| This study consists of two related parts. The first is an examination of the possible empirical relationship between dividends and investment. In particular, simulations are used to examine the limits of the discriminatory power of Smirlock and Marshall's (1983) study employing Granger causality methods on a series of 20 annual observations per firm. Their empirical work is updated using series of 38 annual observations per firm rather than the 20 previously available. Granger causality from dividends to investment, significant at the ;The second part of this study consists of tests of the hypothesis that dividend payout causally precedes price risk. Although causality at statistically significant levels is not found in any substantial proportion of the firms in the sample, there are some interesting observations to be made. Specifically, when longer time series are employed, there is a stronger relationship between changes in OLS beta and changes in dividend payout, than there is between changes in standard deviation of returns and changes in dividend payout. One inference that could be drawn from these findings is that changes in dividend payout policy may contribute to increased systematic risk. |
