THE RECENCY EFFECT OF ACCOUNTING INFORMATION

Abstract

This study tests the joint effects of dividend and earnings information. A study of joint effects is justified for the following reasons. First, dividends and earnings are considered two of the most important signaling devices (Aharony and Swary 1980) that investors use in evaluating stock prices. Second, dividends and earnings are 'garbled' information (Ohlson 1989). Dividends and earnings may contain corroborating or disconfirming news. Third, investors may be have with memory, revising beliefs in complex ways in evaluating a sequence of information. Prior dividend studies that controlling for earnings announcement effects do not address these possibilities.

Using Hogarth and Einhorn's (1992) belief-adjustment theory, this study models the behavior of investor reactions to joint dividend and earnings surprises. The theory predicts that order and timing of dividend and earnings surprises have different effects on stock returns. When dividend and earnings surprises have opposite signs (mixedevidence), the theory predicts that later surprises have a larger impact on stock returns than do earlier surprises (the recency effect hypothesis).

The evidence for the recency  effect hypotheses is relatively strong. In three out of four cases of mixed evidence (positive earnings, negative earnings and positive dividend surprises), the recency effect hypotheses are supported.