Corporate finance has been the subject of numerous studies with varying interpretations economists and investors have sought to explain the correlation between the corporate policies and their impact on company value. Consequently there are many theories and much literature addressing this topic, but perhaps the most controversial issue has been dividend policy and how this can provide valuable information to future investors and current shareholders. According to Black (1976, p. 8) "the harder we look at the dividend picture, the more it seems like a puzzle with pieces that just don't fit together." To facilitate the understanding of this essay it is important to define the meaning of dividend policy and capital gain. The former is defined by Brealey & Myers (2003, p. 154) as "the trade-off between retaining earnings on the one hand and paying out cash and issuing new shares on the other", and the latter can be defined as the positive difference between purchase price and selling price of an asset.
Although some economists have held that dividend policy is irrelevant under certain conditions and assumptions, others have shown strong evidence for the importance of this policy in making investment decisions, some related with market distortions and current prices of shares, others with the ability to predict future earnings of companies. Thus, this essay is an attempt to demonstrate that the dividend policy adopted by a company can provide useful information for making investment decisions. However, given the complexity of the issue and the difficulty of making accurate predictions about future profits of companies, there is no consensus on the impact of dividend policy on profits.
In order to demonstrate this, first there will be a discussion of the validity of the models based on perfect market assumptions given the current conditions that companies and investors must face. The second section will analyse different theories of investor behaviour and will show that in some cases traders can predictably affect dividend policies and these policies in turn provide valuable information to the market. Finally, evidence for and against the idea that dividend policy can be used to predict future earnings of companies will also be discussed.
One of the key arguments against the idea of obtaining useful information on dividends and dividend changes is the work realized by Miller & Modigliani (1961). Assuming perfect capital markets, rational behaviour and perfect certainty, they conclude that dividend policy is irrelevant to the value of the company. This section will focus on the first assumption, which affirms that there are no transaction costs, all types of investors have the same information and there is no difference between the tax rate on dividends and capital gains. Under these conditions the expected benefit will be the same for all financial assets because the demand for higher-yielding assets pressured prices upward and thus it will eliminate the chance of getting a better return (Miller & Modigliani, 1961, p. 412). Given that companies are indifferent between financing their investment plans with equity or debt, dividend policy is irrelevant under these assumptions.
Although this conclusion seems clear and precise, it is based on strong assumptions that are not always valid in reality. Markets, particularly emerging ones, have a number of distortions and consequently make dividend policy, to certain extent, relevant. These can be categorized according to the basic assumptions of perfect capital markets: different tax rates, transaction costs and information asymmetries.
First, there are different tax rates for realized capital gains and dividends, according to Black (1976, p. 9):
In a world where dividends are taxed more heavily (for most investors) than capital gains, and where capital gains are not taxed until realized, a corporation that pays no dividends will be more attractive to taxable individual investors than a similar corporation that pays dividends.
There is indeed some evidence for the presence of such distortions, for example the case of the U.S. where dividends are taxed more heavily than capital gains. Contrary to expectations, companies under this tax regime have a stronger preference for paying dividends (Allen, Bernardo & Welch, 2000, p. 4). Therefore there are other factors discussed below with greater influence on corporate financial policies that require investors to take this cost.
Second, the existence of transaction costs affects both companies and investors. They affect the former because of the cost involved to obtain the resources needed to pay certain level of dividends previously compromised or to delay profitable projects for the business due to limitations in the amount of available cash (Bhattacharya, 1979, p. 262). They affect the latter by the costs involved selling their shares to raise funds to finance consumption rather than receiving dividends (Manos, 2001, p. 31). Although the effect of these costs is not clear because they are opposites, their presence can eventually affect expectations of future cash flows and thereby the value assigned to shares.
Lastly, there are information asymmetries between different groups of investors and between investors and managers. In the former case this phenomenon will represent an advantage for the more informed investors when it is conducting a repurchase because they have better knowledge on future projects of the company and thus of the value of shares. Consequently uninformed investors will prefer dividends to repurchases (Allen, Bernardo & Welch, 2000, p. 2). In the latter managers have privileged information about the firm's prospects and this is communicated to the market by various means. One of them is the dividend policy and, since there is a refusal to reduce the dividend rate, increases are made only if there is safety to be sustained over time. This implies that dividend changes may contain valuable information (Black, 1976, p. 10).
Another point related with information asymmetries among investors is mentioned by Allen, Bernardo & Welch (2000, p. 4). According to them institutional investors, given their size, have greater incentives to be informed and to identify companies with good performance. In addition, these investors are not taxed on dividends. Consequently firms have an incentive to pay more dividends and attract them. This seems to explain the preference shown by the companies to pay dividends instead of repurchases. A further explanation is given by Barclay & Smith (1988, pp. 76-77). According to them repurchases are an opportunity for managers to benefit themselves at the expense of shareholders due to information asymmetries. Therefore, repurchases seem the most expensive means to distribute the surplus of the firm, so that investors have a stronger preference for dividends.
