Unraveling the Dividend Puzzle: A Field Experiment

We conduct a field experiment to test four dividend theories. We enhance managers’ perception of investors’ agency concerns, investors’ risk preference, investors’ information demand, and investors’ tax status respectively in four treatment groups. We find that past payers receiving the agency-related treatment increase their dividends relative to the control group. In contrast, firms receiving the other three treatments do not significantly change dividends. The agency treatment effect is more prominent for firms with weaker governance and robust to various specifications. A post-experimental survey confirms our findings. The evidence suggests that the agency-related motive is most pertinent in explaining dividend payout.