Analysts' Dividend Forecasts and Dividend Signaling

Abstract

This study complements existing research on the information content of dividends by focusing on the use of dividend expectations. We derive a measure of unexpected dividend changes, called dividend surprises, based on Value Line forecasts. Our results highlight a potentially serious sample misclassification arising from the extensively used naive dividend change method. Classifications of unexpected changes in dividends using dividend surprises result in stock price reactions and earnings changes that are consistent with the implications of dividend signaling models. Also, the approach followed in this paper permits the analysis of a significantly "forgotten" sample in previous event studies: Firms announcing no dividend changes in which investors (analysts) are expecting a change. We find that no change in dividends often reflects a negative dividend surprise and is indeed associated with negative stock price reaction and negative earnings changes. We provide evidence that the failure to find a relationship between dividend changes and future earning changes may be due to measurement error arising from misclassification of dividend changes. One implication of this study for future research is that empirical tests of dividend signaling models should incorporate dividend forecasts.

JEL Classification: G35, G12, G29