Dividends, Total Cashflows to Shareholders and Predictive Return Regressions

Abstract

This paper provides evidence on the predictive power of dividend yields for aggregate stock returns using a new annual dataset over the course of the twentieth century. Following Miller & Modigliani (1961), we construct a measure of the dividend yield that includes all cashflows to shareholders. In a cointegrating VAR framework we show that this has strong and stable predictive power for returns. The weak redictive power of standard measures of the dividend yield appears to arise from the strong rejection by the data of the implied restrictions on the impact of non-dividend cashflows.