Why are dividends disappearing? An empirical analysis

Abstract

We investigate the causes of time-series fluctuations in the propensity to pay dividends, including the post-1978 decline documented by Fama and French (2001). We consider explanations based on fluctuations in dividend clienteles, agency problems, information asymmetries, executive stock options, catering incentives, tax code awareness, and short-lived idiosyncratic factors. To evaluate these explanations, we conduct three styles of analysis. First, we count and classify influences on the propensity to pay that were noted in the financial press. Second, we examine time-series relationships between the propensity to pay and proxies for the driving influences in the candidate explanations. Third, we assess whether the candidate explanations are theoretically compatible with related time-series patterns involving dividend policy. Overall, the results are most consistent with the catering explanation. Notably, catering incentives, as measured by the stock market "dividend premium," roughly line up with the four trends in the propensity to pay between 1963 and 2000 and are able to account for the observed magnitude of the post-1978 decline. There is also evidence that idiosyncratic factors, including the Nixon-era dividend controls and the recent growth in options, affected the propensity to pay in specific periods.