However, information symmetries are often caused by agency problems and make signaling model less reliable. Signaling model gives mix conclusions when r studies are utilizing different research methods. Furthermore, signaling model is a less ideal dividend policy when it applies to non-U.S. Market such as Japan. Existed signaling model gives glimpses of possible explanations, but a precise model is still yet to be found. Statement claims about dividend policy as a puzzle seem to be the truest and valid explanation so far.
- Introduction and Background Corporate dividend policy has puzzled economists and researchers for a long time.
Ongoing meticulous researches have been conducted through various empirical studies in order to understand companies’ dividend behavior and shareholders’ preferences towards dividend payouts. Many theories have been stated and established as an effort to understand dividend policy better, however, the relevance is still questionable and the issue is still yet to be resolved (Apothecary 2007, 4). Dividend policy has a strong tie with the decision of utilizing company cash flows. Companies should determine their choices whether to retain cash flows and do reinvestment or distribute it to shareholders as payouts.
- Payouts themselves can be done in two ways which are dividend payments or shares repurchase.
Economists and researchers’ interests concerning dividend policy has led to a continuous study in financial subject and brought many thoughts and arguments. One of the most famous paper researches was written by John Lintier in 1956, and has become the cornerstone of dividend policy study. First of all, Lintier strongly stated that taxes does play an important role in dividend payout as it determines companies’ earnings and consequently influence the amount of dividends paid to shareholders (Lintier 1 956, 97).
In latter time, tax effects on dividend and capital gain were further examined leading to the born of clientele effects theory which were conducted to complement dividend irrelevance theory introduced by Miller and Modeling during earlier research. Further discussion on clientele effects tax and dividend irrelevance theory will be broadly reviewed on the second part of the literature review.
On his paper, Lintier also emphasized that thorough prolonged studies conducted among 28 companies in seven years, has enabled him to understand managers’ beliefs regarding shareholders attitudes.
Managers tend to believe that hardliners will prefer stable dividend stream with gradual growth as this is would suppress negative reaction (Lintier 1 956, 99). Another parameter of dividend policy was also presented, which Was about the dependency of dividends stability is heavily influenced by both current earnings and past dividend payouts (Chem., Dad, and Priestley 2012, 1834). This parameter is also known as the so-called dividend smoothing.
However, Lintier researches were unable to provide reliable reasons of why companies often choose this action.
Economists were trying to make sense of dividend smoothing motives wrought dividend signaling theories brought up by other economists such as Patriarchy’s, Miller and Rock yet the result is unsatisfactory (Lambert and Myers 2012, 1769). Lack of information (both theoretically and empirical studies) has led to new speculation of possible reason presented by economists who believe that dividend smoothing is happened due to information asymmetries (such as Kumar (1988) and Brenna and Thacker (1990)) such as agency cost (Allen, Bernard, and Welch cited in Leary and Michaels 2008, 2).
Agency cost seemed to be plausible factor as it better explained dividend policy according to survey being done by Allen, Leary, and Michaels (Lambert and Myers 2012, 1769).
It is truly seen that dividend policy stills remain a great mystery. The curiosity of economists and researchers has resulted in more varied possibilities of dividend policy motives. New dividend policy theories are proposed in the search process of the most plausible answer. However, the more new theories are introduced and the more economists examined the behavior of dividend policy, the more bewilderment occurs.
Looking through this condition, Frankfurter and Wood 2002) were trying to simplify existed theories so they could be easily studied and understood.
They have successfully grouped the theories and separated them into three major determinants Of dividend policy. Full information model tax factor is the first determinant, followed by information asymmetries in the second place, and behavioral models as the last factor. This literature review is purported to analyses the variety of economists and researchers past studies on dividend policy. This paper will explain two most influential dividend policy theories.
The first dividend policy would be clientele effects followed by signaling models as the second theory. Each policy will be discussed on its own independent section and will be explained through various views as well as cross examination of past studies.
This paper will also try to highlight the common ground of dividend policy theories which are agreed upon economists and researchers while also be believed as relevant in real life business. Overall, this paper will consist of 4 major sections.
The first section will cover introduction and background dividend policy.
Section two will explain about clientele effects in dividend policy continued with section three observing about the signaling model. Finally, the last section would be the overall conclusion of literature review with major and important points being noted.
All investors are subjected to tax by governments, and different taxes policies and regulations shaped different decisions of dividend policy. This problem was also realized by Miller and Modeling, and various tax policies faced by investors result in a condition known as ‘dividend clientele effect’.
To further understand the concept of dividend clientele effects, this paper will use U. S. Market condition as an example.
