Which of stock returns or dividend growth is predictable? A defense of stock return predictability

2018-05-08With CRSP return index widely used to compute the dividend‐price ratio in the finance literature, the lack of dividend growth predictability has been considered as stronger evidence than does the presence of return predictability—“the dog that did not bark: a defense of return predictability”, Cochrane (2008). However, CRSP’s completion of Pre62 Daily Data Series Project made small changes in their historical returns especially before 1947s and a substantial difference in their implied dividends is generated, which dramatically change the ability of the dividend‐price ratio to forecast dividend growth—now it forecasts both future return and future dividend growth. In this paper, I comprehensively re‐evaluate the empirical evidence of predictability of stock returns and dividend growth splitting the sample period pre‐ and pro‐1947 and adopting the two additional market returns: return from Robert J. Shiller online website and return constructed with independently measured NIPA dividends. This paper shows that (1) depends on the time periods (especially pre‐1947s) and data resources taken from, predictability dividend growth is present. (2) all of the long‐horizon return regressions give greater statistical evidence for return predictability. In addition, the same variables that produce dividend growth predictability steadily show the identical patterns of return predictability at long horizons, is stronger evidence of return predictability