University of Greenwich Business School MSc in Finance & Financial Information Systems

Transcription

1 University of Greenwich Business School MSc in Finance & Financial Information Systems Title of the Dissertation: DO DIVIDEND ANNOUNCEMENTS AFFECT THE STOCK PRICES IN THE GREEK STOCK MARKET? STUDENT: SEREFIDIS THEOFILOS (ID Number: ) SUPERVISOR: Dr. S. Athianos September 2009

2 Section TABLE OF CONTENTS page Abstract Introduction Background information on the Athens Stock Exchange and the Dividends Regime in Greece Theories about the relevance between corporate dividend policy and shareholders wealth The Middle of the Road The Conservative Rightists The Radical Leftists Dividends and Taxes: Clientele effect The informational content of dividends and their signals to capital markets Behavioural rationality of investors: the case of dividends Dividend announcements and the Efficient Market Hypothesis Evidence of the Signalling Effect from different capital markets globally US Capital Market European Capital Markets Asian Capital Markets Other Capital Markets the Greek Capital Market Research Design and applied Methodology Sample and Data Selection Hypothesis Development Constructing the Event Study Window Raw Return Measures Abnormal Return Measures Empirical Results Conclusions

3 References

4 DO DIVIDEND ANNOUNCEMENTS AFFECT THE STOCK PRICES IN THE GREEK STOCK MARKET? Serefidis Theofilos (ID Number: ) University of Greenwich, MSc in Finance & Financial Information Systems Abstract The reaction of the stock markets to dividend announcements by firms is a financial issue, which has caused an intense debate between the academics, the corporate officials and the shareholders of many companies for several years. The current study attempts to examine the reaction of the Athens Stock Exchange (ASE) to dividend announcements by a sample of firms negotiating on the indices FTSE/ATHEX 20 and FTSE/ATHEX Mid 40 for a fixed period It also provides analytical information about the Greek Stock Market and the regulations that underlie it, which have been taken into account in the present thesis. Moreover, previous studies of important academic scholars are presented and discussed, in order for the reader to attain the appropriate theoretical knowledge about the examined issue. Finally, significant abnormal activity is documented throughout the multiple event-windows that are employed and therefore, the null hypothesis, which supports the irrelevance theory as introduced by Miller and Modigliani (1961), is rejected. Keywords: Dividend announcements, stock prices, abnormal activity, Athens Stock Exchange, signalling effect. 3

5 1. Introduction There are numerous researchers, during the last decades, that have been concerned in their papers the impact of the dividend announcements on the stock prices. However, it is a matter of intense debate for the academics, the managers and the shareholders of many companies for several years. The theories that have been introduced by significant academics were essentially unable to terminate the above mentioned debate, as the empirical results of various studies, in the most important stock exchanges globally, concluded to different outcomes, supporting different theories. The main issue that occupied the mind of the financial economists was the corporate dividend policy and how if it affects the firm value and thus, the shareholders wealth, as well as the existence of an optimal corporate dividend policy. Lintner (1956) is considered to be a pioneer in the research of the relevance between dividend policy and firm value. According to Lintner (1956), under the assumption that capital markets are imperfect, the firms dividend policy plays a prominent role in managements decision making and hence, in shareholders wealth. He claimed that changes in corporate dividend policy may convey information to the market about company s current and future financial position; given that there are information asymmetries between managers and investors (the former have information advantage over the investors). Therefore, Lintner suggested that increases in the amount of dividends that companies distribute to their shareholders lead to a positive market reaction, while decreases in the amount of dividends lead to a negative reaction of the stock prices. Similar outcomes about the reaction of the market to changes in corporate dividend policy have been resulted from other important researchers as well (Walter, 1956; Gordon, 1959; Gordon, 1962). On the other hand, Miller and Modigliani (1961) postulated in their land-mark study the irrelevance between the dividend policy a firm adopts and the value of the firm. In particular, they argued that under the assumption that perfect capital markets with perfect certainty, no taxes and transaction costs exist, dividend policy does not have any impact on the shareholders wealth. Indeed, they suggested that managers can affect the firm value only by changing the firm s investment policy. Finally, a last group of researchers, the most important of which were Brennan (1970) and Brennan and Thakor (1990), declared that the corporate dividend policy is relevant as well as crucial to the value of 4

