University of Greenwich. Msc in Finance and Financial Information Systems

Transcription

1 University of Greenwich Msc in Finance and Financial Information Systems TSINANI V. ALEXANDRA WHY GREEK INDIVIDUAL INVESTORS WANT DIVIDENDS? ACKNOWLEDGEMENTS First of all I would like to thank my supervisor and instructor, professor of T.E.I. of Kavala Dimitrios Madytinos who helped me in achieving this effort. Additionally I would like to thank my professor in Statistics Efstathios Dimitriadis, who offered me some very useful insight regarding the analysis of the survey s results. Finally I could not forget to thank my family and friends who encouraged me at every single step of this effort. 1

2 ABSTRACT The assumption of why individual investors want dividends is investigated by surveying Greek retail investors in Eastern Macedonia and Thrace, Central and Western Macedonia. The respondents strongly indicated that they want dividends. Their answers provided strong support for the signaling theory of Bhattacharya (1979). Greek individual investors are inconsistent with the uncertainty resolution theory of Gordon (1961, 1962), the agency theories of Jensen (1986) and Easterbrook (1984) and the transaction costs theory of dividends. In addition the behavioural finance theory of Shefrin and Statman (1984) is only confirmed regarding stock dividends. The tax disadvantage theory of dividends is expectedly not confirmed since dividends as well as capital gains are not being taxed in Greece at the individual level. KEYWORDS: Dividends, individual investors, dividend theories, behavioural finance, Greek tax system, questionnaire survey. 2

3 CONTENTS CHAPTER ONE INTRODUCTION 5 CHAPTER TWO LITERATURE REVIEW 2.1 Introduction Theories on why individual investors want cash dividends The dividend irrelevance theory Transaction costs Uncertainty resolution theory Accounting manipulations Behavioural finance Free cash flow Agency costs Signalling The choice between cash dividends and share buy-backs Taxes Theories on why individual investors want stock dividends Stock dividends as small stock splits Transaction costs Taxes Behavioural finance Empirical findings Summary / Conclusions 33 CHAPTER THREE RESEARCH METHODOLOGY 3.1 Introduction 35 3

4 3.2 Empirical methodologies Questionnaire survey Questionnaire design Summary / Conclusions 45 CHAPTER FOUR EMPIRICAL FINDINGS 4.1 Introduction The sample Overview of survey respondents Results for cash dividends Results for stock dividends Summary / Conclusions 64 CHAPTER FIVE CONCLUSIONS / SUGGESTIONS 66 BIBLIOGRAPHY 72 APPENDICES 81 4

5 CHAPTER ONE INTRODUCTION This survey investigates one of the most ambiguous questions in the field of finance regarding investing in stocks. Do individual investors want dividends? And if they do, why do they love dividends that much? The above questions were the basic assumptions we made when we started to investigate our field of research. Our main goal was to explore the preferences of Greek individual investors towards dividends. In order to accomplish our targets we tried to analyse the most important theories related to dividends. Testing these dividend theories provided us with the solution to the questions that were expressed above. Based on the survey of Dong et al. (2005) who investigated the Dutch individual investors preferences towards dividends, we surveyed three large regions of Greece, Eastern Macedonia and Thrace, Central and Western Macedonia. We mailed questionnaires to 150 randomly selected stock exchange companies in order to be distributed to the companies customers. The final response rate was per cent, satisfactory enough for us in order to process our work. We separated dividend theories into theories on cash and stock dividends. First of all our results indicated that the vast majority of our respondents generally want dividends. One of the theories tested is the theory of transaction costs regarding the preference of individual investors for dividends. The basic notion under this theory was the fact that dividends save transaction costs. Additionally, the uncertainty resolution theory (Gordon, 1961, 1962) was also one of the theories we tested, regarding the perception that dividend-paying stocks are less risky. The belief that dividends offer more certainty regarding a company s quality of earnings was 5

6 also discussed. The free cash flow theory of dividends (Jensen, 1986) along with the agency costs theory (Easterbrook, 1984) were presented to test whether Greek individual investors prefer dividends due to these widely known theories. In addition the signalling theory of dividends that was well discussed by Bhattacharya (1979) referred to dividends serving as signals of the future performance and earnings prospects of a company. The theory of tax disadvantage, the behavioural finance theory (Shefrin and Statman, 1984) along with the previously discussed transaction costs theory, was also under investigation in our survey regarding stock dividends too. More specifically the next chapter of our work refers to the literature review of our field of research. All previously discussed dividend theories are presented in detail in the second chapter, regarding cash and stock dividend theories separately. Chapter three refers to the research methodology that was used compared to other past research methodologies. The design and the main targets of the questionnaire are also included in the third chapter. Chapter four includes the results of our survey. Some dividend theories found strong support while others were disproved. The signalling theory of dividends was strongly confirmed in our survey. Greek individual investors firmly believe that fluctuations in the levels of dividends can signal the future performance of a company. Another theory supported and relative to the signalling theory of dividends, is the perception of Greek retail investors that stock repurchases indicate the undervaluation of a stock. Regarding stock dividends, the theories that found strong confirmation are the belief that stock dividends are nothing more than small stock splits along with the behavioural finance theory of Shefrin and Statman (1984) that investors do prefer stock dividends to no dividends at all, even though they are aware of stock dividends being no more than stock splits. On the other hand some important 6

