Assume that n is the number of shares of record at time 0 and that rn is the number of new shares sold at time 1 at a price of PI. Equation (11-1) then can be rewritten as

'Merbn H. Miller and Franco Modigliani, "Dividend Policy, Growth, and the Valuation of Shares," frmrnal of Businrss, 34 (October 1961), 411-33.

312 Part 111 Financing and Dividend Policies In words, the total value of all shares outstanding at time 0 is the present value of total dividends paid at time 1 on those shares plus the total value of all stock outstanding at time 1, less the total value of the new stock issued. The total amount of new stock issued is where I is the total new investments during period 1 and X is the total new profit of firm for the period. Equation (11-3) merely states that the total sources of funds must equal total uses. That is, net profit plus new stock sales must equal new investments plus dividends.

FinancinglDividend Forgone Indifference The total amount of financing by the sale of new stock is determined by the amount of investments in period 1 not financed by retained earnings. By substituting Eq. (11-3) into Eq. (11-2), MM find that the nD, term cancels out and Because Dl does not appear directly in the expression and because X, I, (n + m)Pl, and p are assumed to be independent of Dl, MM conclude that the current value of the firm is independent of its current dividend decision. What is gained by stockholders in increased dividends is offset exactly by the decline in the terminal value of their stock. MM go on to show that nP, is unaffected not only by current dividend decisions but by future dividend decisions as well. Under the assumption of perfect certainty by all investors, the price of the stock at time 1, time 2, and time n is determined solely by Eq. (11-4). Thus, stockholders are indifferent between retention and the payment of dividends (and concurrent stock financing) in all future periods. As a result, stockholder wealth is unaffected by current and future dividend decisions; it depends entirely on the expected future earnings stream of the firm.

If both leverage and dividends are irrelevant, the firm would be indifferent to whether investment opportunities were financed with debt, retained earnings, or a common stock issue. One method of financing would be as satisfactory as the next.

Now one might ask, How does this correspond to our earlier chapters when we said that dividends are the foundation for the valuation of common stocks? Although it is true that the market value of a share of stock is the present value of all expected future dividends, the timing of the dividends can vary. The irrelevance position simply argues that the present value of future dividends remains unchanged even though dividend policy changes their timing. It does not argue that dividends, including liquidating dividends, are never paid, only that their postponement is a matter of indifference when it comes to market price per share.

Irrelevance under Uncertainty In a world of perfect capital markets and the absence of taxation, dividend payout would be a matter of irrelevance even with uncertainty. This argument involves Chapter 11 Dividend Policy: Theory and Pvactice 313 the same reasoning as that for the irrelevance of capital structure and for the irrelevance of diversification of corporate asset decisions. Investors are able to replicate any dividend stream the corporation might pay. If dividends are less than desired, investors can sell portions of their stock to obtain the desired cash distribution. If dividends are more than desired, investors can use dividends to purchase additional shares in the company. Thus, investors are able to manufacture "homemade"

dividends in the same way they devise "homemade" leverage in capital structure decisions.

For a corporate decision to be a thiig of value, the company must be able to do something for stockholders that they cannot do for themselves. Because investors can manufacture "homemade" dividends, which are perfect substitutes for corporate dividends under the above assumptions, dividend policy is irrelevant.

As a result, one dividend policy is as good as the next. The firm is unable to create value simply by altering the mix of dividends and retained earnings. As in capital structure theory, there is a conservation of value so that the sum of the parts is always the same. The total size of the pie is what is important, and it is unchanged in the slicing.

ARGUMENTS FOR DIVIDEND PAYOUT MATTERING The irrelevance of dividend argument assumes an absence of market imperfections.

To the extent imperfections exist, they may support the contrary position, namely, that dividends are relevant. We shall examine various imperfections bearing on the issue, beginning with taxes. After we consider the theory, we look at various empirical evidence bearing on the topic.

Unimportance of Corporate Income Taxes Unlike the capital structure decision, corporate income taxes have no bearing on dividend relevance. Under present law, earnings of a company are taxed at the corporate level, regardless of whether or not a dividend is paid. In other words, it is the profit after corporate taxes that is divided between dividends and retained earnings. Corporate income taxes could affect dividend relevance if tax laws were changed, as would occur under various methods to eliminate the double taxation of dividends.

Taxes on the Investor and Negative Dividend Effect Tax wedge is the difference in tax rates on dividends and on capital gains.

Taxes paid by the investor are another matter. To the extent that the personal tax rate on capital gains is less than that on dividend income, there may be an advantage to the retention of earnings. Apart from explicit tax rates, the capital gains tax is deferred until the actual sale of stock. Effectively, the stockholder is given a valuable timing option when the firm retains earnings as opposed to paying dividends.

