An Investigation into the Dividend Policy of Firms in East Asia Julia Sawicki

Abstract This study investigates the determinants of dividend policy at the firm, industry and country level in seven countries in East Asia: Taiwan, South Korea, Thailand, Malaysia, Hong Kong, Singapore, Indonesia and Japan. The findings indicate that past and expected revenue growth, cash flow adequacy, leverage collateralizable assets and (unexpectedly) systematic risk are positively related to payout ratio. Inter-industry and -country differences are related to legislation, tax and ownership structure, and socio-cultural and political influences.

Keywords: dividend policy, emerging markets JEL classification: G32

Julia Sawicki Nanyang Technological University School of Business Nanyang Avenue Singapore 639798 (65)790.4669 [email protected]

The author would like to acknowledge gratefully research assistance of :Ho Jie Ying, Lee Seet Ling and Teo Harn Kiat.

Introduction For nearly four decades, researchers have been grappling with the dividend puzzle, trying to understand the determinants of dividend policy. Most studies focus on US firms. We expand the investigation by studying dividend policy from an Asian perspective focusing on firms in 8 representative countries of East Asia namely Taiwan, South Korea, Thailand, Malaysia, Hong Kong, Singapore, Indonesia and Japan. We measure the relationship between dividend payout and seven variables: expected growth, historical revenue growth, firm size, leverage, cash flow adequacy, collateralizable assets and systematic risk. We also investigate industry and country differences. The results regarding firm-level differences are consistent with prior work indicating a positive relationship between dividend payout and expected growth, historical revenue growth, cash flow adequacy, and collateralizable assets. Interestingly systematic risk is also positively related to payout ratio, which contradicts the results of prior work. The results for firm size are mixed with the direction of the relationship to dividend payout differing for different countries and industries. Finally, we find inter-industry and inter-country differences. The former can be attributed to the nature of the industries as well as the legislation unique to certain industries, while intercountry differences in dividend payout are due to tax structure, ownership structure and sociocultural and political influences.

Background Miller and Modiglianis (1961) seminal theoretical paper demonstrates the irrelevance of dividend policy and demonstrates in perfect market conditions that dividend payout does not affect firm value. In a frictionless world with no agency costs, information asymmetry, taxes and transaction costs, investors are indifferent between capital gains and dividends.

However such assumptions are unrealistic in the real world. Among the many studies on dividend policy, most investigate the underlying dynamics of dividend policy, investigating questions like: What are the various factors in the real world that affect dividend policy? Do firm characteristics affect dividend policy? How and why do these determinants affect dividend policy? The examination into the effect of dividend policy on firm value examines: the role dividend policy plays in a firm; the informational value of dividend change; investors reaction to dividend changes. A discussion of studies focusing on the determinants of dividend policy follows. Firm-level Effects on Dividend Policy Rozeff (1982) proposes an optimal dividend payout model, which appeals to two market imperfections: agency cost and the transaction cost associated with external financing. He argues that due to agency costs, dividends are increased but on the other hand this raises the costs of external financing. The sum of these two opposing costs determines an optimal payout ratio. The firms beta, past and expected future growth rate of sales as proxy for the transactions associated with external financing He argues that beta is a surrogate for the firms operating and financial leverage, and firms with a high leverage will lower the dividend payout to lower the cost of external financing. Dividend payments are quasi-fixed charges, which are substitutes for other fixed charges. For the other two proxies, he concludes that firms experiencing or anticipating higher revenue growth will lower dividend payout ratios. Firms, in this case would tend to retain funds to avoid external financing. Lastly, he uses the percentage of common stock held by insiders and the number of common stockholders as proxy for agency costs. Firms pay out more dividends when a lower fraction of the equity and or a greater number of stockholders own the outside equity. Jensen (1992) examines the determinants of three policy choices within a system of equations. The three policy choices are insider ownership, debt and dividend policies. In the

dividend equation, which examines the determinants of dividend payout ratios, he finds that investment, insider ownership, debt ratio, growth, and business risk are negatively related to dividends while only profitability is related positively related to dividends. In Mohd (1995), firm size and industry representation function as control variables. Firm size is employed as a control variable for both the transaction cost and agency cost proxies. Industry representation was also used as a control variable as it is an important factor in the payout decision. It was found that dividend payout is positively related to firm size, the amount of institutional holdings, and number of shareholders. It is negatively related to past and future growth, operating and financial leverage risk, intrinsic business risk, and insider shareholdings. Chirinko (1998) exploits the unique initial homogeneity of seven regional phone companies which are created from AT&Ts local operating. This is a result of an anti-thrust libel against AT&T. As the firms originate from the same corporation, there are reasonable grounds for compensation of subsequent heterogeneity in dividend policy of the 7 firms. It is found that investment opportunities and dividend payout are negatively related. Also, increased indebtedness leads to increased contacts with external financial sources, which results in closer monitoring and an increased dividend payout. Fama and French (2000) find that larger and more profitable firms are likely to pay more dividends. This is due to their ability to sustain the high payout. Mollah et. al. (2000) report that the number of common stockholders, the level of collateralizable assets, and free cash flow is positively related to dividend payout ratio. Insider ownership on the other hand is positively related to dividend payout ratio. Inter-industry Effects on Dividend Policy Lintner (1956) argues that there is be positive correlation between dividend policies of firms in the same industry and certain factors within the industry. More than a decade later,

