Time-series implications of aggregate dividend behavior

This article investigates the hypothesis that dividend changes are determined by changes in some measure of permanent earnings. The analysis employs two measures of permanent earnings and takes into account the nonstationarity of dividend and earnings series. This study finds that dynamic dividend behavior is accounted for primarily by changes in permanent earnings, Dividends respond strongly to permanent changes in earnings without any significant over-reaction, whereas they respond little, if at all, to transitory changes in earnings, The findings also suggest that the partial adjustment hypothesis, which assumes managers partially adjust dividends to a target dividend performs better when the target dividend level is proportional to permanent earnings than when it is proportional to current earnings.