• No results found

Catering of dividends : evidence from the Asian Tigers'publicly listed firms

N/A
N/A
Protected

Academic year: 2021

Share "Catering of dividends : evidence from the Asian Tigers'publicly listed firms"

Copied!
35
0
0

Bezig met laden.... (Bekijk nu de volledige tekst)

Hele tekst

(1)

Evid

Name: Ev Student n Universit Faculty E Field: Co Supervis Date: Jun

dence fr

vghenia And number: 104 ty of Amster Economics a orporate Fina or: Drs. Pep ne 29 2015

Ca

rom the

Ba

drusova 417508 rdam and Business ance ijn Trietsch,

atering

e Asian

achel

s , M. Phil.

of divi

n Tiger

lor T

idends:

s’ publ

Thesi

licly list

is

ted firm

ms

(2)

2 Abstract

This paper explores the corporate dividend payment behavior of the publicly listed firms from the Four Asian Tigers: Singapore for the period 1989–2015, Hong Kong for the period 1992– 2015, South Korea for the period 1993–2015, and Taiwan for the period 1996–2015, using the OLS regression model of catering theory of dividends. Empirical examination demonstrates that when corporate managers initiate dividend payments, they are catering to investors’ demands. This result is statistically significant and refers to Singapore, Hong Kong, and South Korea. However, there is no evidence of catering incentives for corporate managers who decide to continue to pay dividends for all four countries.

(3)

3 Table of Contents

Abstract ... 2 

1.  Introduction ... 4 

2.  Literature review ... 6 

2.1. Traditional dividend-policy theories ... 6 

2.1.1.  Competing dividend relevance theories ... 6 

2.1.1.1.  Taxes and clienteles ... 7 

2.1.1.2.  Information asymmetry: signaling theory ... 7 

2.1.1.3.  Agency costs: the free cash flows hypothesis ... 8 

2.2. Behavioral explanation of dividend puzzle: catering theory of dividends ... 8 

2.2.1.  Dividend puzzle ... 8 

2.2.2.  Background of catering theory of dividends and its concept ... 9 

2.3. The model of catering theory of dividends ... 9 

2.3.1.  Basic catering model ... 10 

2.3.1.1.  Assumptions of the model ... 10 

2.3.1.2.  The model ... 10 

2.3.2.  Extension: the catering model based on continuous dividend levels and frequency .. 12 

2.4. Measurement of catering of dividends ... 13 

2.4.1.  Proxies of dividend premium ... 13 

2.4.2.  Dividend payment variables ... 14 

2.4.3.  The empirical model ... 14 

2.5. Evidence from the empirical studies in different countries ... 15 

3.  Methodology ... 18 

3.1. Data collection ... 18 

3.2. The empirical model ... 19 

4.  Empirical results ... 21  4.1. Regression results ... 21  4.2. Discussion ... 21  5.  Conclusion ... 24  References ... 25  Appendixes ... 28 

(4)

4 1. Introduction

According to Modigliani and Miller (1961), who assumed that the capital market is efficient and perfect, dividend policy is irrelevant to firm value. However, the conservative position of these researchers has been criticized by behavioral scientists. These behaviorists relax the assumption of perfect and efficient markets and consider psychological phenomena as explanations for dividend policy decisions.

The behavioral scientists Baker and Wagner (2004a) recently proposed the “catering theory of dividends.” The main idea behind this theory is that corporate managers rationally cater to investor demand. That is, managers pay dividends when investors value dividend paying firms, and they do not pay dividends when market participants prefer firms that are non-dividend payers. Some investors have an uninformed and perhaps time-varying demand for dividend-paying stocks. For this reason, arbitrage fails to prevent the situation in which demand drives apart the prices of payers and non-payers.

As such, investor behavior and the managerial response, such as increasing or decreasing dividend payments, affects firm value in the markets, which is opposed to what Modigliani and Miller (1961) posited. According to them, higher taxes on dividends than on capital gains should lower stock prices. Under the behavioral view, however, the prices of dividend-paying stocks can change in any direction, which depends on investors psychological particularities and mood. If there is investor demand and investors place dividend premium on top of the price of non-payers, the price of dividend-paying stock increases. These competing views about what determines dividend policy make this topic interesting for further study.

The results from testing the catering theory are mixed and controversial. For example, Kuo, Philip, and Zhang (2013) researched both European and US markets and concluded that catering theory does not explain trend of disappearing dividends after controlling for idiosyncratic and systematic risks. This contrasts with the findings of the authors of the theory, Baker and Wurgler (2004a), who investigated US markets, and Lu, Xi, and Lu (2014), who studied Chinese markets. They found support for the catering theory. Thus, further testing and evidence gathering to investigate the catering theory is relevant for understanding dividend-policy determinants.

Although there is some evidence of the catering theory from Europe and the US, there is little evidence from Asia. For example, there is research of catering theory in China, based on data from 2004 to 2009, in Jordan from 1999-2013 and in Japan from 1987 to 2006. For the reason that there is little evidence from Asia and there is no research for the period up to 2015, some Asian countries with developed financial markets have been selected to provide evidence about the catering theory, whether managers cater to dividend demand. The selected countries are the Four Asian Tigers: Singapore, Taiwan, Hong Kong, and South Korea. The main research question is: is catering a determinant for dividend policy in the Four Asian Tigers’ exchange-listed firms?

To answer this question, empirical testing has been performed. The main objects of study are companies listed in the stock exchanges of Singapore, Taiwan, Hong Kong, and South Korea. The sample of at least 30 companies was taken for each year in the period 1992-2015 for Hong-Kong, 1989-2015 for Singapore, 1993-2015 for South Korea, 1996-2015 for Taiwan to create a time series of variables, which is required for testing the catering theory of dividends.

(5)

5 The empirical model and testing procedure follow in general the methodology from the authors of the catering theory of dividends, Baker and Wagner (2004a). However, the model is simplified to receive answer on research question in time. Although there are extensions of this model, the original model and two proxy of it has been used. The model satisfies the scope and goal of the current paper in answering the research question, which requires merely identification of the presence of catering incentives for the managers of the companies.

The paper is structured as follows: Section 2 represents a review of the literature related to the topic; Section 3 explains the methodology, describes the data selection and collection process as well as the empirical model; Section 4 presents the results; and Section 5 provides the conclusion.

(6)

6 2. Literature review

This section is structured as follows: First, the traditional dividend-policy theories are introduced. Next, challenges to the dividend irrelevance theory are discussed along with competing theories and the dividend puzzle. This is followed by behavioral view on dividend policy determinants and discussion of the catering theory of dividends. Further, the measurement of catering effects are explained. Finally, earlier research about catering of dividends in different countries is presented.

2.1. Traditional dividend-policy theories

Determining an appropriate payout policy is difficult, requiring a balance between many conflicting forces. Paying dividends affects both shareholder wealth and a firm’s ability to retain earnings for investment and exploitation of growth opportunities. In practice, many corporate managers carefully consider the choice of dividend policy. According to a survey from Baker and Powell (1999, p. 17), managers are careful about dividend policy because they believe such decisions affect firm value and shareholder wealth. However, the view of these managers, that firm dividend policy is relevant, is in competition with the arguments provided by Miller and Modigliani, who are the authors of the dividend irrelevance theory.

Franco Modigliani and Merton Miller (hereafter MM) showed that, under certain assumptions, such as perfect capital market, rational behavior of economic agents and perfect certainty about future financial benefits, a firm’s mixture of debt and equity does not affect its overall value. In 1961, MM reported similar results for dividend-payout policy, concluding that while investment decisions affect firm value, financing decisions are irrelevant. So, decisions about payout policy don’t impact on value of the company.

2.1.1. Competing dividend relevance theories

DeAngelo and DeAngelo (2006) are critics of MM’s propositions. They argue that dividend irrelevance relies on an unstated condition that requires firms to pay out 100 percent of their free cash flows each period to make the cost of internal and external equity equal. If firms do not distribute all free cash flows, then internal equity is cheaper than external equity. However, MM’s theory lends itself to the conclusion that all feasible payout policies are equally valuable for investors. DeAngelo and DeAngelo (2006, p. 294) point out that the set of possible payout policies is not as limited as assumed by MM. In their opinion, payout policy matters.