As we described in Chapter 4, a differential tax on dividends and capital gains may result in a yield tilt. That is, a dividend-paying stock will need to provide a higher expected before-tax return than will a non-dividend-paying stock of the same risk. This is said to be necessary to offset the tax effect on dividends. ACcording to this notion, the greater the dividend yield of a stock, the higher its expected before-tax return, all other things being the same. Thus, security markets would equilibrate in terms of systematic risk, dividend yield, and, perhaps, other factors.

314 Part 111 Financing and Dividend Policies However, different types of investors experience different taxation on the two types of income. Institutional investors, such as retirement and pension funds, pay no tax on either dividends or capital gains. Corporate investors enjoy what is known as the intercorporate dividend exclusion. Presently, 70 percent of any dividend received by one corporation from another is not subject to taxation. As a result, the effective tax rate on dividends is less than that on capital gains. For example, if Laurel Corporation owns 100 shares of Hardy Corporation, which pays $1 per share dividend, 70 percent of the dividend income is tax exempt. In other words, Laurel Corporation would pay taxes on $30 of dividend income at the corporate tax rate. The overall tax effect will be less than if Hardy Corporation had share appreciation of

$100 and all of this were taxed at the capital gains rate. Accordingly, there may be a preference for current dividends on the part of the corporate investors.

Despite these exceptions, for many investors there exists a differential in tax rate between a dollar of dividend and a dollar of retained earnings. If these investors dominated at the margin, there would need to be an inducement to them to accept dividends. This inducement is simply a higher before-tax return so that after taxes they are as well off. According to this notion, paying dividends has a

"negative" effect on value.

Dividend Neutrality Even with a tax wedge between dividend and capital-gains income, it is not clear that investors at the margin are those who prefer capital gains to dividends. With different tax situations, clienteles of investors may develop with distinct preferences for dividend- or non-dividend-paying stocks. Many corporate investors will Completing markets prefer dividend-paying stocks, whereas wealthy individual investors may prefer impl~etsh at corporations stocks that pay no dividends. Tax-exempt investors will be indifferent, all other tailor dividend payout things the same. If dividend-paying stocks were priced in the marketplace to proto the unf~lledd esires vide a higher return than non-dividend-paying stocks, however, they would not of investors. be indifferent. They would prefer dividend-paying stocks.

If various clienteles of investors have dividend preferences, corporations should adjust their dividend payout to take advantage of the situation. Expressed differently, corporations should tailor their dividend policies to the d i l l e d desires of investors and thereby take advantage of an incomplete market. Suppose two-fifths of all investors prefer a zero dividend payout, one-fifth prefer a 25 percent payout, and the remaining two-fifths prefer a 50 percent payout. If most companies pay out 25 percent of their earnings in dividends, there will be excess demand for the shares of companies paying zero dividends and for the shares of companies whose dividendpayout ratio is 50 percent. Presumably, a number of companies will recognize this excess demand and adjust their payout ratios in order to increase share price. The action of these companies eventually will eliminate the excess demand.

In equilibrium, the dividend payouts of corporations will match the desires of investor groups. At this point, no company would be able to affect its share price by altering its dividend. As a result, even with taxes, dividend payout would be irrelevant.

The neutral position is largely based on corporations adjusting the supply of dividends to take advantage of any mispricing of stocks in the marketpla~e.~

However, another argument against clientele effects is that investors can use combinations of put and call options to isolate movements in stock price. By so doing, the investor effectively strips the dividend from the capital gains component 401 an excellent synthesis of these arguments together with a review of the empirical evidence, see Franklin Men and Roni Michaely "Dividend Policy," in R. A. Jarrow, V. Maksimovir, and W. T. Ziemba, editors, North-Holland Handbook of Operations Research and ManagemPnt Science: Fimc~A. msterdam: North-Holland, 1995.

Chapter 11 Dividend Policy: Theory and Practice 315 of the stock. If investors are able to separate dividends from capital gains to suit their needs, the tax-induced clientele argument for a particular stock is weakened.

Thus, both supply and demand forces may drive the equilibrium toward investors at the margin being indifferent, taxwise, between dividends and capital gains.

Positive Dividend Effect Apart from tax issues, we must recognize an argument for a positive dividend effect.

This is the possibility of a preference for dividends on the part of a sizable number of investors for behavioral reasons. For one thing, the payment of dividends may resolve uncertainty in the minds of some. Also, such payments may be Behavioral reasons useful in diversification of investments in an uncertain world. If in fact investors could produce a net can manufacture homemade dividends, such a preference is irrational. Nonethepreference for dividends, less, sufficient statements from investors make it difficult to dismiss the argument.

Perhaps, for either psychological or inconvenience reasons, investors are unwilling to m&ufacture hokemade dividends.

