However, it generally is agreed that the effect is more important for dividends than it is for capital structure.

EMPIRICAL TESTING AND IMPLICATIONS FOR PAYOUT While we have seen that a number of factors may explain dividends' impact on valuation, many are difficult to test. Most empirical testing has concentrated on the tax effect and on financial signaling. This is not to say that such things as flotation costs, transaction costs, institutional restrictions, and preference for dividends have no effect; only that whatever effect they might have is swamped by the two effects discussed. Empirical testing has taken several forms.

Ex-dividend Day Tests One of the mainstays has involved the ex-dividend behavior of common stock prices. Investors buying the stock before that date are entitled to the dividend declared;

purchases on or after the ex-dividend date are not entitled to the dividend.

In a nontaxable world, the stock should drop in value by the amount of the dividend on the ex-dividend day. If you are a taxable investor, however, and buy the stock before the ex-dividend day, you will need to pay taxes on the dividend. In contrast, if you wait until the ex-dividend day to buy the stock, you will pay no taxes on the dividend, since there is no dividend, and any price movement presumably is subject only to the capital-gains tax. A number of authors reason that if there is a tax effect, owing to capital gains being taxed at a lower rate than dividend income, a stock should decline in price by less than the dividend on the exdividend day. Expressed differently, investors would value a dollar of dividends less than they would a dollar of capital gains.

An early study of the phenomenon was by Elton and Gr~berI.n~ a sample of companies, they found that on average a stock declined by .78 of the dividend on the ex-dividend date. They interpret this result as consistent with a clientele effect where investors in high tax brackets show a preference for capital gains over dividends, and vice versa.

There have been a number of other studies of share price behavior on the exdividend day7 In a number of these studies, the evidence is consistent with the 6Edwin J. Elton and Martin J. Gruber, "Marginal Stockholder Tax Rates and the Clientele Effect," Review ofEc0-

noniics and Statistics, 52 (February 1970), 68-74.

7Avner Kalay, "The Ex-Dividend Day Behavior of Stock Prices: A Re-examination of the Clientele Effect," Journal ofFinance, 37 (September 1982), 1059-70; Kenneth M. Eades, Patrick J. Hess, and E. Han Kim, "On Interpreting Security Retums during the Ex-Dividend Period," Journal of Financial Economics, 13 (March 1984), 3-34; Patrick J. Hess,

"The Ex-Dividend Day Behavior of Stock Rehrns: Further Evidence on Tax Effects," Journal of Finance, 37 (May 1982), 445-56; James M. Poterba and Lawrence H. Summers, "New Evidence That Taxes Affect the Valuation of Dividends,"

Journal of finance, 39 (December 19841, 1397-1416; Costas P Kaplanis, "Gytions, Taxes and Ex-Dividend Day Behavior,"

Journal of Finance, 41 (June 1986), 411-24; Michael J. Barclay, "Dividends, Taxes and Common Stock Prices: The Ex-Dividend Day Behavior of Common Stock Prices before the Income Tax," Journal of Financial Econmflics, 19 (September 1Y87), 31-44; Roni Michaely, "Ex-Dividend Day Stock Price Behavior," Journal of Finance, 46 (July 19911,845-59;

and John H. Boyd and Ravi Jagmathan, "Ex-Dividend Price Behavior of Common Stocks," Review of Financial Studies, 7 (Winter 1994), Tll-41.

318 Part I11 Financing and Dividend Policies foregoing, namely, that stock prices decline on the ex-dividend day but by less than the amount of the dividend. Such findings are consistent with a tax effect where dividends are taxed more heavily than are capital gains, and stock prices reflect this differential. This, of course, is in accord with dividends having a negative effect on value. Other studies look at market microstructure considerations, such as the minimum tick size in a transaction, and find that when these considerations are taken into account the drop in stock price approximates the dividend arn~unt.~

Findings of this sort support the dividend neutrality argument.

Dividend-Yield Approach A second approach to the tax effect question is to study the relationship between dividend yields and stock returns, where other influences on returns are isolated.