7 theories like the tax disadvantage theory, transaction costs, accounting manipulations, free cash flow and agency costs theories as well as the uncertainty resolution theory of Gordon (1961, 1962) were not confirmed. The last chapter contains the conclusions of our survey. Our personal beliefs regarding individual investors in Greece as well as how their investment decisions are being made are presented in the fifth chapter. A mention is also being made regarding the limitations and problems we faced during the procedure of this survey. Finally some ideas are being presented for further research relative to our subject. 7

8 CHAPTER TWO LITERATURE REVIEW 2.1 Introduction Some of the most important decisions a financial manager has to undertake are the capital structure and the payout decisions. The capital structure decisions refer to the financing of a firm s operations with the adequate proportion of debt and retained earnings. On the other hand, the payout decisions refer to the choice between paying dividends and repurchasing shares. The better understanding of the payout depends on the investigation of the payout decisions of a firm in isolation from other financial decisions (see: Brealey, Myers and Allen, 2005). Our understanding of corporate dividend policy is based on the individual investors behaviour regarding their dilemma between cash and stock dividends, from the early work of Miller and Modigliani (1961) and Gordon (1961) to the recent theories of behavioural finance. Many empirical papers have substantiated corporate dividend policy by relating the policies to the theories based on individual investors behaviour. Whereas there is a general agreement that individual investors want dividends, there has been no empirical study until last year (see: Dong et al., 2005) regarding the individual investors preference over cash dividends. Dong et al. (2005) have studied the attitude of Dutch individual investors towards cash and stock dividends. Our work will therefore be based on the empirical study of Dong et al. (2005) in order to examine the behaviour of individual investors towards dividends in Greece. Miller and Modigliani (1961) showed that in a perfect and complete market where there are no taxes, no transaction costs, no asymmetric information and agency problems, dividend policy is irrelevant since individuals can overturn managers 8

9 decisions on dividend policy by buying or selling equity on their own 1. In most countries dividends have been taxed more heavily than capital gains. Thus, the irrelevance theory combined with the inevitable higher taxation of dividends makes corporate dividend policy a puzzle. According to Brealey, Myers and Allen (2005) the dividend dispute is considered to be one of the 10 insoluble problems in Finance. Grullon and Michaely (2002) have stressed the most important factors affecting dividend policy referring to the decision between cash dividends and share repurchases as the substitution hypothesis. For many years the US firms preferred cash dividends to share repurchases although capital gains were taxed advantageously relatively to ordinary income (see: Fama and French, 2001). Nevertheless, almost over the last twenty years the buy back of shares has increased significantly and according to Compustat 2 it reached 41.8 per cent in 2000 while twenty years earlier in 1980 it was only 4.8 per cent. Additionally Fama and French (2001) investigated the same phenomenon from the cash dividends point of view arguing that the number of firms that used to pay cash dividends in the past fell to 20.8 per cent in 1999 while in 1978 was 66.5 per cent. In Greece concerning the years 2004 and 2005, almost per cent of the listed companies in the Athens Stock Exchange have paid cash dividends (see: Athens Stock Exchange, Annual Statistical Bulletin, 2005). Additionally, 6.47 per cent of the companies that have paid cash dividends started paying at Moreover there was observed an increase of 2.96 per cent in the price of dividends from 2004 to When dividends are high, investors can buy more stock and when dividends are low, investors can sell some of the stock, something that makes dividend policy irrelevant (Dong et al., 2005). 2 Compustat belongs to Standard & Poor and provide useful statistical information and data to investment professionals and individual as well as institutional investors 9

10 Finally, according to several Greek finance newsletters, there is a general increasing tendency in 2006 for Greek companies of paying cash dividends instead of paying stock dividends. DeAngelo, DeAngelo and Skinner (2004) have focused on the disappearance of cash dividends by observing the total alteration of practices in corporate dividend policy the last twenty years. According to Baker and Wurgler (2004) who were influenced by Fama and French (2001) there exists a strong relation between the going ups and downs in the tendency of paying dividends and catering incentives. They also argued that the propensity of firms paying dividends is closely related to the stock market divided premium 3. Since corporate dividend policy has for many years given rise to scholars to find possible reasons for the operation of dividends and share repurchases, many surveys have been conducted especially in the last five years in order to specify the factors affecting dividend policy. However, almost all surveys focus on the corporate dividend policy from the managers point of view. As far as we know only Dong, Robinson and Veld (2005) focused on the dividend policy from the individual investors point of view. Literature on corporate dividend policy consists of economic modelling approaches aiming to the development of hypotheses and to the empirical investigation of corporate dividend policy (see: Allen and Michaely, 2003; Frankfurter and Wood, 2002; and Lease et al., 2000). According to Brennan (1970) dividends can be valued using his after-tax model along with the Capital Asset Pricing Model, which was used by other scholars. More specifically, Black and Scholes 3 When the proxy for the stock market dividend premium is positive the tendency to pay cash dividends increases and opposite (Baker and Wurgler, 2004). 10

