Predicting the equity premium with dividend ratios: a matter of balance

Abstract

Goyal and Welch (2003) used recursive residuals and plots of cumulative sum-squared errors to examine the predictive accuracy of dividend ratios in forecasting the equity premium. After a thorough specification search, Goyal and Welch were unable to find evidence in favour of dividend ratios as predictors of the equity premium. This note extends their analysis, further demonstrating the empirical fallacy.