Are Dividends and Stock Returns Predictable? New Evidence Using M&A Cash Flows

Abstract

The lack of predictability of aggregate dividends by the traditional dividend-price ratio has long been considered a puzzle - “the dog that did not bark,” Cochrane (2008a). I show that this evidence is due to the mismeasurement of dividends used in empirical work. If M&A cash dividends are taken into account, the adjusted R2 from a regression of dividend growth on the lagged dividend-price ratio goes from being negative (-1.07%) to positive (17.47%), and coefficients become highly statistically significant. Strong improvements are also found for consumption growth (-1.09% to 10.63%), and out-of-sample return predictability. I also document that dividend-price variation is strongly linked to cash flow news and not only to discount rate news. Lastly, I find stronger predictability in industries with the largest M&A activity.

Keywords: Mergers and acquisitions; M&A cash flows; Dividend and consumption growth predictability; Return predictability

JEL Classification: G11, G12, G17