Dividend Persistence and Return Predictability

Abstract

Evidence of dividend yield return predictability has been presented so widely and consistently that the result has tended to be generally accepted. This paper provides strong evidence that return predictability of the dividend yield is a spurious result, even when the dividend yield definition is broadened to include share repurchases. The paper demonstrates how a spurious regression problem that is due to dividend persistence is compounded when the dependent and independent variables in return predictability regressions are ratios composed of common component variables, and it utilizes a simulation procedure to take account of this problem. The paper's results therefore imply that extreme care should be taken when using ratios as predictor or explanatory variables in time series regression. The paper finds that standard dividend behaviour explanatory models are also affected by the spurious regression problem, and introduces a reformulated Lintner first difference dividend behaviour model that is not subject to spurious regression.