Also, MM suppose that markets are perfect, investors are rational, and the future is not uncertain. Researchers have responded to MM’s conclusion about dividend policy irrelevance by offering competing theories. These theories include hypotheses as to why dividend policy matters, why investors want dividends, and why corporations pay them.

(7)

7 2.1.1.1. Taxes and clienteles

Taxes were among the first imperfections to be researched. As dividends are traditionally taxed at higher rates than capital gains, firms paying smaller dividends should trade at a premium. Farrar and Selwyn (1967) and Brennan (1970, p. 426) discovered that different personal and corporate taxes change MM’s conclusion. Farrar and Selwyn (1967) argue that in the case of higher taxes on personal income than on capital gains, an optimal dividend policy provides zero dividends, otherwise investors are hurt. Brennan (1967, p. 454) came to the same conclusion; however, he stated that dividends are important and reconciled the policy of zero dividends, pointing out that many companies pay dividends and recommending share repurchases.

Several studies have tested the empirical implications of the aforementioned authors. Black and Scholes (1974, p. 21) found no significant relationship between dividends and stock prices. They recommended that corporations choose a dividend policy assuming that it will have no permanent effect on stock price. The author’s recommendation to investors was to ignore dividends and concentrate instead on diversification of portfolio.

Miller and Scholes (1978, p. 333) found no support for the view that the market prefers dividends and for the contrary view, that dividend paying stocks result to a return premium to compensate for the tax penalties on dividend income. They believe that investors are able to offset dividend income tax disadvantages by investing dividends in companies, which are tax protected, such as life insurance and pension funds.

However, Peterson, Peterson, and Ang (1985, p. 280) argued that investors do not try to offset dividend taxes. The estimated marginal and effective tax rates on dividends indicate that individuals pay taxes at higher rates, even if they are higher than capital gains taxes. The authors attributed this to the dividend puzzle.

As for clienteles, some investors demand dividends, and firms adjust dividend policy to attract certain types of investors. MM (1961) recognized the possibility of clientele effects but stated that valuation does not change despite such effects, because the difference in tax rates is small.

In contrast to MM, Shefrin and Statman (1984) argued that the clientele-effect exists because of psychological reasons, as investors prefer cash dividends. One type of clientele is discussed by Allen, Bernardo, and Welch (2000), who argue that a tax clientele exists. The idea of a tax clientele is that investors, who are not affected by change in the dividend tax rate (pension funds), prefer dividends while those who are affected do better if firms repurchase shares. This means that firms can attract a desired type of investor by changing payout policy (John and Williams, 1985).

2.1.1.2. Information asymmetry: signaling theory

The basis of signaling theory is the presence of asymmetric information. It supposes that managers have access to information that the market does not. Hence, corporate financial decisions can be viewed as a signaling tool. Firm managers use dividend policy to communicate information and reduce asymmetry.

(8)

8 Bhattacharya (1979) argued that if taxes on dividends are high and to pay them is costly, dividends can be used as a signal of expected cash flows or future earnings. In support of the signaling theory, Asquith and Mullins (1983) found that stock prices increase by an average of three percent after the signaling.

However, the research of Grullon, Michaely, Benartzi, and Thaler (2005) does not support the signaling theory. They showed that after controlling for non-linear patterns in the behavior of earnings, dividends do not signal information about future earnings. The researchers also mentioned that Watts (1973), Gonedes (1978), Penman (1983), DeAngelo, DeAngelo (2006), and Skinner (1996) found little or no evidence that dividend changes predict abnormal increases in earnings (Grullon et al., 2005, p. 1659). Also, Ang (1987) questioned signaling on a theoretical level, finding that dividends are too costly given the strength of their signal. Ang (1987) thinks that firms can use other signals that are cheaper instead expensive dividend signals.

2.1.1.3. Agency costs: the free cash flows hypothesis

The free cash flow model relies on the assumption that there is an agency problem between corporate insiders and external investors (outsiders), or a divergence of interests between managers and investors. Corporate managers have incentives to keep cash, invest in low-risk projects to continue to use firm’s property in private interests and save their jobs. In contrast, shareholders like to maximize their returns and want managers invested in high-return projects. Increased dividends mean that excess cash for managers is reduced and they cannot misuse funds. For this reason, markets react positively on dividend initiations or increases (Jensen, 1986, p. 323).

Lie (2000) provides arguments and strong empirical support for the hypothesis, proving a relationship between stock-price reaction and cash distribution from firms, which tend to overinvest. Also, he mentions that large, incremental distributions mitigate the agency costs associated with agency funds (Lie, 2000, p. 245).

2.2. Behavioral explanation of dividend puzzle: catering theory of dividends 2.2.1. Dividend puzzle

Black (1956) referred to the controversy between the irrelevance of dividends and the behavior of investors and corporation as if dividends are important as the “dividend puzzle.” (1976, p. 8) He assessed the contributions of the above researchers, which appeared after MM’s irrelevance theory. Following evaluation, he stated that “the harder we look at the dividend picture, the more it seems like a puzzle, with pieces that just don’t fit together” (Black, 1976, p.5). Also, Feldstein and Green (1983, p.17) said, “The nearly universal policy of paying substantial dividends is the primary puzzle in the economics of corporate finance.”

Despite the range of existing explanations, researchers developed additional theories to explain the dividend puzzle, including behavioral explanations. Behaviorists have provided their

(9)

9 own views on what determines dividend policy, why investors like dividends, and why corporations pay them. The most recent behavioral dividend theory is the catering theory of dividends.

2.2.2. Background of catering theory of dividends and its concept

The catering theory of dividends was offered relatively recently by Baker and Wurgler (2004b). The authors were inspired by an article from the well-known researchers Fama and French (2001). Although many previous studies had been dedicated to dividend policy, Fama and French investigated the time series in dividend policy of US industrial companies.

They reported a meaningful decline in the fraction of firms that pay dividends. This is because the population of publicly-traded companies changed to smaller firms with high-growth opportunities but low profits, which are less likely to pay dividends. However, firms were willing to pay fewer dividends, even after controlling for the factors of size and growth. This phenomenon was called the lower propensity to pay disappearing dividends (Fama and French, 2001).

Baker and Wurgler (2002) decided to investigate the causes of the disappearing dividends in the US. They examined different explanations based on the following: the issue of management stock options, asymmetric information, agency costs, taxes, dividend clienteles, and investor sentiment catering. They concluded that catering incentives best explain the change in the propensity to pay dividends. Thus, described above competing theories don’t explain decreasing propensity to pay among companies and this is the reason, why I described all the theories above, and I am discussing catering theory and not another theory.

2.3. The model of catering theory of dividends

The catering theory supposes that firms cater to irrational investor demand for dividend-paying and growth stocks to boost share prices above their fundamental values (Baker, Ruback, and Wurgler, 2005, p. 2). The catering theory relies on the view of behavioral researchers that psychological characteristics of investors play a role in financial markets and that irrational behavior limits the effectiveness of arbitrage. This contrasts with MM’s assumptions about perfect markets, investor rationality, and other traditional theories, which cannot explain, for example, the research results of Long (1978), Poterba (1986), or Hubbard and Michaely (1997). These researchers found that shareholders of Citizen Utilities price cash-dividend share classes differently than stock-dividend share classes, although the payouts are of equal value. Thus, the catering theory of dividends has additional explanatory power compared to other traditional theories. So, it is worth to study this theory.

(10)

10 2.3.1. Basic catering model

The empirical model of the catering theory of dividends represents a discrete model in which investors divide firms as dividend payers and non-dividend payers.

2.3.1.1. Assumptions of the model

Baker and Wurgler (2004a, p. 1125) built the empirical model on three essential assumptions. The first assumption is that investors have uninformed demand for dividend-paying stocks, which varies over time. The main reasons for variation in demand are the psychological particularities of investors and institutional regulations.

The second assumption is that MM’s arbitrage fails to prevent investor demand from driving apart the price of non-dividend payers and dividend payers.

The third assumption is that managers rationally cater to investor demand for dividends by paying dividends when investors prefer dividend-paying stock and place a premium on it. Managers do not pay dividends when investors prefer stock that does not pay dividends, such as when there is economic growth and many growth opportunities (Baker and Wurgler, 2004a, p. 1125). Thus, corporate managers in times of stagnation can initiate dividends to increase stock price or market value.

The difference between the prices of dividend-paying stock and non-dividend-paying stock is called the dividend premium. This premium is high in high-demand years and low in low-demand years (Baker and Wurgler, 2004a, p. 1127).