11 (1979) did not find evidence of a dividend effect while Litzenberger and Ramaswamy (1982) found a notable effect. Several papers focus on this dispute; however there has not yet been reached unanimity regarding the effect of corporate dividend policy on the value of the firm. Notably, Black (1976, p.5) stated that The harder we look at the dividend picture, the more it seems like a puzzle, with pieces that just do not fit together. In order to shed more light on the dividend puzzle that was earlier discussed by Black (1976), Dong et al. (2005) surveyed a Dutch panel of families who answered on individual matters of finance and consumption. They did not include institutional investors in their survey since they wanted to focus on individual investor behaviour and decisions. Our survey focus on Greek individual investors who hold shares of listed on the Athens Stock Exchange companies; directly or indirectly through investment funds. Additionally some demographic details of the respondents are asked in order to test thoroughly the dividend theories. The theories under investigation were developed over more than 40 years and constitute of the corner stone of corporate dividend policy (see: Dong et al., 2005). Research regarding individual investor preferences on dividend policy in Greece provides a significant positive effect, since the Greek tax system concerning the individual level does not tax either dividends or capital gains at all. This tax exemption for the Greek individual investor gives us the potential of testing dividend theories in isolation from the tax effect on dividend policy, which takes effect in other European countries and in US. The purpose of our paper therefore is to specify the Greek individual investor beliefs regarding corporate dividend policy that was also examined by Brav et al. (2005). The main difference is that they asked financial managers to answer instead of 11

12 asking directly individual investors. Nevertheless Brav et al., (2005) concluded that they could only assume regarding investors preference towards dividends. In our survey therefore we do not try to assume but to secure the fact that individual investors want dividends, as Dong et al. (2005) have done one year earlier. 2.2 Theories on why individual investors want cash dividends According to Dong et al. (2005) there appears to be a general opinion that individual investors are not indifferent regarding dividends. Companies have to make a choice between paying cash or stock dividends due to their constant need for cash. Below, our study provides a thorough and detailed presentation of some of the most well known dividend theories like the dividend irrelevance theory (Miller and Modigliani, 1961), cash dividends involving transaction costs, the uncertainty resolution theory (Gordon, 1961), the smaller influence from accounting manipulations, the behavioural finance theory (Shefrin and Statman, 1984), cash dividends involving free cash flow theory (Jensen, 1986), cash dividends involving agency cost theory (Easterbrook, 1984), the signalling theory of dividends (Bhattacharya, 1979), the choice between share buybacks and finally the so-called tax-disadvantage theory of cash dividends. Testing these theories will provide us with useful information regarding the preferences and behaviour of individual investors towards cash dividends The Miller and Modigliani (1961) dividend irrelevance theory. Miller and Modigliani (1961) developed the so-called dividend irrelevance theory supporting that dividend policy can become irrelevant under some specific assumptions. In a perfect and complete market where there are no taxes, no transaction costs, no asymmetric information and agency problems, dividend policy is 12

13 irrelevant since individuals can overturn managers decisions on dividend policy by buying or selling equity on their own. Another important issue raised by Miller and Modigliani (1961) trying to abandon one of their basic assumptions the perfect market conditions is that from the standpoint of dividend policy, what counts is not imperfection per se but only imperfection that might lead an investor to have a systematic preference as between a dollar of current dividends and a dollar of current capital gains (Miller and Modigliani, 1961, p. 431) Transaction costs. The theory of dividends that involves the transaction costs is another one very important issue of corporate dividend policy. A great deal of scholars for many years have tried to analyse and provide useful knowledge regarding transaction costs associated with cash dividends. According to Allen and Michaely (2003) if an investor wants to receive a regular and fixed income, he can choose between two alternative options. The first one is to buy back stocks that pay dividends in order to cash in the dividends and receive income. The second one is to buy stocks that do not pay dividends and simultaneously on a scheduled and regular base to sell part of his portfolio of stocks Uncertainty resolution. Gordon (1961) focused on the so-called uncertainty resolution theory of dividends. Another well-known definition of this theory is the bird in the hand theory that was a challenge for many scholars to defend or reject as a dividend hypothesis. The concept of this theory is that dividends are the preferred ones due to their lower risk opposed to the higher risk of capital gains uncertainty. Therefore Gordon (1961; 13

