Signaling, taxes, and dividend annoucements: evidence from the 2003 dividend tax cut

The Jobs and Growth Tax Relief Reconciliation Act of 2003 (JGTRRA) reduces the maximum personal tax rate of dividends from 38.6% to 15.0%. This unprecedented tax code change enables unique tests of theories of dividend policy. We first study the probability of paying dividends before and after the enactment. We find the probability of paying dividends increases after enactment, and the reduction in dividend tax rate is an important explanatory factor. Second, we implement a new test to distinguish between the dividend signaling hypothesis and free-cash-flow hypothesis. We find the magnitude of cumulative abnormal returns around dividend announcements declines after the enactment of JGTRRA. We also find that the dividend response coefficient remains positive, but the reduction in the dividend tax rate significantly mitigates this effect. These results are consistent with the signaling hypothesis rather than the free-cash-flow hypothesis. The results are also robust to different measures of dividend changes.