2.3.1.2. The model

Baker and Wurgler (2004a, p. 1128) assume that a firm has Q shares outstanding that pay a liquidating distribution of V= F+ε at time t=1. F is the expected fundamental value of the firm, and ε is a standard, normal random variable with a mean of zero. At time t=0, managers decide to pay dividends per share of d {0, 1}. If there is a cost to issue dividends (c), the dividend payment reduces the firm’s liquidation value by d(1+c). c captures the trade off between dividend and investment policy, as imposed by taxes or the cost of external financing.

Also, there are two types of investors, category investors and arbitrageurs, who have aggregate risk tolerance per period of and respectively. Dividend payments are important for category investors. These investors place dividend payers in a different investment category. Baker and Wurgler (2004a) provided several explanations for the demand of category investors. One reason is that dividend clienteles are based on taxes, transaction costs, and institutional investment constraints (Black and Scholes, 1974; Allen, Bernardo, and Welch, 2000; in Baker and Wurgler, 2004a, p. 1128). Other reasons are that investors use dividends to infer firm prospects. If dividends are paid, investors assume that there are no profitable opportunities. Also, dividend-paying companies are perceived to be less risky for some shareholders, who prefer certainty of cash flow at the present time over profitable but uncertain cash flows in the future.

(11)

11 The last reason noted by Baker and Wurgler (2004a) relies on the research of Thaler and Shefrin (1981) and Shefrin and Statman (1984). The authors argued that some investors prefer dividend-paying stock to fight self-control problems.

In the model, Baker and Wurgler (2004a) assume that category investors do not recognize the costs of issuance of dividends (c) and misestimate the mean of liquidating distributions of payers ( ) and non-payers ( ). The arbitrageurs correctly estimate the liquidating distribution to be F if the company does not pay a dividend and F – cd if it pays a dividend. Misestimating of category investors results in the difference between prices of payers and non-payers of dividends, because arbitrageurs are risk averse.

At time t=0, k is the group of investors, k=A if the group includes arbitrageurs, and k=C if the group consists of category investors; ; and . The demand of investor group k equals:

(2.1) Given that the market-clearing condition is and substituting the equation above

in the condition, the authors found the prices for dividend payers ( ) and non-payers ( ): ≡

(2.2) Given these prices, managers decide about distribution of dividends. They are risk neutral and care only about the total value of distributions and the current price of the stock, which they can influence through the cost c. In short-run inefficient markets, managers make a tradeoff to maximize the long-run fundamental value of the firm or its short-run price, depending on category investor demand. Managers will take into consideration time horizon ( ). The managers’ problem is to maximize max 1 ∗ . This results in the following

solution:

≡ . (2.3)

According to the formula, managers pay dividends if premiums are positive and exceed long-run costs, which is the term to the right of the inequality (Equation 2.3). The propensity to pay dividends increases if c decreases as well as horizon (λ) (Baker and Wurgler, 2004a).

Here, the announcement impact of dividend initiations is positive and increases the dividend premium. If the announcement is negative, such as an omission of dividends, such stocks are ignored by category investors and are attractive only to arbitrageurs. Then, price is equal to the following:

(12)

12

. (2.4)

Thus, the decision to initiate dividends is governed by Equation 2.3. The decision of the former payer to continue to pay dividends is governed by Equation 2.5:

(2.5)

The propensity to continue paying dividends increases if c decreases and dividend premiums increase. Even if initiation is interactive, current dividend payers may continue to pay dividends if the price impact of omitting is large. This explains why some firms initiate (reinitiate) dividends despite negative dividend premiums and why such initiations can still have a positive announcement effect (Baker and Wurgler, 2004a).

2.3.2. Extension: the catering model based on continuous dividend levels and frequency

Li and Lie (2006) extended the catering theory of dividends by examining changes in dividend levels in the US. In addition, Ferris, Noronha, and Unlu (2006) studied the relationship between dividend payment frequency and catering for the United Kingdom.

Li and Lie (2006) noted that the corporate managers in the model proposed by Baker and Wurgler (2004a) can only make decisions about whether to pay dividends, not how much to pay. They argued that this is a limitation of the model because, in practice, many corporate managers often need to determine changes in the levels of dividends. Li and Lie (2006) decided to revisit the catering theory due to a lack of a significant relation between stock market reactions and dividend premiums, casting doubt on the catering theory as a whole.

Lie and Lie (2006) investigated US firms between 1963 and 2000, using an extended version of the catering model with continuous dividend levels. They discovered that the decision to change the dividend levels, and the magnitude of the change, relate to the dividend premium assigned by capital markets. They also found that the magnitude of dividend increases is positively related to the dividend premium. So, firms increase dividends when the dividend premium is high and repurchase shares when the dividend premium is low.

In researching US companies, Baker and Wurgler (2004a) did not find a statistically significant relationship between the dividend premium and the average announcement effect of dividend initiations. However, Li and Lie (2006, p. 307) discovered that the reaction of the stock market depends on the dividend premium. When the dividend premium is high, firms are more likely to increase dividends; the increases of dividends tend to be larger, and the stock price reaction on such positive news is favorable. Managers who do not take into account the demand of investors on dividends are penalized with low prices, and those who pay dividends are rewarded with a dividend premium.

In comparing the results of Baker and Wurgler (2004a) with those of Li and Lie (2006), it is worth mentioning that decisions about dividend increases or decreases are less sensitive to the

(13)

13 dividend premium than decisions about dividend initiations and omissions. This is supported by the results of another study by Lie (2005), which concluded that the average announcement returns of dividend omissions are larger than the average announcement returns of dividend decreases.

Ferris, Noronha, and Unlu (2010) investigated 21 civil-law countries and 11 common-law countries between 1996 and 2003 in order to determine the structure and determinants of dividend-payment frequency. They discovered that dividend-payment frequency is higher in civil-law countries than dividend frequency in common-law countries. This means that dividend payments are important to investors in civil-law countries where investors are less protected and managers have higher incentives to use earnings for private purposes. Moreover, the authors mention that catering incentives influence the frequency of dividend payments. They document the positive relationship between dividend frequency and firm value.

2.4. Measurement of catering of dividends 2.4.1. Proxies of dividend premium

Baker and Wurgler (2004a) defined several proxies to measure the presence of catering incentives for managers or the dividend premium. The first proxy is based on market valuations of dividend payers and non-dividend payers. The authors calculated the market-to-book ratio as book assets minus book equity plus market equity minus preferred stock liquidating value plus balance-sheet deferred taxes and investment tax credits minus postretirement assets. Market equity is the product of stock price and the number of shares outstanding at the end of the calendar year. The market-to-book ratio is equally or value weighted across dividend payers and non-payers. The dividend premium ( ) is the difference between the log-normally distributed average market-to-book ratio of dividend payers and non-dividend payers:

ln ∑ ln ∑ (2.11)

where is the weight of firm i in the subset of payers in year t, is the market value of firm i in the subset of payers in year t, is the book-value of firm i in the subset of payers in year t, is the weight of firm i in the subset of non-payers in year t, is the market value of firm i in the subset of non-payers in year t, and is the book-value of firm i in the subset of non-payers in year t.

Baker and Wurgler (2004a, p. 1126) mentioned that investor sentiment partially determines the dividend premium. When investors look for safety, they prefer dividend-paying stocks, and therefore price grows. When investors believe in growth of the economy, they choose stocks that do not pay dividends and bid up the price of such stocks.

(14)

14 Another measure of dividend premium is the difference between the price of Citizens Utilities’ cash-dividend and stock-dividend share classes. Between 1956 and 1986, Citizens Utilities had two classes that were different in form but not in payout.

The third measure of dividend premium is the average three-day announcement effect of recent dividend initiations. This proxy should predict whether investors like to receive dividends and react positively on dividend initiations.

The forth proxy of dividend premium is the future excess return of payers over non-payers. If a firm initiates dividends to benefit from market mispricing, there will be a negative relationship between the return difference and the dividend initiation rate (Baker and Wurgler, 2004a, p. 1126).