14 1962) suggests that investors prefer firms that pay cash dividends to those that do not pay dividends at all. A great deal of studies however shows that this model fails when it is posited in a perfect and complete market where investors are behaved as rational investors (see: Miller and Modigliani, 1961; Bhattacharya, 1979) A smaller influence from accounting manipulations. Companies that pay healthy dividends are often thought to be relatively honest and reliable and less influenced from accounting manipulations. Therefore, for this reason companies prefer paying dividends in order to be considered as healthy and dependable. According to Laing (2002) healthy dividends is an indication that companies generate real earnings that are being grown over a long period of time. Buying dividend paying stocks is a way of getting into growth through the back door in a lower-risk way (see: The above statement is strongly related to the uncertainty resolution theory of Gordon (1961; 1962) Behavioural finance. A quite dissimilar theory of dividends from the previously mentioned ones is the behavioural finance theory. Shefrin and Statman (1984) introduced a new explanation of the individual investor s preference for cash dividends. They proposed the behavioural life-cycle theory that is based on self-control of the rational individual investor. This theory refers mainly to investors who prefer to consume from dividends. Such individual investors are the elderly people who do not anymore have a regular labor income (see: Dong et al., 2005). The definition rational investor has challenged many behavioural researchers for many years. According to Statman (2005) individual investors were never rational. 14

15 In 1960s academics started changing their views defining the individual investor from normal to rational. Rational investors are not affected by emotions and psychological reasons regarding their actions. Normal investors on the other hand are emotionally driven and they do not care only about the risk and return of their portfolio. Statman (2005) claimed that investors used to be normal before Miller and Modigliani (1961) name them as rational, and that they remain normal until today. The behavioural asset-pricing theory of Shefrin and Statman (1994) supports the view of the normal investor since they found that their model is strongly dependent on emotional and cognitive biases. This view is also consistent with Fama and French (2004) who claimed that individual investors care not only about money, but also they are interested in the pleasure of holding high growth stocks. According to Clark-Murphy and Soutar (2004) and to other recent psychological and financial research papers (see: Shefrin, 2000; and Shleifer, 2000) there are internal behavioural factors that may affect the investment decisions of individual investors, like knowledge as well as external behavioural factors, like the framework under which investors investment decisions are being presented. The theory of self-control of Thaler and Shefrin (1981) pointed out the irony of the situation characterising the individual investors as foreseeing planners and at the same time as myopic doers. They linked this conflict with the agency conflicts between managers and ownership of a firm. Miller (1986) responded to the behavioural approach of Shefrin and Statman (1984) by investigating the behavioural rationality of the individual investor regarding the case of dividends. Associated with the behavioural patterns in dividends decision-making, Miller (1986, p.467) simply stated that: Behind each holding may be a story of family business, family quarrels, legacies received, divorce settlements, and a host of other considerations almost 15

16 totally irrelevant to our theories of portfolio selection. That we abstract from all these stories in building our models is not because the stories are uninteresting but because they may be too interesting and thereby distract us from the pervasive market forces that should be our principal concern Free cash flow. Jensen (1986) linked the dividend policy decisions with free cash flow. According to Dong et al., (2005) free cash flow is the cash that a firm is capable of generating after maintaining or expanding its assets. The choice of firms between maintaining their cash flows for investment plans and using their cash flows to pay cash dividends was a topic of research for many scholars. They have tried to explain the effect of free cash flow on corporate dividend decisions. The free cash flow of a firm can be a subject of conflicts between managers and shareholders in order to decide where to invest it. Therefore according to Jensen (1986) managers often prefer undertaking projects with negative net present value in order to grow the magnitude of the firm, rather than paying cash dividends 4. Black (1976) claimed that firms that pay dividends to shareholders, could moderate a possible problem of overinvestment. This is due to the fact that the cash, which is destined for paying dividends, is taken from the amount of free cash flow and as a result free cash flow is reduced and the problem is alleviated Agency costs. According to Easterbrook (1984) a theory relative to the above mentioned free cash 4 The free cash flow dividend theory that was mainly discussed by Jensen (1986) can also be met in the literature of dividend policy as the overinvestment theory. 16

17 flow theory of dividends is agency costs theory. Agency costs are affected by the fluctuations of dividends and stem from the firm s effort to increase its free cash flow. If dividends increase, free cash flow decreases and the agency problem is mitigated. On the other hand if dividends decrease, free cash flow increases and the agency problem is deteriorated (see: Rozeff, 1982). Easterbrook (1984) also argued that dividends could be relatively helpful in order to mitigate the agency problems of management. He suggested that firms have the ability of maintaining themselves in the capital markets with the use of dividends. This may lower agency costs. This monitoring explanation of Easterbrook (1984) takes effect for stock dividends too. Miller and Modigliani (1961) suggested that when income is taxed at a higher level than capital gains, firms should try to conform their dividend policies in order to reach a point of agreement to the shareholders, the so-called market equilibrium. Miller and Scholes (1978) proposed that investors in order to avoid paying taxes on dividends could lever their buying of equities (see also: Miller and Scholes, 1982). Some years earlier, Fama (1980) stressed the agency problems and their relationship with the theory of the firm. He tried to give evidence on how the separation of management and ownership can stand as a profitable form for an organization. Therefore agency theory was a field of analysis for scholars for many years starting from Adam Smith (1937) and his outstanding Wealth of nations. The most contemporary approaches to the agency problem are dated back to Berle and Means (1932). The issue under investigation is how dividend policy may affect the agency theory and consequently increase or decrease the agency problems inside a firm. Dewenter and Warther (1998) in their attempt to compare the dividend policies of US and Japan investigated the implications of agency theory on investors behaviour regarding dividends. Like Easterbrook (1984) and Jensen (1986), Dewenter 17