If higher-dividend-yielding stocks provide higher returns to investors, this would be consistent with a negative effect of dividends on value. For the most part, there is not evidence of a significant relationship between stock returns and dividend yield? A lack of relationship is consistent with dividend neutrality. The evidence is not unambiguous, however. Naranjo, Nimalendran, and Ryngaert find a significant positive relationship between stock returns and dividend yields.10 Moreover, this effect appears to be unrelated to the tax rate of the corporation-the yield tilt hypothesis previously described.

Financial Signaling Siudies Testing for a financial signaling effect has involved a different methodology. Typically, an event study is employed where daily share price changes, relative to the market, are analyzed around the announcement of a dividend change. A number of studies report findings consistent with a dividend announcement effect: increases in dividend leading to positive excess returns and decreases leading to negative excess returns.ll The effect seems to be more pronounced for companies

'Src Rahcsl! Ball a~Gdsi len L. H~te,' tx Ulv~dendI Jay Stork I'rio. Hehavlor Oiirrerenrsi or T,~.lr-lnd.~Lrd iientclcs!' Ir:v*ra! fl'Fl~lanna: rrt.?zonri:s. 17 (February 19481. 127 50, XIurrav Frdnk and I

Stuck Pncvs Drop b\. Less rhan the Value nt rhr I)~v~dendF?v iilenrp irnm .? Cot~nrnr,v ~tho.~Tt t~.o ,.I ~ t~r,,*,':.~F.,i~,t cial Economics, 47 (~Lbruary1 998), 161-88; and M. Ameriane Lasfer, "Ex-Day Behavior: Tax or Short-Term ~raiiEnf-~

&c~s,"J OUTo~f F~iInu nce, so (July 1995), 875-97.

'See Fischer Black and Myron Scholes, "The Effects of Dividend Yield and Dividend Policv on Common Stock Prices and Returns," Jorrnwl of Financial Economics, 1 (May 1974), 1-22; Robert H. ~itrenberger and Krishna Ramiswamy, "The Effect of Personal Taxes and Dividends on Capital Asset Prices," Journal ofFinancia1 Econoniics, 7 Uune 1979), 163-95; Nai Fu Chen, Bruce Grundy, and Robert F. Stambaugh, "Changing Risk Premiums and Dividend Yield Effects," Journal of Business, 63 (Spring 1990). 51-70; William G. Christie, "Dividend Yield md Expected Returns:

The Zero Dividend Puzzle," Journal of Financial Economics, 28 (November-December 19901, 95-126; and AUen and Michaely, "Dividend l'olicv."

'"dy Naranjo, M. Nimalendran, and Mike Ryngaert, "Stock Returns, Dividend Yields, and Taxes," Journnl of Fznnnce, 53 (December 199B), 2029-57.

"See, for example, Joseph Ahamny and Itzhak Swaly, "Quarterly Dividend and Earningfi Announcements and Stockholders' Returns: An Empirical Analysis," Journal of Finance, 35 (March 1980). 1-12; Kenneth M. Eades, Patrick H. Hess, and E. Han Kim, "Market Rationality and Dividend Announcements," lournal oiFinnncia1 Economics, 14 (December 1985), 581-604; Aharon R. Ofer and Daniel R. Siegel, "Corporate Financial Policy, Tnfonnation, and Market Expectations:

An Empirical Investigation of Dividends," Journal oJFinancr (September 1987), 889-911; Paul M. Healy and Krishna G. Palepu, "Earnings Information Conveyed by Dividend Initiations and Omissiol=." journal of Financial Ecunomics, 21 (September 1988), 149-75; Paul Asquith and David W. Mullins It, "The Impact of Initiating Dividend Payments on Shareholders' Wealth," lournal ofBusiness, 56 (January 1983), 77-96; Larry H. P. Lang and Robert H. Litzenberger,

"Dividend Announcements: Cash Flow Signalling vs. Free Cash Flow Hypothesis," Journal of Financial Economics, 24 (September 1989). 181-91; Chihwa Kao and Chunh Wu, "Tests of Dividend Signaling Using the Marsh-Merton Model: A Generalized Friction A~~roach.T"o urnal of Business. 67, No. 1 11994). 45-68: and Roni Michaely, Richard H. Thaler, and Kent Womack, "fihce ~eactionsto ~ividendin itiations and)miasions: derreaction or Drift?" Journal of Finance, 50 (June 1995), 573-608.