2.4.2. Dividend payment variables

In the catering model of dividends from Baker and Wurgler (2004a), investors divide firms into dividend payers and non-payers. The authors define three variables in capturing dividend payment dynamics:

(2.12)

(2.13)

Payers are the total number of payers (new payers, old payers, and list payers at time t). Old payers are payers who paid last year and continue to pay this year. New payers are the number of initiations from among last year’s non-payers. List payers are the number of payers this year who were not on the list last year. New non-payers are the number of firms that omit dividends although they paid dividends last year. Delist payers are the number of last year’s payers who are not in the sample this year (Baker and Wurgler, 2004a, p. 1133).

2.4.3. The empirical model

The empirical models that Baker and Wurgler (2004a) offered for empirical testing are represented by regression equations:

(2.14)

(15)

15 If the estimated coefficient is positive and significantand there is a positive dividend premium ( ), managers have an incentive to initiate or continue to pay dividends. A is announcement effect, is the log of the ratio of the annual average cash dividend class share price to the annual average stock dividend class share price. According to Baker and Wurgler (2004a, p. 1145) results, coefficients of and are statistically significant, while coefficient c is not. They also mentioned that the value-weighted dividend premium explains 60% variations in the initiate rate and the equally weighted dividend premium explains 52% of variations, while explains only 11%.

2.5. Evidence from the empirical studies in different countries

Several studies have investigated the catering theory of dividends. Researchers have used samples from the United States (U.S.), the United Kingdom (U.K.), Europe, China and Japan as well as cross-country samples to find the presence of incentives for managers to cater to investors. Many of these studies included mixed evidence for the catering theory. Some of them found catering effect and others don’t.

Baker and Wurgler (2004a) investigated U.S. firms from 1962–2000. They related the dividend premium to aggregate levels of dividend initiations and continuations. The authors discovered that the first three dividend premium proxies positively correlated with the rate of dividend initiation. But in a multivariate framework, which included these three proxies, only the first proxy showed a significant positive relationship. Also, there was a significant positive relationship between the dividend premium and continuation rate, signifying that dividend payers are more likely to continue paying dividends when the share price is higher.

The authors of the theory received the following quantitative results. They found coefficient 3.90 for the value-weighted dividend premium, 3.63 for the equal-weighted dividend premium, 1.70 for the Citizens Utilities dividend premium and 2.15 for announcement effects in the model with initiation rate as dependent variable. The first three coefficients were statistically significant and explained respectively 60%, 52%, 11% of variations in rate of initiations of dividends.

In the model with rate of continuations of dividend payments, Baker and Wurgler (2004) found coefficient 0.85 for the value-weighted dividend premium, 0.93 for the equal-weighted dividend premium, 0.44 for the Citizens Utilities dividend premium and 0.03 for announcement effects. The value-weighted dividend premium explained 26% of variations in rate of continuation of dividend payments and equal-weighted dividend premium explained 30% of variations.

Baker and Wurgler (2004a) also performed returns analysis and found evidence consistent with catering theory. Returns on dividend payers are relatively low during the next one to three years, when dividend initiations prevail relative to dividend omissions. This is supported by the fact that when investors are positive about the growth stocks, dividend premiums are negative and non-dividend paying firms attract investors. When investors are in search of stable income and there are no growth opportunities, the high premium reflects this low sentiment of investors. Thus, corporate managers cater to investor demands based on investor sentiment.

(16)

16 Table 2.1

Empirical Studies Investigating the Catering Theory of Dividends

Country Author(s) Sample Period

Evidence for or against Catering Theory

Evidence Against Catering Theory

U.S. Hoberg and

Prabhala (2006) 1963–2000

Evidence against

ˉ The catering theory does not explain disappearing and reappearing dividends.

ˉ The main explanation is that risk, cash flows, and information affect dividend premiums and that idiosyncratic risk significantly explains the decreasing propensity to pay dividends.

ˉ The catering effect disappears if factors such as risk in the mode are controlled.

U.S.

Bulan,

Narayanan, and Tanlu (2007; in Baker and Kolb, 2009)

1963– 2001

Evidence for

ˉ The catering theory explains disappearing dividends.

ˉ Even after controlling for idiosyncratic risk, the dividend premium affects the propensity to initiate dividends.

U.S. Kale, Kini, and Payne (2006, p. 365)

1979– 2005

Evidence against

ˉ The primary goal was to determine how weather information signaling explains dividend initiation.

ˉ However, the additional finding supported the catering theory of dividends.

U.K., France,

Germany Denis and Osobov (2008, p. 63)

1989– 2002

ˉ The authors studied the relationship between propensity to pay dividends and the dividend premium.

ˉ The main finding is that there is little evidence that changes in propensity can be explained by changes in investor sentiment toward dividend payers. ˉ The findings cast doubt on signaling and catering explanations for dividends

but support agency cost-based life cycle theories.

Germany, France, Netherlands, Spain, Belgium, Portugal, Ireland, Austria, Italy Neves (2009, p. 139) 1986– 2003 Evidence for

ˉ The conclusion is that the catering theory is applicable in the Eurozone and that the institutional environment moderates the catering theory of dividends.

ˉ The main discovery comprises three relationships.

First, firms with higher levels of ownership concentration and more efficient boards of directors cater to a larger extent to investor demand than firms with lower levels of concentrated ownership and less effective boards of directors.

Second, stronger investors are legally protected, the lesser the extent to which firms cater to investor sentiment.

Third, corporate managers in market-oriented financial systems do not cater to investors with dividends.

China Lu, Xi and Lu

(2014, p. 508) 2004–2009

ˉ A significant effect of dividend catering in the Chinese stock market: although the stock investment income of cash dividend portfolios is significantly low, investors still irrationally preferred cash-dividend-paying firms.

Jordan Imad (2015, p. 229)

1999– 2013

ˉ Corporate managers cater to investors’ demands for dividends. This provides confirmation of the validity of the catering theory of dividends in the Jordanian market.

Japan Tsuji (2011, p.

1) 1987–2006

Evidence against

ˉ The researcher found evidence that Japanese corporate managers do not cater to investor demand on dividends in the cases of initiation or continuation decisions about dividend payments.

ˉ Japanese managers make decisions based on value-weighted market-to-book ratio, size, dividend yields, and after-tax earnings-to-asset ratios of previous years. Indonesia, Malaysia, Thailand, South Africa, Australia Ferris, Jayaraman, and Sabherwal (2009) 1996– 2000

ˉ The main findings are that firms in common law countries cater more to investors’ demands for dividends than do those in civil law jurisdictions.

(17)

17 Given the findings of Ferris, Jayaraman, and Sabherwal (2009), the following hypothesis was formulated: Support for the catering theory of dividends will be found in Singapore and Hong Kong, which are common law countries, and no support for the catering theory of dividends will be found in South Korea and Taiwan, which have civil jurisdictions.

(18)

18 3. Methodology

This section will clarify the sample received from the databases of Compustat Global and its characteristics and describe the empirical model, the sample’s variables, data construction and empirical results.

3.1. Data collection

According to the research question, the data collection was made for four countries: Hong Kong, Taiwan, Singapore, and South Korea. These countries were selected for several reasons. First, there is little evidence of the catering theory of dividends in Asia. So, it is interesting to investigate them and make new contributions to discussion of this theory. Moreover, these countries are known as growth miracles. Hong Kong, Taiwan, Singapore, and South Korea have consistently maintained high levels of growth since the 1960s based on exports and rapid industrialization (Sarel, 1999). Also, Singapore and Hong Kong are two of the biggest financial centers in the world with a developed stock exchange (Qatar Financial Center, 2015, p. 4). South Korea and Taiwan are two of the leading manufacturing centers in the world (Council on Competiveness, 2013, p. 2). These countries also have stock exchanges and can therefore serve as information sources for this study.

Baker and Wurgler (2004a) used Compustat as a main source of necessary data. The data retrieved from the database are the maximum available data for each country’s stock exchange. Compustat Global has two sections: “Fundamentals Annual” and “Security Daily.”

The data retrieved from the “Fundamentals Annual” include total assets (TA, abbreviation in the system), total liabilities (LT), preferred stock liquidating value (PSTK), deferred taxes, and investment tax credit (TXDITC). The data are used for calculation of book equity according to Baker and Wurgler (2004a, p. 1135). All data are retrieved for the fiscal year end.

In addition, the following data were retrieved: ISO currency (CURCD), stock exchange code, company name, and SEDOL. All retrieved data were filtered by currency type so that all variables were expressed in one currency. Filtering by stock exchange code allows exclusion of information about securities traded in several stock exchanges. The Singapore stock exchange code is 251, Hong Kong’s is 170, South Korea’s are 248 and 298, and Taiwan’s are 260 and 303. SEDOL is an individual company code, which is used for retrieving information from “Security Daily.” The company name is required to match data between two sections of Compustat Global.