6 a corporation. Nevertheless, this group of academics claimed that an increase in dividends has a negative effect on the stock prices due to the existence of taxation. According to Brennan and Thakor (1990), most of a company s shareholders prefer dividend payments when distributions are small, while they prefer tender offer stock repurchases when distributions are quite larger. The level of taxation seems to affect stock prices considerably in many stock exchanges all over the world. The taxation issue and the clientele effect will be discussed in a distinct chapter of the current dissertation. The aim of the current thesis event study is to investigate the reaction of the stock prices to the announcements of dividends by Greek listed companies. It applied the classical event study methodology used by Brown and Warner (1985), in order to measure the abnormal returns of companies stock prices that occurred during a fixed time period, before and after the day of the announcement (event day). The sample consists of Greek firms listed on the Athens Stock Exchange (ASE), and more particularly negotiating on the indices FTSE/ATHEX 20 and FTSE/ATHEX Mid 40. FTSE/ATHEX 20 is an index dealing with the 20 largest blue chip companies on the ASE, in terms of total capitalization, while FTSE/ATHEX Mid 40 is an index comprises the next 40 listed companies that are classified as medium capitalization companies on the ASE. The selected period refers to the 1 st of January 2004 until the 31 st of December It is a considerably important study, as it examines the dividend signaling hypothesis in a capital market that is quite different from other capital markets, in market capitalization, in the number and the size of listed companies and in the regulations underlying it. Analytical information about the Athens Stock Exchange as well as its special characteristics, having regard on the amount of dividends that companies distribute, will be discussed in the next part of the current study. The rest of the dissertation is organized as follows. Chapter 2 introduces an analytical background of the Athens Stock Exchange (ASE) and the regulations that underlie it, in regard to the dividend amounts distributed by the listed companies. Chapter 3 presents the most important theories that have been developed about the relevance among the corporate dividend policy and the firm value. Moreover, chapter 4 provides extensively the clientele effect, the theory which assumes that stock prices move accordingly to the aims and the demands of investors, in reaction to a corporate policy and the taxation of dividend gains. Furthermore, chapter 5 discusses 5

7 the informational content of dividends and the signals they convey to the capital markets. The behavioural rationality of investors is discussed in chapter 6, as it is considered to be a very important issue, which has stimulated the interest of the researchers for several years. Chapter 7 relates dividend announcements with the efficient market hypothesis (EMH), an issue that firstly introduced by Fama (1965). Finally, before presenting the methodology applied in the current event study, chapter 8 provides some of the most important empirical findings of various researches on the largest capital markets globally. In chapter 9 the applied methodology is presented, as well as the research design of the present event study. Ultimately, chapter 10 demonstrates extensively the empirical findings of the study, in order to support or to reject the developed null hypothesis. 2. Background information on the Athens Stock Exchange and the Dividends Regime in Greece The Athens Stock Exchange (ASE) is considered to be a medium-sized European capital market, quite smaller than the others in Europe, USA and Asia, in terms of market capitalization, number of listed companies and trading volume. The number of listed firms on the Athens Stock Exchange has reached the 350 at the beginning of Its market capitalization witnessed an extraordinary upward slope after 1995, having a remarkable growth since 1999, when the general index (ASEI) of the ASE raised at historical levels. At this time, a lot of individuals invested their savings on shares of stocks. However, a striking decline on the prices that began in 1999 and lasted until the middle of 2003, turned the dream of thousands of investors into a nightmare (Dasilas, 2007). As it was stated in chapter one, the current thesis aims to investigate the reaction of the stock prices to the announcements of dividends by the Greek listed companies negotiating on the indices FTSE/ATHEX 20 and FTSE/ATHEX Mid 40. Although the Athens Stock Exchange is an emerging medium-sized stock market, it seems to be an ideal candidate for this kind of study, due to some characteristics that other stock exchanges do not have. First of all, companies that have their shares listed on the Athens Stock Exchange distribute dividends to their shareholders on a yearly basis, contrary to the 6

8 common practice of firms listed on many other developed stock exchanges worldwide. Secondly, the majority of the companies in Greece are individual-family owned companies, and hence, they have a more concentrated ownership structure and less information asymmetries between owners and managers. Therefore, the agency costs are generally lower in Greek firms and thus, there is no motivation for the board of directors to use the corporate dividend policy as a mechanism for monitoring the managers, in order to avoid wasting capital in low-return investments (Jensen, 1986). Thirdly, in Greece it is determined precisely by the law the minimum percentage of net profits that companies have to distribute as dividends to their shareholders. Anastasiou (2007) implied in his study that dividends in Greece are proportional to earnings. In particular, companies have to distribute as cash dividends the larger amount between the 6% of their equity and the 35% of their net profits (Law 2190/1920). Nonetheless, the general assembly of shareholders can require the lower amount to be distributed, taken for granted that the difference will be capitalized and distributed as stock dividends in a subsequent financial year (Asimakopoulos et al., 2007). Finally, according to the Greek Law (2065/1992), dividends are not taxed both at personal and at company level. Dividends are taxed only at personal level and capital gains are not taxed at all. Therefore, the tax system in Greece does not impose the double taxation on dividends that it is imposed on many of the biggest stock exchanges all over the world, such as on the US capital market. Only a tax of 0.3% is imposed on every transaction on the Athens Stock Exchange (Dasilas et al., 2008). The above unique characteristic of the Greek Stock Market makes it quite attractive to potential investors and an ideal case of study, as it allows examining the market reaction to dividend announcements without considering the effects of double taxation. 3. Theories about the relevance between corporate dividend policy and shareholders wealth The possible impact of corporate dividend policy on the price of a company s shares and, therefore on the shareholders wealth, is a significant issue of the financial management that has puzzled the financial researchers throughout the last decades. This issue is crucial not only for the corporate managers that seek to find the optimal 7

