A logical starting point for any discussion about dividends is the Irrelevance Theorem developed by Modigliani & Miller. According to this theorem, dividend policy does not affect the firm's value if its investment policy is predetermined. In practice, however, the market does not behave in this manner. Firms do distribute dividends, and increases in dividends usually lead to increases in share prices. Given the inferior tax treatment of cash dividends as opposed to capital gains and the high costs involved in raising new funds in the market, this suggests, contrary to the irrelevance theorem, that investors and managers do care about dividend policy. This phenomenon is known as the “Dividend Puzzle”.
The literature on dividend policy revolves around this “puzzle”. Why do managers distribute dividends at all? Why do investors care about dividends? Various explanations have been offered suggesting some benefits to compensate for the extra costs associated with dividend distribution: information or signaling effects (managers use dividends to credibly signal their forecast of the firm's future performance through changes in the level of distribution); reduction of agency costs (by both driving the firm into the capital market and diminishing the internal cash flow available to management).