The notion that public dividend announcements contain relevant information was originally offered to reconcile the dividend irrelevancy proposition with the observation that a change in the dividend rate is often followed by a change in the market price. While empirical analyses of the informational content of dividends hypothesis have resulted in conflicting conclusions, there is substantial evidence which indicates that the market generally reacts favorably when the dividend rate is increased and unfavorably when dividends are reduced. The primary purpose of this study is to examine the market's reaction to specific types of announced dividend changes. By analyzing relatively homogeneous subsets of dividend announcements, we ·provide additional insight into the dividend information controversy as well as to how the market response varies with different types of dividend changes.

Dividend announcements examined in this study include (1) initial dividend declarations, (2) increases by traditionally high-yielding and high-payout firms, (3) increases of less than twenty-five percent, (4) increases of twenty-five percent and greater, (5) decreases of twenty-five percent and greater, and (6) dividend omissions. A standard daily residual analysis is conducted for each sample using four different sets of estimates for expected returns. Performance indicies are computed for the forty-one days symmetric to the announcement date and tests of significance are provided.

The primary results of the study can be summarized as follows:

1. In all cases, the evidence supports the informational content of dividends hypothesis.

2. The market's reaction to initial dividend declarations and dividend increases by high-yielding firms is much more pronounced than it is for other increases.

3. Substantial price adjustments occur in the month immediately preceding dividend reductions, but no such adjustments occur for dividend increases.

4. The market, on average, reacts efficiently to both dividend increases and decreases in the short run.