9 payout policy if exists for their companies, but also for the shareholders and the potential investors that seek to find whether the decision of the corporate dividend policy affects the current price of their shares. In order to interpret this matter and answer to some basic questions that had been arisen, copiousness of theories has been developed during the last fifty years. The current part of the present thesis aims to introduce the most important of these theories about the firm s dividend policy and their empirical results, which keep on influencing the economists until today. It aims for the reader to fully understand the theoretical basis of dividend policy, in order to present on subsequent chapters the meaning of information content hypothesis and the dividends signaling effect. Brealey et al. (2006) suggested that investors can use dividends, in order to separate a profitable company from a troubled one. They believe that generous firms, in terms of cash dividends distribution, have reported high earnings at the end of the respective financial period. As a result, investors may use dividend policy, in order to learn about firms future prospects or other information that is available only to corporate officials. Brealey et al. (2006) also claimed that the information that dividends convey to the market is fairly reliable, as firms can overstate their earnings, to enhance their financial position, only for a short period of time and not in the long run. To recapitulate, changes in corporate dividend policy can help market participants to make presumptions about the firms future prospects. The question that arises, however, is whether these changes in the payout policy can affect the value of a firm or is just a signal of its stocks intrinsic value. The academic scholars that tried to answer the previous question can be classified into three different groups, according to the results of their researches. The first group, the so-called rightists, asserts that increases in the payout ratio have a positive effect on the firm value. On the other side, the so-called leftists declare that an increase in the dividend payments lead to a decline in the stock prices, due to taxation issues. This group believes that investors preferences do not centre upon high cash dividends, as they are imposed to higher taxes, compared to capital gains. Therefore, they prefer firms to increase their plowback ratios and retain their earnings, which are imposed to lower taxation. The third party of theoreticians, who are placed in the middle of the road, asserts that there is no relevance between the corporate dividend policy and the firm value. Thus, 8

10 changes positive or negative in the amount of dividends can not affect shareholders wealth. The figure below summarizes the three different groups of researchers, on how changes in dividend policy affect the firm value. 3.1 The Middle of the Road This group of theoreticians supporters of the irrelevance theory is mainly represented by great academics, such as Modigliani, Miller, Black and Scholes, the so-called middle of the roaders. Modigliani and Miller (M-M) had a profound influence in the corporate capital structure theory, since they published their two landmark studies (Miller & Modigliani, 1958; 1961). Their first study (1958), based on the assumptions of the perfect capital markets and the inexistence of taxation, concluded that the firm value is not affected by the choice of the firm s capital structure. Thus, the firm value is irrelevant to the debt to equity ratio, which a firm selects to adopt. At this point of the study it is worth to present the assumptions that exist in the ideal world of Modigliani and Miller. First of all, companies operate in perfect capital markets. More particularly, in perfect capital markets all investors are rational; the information is symmetrical and available to all market participants. Therefore, everyone who participates in the capital markets has access to the same information and shares the same expectations about the future prospects of the 9

11 companies. Furthermore, in perfect capital markets there are no transaction costs for the buy or the sale of securities and no individual investor has the power to influence the stock prices by buying or selling securities. Additionally, the companies net profits are not imposed to any taxation. Moreover, companies are classified in risk classes according to their business risk, and finally, the individual and institutional investors can borrow and lend capital at the same interest rate that exists in the market place, the risk-free interest rate. Taken all the above into account, Modigliani and Miller (1958) revealed their Proposition I: the value of the (i) company does not depend on its capital structure and it is determined by the expected future return on assets (EBITi), discounted by a rate of return (e) that is required by the shareholders and is the appropriate for the risk class, which the company belongs. The mathematical expression of the above proposition is depicted on the following formula. Firm Value 1 Debt Equity (EBIT ( )) e, Source: Miller, M., Modigliani, F., (1958), The Cost of Capital, Corporation Finance and the Theory of Investment, The American Economic Review, Vol. 48, No. 3, pp As a result, any increase in the firm s debt is immediately offset by the increase in the required rate of return by the shareholders, due to an increase in the amount of the undertaken risk (M-M s Proposition II). Later on, in 1961 Modigliani and Miller stated in their ground-breaking study that a firm s value is not affected in any way by the adopted dividend policy by the firm. They challenged the validity of the existence of an optimal payout ratio, by suggesting that such a ratio if exists would be adopted by the managers of every company, in order to maximize its stock value and hence, the shareholders wealth. On the other hand, they declared that the amounts of dividends a company distribute is where they are, in order to gratify their shareholders and because the corporate officials do not believe that they can add to shareholders wealth by increasing or decreasing the payout ratio. Instead, Modigliani and Miller pointed out in this article that corporate managers can affect the firm value only by changing its investment 10