The data retrieved from the “Security Daily” are number of shares outstanding (CSHOC), dividends per share by ex-date (DIV), and closing stock price (PRCCD). The data are used to calculate market equity according to Baker and Wurgler (2004a, p. 1135). The prices and shares outstanding are the last values at the end of a fiscal year. In addition, ISO currency code, company name, and SEDOL are retrieved for the matching and filtering of information.

As a result of processing the data, several sets of data were received. The sample for Singapore includes data for the period 1989–2015. The years 1987 and 1988 were excluded from the sample because the companies included during those years were fewer than thirty. According to the Central Limit Theorem, the sampling distribution of the mean of a random sample is

(19)

19 approximately normal for a sufficiently large sample size. In many practical situations, a sample size of 30 may be sufficiently large to use the normal distribution as an approximation to the sampling distribution (Keller, 2009, p. 300). On the basis of both this argument and the need to receive statistically satisfactory results for regression, the years 1987 and 1988 were excluded from the sample. The same was done for other countries as necessary.

The sample for Hong Kong includes data for the period 1992–2015, for South Korea from 1993–2014 except 1998 (for which no data about dividends are available), and for Taiwan from 1996–2015.

3.2. The empirical model

Despite Li and Lie (2006)’s extension of the Baker and Wurgler’s (2004a) empirical model, the current paper is based on the original model of the catering theory of dividends. The extended model is more complicated and supposes that corporate managers make decisions about increasing and decreasing the dividends more often than about initiating or omitting the dividend payments. Although this is true, the paper’s goal is merely to determine whether the catering effect is present in the Four Asian Tigers’ publicly listed firms. Baker and Wurgler’s (2004a) model can be used to answer such questions.

Baker’s and Wurgler’s (2004a) empirical model is based on four proxies of dividend premium, two of which are the focus of this research: equal-weighted market-to-book ratio and value-weighted market-to-book ratio. There are several reasons for this. First, the proxy ), announcement effect, appeared to be insignificant in Baker and Wurgler (2004a). The proxy, the difference in prices of Citizens Utilities’ cash dividend and stock dividend share classes ( ), explained 11% of variations in the initiate rate of dividends, while equal-weighted and value-weighted dividend premium explained more than 50% of variation. Because the calculation of

proxy is complicated and requires access to other databases, CRSP and the time and scope of this research were limited, and this proxy was abandoned. Second, the research question requires ascertaining where there is a catering effect in four Asian countries. Such a question can be answered using the following model.

(3.1) . (3.2) If the beta received is significantly different from zero, given the 95% confidence interval,

there is evidence that corporate managers cater to investor demand for the dividends. This supposes that presence of the dividend premium in the previous period explains why corporate managers begin to initiate dividends in the current period.

The dependant variable is initiate or continue, the explanatory variable is dividend premium. The dependent variables are calculated according to Baker and Wurgler (2004a, p. 1132).

(20)

20 Equation 2.12 was used to calculate the rate of initiation of dividend, and Equation 2.13 was used to compute the rate of continuation of dividend payments. The explanatory variables were calculated by equation 2.11, where the equal weight is determined as one divided on a number of dividend payers or nonpayers. The value weight is determined as the market-to-book ratio of certain companies divided by the total sum of market-to-book ratios of all companies listed in the stock exchange for the year. The calculations for payers and nonpayers are made separately.

The results of calculations of dependant and explanatory variables are presented in Tables A1 and A2 for Singapore, in Tables A3 and A4 for Hong Kong, in Tables A5 and A6 for South Korea, and in Tables A7 and A8 for Taiwan.

(21)

21 4. Empirical results

4.1. Regression results

The results of OLS regression for four countries, the Four Asian Tigers, are presented in Table 4.1. Panel A of Table 4.1. indicates that the increase in equal-weighted dividend premium in time t-1 leads to an increase in the initiation rate in the following year for Singapore, Hong Kong, and South Korea. The results are highly statistically significant at a 99% confidence interval. However, no significant coefficient was found for Taiwan. Also, the equal-weighted dividend premium explains 43% of variations in the initiate rate for Singapore, 49% for Hong Kong, and 42% for South Korea, but it does not provide an explanation for Taiwan.

Moreover, it is possible to see from Panel A that the value-weighted dividend premium increase is associated with an increase in the initiation rate in the following year for Singapore, Hong Kong, and South Korea. However, there is no significant coefficient for Taiwan.

The results are highly statistically significant for Singapore and Hong Kong at a 99% confidence interval and statistically significant for South Korea at a 5% significance level. In Hong Kong, value-weighted dividend premium explains 65% of variations in the initiate rate. This is consistent with Baker and Wurgler’s (2004a) results, where this proxy explains 60% of variations in the initiate rate in the United States. The value-weighted dividend premium explains 39% of variations in the initiate rate for Singapore, 26% for South Korea, and 3% for Taiwan.

Panel B reports regression for the rate of continuation. For all countries, there is no statistically significant coefficient. That means that increase in equal-weighted or value-weighted dividend premiums is not associated with the increase in the rate of continuation. According to Baker and Wurgler (2004a, p. 1144), 30% of variations in the continuation rate can be explained by this proxy in the United States.

4.2. Discussion

The results of this paper suggest that there is a catering incentive for managers of Sinagapore, Hong Kong and South Korea, who determine dividend initiation. These results are consistent with Baker and Wurgler’s (2004a) findings in the United States. The equal-weighted dividend premium explains from 43% to 49% of the variation in initiation rate, while according Baker and Wurgler (2004a), it explains 52%. The value-weighted dividend premium explains 23% to 65% of variation in the initiation rate, while Baker and Wurgler’s (2004a) results explain 60%. The beta-coefficients of both proxies are as statistically significant as in Baker and Wurgler (2004a).

However, the coefficient for both proxies is insignificant for all countries studied. Moreover, it seems that the model does not provide any explanation for why managers continue to pay dividends. The coefficient of determination in the model with value-weighted dividend premium in Baker and Wurgler’s (2004a) research for the United States is 0.26, while it is 0.06 for Singapore, 0.01 for Hong Kong, 0.05 for South Korea, and 0.02 for Taiwan. The coefficient of determination for the model with equal-weighted dividend premium is 0.30 in Baker and Wurgler’s (2004a) findings, while it is 0.03 for Singapore and Taiwan, 0.06 for Hong Kong, and 0.11 for South Korea.

(22)

22 Table 4.1.

Basic Relationships Between Dividend Payment and Demand for Dividends Countries

Measure Correspondence of results to expectations N n Panel A: Singapore 0.50** [0.0004] 25 0.43 7388 yes 0.11** [0.0009] 25 0.39 7388 yes Hong Kong 0.83** [0.0003] 21 0.49 2335 yes 0.38** [0.0000] 21 0.65 2335 yes South Korea 1.07** [0.0021] 20 0.42 10694 no 0.25* [0.0207] 20 0.26 10694 no Taiwan -0.02 [0.8409] 18 0.00 8469 yes -0.03 [0.4761] 18 0.03 8469 yes U.S. 3.63** [5.10] 38 0.52 113975 yes 3.90** [6.56] 38 0.60 113975 yes Panel B: Singapore -0.06 [0.3829] 25 0.03 7388 no -0.02 [0.2405] 25 0.06 7388 no Hong Kong 0.01 [0.9122] 21 0.00 2335 no -0.01 [0.7137] 21 0.01 2335 no South Korea [0.1601] -0.14 20 0.11 10694 yes -0.03 [0.3592] 20 0.05 10694 yes Taiwan 0.33 [0.4601] 18 0.03 8469 yes -0.10 [0.6216] 18 0.02 8469 yes U.S. 0.93** [2.96] 38 0.30 113975 yes 0.85** [2.83] 38 0.26 113975 yes Note. values are given in brackets []. value less than 0.05 means the result is statistical significant and noted with a star*.

