A Tax-Based Test of the Dividend Signaling Hypothesis
B. Douglas Bernheim, Adam Wantz
We propose and implement a new test of the dividend signaling hypothesis that is designed to discriminate between dividend signaling and other theories that would account for the apparent existence of a dividend preference. Our test refines the use of data on stock price responses to dividend announcements. In particular, we study the effect of dividend taxation on the bang-for-the-buck, which we define as the share price response per dollar of dividends. Most dividend signaling models imply that an increase in dividend taxation should increase the bang-for-the-buck. In contrast, other dividend preference theories imply that an increase in dividend taxation should decrease the bang-for-the-buck. Since there have recently been considerable variation in the tax treatment of dividends, we are able to study dividend announcement effects under different tax regimes. Our central finding is that there is a strong positive relationship between dividend tax rates and the bang-for-the-buck. This result supports the dividend signaling hypothesis, and is consistent with alternatives. The paper also provides corroborating evidence based on the relationship between the bang-for-the-buck and bond ratings.
Published: American Economic Review 85(3), june 1995, 532-551.