12 policy, which is considered to be its only determinant. These two studies by Miller and Modigliani are extremely significant to financial theory, as they have influenced several researchers later. Another, essential study about the relationship between the dividend policy of a firm and its common stock prices was accomplished by Black and Scholes (1974), which challenged openly the bird in the hand theory by Graham and Dodd (1951) that will be discussed in the next part of this thesis. They supported the irrelevance theory demonstrated by Miller and Modigliani, as they found no significant relationship between the corporate dividend policy and the shareholders wealth. They argued that the most correct manner to examine the impact of the dividend policy on the firm value is to determine the effects of dividend yield on stock returns (Black & Scholes, 1974). Finally, Miller and Scholes (1982) provided empirical evidence about the irrelevance between the dividend policy and the firm value that proposed by Modigliani and Miller, by assuming, however, existence of taxation in dividends and capital gains. Nevertheless, their research will be discussed comprehensively in a subsequent part of the current thesis. 3.2 The Conservative Rightists The above stated Modigliani and Miller s argument entails that the value of any firm is irrelevant to its corporate dividend policy. By saying irrelevant they mean that all feasible payout decisions for the company are optimal (DeAngelo & DeAngelo, 2004). Therefore, corporate officials can not change the shareholders wealth by increasing or decreasing its distributed dividends. However, a few years before the publication of MM s article, in the early fifties, a more conservative group of important academic scholars, the so-called rightists, revealed their own view about the relationship between the corporate dividend policy and the firm value. This party of theoreticians believed that the firm value is positively related to the dividend payments and hence, higher dividend payments can result in an increase to the company s stock price. The first researchers that tried to demonstrate the above statement were Graham and Dodd (1951). According to them, the only reason of companies existence is to pay dividends and thus, individual investors who evaluate and plan the purchase of common stocks should take into account that companies with liberal dividends are better than those with niggardly ones (Graham & Dodd, 1951). In 11

13 addition, institutional and individual investors seem to prefer stocks that pay high dividends, because they consider dividends as immediate spendable gain, while capital gains are considered as additions to the existent capital. Graham and Dodd developed the following traditional model, as a standard practice for investors to evaluate any common stock. P M (D 1 E) 3, Source: Graham and Dodd (1951), as found at Walter, J., (1956), Dividend policies and common stock prices, The Journal of Finance, Vol. 11, No. 1, p. 36. where M is the multiplier, D is the expected dividend per ordinary share and E is expected earnings per ordinary share. Two years later, Harkavy (1953) demonstrated empirically that common stock prices move at the same direction with the proportion of earnings distributed as dividends. Therefore, all other things been equal, investors would pay a higher price to purchase a stock that pays higher dividends from another that pays lower ones. On the other hand, in the same article Harkavy (1953) implied that in the long run, companies that distribute lower percentage of their earnings, having thus higher plowback ratios, demonstrate finally the higher increases in their stock prices. This statement applies in cases of small and developing companies that retain the higher proportion of their earnings, in order to finance their expansion. However, it does not apply at all events, as distributing low amounts of dividends and retaining high proportion of earnings does not ensure a remarkable enhancement on the stock prices in the future. Based on Harkavy s conclusion about the relationship between the common stock prices and the proportion of retained earnings, James Walter (1956) is another significant academic scholar belonging at the rightists party, who attempted to develop a theoretical model, in order to explain the relationship between the different corporate dividend policies and the firm value. His basic assertion, upon which he based his theoretical model, is that in the long run the present values of expected 12

14 dividends are reflected on the stock prices. The Walter s Model for the stock price evaluation is illustrated below in a quantitative manner. P D Ra (E D) Rc Rc, Source: Walter, J., (1956), Dividend policies and common stock prices, The Journal of Finance, Vol. 11, No. 1, p. 32. where D is cash dividend per ordinary share, E is earnings per ordinary share, Ra is the internal rate of return on the investments and Rc is the market capitalization rate. Walter s Model was a revolution in financial theory and one of the most important equity valuation models. By using the above equation Walter (1956) investigated not only the relationship between the stock prices and the dividend payout ratio, but also the relationship between the internal rate of return on additional investments (Ra) and the market capitalization rate (cost of capital - Rc). However, the former relationship is highly related with the latter one. Walter (1956) classified the companies stocks as growth, intermediate and creditor, according to the relationship between the rate of return and the cost of capital. According to his assertion, when the internal rate of return is greater than the market capitalization rate (Ra > Rc), then the stock price increases as the dividend payout ratio declines. Therefore, corporate officials could maximize shareholders wealth only by retaining the earnings and reinvesting them in the company, rather than by distributing them to the shareholders. This is the case of growing companies, which have a lot of lucrative investment opportunities (NPV>0) and have immediate needs of available cash. If such a company distributes the earnings as dividends, it relinquishes lucrative projects (Ra) and let the shareholders to invest the dividends in investments with lower rate of return (Rc). The optimal payout ratio in this case is equal to nil and all earnings have to be reinvested in the company. Another category of stocks, according to Walter (1956), are the intermediate stocks. Companies with intermediate stocks are characterized by medium to high dividend ratios. In this case the internal rate of return is lower than the market 13