Chapter 11 Dinidend Policy: Theory and Practice 319 that previously overinvested free cash flow in projects with returns less than what the financial markets require. After all, those cash flows belong to stockholders and should not be invested in negative NPV projects. The signaling effect is particularly pronounced for companies that initiate dividends for the first time or after a long hiatus (positive share price effect) and for companies that omit dividends

(negative effect). The two effects are not symmetric: Omitted dividends have a greater negative effect than the positive effect associated with dividend initiations.

Conceptually, dividend changes should provide information about future cash flows and earnings of a company. In an interesting article Benartzi, Michaely, and Thaler find little evidence of a positive relationship between dividend changes and earnings changes 2 years out." Expressed differently, dividends do not appear to contain information about future earnings. At the time of the dividend change, however, the authors find that there is a significant earnings change.

bEexlocnegsss to t h Implications for Corporate Policy As reflected in the foregoing discussion, the empirical evidence is mixed as to there being a tax effect that requires a higher before-tax return the greater the dividend.

In recent years, however, the evidence is largely consistent with a di-~idend neutrality. No studies of recent vintage support a positive dividend effcct. In contrast to the mixed results for a tax effect, most empirical studies suggest that dividends convey information and that there is a signaling effect. Where does this leave us with respect to directions for a corporation?

A company should endeavor to establish a dividend policy that will maximize shareholder wealth. Almost everyone agrees that if a company does not have sufficiently profitable investment opportunities, it should distribute excess funds to its stockholders. As we have said before, these funds belong to them, not to management to squander on wasteful expenditures. The firm need not pay out the exact unused portion of earnings every period. Indeed, it may wish to stabilize the absolute amount of dividends paid from period to period. But over the longer run the total earnings retained, plus the senior securities the increasing equity base will support, will correspond to the amount of profitable investment opportunities.

Dividend policy still would be a passive residual determined by the amount of investment opportunities.

For the firm to be justified in paying a dividend larger than that dictated by the amount left over after profitable investment opportunities, theze must be a net preference for dividends in the market. It is difficult to "net out" the arguments just discussed to arrive at the bottom line. Only institutional restrictions and some investors' preference for dividends argue for dividends. The other arguments suggest either a neutral effect or a bias favoring retention. There does appear to be some positive value associated with a modest dividend as opposed to none at all.

This occurrence may be due to institutional demand and/or a signaling effect.

Beyond that, thc picture is cloudy, and some argue that even a modest dividend has no effcct on valuation. Few academic scholars argue that dividends significantly in excess of what a passive policy would dictate will lead to share price improvement. This is particularly true when there is information asymmetry between management and investors, as the excess dividend ultimately must be replaced with a stock offering. Still, many corporations pay out a significant portion shareholders should be distributed to 12Shlomo Benartzi, Roni Michaely, and Richard Thalcr, "Do Changes in Tlividends Signal the Future or the Past?"loumnl $Finance, 52 (July 1997), 1007-34.

and them.

320 Part 111 F i n a n c i n g a n d D i v i d e n d Policies of their earnings, and many high-taxed stockholders receive and pay taxes on these dividends. These facts are difficult to reconcile with the theory.

SHARE REPURCHASE In recent years, the repurchase of stock by corporations has grown dramatically.

When you look at the total cash distributed to shareholders-cash dividends, Share repurchase share repurchases, and cash tender offers in connection with acquisitions-divihas increased in dends remain the primary mechanism for such distribution but have diminished importance relative relative to the other two. Stock repurchase often is used as part of an overall corpoto dividends. rate restructuring, a topic taken up in Chapter