The Information Content of Dividend Changes and Earnings: A Test of Signal Mitigation

Abstract

If reported earnings and dividends convey similar information regarding economic earnings, the second of the two announcements, be it dividends or earnings, should be less informative than had it been the first announcement. We refer to this phenomenon as "signal mitigation," and show that: (1) dividend changes which follow earnings reports are less informative than those which precede earnings reports for both large and small firms; (2) earnings reports which follow dividend changes are less informative than those which precede dividend changes only for small firms; (3) the extent of signal mitigation differs between small and large firms; (4) signal mitigation exists only when the earnings and dividend announcements convey similar messages. Our results are consistent with the Miller and Rock(1985) model as modified to allow reported earnings and dividends to be noisy measures of current economic earnings, and investors to revise their expectations of economic earnings when they observe reported earnings and dividends. Our results also suggest that dividend changes(earnings reports) convey value pertinent information when they reflect additional information that: (1) becomes available subsequent to earnings reports(dividend-change announcements); and (2) contradicts the news conveyed by the preceding earnings announcement(dividend announcement).