15 capitalization rate (Ra < Rc) and thus, the stock prices increase as the payout ratio increases as well. In such a case it would be irrational for a company to retain its earnings, without distributing them to the shareholders, as they have more profitable investment opportunities (Rc) than the company has (Ra). The optimal payout ratio for intermediate companies is 100%. Finally, creditor stocks are those that are characterized by fixed dividend payout ratios. In this case the internal rate of return is equal to the market capitalization rate (Ra = Rc) and hence, the payout decision makes no difference to the firm value. This case is accordant to the irrelevance theory suggested by Modigliani and Miller (1961) and no payout ratio appears to be optimal for the firms. The following figure summarizes the relation between dividend per share and period of time, suggested by Harkavy (1953) and Walter (1956), illustrating the dividend growth for two earnings reinvestment policies. Figure 1. Source: Bodie, Z., Kane, A., Marcus, A., (2009), Investments, McGraw-Hill International Edition, New York, p Continuing presenting, in a chronological order, the results of the empirical researches and the theoretical studies, Gordon and Lintner were another two important theoreticians belonging in the conservative rightists group, who concerned in their studies the relationship between the corporate dividend policy and the shareholders wealth. Firstly, Gordon (1959) implied that an investor by purchasing a common 14

16 stock, she actually buys its future price, hoping to be higher than its current one. He also provided empirical evidence that its future price is related someway with the expected dividend payouts and the expected earnings of the company. Therefore, he demonstrated the relevance between the dividend policy and the stock prices. However, he failed to fabricate at that time a theoretical model that links stock prices and dividends, in order to support the results of his research. In the early sixties, it was Gordon (1962) and Lintner (1962) who proposed the so-called bird in the hand argument that came to be one of the most momentous propositions of the rightists party. Gordon (1962) accomplished the construction of a theoretical model that, contrary to Walter s Model (1956), took into account the risk incorporated in the stocks. He stated that current dividends are considered by the investors to be less risky than the expected dividends and the future capital gains and due to the low uncertainty, current dividends are discounted in the present with a lower discount factor resulting in a higher stock value. The formula below illustrates quantitatively the theoretical model developed by Gordon (1962). P E (1- b) k - b r, Source: Gordon, M., J., (1962), The Savings Investment and Valuation of a Corporation, The Review of Economics and Statistics, Vol. 44, No. 1, pp. 38. where P is the stock price, E indicates the earnings per ordinary share, b is the plowback ratio and hence, E * (1-b) = D, where D is the dividend per share, b * r = g is the growth rate, and k is the rate of return required by the shareholders. Thus, the stock value is equal to the ratio of current dividend and the difference between the rate of return required by the shareholders and the growth rate of dividends. It is quite straightforward from the above equation that Gordon (1962) based his Model on Walter s Model (1956). Additionally, he took similar assumptions to Walters, in order to construct his model. He assumed that companies can finance their operations by using only their retained earnings. However, he dissociated his 15

17 model from Walters, as he assumed that companies can also use a stable debt to equity ratio. Furthermore, he assumed that some of his model variables, such as b, g and k would not be changed by the corporation in every future financial period. His final assumption was that the model is valid only when the rate of return required by the shareholders (k) is higher than the dividends growth rate (g). If future dividends are expected to have a growth rate higher than the firm s cost of capital, then the stock price would tend to be infinite. To conclude, the group of rightists comprises of major theoreticians who suggested that there is a positive relationship between the corporate dividend policy and the firm value, and constructed models that despite their shortcomings they have managed to shed some light to the deep tunnel of the dividend puzzle. 3.3 The Radical Leftists Contrary to what Modigliani and Miller (1961) and Black and Scholes (1974) have said in their seminal works about the irrelevance theory, the group of radical leftists, supporting the relevance theory, argued for a negative relationship between the corporate dividend policy and the firm value, due to taxation issues. In particular, they stated that whenever dividend payments are taxed more heavily than capital gains, it is more beneficial for companies to transmute dividends into capital gains, while the existing profit has to be either retained, or to be used in share repurchases. Therefore, companies should adopt a very low payout ratio, in order to pay low cash dividends, which results in lower tax payments. As regards to investors, they would be satisfied to pay lower taxes, as in cases where dividends are taxed more heavily than capital gains, investor seek to purchase stocks with low dividend yields (Brealey et al., 2006). The pioneer of this group of theoreticians and the most important of them is Brennan (1970). He is considered to be a pioneer, as he was first developed the After- Tax Capital Asser pricing Model, applied for differential corporate and personal taxes (Litzenberger & Ramaswamy, 1980). The assumptions that Brennan (1970) based his model are of great importance. He assumed that corporations and individual investors can borrow and lend capital illimitably at the same interest rate which is equal to the risk free rate of interest. Additionally, he assumed no restrictions on short sales and that dividend payments are known precisely by the investors (Litzenberger & 16