P-value less than 0.01 means that the result is highly statistical significant and noted with two stars**. The result of Baker and Wurgler (2004) are presented together with t-statistics in brackets[]. N denotes sample size for regression and n total number of observations. is an equal-weighted dividend premium calculated as the difference between the logs of the dividend payers’ and non-payers’ equal-weighted market-to-book ratios. is a value-weighted dividend premium as the difference between the logs of the dividend payers’ and non-payers’ value-weighted market-to-book ratios. R² is a coefficient of determination, which explains variation in rate Initiate or Continue. The column correspondence of results to expectations shows whether a significant or insignificant result received is in line with the following hypotheses formulated: Support for the catering theory of dividends will be found in Singapore and Hong Kong, which are common law countries, and no support for the catering theory of dividends will be found in South Korea and Taiwan, which have civil jurisdictions. The results for U.S. represent findings of Baker and Wurgler (2004), which are presented for the comparison with the results of this paper.

(23)

23 It is likely that these countries continue to pay dividends for reasons other than investor preference. For example, managers could continue to pay dividends to signal future earnings, which is in line with Bhattacharya’s (1979) hypothesis about dividend policy relevance. Another theory that explains why managers can continue to pay dividends is the cash-free hypothesis, supported by Lie (2000). This question requires further investigation.

In general, the results are in line with the hypothesis of Ferris, Jayaraman, and Sabherwal (2009, p. 1737) that managers in common law countries are more likely to cater to investors’ demands than are managers in civil law countries. This is probably due to the extensive set of rights and protections provided to shareholders in common law countries, which makes managers more responsive to investor preferences for dividends.

By contrast, managers in civil law countries are subject to controlling insiders who hold the majority of shares and have little if any interest in pleasing minority shareholders. Thus, insiders in civil law firms have less interest in dividend catering. Ferris, Jayaraman, and Sabherwal (2009) think that this might be due to either the desire of large shareholders to enjoy the private benefits of control or an unwillingness to respond to what they perceive as temporary market misevaluations in the firm’s equity due to investor preferences. However, the managers of South Korea, a civil law country, initiate dividends based on investor demand. The question of why managers cater to investor demands in South Korea is beyond the scope of this paper and requires further investigation; the answer may involve South Korean legislative particularities.

(24)

24 5. Conclusion

This paper aimed to investigate the presence of catering incentives for corporate managers of stock exchange-listed firms of the “Four Asian Tigers”—Singapore, Hong Kong, South Korea, and Taiwan—using an empirical model with two proxies and OLS regression. According to literature review performed, there were few investigations of the catering theory of dividends in Asian countries. Mainly, these investigations covered the period from 1987 till 2013. In this paper, four Asian countries are researched and the investigation is made for the period from 1989 till 2015.

The empirical model of the catering theory of dividends was used to investigate whether catering determines the dividend policy in the Four Asian Tigers’ publicly listed firms. The regression results suppose that there is catering incentive for the corporate managers to pay dividends if there is a dividend premium on the market, which reflects investors’ wishes to hold dividend-paying stocks. The coefficient of value-weighted and equal-weighted dividend premiums is statistically significant in the case of initiation rate. Thus, initiation of dividend payments is explained by the catering theory of dividends for Singapore, Hong Kong and South Korea. This result is in line with the results of Baker and Wurgler (2004a). The results for Singapore and Hong Kong are also in line with the findings of Ferris, Jayaraman, and Sabherwal (2009). As was expected, the evidence for catering theory of dividends was found in the common law countries Singapore and Hong Kong, and there was no evidence found in Taiwan, a civil law country.

However, the regression of rate of continuation on equal-weighted and value-weighted dividend premiums did not show significant results for all countries. In general, the model does not seem to explain why the corporate managers continue to pay dividends; further investigation is required. The suggestion is that if the catering dividend model does not explain why the managers continue to pay dividends, other theories such as signaling theory or the cash-free hypothesis of dividends could provide an explanation.

Also, the results for South Korea are not in line with the findings of Ferris, Jayaraman, and Sabherwal (2009). According to their findings, the managers of civil law country such as South Korea should not have incentives to cater to investors’ demand of dividends. So, the results for South Korea do not correspond to expectations because evidence of catering theory of dividends was found. Further investigation is required to account for the existence of catering incentives in South Korea despite its status as a civil law country. It is likely that the answer can be found in the particularities of Korean legislation. Namely, investors should enjoy a wider set of rights and protections by law, which makes it easier for them to discipline corporate managers who fail to satisfy shareholders. So, the managers of South Korea are supposed to be more responsive to investors and cater to their demand of dividends if they wish to remain employed.

(25)

25 References

Allen, F., Bernardo, A. E., & Welch, I. (2000). A theory of dividends based on tax clienteles. Journal of Finance, 55(6), 2499-2536.doi: 10.1111/0022-1082.00298

Ang, J. S. (1987). Do dividends matter? A review of corporate dividend theories and evidence. Monograph series In Finance and Economics. New York: Salomon Brothers Center for the Study of Financial Institutions and Graduate School of Business Administration, New York University.

Asquith, P., & Mullins, D. W. (1983). The impact of initiating dividend payments on shareholders’ wealth. Journal of Business, 56(1), 77-96.

Baker, M., & Wurgler, J. (2004a). A catering theory of dividends. The Journal of Finance, 3,1125-1166.doi: 10.1111/j.1540-6261.2004.00658.x

Baker, M., & Wurgler, J. (2004b). Appearing and disappearing dividends: The link to catering incentives. Journal of Financial Economics, 73(2),271-288.

doi:10.1016/j.jfineco.2003.08.001

Baker, H. K., & Kolb, R. W. (2009). The catering theory of dividends. In M. de Rooij, L.

Renneboog (Eds.), Dividends and dividend policy (pp.215-238). Hoboken, NJ: John Wiley & Sons, Inc.

Baker, H. K., & Powell, G. E. (1999). How corporate managers view dividends policy. Quarterly Journal of Business and Economics, 38(2), 17-35.

Baker, M., & Wurgler, J. (2002, November). Why are dividends disappearing? An empirical analysis (New York University Working paper No.S-CG-02-09). Available at SSRN: http://ssrn.com/abstract=1295257

Baker, M., Ruback, R. S., & Wurgler, J. (2005). Behavioral corporate finance: A survey (NBER Working Paper No. 10863). NBER. doi: 10.3386/w10863

Bhattacharya, S. (1979). Imperfect information, dividend policy, and "The Bird in the Hand" fallacy. The Bell Journal of Economics, 10(1), 259-270. doi: 10.2307/3003330

Black, F. (1956). The dividend puzzle. Journal of Portfolio Management, 2(2), 5-8. doi: 10.3905/jpm.1976.408558

Black, F., & Scholes, M. (1974). The effects of dividend yield and dividend policy on common stock prices and returns. Journal of Financial Economics, 1(1), 1-22. doi: 10.1016/0304-405X(74)90006-3

Brennan, M. (1970). Taxes, market valuation and corporate financial policy. National Tax Journal, 23(4), 407-416.

Council on Competiveness (2013). Global manufacturing competiveness index (Deloitte report). Retrieved from

https://www2.deloitte.com/content/dam/Deloitte/global/Documents/Manufacturing/gx_2013 %20Global%20Manufacturing%20Competitiveness%20Index_11_15_12.pdf

(26)

26 DeAngelo, H. & DeAngelo, L. (2006). The irrelevance of the MM dividend irrelevance theorem.

Journal of Financial Economics, 79(2), 293-315. doi: 10.1016/j.jfineco.2005.03.003

Denis, D. J., & Osobov, I. (2008). Why do firms pay dividends? International evidence on the determinants of dividend policy. Journal of Financial Economics, 89(1), 62-82. doi: 10.1016/j.jfineco.2007.06.006

Fama, E., F., & French, K., R. (2001). Disappearing dividends: Changing firm characteristics of lower propensity to pay? Journal of Financial Economics, 60(1), 3-43. doi: 10.1016/S0304-405X(01)00038-1

Farrar, D. E., & Selwyn, L. L. (1967, January 1). Taxes, corporate financial policy and return to investors. National Tax Journal, 20, 444-455.

