Dividend announcements under asymmetric information: An empirical study

Previous empirical studies have documented that equity prices react to announcements of unexpected dividend changes. The most promising hypothesis forwarded as an explanation of this phenomenon is the dividend signalling hypothesis. Building on a foundation of dividend signalling models, this study employs an expanded set of explanatory variables to explain excess returns documented around dividend announcements. This set includes variables found in previous studies to have an impact on excess returns around dividend announcements, as well as other variables considered by some to be important (but previously untested) determinants of equity price movements. This study utilizes a cross-sectional regression approach to extend the traditional event study methodology and temper the effects of the limitations of residual analysis. This approach is used to examine the cross-sectional variation in excess returns around dividend announcements with a set of explanatory variables which represent firm-specific characteristics and information available to investors at the time of the dividend announcement. Furthermore, this approach allows us to actually test whether the unexpected dividend change is the only influence on the announcement day excess returns or if there are other factors which explain these excess returns. This study has several empirical findings which expand our present knowledge about the factors that influence excess returns around dividend announcements. First, the key assumption of residual analysis is violated. The cross-sectional variation in excess returns does not average out even after the sample of firms is screened using the strictest procedures outlined in previous studies. Second, there are other variables along with the unexpected dividend change that significantly explain the cross-sectional differences in the excess returns. These results suggest that previous dividend studies using residual analysis which document excess returns around announcements of unexpected dividend changes have erred by attributing all the excess return to the unexpected dividend change alone