18 Ramaswamy, 1982). Thus, the mathematical relationship that derives from the above mentioned statements is the following. E( Rt r f) b c0( d i f), 0 r Source: Litzenberger, R., Ramaswamy, K., (1982), The Effects of Dividends on Common Stock Prices Tax Effects or Information Effects?, The Journal of Finance, Vol. 37, No. 2, p where Rt is total rate of return on asset i before taxes, βi is the systematic risk, di is the dividend yield and rf is the risk free interest rate. Brennan (1970) used this model in his study to result that any increase in dividends causes a decrease in firm stock prices, as investors do not have preference on dividend payments, except for the case that taxes are inexistent. Later on, Litzenberger and Ramaswamy (1979) developed a more extended version of Brennan s model by adding some constrains on unlimited borrowing by individuals and corporations with risk-free rate. However, all other things that Brennan assumed in his model were taken as given. The mathematical approach of Litzenberger and Ramaswamy s (1979) Model is the following: E( Rt r f) b1 c1( i d, 1 i r f ) Source: Litzenberger, R., Ramaswamy, K., (1982), The Effects of Dividends on Common Stock Prices Tax Effects or Information Effects?, The Journal of Finance, Vol. 37, No. 2, p where α 1 > 0 is the risk premium in a portfolio that has a beta equal to nil and a dividend yield equal to the risk-free interest rate. The results of their research are compatible with Brennan s ones, although based on their altered assumptions about unlimited borrowing constrains (Litzenberger & Ramaswamy, 1982). Nevertheless, Litzenberger and Ramaswamy (1980; 1982) on their subsequent articles argued that when there are restrictions on sales and the corporate dividend 17

19 policy is precisely known by the investors, there is a positive but not linear relationship between the firm value and the dividend yields. Dividends can be forecasted only by using information that is available to investors in advance. It seems that there is no unambiguous conclusion derived from the above discussed theories about the relationship among corporate dividend policy and firm value. Far from it, the supporters of the three most important groups result in clearly contradicting theoretical and empirical outcomes. Several recent researchers have attempted to shed some light to the dividend controversy without giving, however, any remarkable explanation. One of the academics major concerns all these years was the role that taxation issues play in the decision of dividend policy by the corporations. The next chapter of the current thesis deals with the relationship between dividends and taxes and the so-called clientele effect, by presenting the most significant studies, which have dealt with this subject. 4. Dividends and Taxes: Clientele effect There are several financial academic scholars who dealt in the past, even in separate studies, with the impact that taxes imposed on dividends have on the stock prices and whether they can modify shareholders and investors preferences on stock purchases. However, it seems that there is no coincidence of views even in this topic, as the researchers that concerned with the clientele effect belong equally in the three parties of theoreticians noticed above. Although the economists who were involved in this matter belong in the three groups as well, their studies are presented in the current chapter following a chronological order. Modigliani and Miller (1961) are considered to be the first economists who attempted to provide some insight in this matter. It is already known that both authors are supporters of the irrelevance theory middle of the road and stated that the value of a firm is not dependent in any way to its dividend policy. In order to corroborate the validity of their theory, they assumed that capital markets are perfect (costless access to information, no brokerage and transaction costs and no tax differentials between dividends and capital gains), investors act rationally and are perfectly certain about the future dividend policy of the firms. However, according to Modigliani and Miller (1961), the only assumption that would barely diverge from their theory is that there are no personal income taxation differentials between dividends and capital gains. They suggested that in capital markets, where 18

20 imperfections at regard of taxes imposed on dividend and capital gains exist, investors who behave rationally prefer to purchase stocks with high or low distributions with reference to whether capital gains are taxed more heavily than dividends respectively. Thus, the appearance of the clientele effect comes to be true. In addition to the tax differentials between dividends and capital gains, market imperfections comprise the existence of considerable transaction and brokerage fees, as well as gaps between the interest rates. These imperfections are certainly not consistent with the irrelevance theory suggested by Miller and Modigliani (1961). Clientele theory claims that there is an inclination of dissimilar types of securities to allure different types of investors, according to the dividend policy adopted by each corporation. Hence, investors are not willing to purchase stocks with low dividend payments, if capital gains are taxed more heavily than gains accruing from dividends and when a company change its dividend policy, investors transform their stock portfolio accordingly. Few years later, Elton and Gruber (1970) made a quite important research, resulting to outcomes consistent with Modigliani and Miller s. In particularly, they attempted fruitfully to determine the marginal shareholders tax brackets and their relation if any with the corporate dividend policy, an issue that had not been investigated till then, as well as the validity of the clientele effect, proposed by Modigliani and Miller (1961). Their results demonstrated that shareholders tax brackets are indisputably related with firm s dividend payments, and consequently, changes in dividend policy lead to changes in the firm value. Corporations having high payout ratios would attract risk-averse investors from a comparatively low tax bracket, who prefer a stable cash income from dividend payments. On the other hand, growth companies, which need cash to finance their expansion in fruitful investment opportunities, usually adopt more limited payout policies and attract investors who belong in the high tax brackets, do not have immediate cash requirements, and hence prefer capital gains over dividends. Correspondingly, Black and Scholes (1974) supported in their article the clientele effect. However, they postulated that the expected returns of high yield ordinary stocks are the same with the expected returns of low yield ones, even whether taxation exists. An investor, who has imposition of taxes, certainly adjusts her portfolio to low-yield stocks and a tax exempt investor adjusts his portfolio to high-yield securities. Nevertheless, both of them are not in the position to demonstrate 19