U. S. Investors are mainly made up of individuals and institutions. Both individuals and institutions are facing different taxation system. Individuals are facing different tax rate based on one people’s tax bracket, people with higher income are usually taxed at higher rate. As for institutions, companies such as pension funds companies, schools, and foundations are either tax free or taxed in lower rate.
This condition has shaped companies clientele which in a simpler term can be defined as groups of investors in companies who find companies’ dividend policy suitable for their preferences and taxation condition. To simplify the assumptions, these groups of investors will be categorized as ‘untaxed institutions’ and ‘taxed individuals’ (Allen, Bernard, and Welch 2000, 2500). First, this paper will try to examine the relationship between taxed individuals and companies’ dividend policy. This relationship was beautifully explained through survey done by Edwin J. Elton and Martin J. Grubber in 1970.
Two hypotheses are used to help researchers conduct their study:
- The first one is related with companies’ dividend yield. It is believed that the smaller companies’ dividend yield, investors will only gain small percentage n dividend, and should expect larger return on capital gain. The most likely investors on this stock are people who fall in the high tax bracket, and consequently high dividend yield stock should be investors who fall in the low tax bracket.
- The second one is related with the amount of dividend payout.
It is believed that companies who paid out high dividend, in equilibrium will grow at slower rate compare to companies who paid out low dividend. Hence, companies with high dividend payments should attract investors in low tax bracket. To test the hypothesis, researches were carry out by studying the behavior f all listed stock in New York Stock Exchange who paid out dividend in one year time (from 1966-1967).
To examine the relationship, researchers utilize test statistic technique which in the end did prove the existence of ‘dividend clientele effect’.
In favor of the first hypothesis, it is found out that as dividend yield went higher, shareholders’ tax bracket went lower. Therefore it is concluded that there is a positive correlation between shareholders’ tax bracket level and dividend yield (Elton and Grubber 1970, 72). Although the first decision seemed to be out of place, researchers had little infinite of its accuracy which was due to the extreme gap of dividend payment Of some companies amongst others in the first decision.
Positive relationship between tax bracket level and dividend payouts can also be seen at the result of second hypothesis test.
It is true that this condition applied to individuals, but it is also attributable to constitutions as the investors/ stockholders. Although institutions such as pension fund companies and foundations are often tax exempted, which should make them indifferent on dividend policy, other companies are actually facing lower dividend tax than capital gain tax. Therefore, a high dividend payout seems to be more attractive to companies (as investors) rather than a high return on capital gain (Elton and Grubber 1 970, 56).
Further studies of tax on clientele effects were also being conducted by Franklin Allen, Antonio E. Bernard, and Vivo Welch under two assumptions which were the ‘untaxed institutions’ and ‘taxed individuals’ followed with the assumption that dividend is a way of attracting institutions. This study by Allen and his colleagues was apparently more focused on how dividend payments attract institutions. First of all, it is agreed that institutions prefer ND attracted to dividend payment because they are mostly tax-free or largely tax-exempt.
But it is also believed that the same reason has made institutions more inclined to be better informed about the firm (Allen, Bernard, and Welch 2000, 2500). Institutions that are better informed have the tendency to be able to perceive the quality of companies, which in this case is reflected through companies’ amount of dividend payment.
Discovery of this finding by Allen and his colleagues has brought new plausible reason of dividend policy, that effect on tax is actually related to signaling theories information asymmetries and agency effect).
Looking from the overall point of view, all Linter, Elton and Grubber, and Allen along with his colleagues’ findings are actually related. But how does all these findings precisely connected?
- Lintier stated that companies prefer to payout stable dividend with gradual growth.
- Elton and Grubber stated that companies with high payouts tend to experienced more growth compare to companies with low dividend payouts.
- Allen, Bernard and Welch stated that dividend payouts signal quality of firm.
All these three dots can be connected into one line. Dividend is used to signal ‘good’ management hence reflects quality of firms, which helps to forecast the reason behind companies preference of a stable dividend payments and hardly ever reduce their dividend payments (Lintier 1 956), as it will be perceived as a reduction in quality of companies performance (Tangelo and Deadening cited in Allen, Bernard, and Welch 2000, 2518).
Untaxed/low taxed earning on dividends has made constitutions to become major shareholders of a company, and when companies’ dividend payout reduce, institution investors will probably perceive that as a negative signal and sell big chunk of heir shares which in the end could reduce companies’ share price. Thus, the condition allows company who paid high dividend tend to grow at faster rate because institutions as major shareholders are co-exist with companies’ quality, it will force company to defend or even improve its performance to avoid unwanted situation.
Therefore, the clientele effect will allow companies to gauge their clientele, hence companies could optimism its dividend policy accordingly so it would suitable with clientele tax preference.
Dividend as signaling model is believed to be often used by companies when here are information asymmetries (which in the end often leads to agency problems). Information asymmetries assumed to happen when companies’ managers have insider information unknown to outside investors or shareholders (Demented and Whereat 1998, 880).
