Dividends and Investment: Further Empirical Evidence

This study further examines the relationship between corporate dividend and investment policies. Under Modigliani-Miller perfect capital market assumptions, these policies should be independent, but previous research finds conflicting results. Independence claims have been made based on Granger causality tests with 20 year time series. When we construct simulated series with a strong dependency between dividends and investment, however, causality tests are unable to detect it. We obtain a sample consisting of firms for which 37 years of data are available. Granger causality in the direction from dividends to investment is found in approximately 33 percent of these firms. This finding complements survey results showing dividends are given priority in some firms, a dependency that violates the separation principle.