21 that these prompt adjustments add value on their portfolios. Therefore, the independence between the dividend policy and the firm s stock price is undoubtedly supported. Miller and Scholes (1978; 1982) published two successive studies in an intense endeavour to examine the relationship between dividend policy and firm value, under the existence of tax differentials between dividends and capital gains. They concurred with Modigliani and Miller s (1961) proposition that only investment policy can affect the firm value, while changes in dividend policy can not provide added value to shareholders wealth (Miller & Scholes, 1978; as found at Miller & Scholes, 1982). In their subsequent research they attempted to use measures, which subtract the tax burden from stocks that distribute dividends, in order to minimise in the short-run the differences form the stocks that do not pay dividends (growth stocks). However, they stipulated that such measures were not suitable for this reason, as they provide biased results in the short-run, while in the long-run taxes do not affect investors preferences for dividend policy, and hence irrelevance theory stands true (Miller & Scholes, 1982). Conversely, there were other researchers who investigated the impact that dividend taxation has on the firm value, but their results were quite different from the above mentioned results, since the former researchers belong either in the group of rightists or in the group of leftists. Masulis and Trueman (1988) examined how differential personal taxation affects the dividend and the investment policy of a firm. They suggested that differential personal taxation can be harmful even for a company, as shareholders who are differently imposed it have deviated opinions about the optimal use of the internal funds in investment and dividend policies. More particularly, shareholders in low tax brackets show a preference for high dividend payouts and less reinvestment of the internal funds and vice versa for high tax bracket shareholders. However, their model assumes that shareholders disagreement exists when companies are not permitted to invest their internal funds in securities, but any deviation of their views is shrunk when companies can invest internal capital in securities. In chapter three it was cited that Brennan (1970) was one of the most important theoreticians and a pioneer of the radical leftists group. In a more recent article of him about shareholders preferences among different cash distributions by firms, Brennan and Thakor (1990) took into account the impact of taxation on realized 20

22 investors gains and implied that in spite of the differential taxation between dividends and capital gains, shareholders, in an attempt to maximize their wealth, tend to prefer cash dividends in cases of small distributions. For medium-sized distributions they prefer shares repurchases by the company, while for large ones they desire tender offer repurchases. Moreover, if the actual personal dividend tax is low, then shareholders with small stakes of ownership would prefer cash dividends, while shareholders with substantial participation in the firm s ownership structure would prefer a share repurchase. In an attempt to scrutinize the dividend puzzle under the existence of taxation, Bernheim (1991), a year later, published his article based on a model in which dividend payments convey information to shareholders about firm s profitability. He assumed that although dividends are taxed more heavily, they are indistinguishable with share repurchases. His results, which were consistent with the irrelevance theory, revealed that the higher the dividend taxes, the lower the amount available for distribution, without affecting the firm value at all. Allen et al. (2000) questioned why companies continue to distribute dividends, despite the fact that dividends are taxed more heavily than other forms of cash distribution. In order to answer the preceding question, Allen et al. (2000) developed a model, which classified investors as institutional and individuals and by taking into consideration the ownership clientele effect, it attempted to investigate investors preferences. They stated that institutional investors have an indisputable preference to dividends, and hence they are attracted by companies with high payout ratios. Additionally, Allen et al. (2000) took in account the agency theory as well, by suggesting that firms distribute dividends, in order to attract institutional investors. Agency theory implies that institutional investors have generally greater incentives to monitor the top managers, than the diffused small individual investors. Therefore, this kind of ownership might have good effects on the financial performance of a company (Barako & Tower 2007), and companies would like to magnetize this kind of investors by distributing dividends, in order to sign a positive image outside. To summarize, most of the above mentioned theories concluded that although taxed dividends are in the position to change investors preferences, they can not remarkably modify the firm s stock prices. However, recent empirical findings (mentioned in a subsequent chapter of the current study) report different results. Anyway, as it was discussed in chapter one, dividends are not imposed to double 21