The so called unknown information is often associated with future cash flow known by managers, but closed to outside shareholders (Apothecary 2007, 6) which in the end results in different interest priority thus led to agency cost. The information about possible future free cash flow is connected with the amount of paid evident.
The higher the possible future cash flow, the higher the dividend payment will be (Apothecary 2007, 7). Hence, generally, economists suggested when an information gap occurs between firms and shareholders, companies are usually giving signals through dividend payments.
Majority of researchers and economists have given a great thought about this, and reach to conclusion that the great presence of asymmetric information has made companies become more active in paying dividends (Lie and Chaos 2008, 674). This statement is basically valid under the condition of:
- Companies truly possess superior quality (Allen, Bernard, and Welch 2000, 2505).
- Managers are willing to bridge information gap by releasing insider information through dividend payments (Malawi, Rafter, and Pillar 201 0, 187).
Firms’ true quality is important in signaling model because dividend announcement is usually perceived as good sign thus often follows by increasing share price (Malawi, Rafter, and Pillar 2010, 187). Analyzing firms’ true companies are generally fall into managers’ hand, who will consequently determine companies’ dividend policy. Empirical studies by Allen and his colleagues has shown that firms who are truly superior in value and quality Will eventually choose to conduct dividend announcement as ‘signal’ even though capital gain is a more preferable option.
However, if managers of firms fail to understand companies’ true value yet still decide to pay dividend, share price could increase in consequence of market response to wrong signal. But when firms’ true value are revealed, disappointment and disadvantages would be felt by investors due to taxed dividend meanwhile companies’ true value revealed to be poor, hence they will choose to sell their hares and firms’ capital paid for dividends would be a waste (Allen, Bernard, and Welch 2000, 2509).
Condition above will stay true, when agency problem is out of consideration. However, when agency problems are being inserted to the scene, it is possible that manipulation of free cash flow would happen, because of the priority difference between managers and shareholders.
Various researches have been done to determine underlying firms’ characteristics that make them prone to agency cost. There are two characteristics believed to be possessed by companies, which uniquely, seem o be contradictory in nature.
Jensen-Neckline argued that when companies’ use internal equities rather than external equities, it presented a company with condition where ownership of company are mainly dominated by inside shareholders, which in the end could reduce agency cost because companies’ insiders as shareholders’ could better aligned the interests (Kink 2001, 82). On the opposite, when external shareholders dominate ownership of company, it would incur shareholders’ monitoring cost because they need to monitor managers’ behavior due to the possibility of conflicted interest (Malawi, Rafter, and Pillar 2010, 190).
Grossman and Hart (cited in Allen, Bernard, and Welch 2000, 2510) suggested that dominant shareholders such as institutional investors could actually be helpful in controlling agency problem. It was written above that outside shareholders dominance will increase shareholders’ cost due to constant surveillance of possible conflicted interest, which according to Staterooms (1984) is necessary because in the end it would lead to lesser agency problems. Institutional shareholders as major ownership has been a trend because they are mainly tax free or taxed in small rate.
Existence of institutions as major shareholders are usually able to influence the board decision, thus is believed could be helpful in controlling agency problems.
Conquering agency cost problem in companies has been a tricky issue, Demented and Harsher (1 998, 881 ) believed that reducing the amount of cash flow in firms’ managers’ hand, therefore it is less likely for managers to waste the free cash. This method is also trusted to make dividend payment become more transparent therefore reliable source to predict future cash flows.
In short, dividend will disciplined managers and force them to obey the rule of uncial market. Utilizing dividend as solution is also agreed by Allen and his colleagues who proposed the same solution to agency cost on their paper research. Looking to aggregate past research above, it is pretty clear that asymmetric information does impact companies’ dividend policy. Presence of agency cost and managers’ failure in detecting true firms’ value has made dividend less ideal as signaling model.
To overcome the overall dilemma, Allen and his colleagues proposed an idea of high firm value and low firm value separation. Separation of firms’ value can be done through giving trotter signal, in simpler term means pay higher dividends. Yet, if the cost of separation is felt to burden companies and too troublesome, it would be better if companies do not pay out dividend at all. However, another interesting research conducted by Lie and Chaos in 2008, surprisingly yet interestingly concludes opposite outcome, although U.S.
Companies are used as subject of research (it is stated that Lie and Chaos gather information from CRISP and BIBS on page 674). Employing earning forecast error and dispersion (to estimate the degree of information gap between companies and shareholders) as research method has led them to conviction hat dividend signaling has negative relationship with information asymmetries. Companies exposed to information asymmetries are believed to be less likely to announce, omit, and change dividend payments.