Feldstein, M., & Green, J. (1983). Who do companies pay dividends? American Economics Review, 73(1), 7-30. doi: 10.3386/w0413

Ferris, S. P., Noronha, G., & Unlu, E. (2010). The more the merrier: An international analysis of the frequency of dividend payment. Journal of business finance and accounting, 37(1-2), 148-170. doi: 10.1111/j.1468-5957.2009.02174.x

Ferris, S. P., Jayaraman, N., & Sabherwal, S. (2009). Catering effects in corporate dividend policy: The international evidence. Journal of Banking and Finance, 33, 1730-1738. doi:10.1016/j.jbankfin.2009.04.005

Grullon, G., Michaely, R., Benartzi, S., & Thaler, R. H. (2005). Dividend changes do not signal changes in future profitability. Journal of Business, 78(5). 1659-1682. doi: 10.1086/431438 Hoberg, G., & Prabhala, N. R. (2006, March). Dividend policy, Risk, and Catering (Working

paper).University of Maryland. Retrieved from

https://business.rice.edu/uploadedFiles/Faculty_and_Research/Academic_Areas/Finance/Se minar_PDFs/pdf05_06_Prabhala_rice.pdf

Imad, R. Z. (2015). The Jordanian Catering Theory of Dividends. International Journal of Business and Management, 10(2). 226-230. doi: 10.5539/ijbm.v10n2p226

Jensen, M. C. (1986). Agency costs of free cash flows, corporate finance, and takeovers. American Economic Review, 76(2), 323-329. http://dx.doi.org/10.2139/ssrn.99580

John, K, & Williams, J. (1985). Dividends, dilution, and taxes: A signal equilibrium. Journal of finance, 40(44), 1053-1070. doi: 10.1111/j.1540-6261.1985.tb02363.x

Kale, J. R., Kini, O., & Payne, J. D. (2012). On dividend initiation decisions of newly public firms: some evidence on signaling with dividends. Journal of Financial and Quantitative analysis. 47(2), 365-396. doi:10.1017/S0022109012000063

Keller, G. (2009). Managerial Statistics. Cengage: South-Western.

Kuo, J., Philip, D., & Zhang, Q. (2013). What drives the disappearing dividends phenomenon? Journal of Banking and Finance, 37,3499-3514. doi: 10.1016/j.jbankfin.2013.05.003

Li, W., & Lie, E. (2006). Dividend changes and catering incentives. Journal of Financial Economics, 80, 293-308. doi: 10.1016/j.jfineco.2005.03.005

(27)

27 Lie, E. (2000). Excess funds and agency problems: An empirical study of incremental cash

disbursements. The Review of Financial Studies, 13(1), 219-248.doi: 10.1093/rfs/13.1.219 Lie, E. (2005). Operating performance following dividend decreases and omissions. Journal of

corporate finance, 12(1), 27-53.doi: 10.1016/j.jcorpfin.2004.04.004

Lu, X., Xi, Y., & Lu, D. (2014). An Empirical Study about Catering Theory of Dividends: The Proof from Chinese Stock Market. Journal of Industrial Engineering and Management, 7(2), 506-517. doi: 10.3926/jiem.1013

Miller, M. H., & Scholes, M. S. (1978). Dividends and taxes. Journal of Financial Economics, 6 (4), 333-364. doi:10.1016/0304-405X(78)90009-0

Modigliani, F., & Miller, M. H. (1961, October). Dividend policy, growth, and the valuation of shares. The Journal of Business, 34(4), 411- 433.

Neves, M. E. D. (2009). The catering theory of dividends: the moderating role of the firm characteristics, corporate governance factors and corporate ownership (Doctoral dissertation, University of Salamanca). Retrieved from

http://gredos.usal.es/jspui/bitstream/10366/76429/1/DAEE_Duarte_Neves_ME_The_catering _theory.pdf

Peterson, P. P., Peterson, D. R., & Ang, J. S. (1985). Direct evidence on the marginal rate of taxation on dividend income. Journal of Financial Economics,14(2), 267-282.

doi: 10.1016/0304-405X(85)90018-2

Qatar Financial Center (2015, March). The global financial centers index 17. Retrieved from http://www.longfinance.net/images/GFCI17_23March2015.pdf

Sarel, M. (1999, September). Growth in East Asia: What we can and what we cannot infer (Economic Issues No.1). International Monetary Fund, Washington, DC. Retrieve from http://www.imf.org/external/pubs/ft/issues1/issue1.pdf

Shefrin, H. M., & Statman, M. (1984). Explaining investor preference for cash dividends. Journal of Financial Economics, 13(2), 253-282.doi: 10.1016/0304-405X(84)90025-4

Tsuji, C. (2011). A test of dividend policy: The case of Japanese machinery industry firms. International Journal of Economics and Finance, 3, 1-13. doi:10.5539/ijef.v3n3p119

(28)

28 Appendixes

Table 1.

Dividend premium for Singapore

Payers Nonpayers Dividend premium

Year EW M/B VW M/B EW M/B VW M/B EW M/B VW M/B 1989 1.57 1.87 1.62 1.89 -0.03 -0.01 1990 1.15 1.38 1.37 1.77 -0.18 -0.25 1991 1.49 1.85 1.31 1.54 0.13 0.18 1992 1.28 1.54 1.55 1.77 -0.19 -0.14 1993 1.93 2.83 1.93 2.82 0.00 0.00 1994 1.59 2.16 1.37 1.44 0.15 0.41 1995 1.57 2.19 1.58 1.85 -0.01 0.17 1996 1.52 2.09 1.23 1.43 0.21 0.38 1997 3.30 15.58 2.26 9.39 0.38 0.51 1998 1.01 1.34 1.06 1.40 -0.05 -0.04 1999 1.36 1.87 1.77 2.81 -0.26 -0.41 2000 1.06 1.37 1.06 1.33 0.00 0.03 2001 1.08 1.38 1.13 1.46 -0.05 -0.06 2002 1.06 1.36 1.15 2.13 -0.08 -0.44 2003 1.38 1.99 1.72 3.43 -0.22 -0.54 2004 1.28 1.71 1.44 3.25 -0.12 -0.64 2005 1.34 2.51 1.45 6.35 -0.08 -0.93 2006 1.50 3.64 2.08 16.48 -0.33 -1.51 2007 1.54 2.39 2.18 8.58 -0.35 -1.28 2008 0.95 1.24 1.58 15.78 -0.50 -2.54 2009 1.15 1.59 1.63 7.68 -0.35 -1.57 2010 1.19 1.68 1.69 5.79 -0.35 -1.24 2011 1.00 1.33 1.45 13.94 -0.38 -2.35 2012 1.11 1.46 2.00 16.14 -0.59 -2.40 2013 1.57 1.87 1.62 1.89 -0.03 -0.01 2014 1.15 1.38 1.37 1.77 -0.18 -0.25

Note. A firm is defined as a dividend payer at time t if it has positive dividends per share by the ex date. The market-to-book ratio

is the ratio of the market value of the firm to its book value. The market-to-book ratio reported is an equal-weighted (EW) or value-weighted (VW) average, by book value across dividend payers and non-payers. The dividend premium is the difference between the logs of the dividend payers’ and non-payers’ average market-to-book ratios.

(29)

29 Table 2.

Measures of Dividend Payment for Singapore

Payers Nonpayers Payment Rates, %

Year Total New Old List Total New Old List Initiate Continue

1990 41 16 24 1 13 1 8 4 0.67 0.96 1991 38 1 33 4 18 5 11 2 0.08 0.87 1992 48 7 38 3 11 0 9 2 0.44 1.00 1993 66 5 47 14 27 1 5 21 0.50 0.98 1994 120 22 64 34 14 1 5 8 0.81 0.98 1995 111 5 97 9 30 12 9 9 0.36 0.89 1996 81 9 62 10 61 36 18 7 0.33 0.63 1997 146 31 64 51 70 14 25 31 0.55 0.82 1998 159 37 118 4 51 17 26 8 0.60 0.88 1999 168 18 127 23 90 19 33 38 0.35 0.87 2000 138 27 104 7 75 28 39 8 0.41 0.79 2001 161 26 97 38 146 16 33 97 0.44 0.86 2002 174 34 118 22 128 20 75 33 0.31 0.86 2003 218 29 145 44 166 18 88 60 0.25 0.89 2004 232 33 181 18 171 23 104 44 0.24 0.89 2005 235 35 184 16 164 29 100 35 0.26 0.86 2006 259 48 198 13 179 22 100 57 0.32 0.90 2007 284 52 216 16 134 14 99 21 0.34 0.94 2008 223 14 194 15 114 26 72 16 0.16 0.88 2009 262 22 185 55 207 30 88 89 0.20 0.86 2010 307 65 235 7 153 13 127 13 0.34 0.95 2011 320 33 280 7 138 14 112 12 0.23 0.95 2012 322 29 283 10 134 26 102 6 0.22 0.92 2013 306 13 284 9 147 24 116 7 0.10 0.93 2014 311 22 286 3 144 14 122 8 0.15 0.95

Note. A firm is defined as new dividend payer at time t if it has positive dividends per share by the ex-date at time t and zero

dividends per share by the ex-date at time t-1. A firm is defined as an old payer at time t if it has positive dividends per share by the ex-date at time t and positive dividends per share at time t-1. A firm is defined as a new list payer if it has positive dividends per share by the ex-date at time t and is not in the sample at time t-1. The opposite is true for non-payers. Initiate expresses payers as a percentage of surviving non-payers from t-1. Continue expresses payers as a percentage of surviving payers from t-1.