23 taxation in Greece, and thus it is assumed in the present event study that they would not affect investors preferences. 5. The informational content of dividends and their signals to capital markets In this part of the study it is worth to discuss about the various information conveyed by dividends to market participants, the so-called information content hypothesis, as well as the dividend signalling effect, which are extensively cited in the financial literature. Modigliani and Miller (1961), in an attempt to bring together the irrelevance proposition of their ideal world with the dividend policy, under circumstances of uncertainty, introduced the informational content of dividends. They suggested that dividends provide useful information about managers views of company s future profitability prospects. However, they also stated that a change in the dividend policy is only the handle for a change on the stock prices and not its cause. He was Bhattacharya (1979) who reconciled in his study the Modigliani and Miller s information content hypothesis with the signaling effect. He assumed existence of information asymmetries between outside investors and corporate officials. In particular, he stated that managers have inside information that is not available to outside investors and can be important if firm s current investments would have positive impact on its value. His model revealed that, in this case, dividends serve as signals to investors about firms expected profitability. Taking into consideration the valuable inside information about firms future plans possessed by managers, Aharony and Swary (1980) named dividend announcements as one of the various mechanisms that managers incorporate, in order to signal information to market participants. According to the same authors, another significant mechanism is the earnings figures. However, dividends offer more dubious information than earnings, as their distribution is in managers discretion and can be quite effortlessly manipulated by them. Numerous other researchers suggested that dividends convey a substantial amount of information to markets, when changes in dividend policies are observed. Lintner (1962) provided empirical evidence that, in any case, managers in large firms are likely to increase the payout ratios whenever they are convinced that future cash inflows would be adequate to support the payments, and they decrease dividends when they are not confident about the expected cash inflows. Watts (1973) applied 22

24 annual data to test the information content hypothesis, resulting to a positive, however trivial, relationship between changes in future earnings and unanticipated changes in dividends presently. On the other hand, Pettit (1972; 1976) used quarterly data to conclude that the intrinsic value of any security can be accurately assessed by evaluating the changes in dividend policies. His results are consorted with Lintner s proposition that managers hesitate to increase dividend payments before being certain that the amount of expected future cash inflows would be sufficient to cover the dividend distributions. Definitely, the majority of financial researches concerning with this issue conclude that dividend changes convey information to market participants, sometimes beyond the information that has been already available by earnings figures. Nevertheless, the main dispute between the academic scholars is whether market participants comprehend the dividend signals and accordingly adjust their portfolios. As noticed above, the group of rightists assumes a positive relationship between changes in dividends and investors reaction, while leftists take for granted that investors have a negative reaction to dividend changes, and finally, the middle of the roaders presume no relevance between them. 6. Behavioural rationality of investors: the case of dividends In the previous part, it was mentioned that one of the main disputes of the researchers, concerning the information content hypothesis, is whether market participants comprehend the dividend signals and adjust their portfolios accordingly. Unquestionably, it is an issue that depends on the behavioural rationality of investors, which has been occupied the mind of many researchers until nowadays. Miller and Modigliani (1961) first introduced the concept of rational investors, when they assumed in their seminal study that investors behave rationally when they function in capital markets. Particularly, they stated that rational investors are those who prefer to maximize their wealth than minimizing it, either by receiving high cash dividends or by enjoying high capital gains. Miller (1986) developed an important study, in order to demonstrate that rationality-based models are as significant as the other economical and financial models. After classifying investors in institutional - large investors - and individual - small investors -, he revealed that the behavioural elements in decisions concerning dividend payments differ between institutional and individual investors. More specifically, unlike institutional, individual investors carry 23

25 small amounts of securities and the majority of them do not count the experts advises when they function in capital markets. He based his assumption in the fact that individual investors do not view the purchase of securities solely as an immediate way to maximize their wealth, but they relate securities with their families, their business and other factors that financial models can not take into account. Alternatively, he was Gordon (1959) who first established the so-called bird in the hand argument. Gordon (1959) allied with Modigliani and Miller s assumption that investors behave rationally in capital markets, however, he implied that investors generally show a preference in current cash dividends, as they consider being less risky than future capital gains. Therefore, he dissociated his proposition from Modigliani and Miller s, suggesting that investors show a preference in the source they use to maximize their wealth. He finally concluded that due to low uncertainty that dividends involve, companies which distribute a high proportion of their earnings tend to enjoy higher increases in their stock prices than those who adopt more restricted dividend policies. However, the bird in the hand theory has been challenged by several researchers. One of those was Bhattacharya (1979), who called Gordon s argument as the bird in the hand fallacy. According to him, the low uncertainty of current cash dividends is not a so important reason to convince the investors to show a preference in current dividends against the future capital gains. Consequently, it is rather straightforward that the behavioral rationality of investors is another issue of debate between the financial scholars and there is no room for convergent views. 7. Dividend announcements and the Efficient Market Hypothesis The efficient market hypothesis is an issue that firstly developed by Fama (1965) and it has been the source for an incessant dispute in economic and academic circles between its proponents and its rivals. Fama (1965) asserted that markets are efficient when the current market prices of the securities reflect, without bias, all the publicly available information. As a result, market prices reflect only information that is available currently and not information that was accessible in previous economic periods. In addition, it is impossible for any investor to use the publicly available information, in order to outperform the market and to enjoy excess profits repetitively, but only randomly. Moreover, Fama (1965) rejected the proposition that future security prices can be predicted using past information and concur the theory of 24