A Theory of Dividends Based on Tax Clienteles

This paper offers a novel explanation for why some firms prefer to pay dividends rather than repurchase shares. It is well-known that institutional investors are relatively less taxed than individual investors, and that this induces “dividend clientele” effects. We argue that firms determine the proportions of institutional and individual owners when they set dividends. Institutions have a relative advantage in detecting firm quality and ensuring firms are well managed. Firms paying dividends attract relatively more institutions and perform better. The theory is consistent with some documented regularities, such as a reluctance of firms to cut dividends, and offers novel empirical implications, such as a prediction that it is the tax difference between institutions and retail investors that determines dividend payments, not the absolute tax payments. This paper will be available from http://linux.anderson.ucla.edu/research.html. We thank Laurie Bagwell-Hodrick, Shlomo Benartzi, Charles Calomiris, Mark Grinblatt, Avraham Kamara, Hamid Mehran, Avanidhar Subrahmanyam, Richard Vines and Chris Hayden (Georgeson Investor Services) and seminar participants at Northwestern, UC/Davis, Columbia, UC/Riverside, Wharton, and Yale for helpful comments. This paper is UCLA-Anderson working paper #13-98 and Wharton working paper #15-98. Contact: mailto:[email protected]. US mail address: The Wharton School, 2300 Steinberg Hall — Dietrich Hall, Philadelphia, PA 19104-6367, (215) 898-3629. Contact: mailto:[email protected]. US mail address: Anderson Graduate School of Management at UCLA, 110 Westwood Plaza, Box 951481, Los Angeles, CA 90095-1481, Tel: (310) 825-2508. Contact: mailto:[email protected]. US mail address: Anderson Graduate School of Management at UCLA, 110 Westwood Plaza, Box 951481, Los Angeles, CA 90095-1481, Tel: (310) 825-2508.