A clear example of the effects of information asymmetries can be seen in the case of emerging markets. According to Travlos, Trigeorgis & Vafeas (2001, p. 109) in developing countries where, unlike developed countries, companies aim to generate the necessary credibility to raising capital in a scenario characterized by the "lack of credible means for the dissemination of financial information". There is evidence of a positive reaction in the market to growth in the dividends rate and this could be the result of the companies intentions "to bridge the information asymmetry gap with investors via their dividend policy". In sum, it seems clear that while the market is smaller and less developed, there are more inefficiencies in the dissemination of information. Hence, is enhanced dividend policy as a means of communication between managers and market. However it is unclear whether investors use this information properly and whether misuse of this can lead to other opportunities. Thus, it is important to analyse the investor behaviour.
A key argument in favour of the relevance of dividend policy is the sometimes irrational behaviour from investors that affect stock prices and thus generate inefficiencies in the market. Shiller (2000, pp. 149-150) described three useful experiments. The first one realized by social psychologist Asch (1952, cited in Shiller, 2000, p. 149) attempted to demonstrate the power of social pressure over individual decisions. For this purpose created groups of people with several members previously prepared to give wrong answers to obvious questions, as a result of this a third of the time the new members gave incorrect answers. Asch interpreted this result as a consequence of social pressure. The second one realized by psychologists Deutsch and Gerard (1955, cited in Shiller, 2000, p. 150) was a variant of Asch's experiment where individuals answered the questions anonymously, but knowing the group's responses, the results were the same a third of the time the new members gave wrong answers. The psychologists concluded that a large proportion of people act thinking that the other acts cannot be wrong, "reacting to the information that a large group of people had reached a judgement different from theirs, rather than merely the fear of expressing a contrary opinion in front of a group". The third and final experiment was performed by Milgram (1974, cited in Shiller, 2000, p. 150), in which a previously prepared victim was receiving electric shocks by a new member. The latter, seeing the suffering of the 'victim' asked to stop the experiment. However, the leader of the experiment refused and in many cases the 'perpetrator' continued. This was interpreted "as demonstrating the enormous power of authority over the human mind". As a result, Shiller (2000, p. 151) concludes that people trust in the opinion of the majority and the authorities on the subject even when these contradict their own convictions, especially in a complex issue as the shares value. This type of behaviour seems to be the explanation to certain phenomena observed in financial markets such as the market overreaction caused by changes in the global economic scenario, or specifically to changes in dividend rates.
Nevertheless, economists have held that investors behave rationally. As Miller & Modigliani (1961, p. 427) pointed out an individual investor acts thinking that other people have rational behaviour. In other words they always prefer more wealth independent of how they get it. Furthermore, an individual investor believes that other investors think that he has a rational behaviour. Finally, the authors assume that all traders comply with these conditions, thus giving rise to the concept of "symmetric market rationality". These assumptions imply that each individual maximizes his wealth and is indifferent to dividend policy, and "that he believes the same thing to be true of all other participants in the market" (Baumol, 1963, p. 113). Moreover, according to Miller & Modigliani (1961, p. 428), an investor who thinks that another will act irrationally will choose a strategy that, without this assumption, would seem irrational. Therefore, Miller & Modigliani accept the possibility of irrational behaviour among market participants, and this remains an open discussion on the influence of this irrationality on the value of the information contained in dividend policy.
A further theory was developed by Figlewski (1979) analysing the betting market in horse races. This was chosen because of its similarities with financial markets. There is much historical information available, specialists produce public reports and different types of bettors exist. The results obtained indicate that, in this specific market, the process of incorporation of public information is efficient and thereby this does not represent any help for making accurate predictions. This, extrapolated to the financial markets, implies that rational speculators counter the effects of irrational. Hence, market distortions related information and behaviour are eliminated and thus the possibility to 'beat' the market.
However, it seems clear that his argument is built on evidence based on a market that, despite being similar, is less complex. According to De Long et al. (1990) in the real world rational investors have incentives to follow it rather than counteract irrational behaviour. In other words, and given that an irrational investor for varying reasons sells at a low price and buys at a higher price, a rational speculator will expect that the effect of irrational speculation in the market move the stock prices away from their real values. Hence, a rational investor would have incentives to continue buying despite the fact that the shares are overvalued. Because he knows that due to the irrational investors those prices continue to rise. These attitudes seem reveal that market complexities are higher than expected and generate an atmosphere of uncertainty where investors must make their decisions.
