economic environment is investigated.Three extrapolatory forecasting models are employed: 1. an average growth model, 2. an exponential smoothing model, and 3. a random walk model. Supplementing these are 3 lead-indicator models: 1. a money supply model, 2. a stock index model, and 3. a bank loan model. The models are used to generate one-year-ahead forecasts of EPS for 218 companies for the years 1968-77. The results suggest that predictive information concerning the general behavior of annual EPS existed over 1968-77 for the firms tested. Due to sample and forecasting model limitations, the results are to be considered preliminary.