18 and Warther (1998) treated dividends as a disciplinary method of agency costs. All these different views of the same phenomenon converged to the point that dividend changes are positively related to stock returns due to the fact that an increased dividend shortens the margins of managers to use a firm s cash flow in the investing plans of a firm. Moreover Miller and Rock (1985) argued that the optimal dividend policy, with no existence of asymmetric information, is the complete payout of excess cash flow to stockholders while maintaining the firm s investment policy according to the Fisher s separation theorem 5. Finally more recently DeAngelo, DeAngelo and Stulz (2004) regarding the theory of dividends as a monitoring mechanism of agency costs found strong confirmation. When the costs of excess upkeep are considered together with agency costs and taxes then firms try to avoid paying cash dividends to stockholders if the earned capital is low and decide to pay cash dividends only if they achieve in earning a significant amount of equity (see: DeAngelo, DeAngelo and Stulz, 2004) Signalling. The information signalling theory of dividends is the dividend theory that explains the dividend puzzle from the information content point of view. According to Brealey et al., (2005) the underlying of this theory is the support of a sticky dividend policy, which means that managers in order to avoid reducing dividends prefer a lower payout ratio and only if the future seems propitious they decide to increase dividends. Cutting off dividends can destroy a firm s value while increasing dividends can 5 This theorem states that the value of a firm is not affected by the dividend policy and consequently investment decisions do not affect the firm s financial decisions. 18

19 strengthen the value of a firm. However the significant part of this implication is the information that is going to be conveyed to investors due to changes in dividends. Therefore, dividends may well become a signal for investors because of the existence of management s asymmetric information. A confirming and supporting view of this theory belongs to Bhattacharya (1979) who claimed that the existence of information asymmetries between managers and outside investors could play a signalling role regarding dividends. This view regarding the signalling effect is also consistent with Miller and Rock (1985). The signalling theory of Bhattacharya (1979) has drawn the most attention. According to Bhattacharya (1979) cash dividends can operate as a function to signal the expected cash flows of the firm. In order for such statement to be in effect two assumptions may exist. First of all the outside investors are assumed to have imperfect information about the firm s progress and profitability and second the cash dividends have to be taxed higher than capital gains. If these two assumptions exist in the market then cash dividends may serve as a signal regarding the firm s future performance. Bhattacharya (1979) adapted the signalling model that was first investigated by Lintner (1956), and Miller and Modigliani (1961). One year later Aharony and Swary (1980) examined the impact of quarterly dividend and earning announcements on the signalling effect of dividends. They claimed that if cash dividends indeed convey the future performance of a firm in an efficient capital market, this situation will have as a result the immediate reaction of stock price changes in order to include the dividend change directly after the announcement is made. The outcomes of Aharony and Swary (1980) indicate that they prop up for the semi strong form of market efficiency. Consequently Aharony and Swary (1980) found evidence 19

20 regarding the signalling hypothesis of Bhattacharya (1979) underlying the fact that changes in quarterly cash dividends indeed convey information regarding the expected cash flows of the firm. Regarding the validity of market efficiency relative to announcements of dividend changes, Pettit (1972) offered some significant insights. He suggested that even though market uses changes in dividends in order to convey information to outside investors, this does not seem to be efficient since it is not perfect in describing the future expectations of a firm. Dielman and Oppenheimer (1984) tried to examine the behaviour of investors around the dates of the announcement having as a target to provide confirming evidence regarding the signalling hypothesis and the information content of dividends. Their results show that there is a strong supporting view for the information content of dividends. Miller and Rock (1985) provided confirmation to the previous studies regarding the information content of dividends serving as signals. They argued that information asymmetries among firms and investors have the ability to challenge a signalling role regarding dividends. They finally stressed the importance of the information content of dividends by stating that In fact, the best place for empirical researchers to look for evidence of dividend signalling may well be among firms falling into adversity, not because they then start signalling, but because they stop (Miller and Rock, 1985, p. 1046). A quite dissimilar view regarding the signalling hypothesis is the one discussed be John and Williams (1985). They investigated the signalling equilibrium based on dividends, dilutions and taxes. Their major target was to find solutions to several unanswered questions derived from previous research (see: Bhattacharya, 1979; Hakansson 6, 1982; and Miller and Rock, 1985). Amongst the number of studies 20

