The bird-in-hand theory of dividends argues that investors prefer a dollar of dividend today over a dollar of capital gains tomorrow. This means that a company can use its dividend policy to increase its stock price. On the other hand, the tax-clientele theory argues that investors in high tax brackets prefer capital gains over dividends as the latter are taxed at a lower rate, and by deferring capital gains, the present value of taxes is reduced. These are two opposing theories. Should a company pay higher or lower dividends?

The answer to this dilemma is not easy when you consider the economic argument that says that the intrinsic value of a stock is determined by a firm's investment policy. If a firm invests optimally (marginal rate of return is in excess of marginal cost of capital until the two are equal), a company's stock price is maximized. So, under this model, if a company increases its dividends, it will reduce its investments and lower its stock price by giving up positive NPV projects.

If you think this settles the issue, wait till we talk about financial leverage. Someone may say that if a company borrows more to pay higher dividends without cutting dividends it should be okay. But it is not that simple. If a firm is already at its optimal capital structure (where bankruptcy costs are balanced by tax savings of debt), then by borrowing more a firm increases its financial risk which would raise it discount rate, lowering the stock price offsetting the benefit of higher dividends.

In short, the bird-in-hand theory is a fallacy. If a firm optimizes its capital structure and investment opportunities, dividend policy is irrelevant. However, there is another wrench in all of this. It has been argued that investors prefer steady dividends over erratic dividends. In the real world unforeseen but profitable projects may crop up when the markets may not favor equity issuance. A company would do well to maintain a reserve borrowing capacity. Since dividends may also signal higher future cash flows in an asymmetirc world, a company should pay steady dividends which grow with future prospects. Abnormal earnings should be paid out as a special dividend which the market knows may not be repeated. This enables a company to maximize its stock price and satisfy invstors' preference for predictable and stable dividends.

The above arguments are valid under different assumptions and we have tried to put everything togther for a candidate who is lost and does not know what is the final conclusion relating to dividend policy. Candidates must be able to differentiate the economic arguments from those based on signaling and investor preference assumptions.