(30)

30 Table 3.

Dividend premium for Hong Kong

Payers Nonpayers Dividend premium

Year EW M/B VW M/B EW M/B VW M/B EW M/B VW M/B 1992 1.53 2.25 0.99 1.05 0.43 0.76 1993 1.89 3.11 1.56 1.60 0.19 0.67 1994 1.33 2.16 0.71 0.81 0.62 0.98 1995 1.17 1.78 0.95 1.13 0.20 0.45 1996 1.26 1.84 1.20 1.34 0.05 0.32 1997 1.00 1.37 1.03 1.23 -0.03 0.11 1998 0.94 1.22 0.96 1.12 -0.02 0.09 1999 1.04 1.45 0.95 1.77 0.08 -0.20 2000 1.01 1.49 1.15 3.76 -0.13 -0.93 2001 1.02 1.41 1.07 1.72 -0.05 -0.20 2002 0.99 1.29 1.07 3.05 -0.08 -0.86 2003 1.15 1.62 1.00 1.62 0.14 0.00 2004 1.26 1.88 1.29 2.08 -0.02 -0.10 2005 1.10 1.42 1.35 2.86 -0.20 -0.70 2006 1.16 1.51 1.09 1.61 0.06 -0.06 2007 1.37 1.86 1.83 3.59 -0.29 -0.66 2008 0.82 0.99 0.77 1.10 0.07 -0.11 2009 0.41 0.55 0.39 0.80 0.03 -0.37 2010 1.20 2.17 1.31 2.33 -0.09 -0.07 2011 0.90 1.17 1.15 3.35 -0.25 -1.05 2012 0.92 1.23 0.98 1.61 -0.05 -0.27 2013 0.94 1.40 1.40 3.53 -0.40 -0.93 2014 1.02 1.57 1.47 3.70 -0.37 -0.86

Note. A firm is defined as a dividend payer at time t if it has positive dividends per share by the ex date. The market-to-book ratio

is the ratio of the market value of the firm to its book value. The market-to-book ratio reported is an equal-weighted (EW) or value-weighted (VW) average, by book value across dividend payers and non-payers. The dividend premium is the difference between the logs of the dividend payers’ and non-payers’ average market-to-book ratios.

(31)

31 Table 4.

Measures of Dividend Payment for Hong Kong

Payers Nonpayers Payment Rates, %

Year Total New Old List Total New Old List Initiate Continue

1993 29 3 24 2 3 1 1 1 0.75 0.96 1994 37 3 27 7 6 2 0 4 1.00 0.93 1995 42 5 35 2 4 1 1 2 0.83 0.97 1996 43 2 37 4 8 5 2 1 0.50 0.88 1997 51 4 32 15 17 0 2 15 0.67 1.00 1998 56 7 41 8 16 4 7 5 0.50 0.91 1999 49 3 36 10 41 18 10 13 0.23 0.67 2000 54 4 45 5 44 1 34 9 0.11 0.98 2001 52 5 43 4 34 2 28 4 0.15 0.96 2002 55 3 49 3 31 0 24 7 0.11 1.00 2003 91 3 52 36 61 1 22 38 0.12 0.98 2004 73 11 57 5 27 0 23 4 0.32 1.00 2005 69 4 63 2 34 4 19 11 0.17 0.94 2006 113 4 62 47 49 5 24 20 0.14 0.93 2007 82 6 71 5 37 1 24 12 0.20 0.99 2008 86 8 77 1 31 1 22 8 0.27 0.99 2009 85 6 71 8 50 15 25 10 0.19 0.83 2010 88 11 76 1 42 5 37 0 0.23 0.94 2011 91 5 83 3 40 3 36 1 0.12 0.97 2012 123 4 83 36 63 8 35 20 0.10 0.91 2013 126 8 118 0 64 4 55 5 0.13 0.97 2014 94 8 85 1 44 4 36 4 0.18 0.96

Note. A firm is defined as new dividend payer at time t if it has positive dividends per share by the ex-date at time t and zero

dividends per share by the ex-date at time t-1. A firm is defined as an old payer at time t if it has positive dividends per share by the ex-date at time t and positive dividends per share at time t-1. A firm is defined as a new list payer if it has positive dividends per share by the ex-date at time t and is not in the sample at time t-1. The opposite is true for non-payers. Initiate expresses payers as a percentage of surviving non-payers from t-1. Continue expresses payers as a percentage of surviving payers from t-1.

(32)

32 Table 5.

Dividend premium for South Korea

Payers Nonpayers Dividend premium

Year EW M/B VW M/B EW M/B VW M/B EW M/B VW M/B 1993 1.00  1.02 0.97 1.00 0.03  0.02 1994 1.14  1.21 1.08 1.10 0.05  0.09 1995 0.99  1.00 1.02 1.04 ‐0.04  ‐0.04 1996 0.97  0.99 0.97 0.99 0.00  0.00 1997 0.89  0.90 0.90 0.92 ‐0.01  ‐0.02 1999 0.94  1.50 0.91 1.02 0.03  0.39 2000 0.75  0.81 0.87 0.92 ‐0.14  ‐0.13 2001 0.85  0.97 0.99 1.18 ‐0.14  ‐0.19 2002 0.80  0.89 0.92 1.08 ‐0.14  ‐0.19 2003 0.89  1.08 1.04 1.44 ‐0.16  ‐0.29 2004 0.87  1.01 0.95 1.23 ‐0.09  ‐0.20 2005 1.08  1.33 1.61 5.84 ‐0.40  ‐1.48 2006 1.05  1.29 1.49 3.14 ‐0.35  ‐0.89 2007 1.11  1.34 1.37 2.23 ‐0.21  ‐0.51 2008 0.84  0.97 0.90 1.08 ‐0.07  ‐0.10 2009 1.02  1.30 1.16 1.56 ‐0.12  ‐0.18 2010 1.05  1.35 1.19 1.75 ‐0.13  ‐0.26 2011 1.03  1.47 1.31 2.93 ‐0.24  ‐0.69 2012 1.08  1.40 1.34 2.58 ‐0.22  ‐0.61 2013 1.10  1.61 1.27 1.79 ‐0.14  ‐0.11 2014 1.20  1.86 1.38 2.04 ‐0.14  ‐0.09

Note. A firm is defined as a dividend payer at time t if it has positive dividends per share by the ex date. The market-to-book ratio

is the ratio of the market value of the firm to its book value. The market-to-book ratio reported is an equal-weighted (EW) or value-weighted (VW) average, by book value across dividend payers and non-payers. The dividend premium is the difference between the logs of the dividend payers’ and non-payers’ average market-to-book ratios.

Referenties

GERELATEERDE DOCUMENTEN

With the results of the positive or negative influence of firm size, leverage, bank debt, cash flow, cash flow volatility, liquid assets, investment opportunity,

However, there is also some contradicting evidence that the mandatory quota, and so the number of women on the board, does not have a positive, but a negative effect on

The research objective of this thesis is to find out if the firm characteristics Firm Size, Leverage, Bank debt, Cash flow, Cash flow volatility, investment opportunity and

To give an answer to the research question: “To what extent does the capital structure of listed firms in Germany influence the performance during the

This thesis investigates the influence diversification on firm value of publicly listed firms in Germany and the Netherlands. This influence is investigated by examining the

Finally, no evidence is found in favor of the hypothesis that dividend and R&D expenditure have a negative interaction effect on stock performance, despite

23 Different as the outcome of firm size and interest coverage ratio evaluation, we can see from Table 6 that the average cash holding in the proportion of market value of equity

In this section we present the data and methodology used to answer the research question and hypotheses. First the conceptual model is presented and the variables are