21 regarding the information content of dividends is the signalling approach of Bar- Yosef and Huffman (1988) who used the incentive-signalling framework in order to provide justification and testing for some unanswered questions of past papers. They focused on the managers incentive finding that the informative dividends are a function that increases anticipated cash flows for a firm. A more recent research providing extensions in the field of the information content of dividends and their signalling role belongs to Denis, Denis and Sarin (1994) who examined the signalling, tax-clientele and overinvestment hypotheses into a single framework finding support for the signalling hypothesis. Bernheim and Wantz (1995) conducted a tax-based test of the signalling role of dividends. In addition Benartzi, Michaely and Thaler (1997) stressed to investigate whether changes in dividends convey the future or the past, finding strong support only in that dividend cuts are signalling an increase in a firm s future earnings. Finally, Eades (2005) tried to extend the dividend-signalling model of Bhattacharya (1979) and consequently came up with a new hypothesis, the relative signalling strength (RSS) hypothesis The choice between cash dividends and share buy backs. According to Dong et al. (2005) share buy back serves as an alternative for cash dividends. Either cash dividends or common stock repurchases are useful in order to mitigate the agency problems (see: Easterbrook, 1984; and Jensen, 1986). A great deal of academic papers indicates that common stock repurchases may also signal that the stock of a company, which buys back its shares, is undervalued. According to 6 According to Hakansson (1982) dividends may serve as a function of improving efficiency only when they have information content in order to provide investors with the necessary information regarding a firm s expected cash flows. 21

22 Comment and Jarrell (1991) and Ikenberry et al. (1999; 2000) the common stock repurchase announcements are closely related to significantly positive abnormal returns, a result that supports the undervaluation hypothesis of common stock repurchases Taxes. Miller and Modigliani (1961) suggested that when income is taxed at a higher level than capital gains, firms should try to conform their dividend policies in order to reach a point of agreement to the shareholders, the so-called market equilibrium. Miller and Scholes (1978) proposed that investors in order to avoid paying taxes on dividends could lever their buying of equities (see also: Miller and Scholes, 1982). Perhaps two of the most discussed theories in the dividend literature are the tax effect theory and the clientele effect theory regarding the dividend preference of investors. According to Litzenberger and Ramaswamy (1980) regarding the consequences of tax-induced clientele effects on capital asset prices, stockholder clienteles exist, if investors in high tax brackets hold stocks with low yield and opposite. Another study consistent with the tax-induced clientele effect of Litzenberger and Ramaswamy (1980) is that of Lakonishok and Vermaelen (1983). They provided evidence regarding the implications of dividend change announcements to stock prices based on the so-called tax-clientele hypothesis. They also examined the consequences of the US tax reform on the preference of individual investors to cash dividends or capital gains. They finally concluded that stock prices are indeed dependent on taxes and dividend policy and that their empirical findings are somewhat inconsistent with the innovative evidence of Elton and Gruber (1970) regarding the tax-clientele effect. 22

23 According to Dong et al. (2005) taxes and the taxation of dividends and capital gains is one of the most important factors that affect dividend policy and the preference of individual investors according to dividends and share repurchases. The tax effect on dividends and capital gains have been troubling scholars for many years. A significant part of the dividend literature refers to the taxation of dividends according to the tax system of US (see: Fama and French, 2001; Grullon and Michaely, 2002; and DeAngelo, DeAngelo and Skinner, 2004). The old tax system taxed dividends at a higher rate than capital gains and this had as a result investors to prefer selling part of their portfolio instead of receiving shares. The tax effect on dividends due to its negative impact on dividends preference is often called as tax disadvantage. This situation has been significantly altered since the new US tax reform imposed an equal tax rate for both dividends and capital gains. According to the Organisation for economic Co-operation and Development and the Greek Tax reform of 2002, capital gains and dividends are tax exempt regarding the individual investor while they are being taxed only at a company level. The tax exemption of dividends in Greece was discussed a lot by scholars and especially Edwards (2003, p. 281) who claimed that: Take Greece. Not only did that country invent democracy and give us the word "economics," but it has a tax code that does not tax domestic dividends at the individual level at all. Therefore since the tax system in Greece does not tax dividends and capital gains at all for the Greek individual investors the tax effect on dividends and capital gains may not be that much important regarding the preference of the individual investor. According to Dong et al. (2005) the tax effect that was previously discussed is strongly related to another well-known theory of dividends the tax-clientele theory. Elton and Gruber (1970) argued that the tax-clientele theory stems from the difference 23