Shefrin and Statman reason that some investors are reluctant to sell shares because they will experience regret if the stock subsequently rises in prim3 For them, dividends and the sale of stock for income are not perfect substitutes. A second argument the authors advance is that although many investors are willing to consume out of dividend income they are unwilling to "dip into capital" to do so. Again, dividends and the sale of stock are not perfect substitutes for these investors. For behavioral reasons, then, certain investors prefer dividends. Whether they are numerous enough to make a difference is the Ultimately, the question of whether a negative, neutral, or positive dividend effect prevails is an empirical matter. We investigate this later in the chapter. First, however, we need to consider other influences on dividends.

Impact of Other Impeifections Although the factors we have examined are the most important when it comes to establishing whether or not dividend payout matters, there are other imperfections.

Flotation Costs The irrelevance of dividend payout is based on the idea that in accordance with the investment policy of the firm, funds paid out by a company must be replaced by funds acquired through external financing. The introduction of flotation costs favors the retention of earnings in the company. Flotation costs are the legal, investment banking, and other costs incurred by the corporation in a security issue. This means that for each dollar paid out in dividends, the company nets less than a dollar after flotation costs per dollar of external financing. Moreover, the smaller the size of the issue, the greater in general the flotation costs as a percentage of the total amount of funds raised. In addition, stock financing is

"lumpy" in the sense that small issues are difficult to sell even with high flotation costs.

Transaction Costs and Divisibility of Securities Transaction costs involved in the sale of securities tend to restrict the arbitrage process in the same manner as that described for debt. Stockholders who desire current income must pay brokerage fees on the sale of portions of their stock if the dividend paid is not sufficient to satisfy their current desire for income. This fee varies inversely, per dollar of 3Hersh M. Shefrin and Meir Statman, "Explaining Investor Preferences for Cash Dividends," Journal of Financial Economics, 13 uune 1984), 253-82.

316 Part 111 Financing and Dividend Policies stock sold, with the size of the sale. For a small sale, the brokerage fee can be a rather significant percentage. Because of this fee, stockholders with consumption desires in excess of current dividends will prefer that the company pay additional dividends. Perfect capital markets also assume that securities are infinitely divisible.

The fact that the smallest integer is one share may result in "lumpiness" with respect to selling shares for current income. This, too, acts as a deterrent to the sale of stock in lieu of dividends. On the other hand, stockholders not desiring dividends for current consumption purposes will need to reinvest their dividends.

Here again, transaction costs and divisibility problems work to the disadvantage of the stockholder, although in the opposite direction. Thus, transaction costs and divisibility problems cut both ways, and one is not able to draw directional implications about dividends versus retained earnings.

Institutional Restrictions Certain institutional investors are restricted in the types of common stock they can buy or in the portfolio percentages they can hold in these types. The prescribed list of eligible securities is determined in part by the duration over which dividends have been paid. If a company does not pay a dividend or has not paid dividends over a sufficiently long period of time, certain institutional investors are not permitted to invest in the stock.

Universities, on the other hand, sometimes have restrictions on the expenditure of capital gains from their endowment. Also, a number of trusts have a prohibition against the liquidation of principal. In the case of common stocks, the beneficiary is entitled to the dividend income, but not to the proceeds from the sale of stock. As a result of this stipulation, the trustee who manages the investments may feel constrained to pay particular attention to dividend yield and seek stocks paying reasonable dividends.

If institutional investors are better monitors of corporate performance than are individual investors, their presence may help ensure that the company is managed efficiently. In this context, paying dividends would be a positive signal to the market because such dividends will attract institutional investors seeking investment in quality c~mpaniesF.~ro m the standpoint of increased demand for the stock and signaling, the argument works in the direction of a preference for dividends as opposed to retention and capital gains.

FINANCIAL SIGNALING Signaling may occur if the dividend is more or less than expected.

Cash dividends, then, may be viewed as a signal to investors. Presumably, companies with good news about their future profitability will want to tell investors.

Rather than make a simple announcement, dividends may be increased to add conviction to the statement. When a firm has a target-payout ratio that is stable over time and it changes this ratio, investors may believe that management is announcing a change in the expected future profitability of the firm. The signal to investors is that management and the board of directors truly believe things are better than the stock price reflects. In this vein, Miller and Rock suggest that investors draw inferences about the firm's internal operating cash flows from the dividend anno~ncementT.~h e notion is based on asymmetric information. Management 4This argument is made by Franklin Allen, Antonio Bernardo, and Ivo Welch, "A Theory of Dividends Based on Tax Clienteles," workin-g p.a .p er, Wharton Wool (May 1999).

jMerton H. Miller and Kevin Rock, "Dividend Policv under Asvmmehic Infomation." Tournal of Fmance. 40

(Scpterl~ber 13351,