Moreover, Baumol (1963) associated this uncertainty with the relevance of dividend policy because, given this situation, traders should look for logic signals on the market. He showed a clear example of this: the increase in stock prices experienced when the first U.S. manned satellite orbited the earth despite its irrelevance to the value of firms. Then, one could deduce that the market interprets this type of objective signals and incorporates them into their valuations at least temporarily. In addition, the general view is that the shares of low-payout companies suffer a reduction in their value (Miller & Modigliani, 1963, p.432). The explanation behind this phenomenon seems to be that the rate of dividends is taken by the market as one of the aforementioned objective signals (Baumol, 1963, p.114). This clearly demonstrates that traders influenced by their fears or optimism overreact to the arrival of new information, which includes changes in dividend policy. In consequence, investors have the possibility of making temporary profits analysing these news and anticipating the irrational market behaviour. Further examination, however, is required in order to determine whether dividend policy can provide information about future earnings of companies.
There has been much discussion whether dividend policy can signal the future, a number of studies have attempted to demonstrate that there is a positive correlation between dividend changes and future earnings changes. According to Bhattacharya (1979, p. 261) dividend policy not only provides useful information to shareholders but also is sent intentionally by managers that "optimize the after-tax objective function of shareholders, possibly because their own incentive is tied to the same criterion". In addition, he assumes that only managers manage the information about cash flow and thus there is a link among the information disseminated and future earnings. Nissim and Ziv (2001) went further and developed a model that demonstrate a positive and significant correlation between dividend increases and future profitability for a period of at least four years after the change occurs. However, they did not find significant correlation between dividend decreases and future profitability. Consequently, they argue that these results are justified by the obligation to be conservative in accounting, recognizing losses in advance and profits only when they are realized. These results together with the information provided by Bhattacharya (1979) show some evidence to support the hypothesis that dividends can convey information to predict future earnings.
However, although Nissim and Ziv (2001) demonstrate with empirical evidence the validity of their model, there are others that not only differ from these results but also demonstrate the opposite. This diversity of outcomes seems to be explained by the different assumptions and variables used in the process of modelling the problem. The study presented by Benartzi, Michaely & Thaler (1997) illustrates this contradiction; their findings show that there is an insignificant correlation between dividend changes and future earnings. However, they recognize an effect in the share value on the announcement. According to them this implies that investors act thinking that dividends convey information about the future, this conclusion is related to the results obtained in the previous section because suggests that investors behave irrationally. Finally, they suggest that dividend policy reflects what has occurred and "if there is any information content in this announcement, it is that the concurrent change in earnings is permanent rather than transitory" (Benartzi, Michaely & Thaler, 1997, p.1032). On the same line of reasoning, Garret & Priestley (2000) show that dividend policy is not useful to predict future profits. Furthermore, they argue that dividends provide information only about the positives changes from current earnings. The evidence shown by these studies discards the possibility of foreseeing the future profitability of companies based on dividend policy. However, they recognize the existence of correlation between dividend policy and permanent earnings and in the case of Garret & Priestley (2000) a lagged stock price change. It seems that the controversy is not closed.
Another interesting point of view is the contribution of Koch & Sun (2004), they hypothesize that changes in dividends are explained by "the persistence of past earnings", this implies that there is a lag in the market to incorporate the information supplied by firms. The problem faced by this research is that finds a negative correlation between price changes and past earnings in the case where positive profits are followed by a decrease in dividends. Generally the pieces of research that have addressed this issue seem to encounter the same limitation, not all the cases can be explained with the model without making assumptions that are quickly countered by the following studies. Therefore, the answer to the dilemma of signalling seems unclear, in other words we cannot affirm or rule out with certainty that the dividend policy can be used to predict future earnings.
In conclusion, this essay has attempted to demonstrate that dividend policy under certain circumstances can convey useful information to the market allowing investors to enrich their analysis of the value of the companies. Despite the fact that some economists argue that dividend policy is irrelevant, the evidence showed weakens the assumptions on which this assertion is based. Transaction costs and different tax rates undermine the foundations of the theory of perfect capital markets. Moreover, information asymmetries and irrational behaviour generate a dynamic and uncertain environment where investors seek to find some objective signs. As a result dividend policy takes a significant value within a range of financial indicators that can be used to make investment decisions.
However, although the usefulness of dividend policy for making investment decisions seems clear, because of distortions in financial markets and investor behaviour, it is not a powerful tool for prediction and therefore seems to need to be supplemented with more market information. Furthermore, there is a lack of clarity about its relationship to future earnings and this uncertainty leaves open the possibility for future research addressing this issue.