24 in the taxation of dividends and capital gains. This situation of course is not in effect regarding Greece because capital gains and dividends regarding the individual investor are tax-exempt. However the clientele hypothesis is troubling scholars especially in the countries were the tax disadvantage exists. Due to the fact that dividends are taxed differently than capital gains Elton and Gruber (1970) claimed that the cost of retained earnings can constitute of a function of the marginal tax bracket of the stockholders. They also tried to prove that if there is a relationship between marginal stockholders tax brackets and corporate dividend policy, the existence of the clientele effect could be justified as there was established in Miller and Modigliani (1961). The latter stated that even though disparities in taxes and risk levels exist among firms, dividend policy still remains irrelevant regarding firms but may not be that irrelevant regarding the individual investor. 2.3 Theories on why individual investors want stock dividends The other alternative option companies have in order to distribute their earnings to stockholders is paying stock dividends (Dong et al., 2005). The issues that are going to be examined in our paper regarding the behaviour of individual investors towards stock dividends are the consideration of stock dividends as small stock splits, stock dividends involving transaction costs, the taxation of stock dividends and the behavioural finance theory on stock dividends (Shefrin and Statman, 1984) Stock dividends as small stock splits. First of all as DeBondt and Thaler (1995) claimed that besides firms paying cash dividends there are other firms that prefer paying stock dividends (Dong et al. 2005). Stock dividends are much more similar than dissimilar to stock splits. DeBondt 24

25 and Thaler (1995) report relatively to stock dividends that they consist of one of the biggest anomalies in finance. Moreover as Dong et al. (2005, p. 129) have stated, as every standard textbook in Finance teaches us, stock dividends are nothing more than a small stock split Transaction costs. Stock dividends can prove more advantageous relatively to the lower transaction costs. With a stock dividend, the transaction costs are lower because the dividend is invested again in the same stock. With a cash dividend, the transaction costs are higher because the dividend is invested again in other stocks. However, the disadvantage of a stock dividend is that for the small individual investor it could prove a more expensive solution (see: Dong et al., 2005) Taxes. In the Greek tax system, stock dividends as well as cash dividends are not being taxed at the individual level. Therefore, it should be noted that the fact that stock along with cash dividends are not taxed, does not give to either way of payment an advantage regarding the individual investor preference Behavioural finance. According to Dong et al. (2005) relatively to the behavioural finance theory (see: Shefrin and Statman, 1984) stock dividends are preferred over cash dividends especially when a firm does not have free cash flow available for paying cash dividends. 25

26 2.4 Empirical findings Studying the dividend literature from both sides of managers and individual investors, and according to Dong et al. (2005), the major part of the research focus on dividend policy, theories and tests form the managers, institutional and professional investors point of view. Only a small part refers to the impact of corporate dividend policy on individual investors 7. De Jong, Dijk and Veld (2001) empirically investigated the dividend and share repurchase policies of firms in Canada from the managers point of view by testing three separate models regarding cash dividends and share buy backs. With the use of standard, simultaneous and logit models as well as questionnaire data they focused on the firms decisions regarding paying cash dividends and on the type of payout (cash dividends, share repurchases or both). Their survey was based on the free cash flow, behavioural and tax effect theories as considered being the most important determinants of dividend policy. They involved the signalling hypothesis finding evidence for the model of Brennan and Thakor (1990) 8. Their results indicate that 35 per cent of the firms used share buy backs as a form of payout while the rest used dividends or both. They also confirmed their expected significantly negative relationship between dividend payments and transaction costs. A significantly positive relationship between dividend payments and the behavioural hypothesis was confirmed as well. This significantly positive relationship between dividend payments and the behavioural hypothesis indicated the existence of different clienteles, as well as tax clienteles that can also be found in the survey of Baker, Farrelly and Edelman (1985). 7 The definition individual investor is often met in the literature as retail investor too. 8 Brennan and Thakor (1990) argued that the information asymmetry among firms and outsiders is strongly linked to a preference of paying cash dividends rather paying stock dividends. 26

27 As far as the stock repurchases are concerned, De Jong, Dijk and Veld (2001) found a significant positive indication for the free cash flow hypothesis while they found a significant negative indication regarding the existence of asymmetric information among firms and outsiders. In general, De Jong, Dijk and Veld (2001) found strong empirical evidence that the Canadian firms first decide on the payout decisions and then on the type of payout (cash dividend, share buy back or both). Baker, Veit and Powell (2001) focused on factors influence the dividend policy decisions of Nasdaq firms, developing a questionnaire based on previous studies of Baker, Farrelly and Edelman (1985); Farrelly, Baker and Edelman (1986); and Baker and Powell (2000). The respondents were all members of the senior management of these firms. Therefore based on the answers of the 188 respondents, the most important factor for these firms affecting their dividend policy is the phenomenon of past dividends, the stability of earnings and the level of current and future earnings. The latter two factors are viewed very often in the literature regarding their impact on corporate dividend policies (see: Gordon, 1959; Black and Scholes, 1974; Blume, 1980; and Nissim and Ziv, 2001). The importance of dividend policy decisions regarding managers is obvious if we look at their revision and re-examination of the firm s dividend policy once every year. Additionally Baker, Veit and Powell (2001) stressed the importance of dividend decision making for managers arguing that managers care a lot about these decisions due to the fact that they affect the firm s value and consequently the wealth of stockholders. Baker, Powell and Veit (2002) focused on the managerial perspectives on dividend policy investigating the opinions of a group of Nasdaq firms that regularly paid cash dividends. The hypotheses that they investigated here are the signalling 27