They believe that companies with high transparency (less asymmetric information) are the one who more likely to announce dividends. Findings such ultimately cast doubt regarding signaling model accuracy. Signaling model also seemed to be inaccurate in different market.
Most of economists who considered dividend as signal of quality or existence of information asymmetries are probably subjected to biasness because U. S. Based companies are utilized as study subjects (except for Lie and Chaos who stated otherwise).
Fresh perspective of dividend policy in different market were brought to attention by Demented and Wanted in 1 998 who observed the effect of dividend signaling between Japanese and American firms.
Subject of study were taken in 1983 from S&P 500 index (for American companies) and Morgan Stanley index (for Japanese companies) and observed using descriptive statistics, two tailed t-test, and chi-square test static (Demented and Wanted 1 998, 883). The result of studies has given very important knowledge, that in Japan, interaction between managers are shareholders are more intense compare to U. S. Market.
Anent communication between two parties has successfully minimized information gap as well as agency cost which makes dividend announcement and dividend payment are no longer indicators of information asymmetries and do little influence to companies’ share price. Thus, Japanese companies are rarely hesitates in altering dividend payment (including dividend cut) as response to earnings change due to the fact that dividends in Japan tend to heavily rely on companies’ earnings.
Long time studies have given various results and choices for the most plausible answers of dividend policy puzzle.
Financial study, investors, and even economist get glimpses here and there in process of understanding dividend policy. For example, famous studies by economists such as Lintier, Modeling, Miller, and Patriarchy have become benchmark of further observation in dividend policy study.
It has challenged other economists to conduct more meticulous study which considers different variables also hypothesis in process of observation. Plausible answers are many in variation UT it point to some most possible understanding, which two are being discussed in this paper.
The first theory discussed above is the clientele effect. Clientele effect is strongly related with Modeling and Miller dividend irrelevance theory (assume market is a frictionless place therefore dividend policy to become irrelevant) which does not explain the reason of dividend payments by firms. Therefore, clientele effect is proposed as complement theory.
Tax is one of the most important factors that cannot be ignored in real life business, so tax on firms’ dividend policy is thoroughly examined.
The result explains that different level of tax faced by shareholders shape different dividend policy preferences. The most important point to be noted is that people who fall in low tax bracket or tax exempt (or small taxed) institutions tend to prefer dividend payments because dividend are taxed lower compare to capital gain. And so does the opposite, people who fall in the higher tax bracket would favor small or no dividend payment as they expected high return on capital gain. These two different groups of shareholders are called as ‘clientele.
In the end, this clientele will shape impasses’ dividend policy so it could suit their tax preference.
But, past study has said that dividend payments are usually attracting institutions because it gives them tax advantages which cause them to become the major shareholders Of companies and make companies are more inclined to make institutions as clientele. The reason of why companies rather pay dividend than do repurchase has been studied and discussed, unfortunately not under the clientele effect. The clientele effect has explained that type of clientele will determine the amount of paid dividend.
Another study has pointed out hat the amounts of dividend paid by companies are actually reflecting the quality of companies, and this study is commonly known as signaling models. Dividend is often used as signal of quality when there is asymmetric information between firms and shareholders.
When firms have information unknown to outside shareholders, it is assumed that firms are trying to signal shareholders through dividend payment. Under signaling theory, high dividend is often associated with high quality while low dividend is often associated with low quality firm.
This statement is true when companies who announce dividend payment are truly high quality firm, and firms are willing to bridge the information gap. The bad news is under signaling model, asymmetric information often arises due to agency problem. Firms’ managers who hold the free cash flow have conflicted interest (considering outside shareholders is own majority of ownership). Managers want to use the free cash flow to invest in somewhere else instead of using it as dividend payment, thus, the amount of paid dividend has become less reliable quality of the firm.
- This is where the problem starts to get complicated.
Jensen Neckline stated that major inside ownership to be the solution of problem because managers can align their interests together. On the other hand, Grossman, Hart, and Staterooms said major outside ownership could overcome agency problem because outside shareholders are aware Of the possibility of conflicted interests, therefore do strict surveillance, plus become major party of companies’ ownership give outside shareholders higher chance of influencing boards’ decision. Overcoming problems under signaling model can be done through higher dividend payment, according through Allen and his colleagues.
However, if this is too burdensome, it might e better to pay no dividend at all. The overall point from signaling model is so far valid.
Surprisingly, the signaling model might be biased because U. S. Companies are used as research subject. Demented and Harsher said in some market such as Japan, dividend as signaling model is simply not true because agency problem and information asymmetries are small.
Other surprising research result was carried out by Lie and Chaos who use U. S. Companies as subject research ( but method of research is different from Allen’s’) yet still find that dividend as signaling model is invalid.