The Variability of Earnings and Non-Earnings Information and Earnings Prediction

This paper examines the performance of a 'composite' model of earnings prediction that integrates current earnings and current price as predictors of next year's earnings. The results show that current earnings (current price) play a key role in predicting future earnings when the ratio of earnings variance to price variance is low (high). The composite model is superior to univariate time-series models in out-of-sample predictive accuracy for the overall sample, and is substantially so for the group of firms with a high ratio of earnings variance to price variance.