28 hypothesis developed by Bhattacharya (1979), John and Williams (1985) and Miller and Rock (1985), the tax effect hypothesis based on Miller and Scholes (1978) as well as on Litzenberger and Ramaswamy (1979), the impact of agency theory on dividend policy of Easterbrook (1984) and Jensen (1986) and finally the uncertainty resolution theory of Gordon (1961). The first important result is that managers of these Nasdaq firms do strongly believe that dividend policy matters, something, which means that dividend policy influences stock prices. This result is inconsistent with the irrelevance theory of Miller and Modigliani (1961). The most possible explanation is due to the imperfections of market that led to the relevance of dividend policy. Respondents also agreed with the signalling hypothesis of Bhattacharya (1979) arguing that investors believe that dividends serve as signals of the firm s future prospects and profitability. However respondents were not that positive regarding the increase in stock prices due to an unexpected increase in dividends. The responding managers did not give evidence to the bird in the hand theory. This theory claims that investors prefer cash dividends today to an uncertain higher income in the future 9. Regarding the tax effect hypothesis and the agency cost explanations surprisingly managers gave little or no support. Baker, Mukherjee and Paskelian (2005) investigated Norwegian managers on how they view dividend policy. They used a sample of Norwegian firms listed in the Oslo Stock Exchange. They found that the most significant factors affecting dividend policy are the stability of earnings, the level of current and future earnings, the financial leverage and the liquidity limitations. The first two factors are consistent 9 From the responding managers almost 54.9 per cent disagree with the bird in the hand theory and 28.0 per cent didn t have any opinion regarding the bird in the hand statement. 28

29 with the survey of Baker, Veit and Powell (2001). Nevertheless Norwegian managers were not that consistent with the view that the firm s dividend policy affects its value. This study was based on four hypotheses that were tested in order to examine their validity and cogency. The first one is whether earnings are the major disciplinary force of dividends. The second one is that there are important differences between US and Norwegian dividend policies. The proposal focuses on the fact that the precise and severe government standing orders may reduce the signalling role of dividends since Norwegian managers care most about the legislation and regulations and not about affecting the stock prices. The third hypothesis refers to the effect of dividend policy on firm value and the last one is that managers seem to prefer the signalling theory to the tax effect theory regarding paying cash dividends. Three of the most important factors of the survey involve the stability of earnings, the level of current and future earnings. This result gives support to the first hypothesis regarding the importance of earnings. Relatively to the second hypothesis Baker, Mukherjee and Paskelian (2005) found that Norwegian managers express a strong confirmation for this hypothesis since they care most about regulations and also they are afraid of giving a false signal to investors. Regarding the third hypothesis of firm value Norwegian managers seem a little undecided about whether dividend policy is important and relevant. This happens because most of the respondents were undecided to whether dividend announcements are used as information to help estimate a firm s value. Finally the last hypothesis is confirmed since almost 70 per cent of the respondents agreed that dividend changes operate as signals regarding the firm s profitability. Additionally the survey provides support for the bird in the hand theory since managers prefer cash dividends today to an uncertain higher income in the future. 29

30 Brav, Graham, Harvey and Michaely (2005) investigated the payout policy in the 21 st century in order to specify the factors that affect payout policy. More specifically they investigated cash dividends and share repurchases asking managers for the view of individual investors. The main issues raised from Brav et al. (2005) were taxes, agency costs and signalling theory regarding the propensity of firms to pay dividends; why paying cash dividends is a phenomenon mainly observed at the bigger firms; why still many firms pay large and increased dividends. They used interviews to a sample with responses from 384 managers, mainly financial officers who expressed the view that market is more prepared to accept a decrease in share buy backs rather than in dividends, in order for firms to be more flexible and liberated regarding their repurchase policy. To shed light to the factors affecting payout policy we take a look at the results of each factor. According to the survey tax effect theory regarding the tax disadvantage of dividends seems to be of second-order importance for the respondents either for cash dividends or for share repurchases. Only 21.1 per cent of dividend payers reported tax effect theory as very important factor affecting payout policy, and similarly only 29.1 per cent of repurchasing firms have cited tax disadvantage as a very important factor. A very important result of this survey is regarding individual investors. Managers believe that the only class of investors that would prefer cash dividends instead of share buy backs is retail investors. Therefore the importance and preference of cash dividends is stronger for individual investors than for other forms of investors. Regarding the signalling role of dividends and share repurchases and the issue raised about whether their changes convey information to investors, almost 80 per cent of the managers claimed that cash dividend changes do convey information 30