Harkins (1971) also posits a relationship among dividend of firms in the same industry due to their comparable investment opportunities. Michel (1979) uses data from 13 industries in America over the period 1967-1976, to determine if a systematic relationship exists within an industry. The null hypothesis, that dividend yield are generated from the same population or identical populations, was rejected on the basis of significant differences in dividend payout ratios across industries. Michel (1986) did an inter-industry analysis of US and Japan to find out if a systematic relationship exists between a firms dividend policy and the industry it is in. The null hypothesis that across-industry dividend yields are generated from the same population is rejected for both the USA and Japanese samples. The results conform to those reported by Michel (1979). This present study however, suggests that industry-influence phenomenon also exists in Japan. Baker (1988) updates Michel (1979) study by using data from 1977 to 1981, which he too finds support for industry effects on dividend payout. Collin et. Al. (1996) and Saxena (1999) recognize the differences in dividend policy between regulated firms and unregulated firms. Collins focuses on agency-cost and monitoring explanations for the relevance of dividend and examines the role of insiders in determining dividend policy for unregulated firms, utilities and financial firms. Saxena, on the other hand, investigates if determinants of dividend policy differ between regulated and unregulated firms. He concluded that some of the determinants of dividend policy are different for regulated and unregulated firms. Specifically, the percentage of common stock held by insiders and expected future growth rate do not play a key role in a regulated firms payout ratio as compared to an unregulated firm. Inter-country Effects on Dividend Policy Michel (1986) investigates whether a systematic relationship exists between a firms dividend policy and the country it operates. The results indicate a difference in the payout ratios

of US and Japan. Reasons cited include bank ownership of companies, growth rate of the economy as well as accounting, tax and risk characteristics. La Porta (1998) finds that firms in common law countries, where investment protection is higher, tend to make higher dividend payout than those in civil law countries as investors in protected countries use their legal rights to demand higher dividends. Also, Buchanans (2000) study of the G7 countries reports that there is a relationship between the dividend policy and the country in which the firm operates. One reason cited is that the corporations in one country may be leveraged differently from their counterparts in another country. Kang (2001), it is argues that institutional structure may be a reason for differences in dividend payout in different countries, noted that the corporate governance systems in UK and the US are characterized as a market based system while the governance systems in both Australia and France are governed by relationship-based system. UK and US firms generally have lower insider ownership and agency problems may be more widespread, thus the firms have relatively higher dividend payout ratios.

HYPOTHESIS: Relationship between firm characteristics and payout ratio. Expected Growth: negative Firms establish lower dividend payout ratio when they anticipate higher growth, presumably because growth entails higher investment expenditures. Due to the high cost of external financing, firms will retain a higher proportion of earnings to finance future investment needs, hence reducing their dividend payout in anticipation of future growth. Hence, a negative relationship between dividend payout and expected growth is expected. Historical Growth: negative In order to support rapid past growth, firms would have retained more funds for investment to generate the growth as opposed to raising the more costly external funds. Hence, this implied that

firms would have adopted a low payout ratio in periods when past growth was high. Hence, it is hypothesized that dividend payout is negatively related to historical growth. Firm Size: positive As the size of a firm increases, shareholders are not able to monitor the firm effectively and there is a higher tendency of agency problems. Thus, shareholders will demand higher dividend payout, which acts as an indirect monitoring tool. Firms in current or potential need of external financing will use their funds more prudently as they will be monitored by both existing and potential creditors. Leverage: positive or negative On one hand, firms trade off dividend payments with fixed financial charges. A highly leveraged firm would tend to lower its dividend payout ratio because of high fixed financial commitments. On the other, increased indebtedness leads to increased contacts with external financial sources, which results in closer monitoring and an increased dividend payout. Hence, we do not expect any particular direction of relationship between leverage and dividend payout. Cash Flow Adequacy: positive A firm will not be able to distribute cash dividends without the means to do so. Also, a firm with high cash flow may tend to have higher agency problems if the cash are misused in improper ways. A higher dividend payout will reduce the free cash flow, and in turn reducing the agency problems. Collateralizable Assets: positive Shareholders may expropriate wealth from bondholders by paying themselves dividends. Bondholders try to contain this problem through restrictions on dividend payments in the bond indenture. However, fewer restrictions are placed on the firm if debt can be collateralized as the borrower is restricted to use the funds for a specific project. Hence, a positive relationship between collateralizable assets and dividend payout is expected.

Systematic Risk: negative Beta is a proxy for systematic risk. Risky firms tend to have higher volatility in their cash flows. In this case, the firms reliance on external financing would be increased. A firm with high beta would tend to adopt a lower dividend payout to avoid costly external financing. Hence, a negative relationship between systematic risk and dividend payout is expected. Industry-specific Differences in Dividend Payout Dividend policies and certain factors like sales volume, internal funds flow may be positively correlated for firms in the same industry, indicating some relationship between industry and dividend payout. Due to the structural characteristics of an industry, it is likely firms within the same industry faced comparable investment opportunities while firms in different industries may have different investment opportunities. The extent of regulation may also have an impact on dividend payout as regulated firms face a different set of factors from unregulated firms in the determination of dividend payout. Due to the difference in structural characteristics and investment opportunities, we expect that dividend payout among industries would exhibit differences. Hence, we propose the hypothesis that there are industry-specific differences in dividend payout. Country-specific Differences in Dividend Payout Although there are limited previous studies on inter-country differences, we expect some differences in the typical dividend payout in different countries due to the fundamental differences in these countries. Examples of fundamental differences include legal frameworks, culture and political environment.

DATA Data Source All the financial data used in our study are obtained from the PACAP database.

Sample Coverage Our study covers a period of 10 years from 1989-1998. We have included firms from 8 countries including:
Hong Kong Indonesia Japan Malaysia Singapore South Korea Taiwan Thailand

Furthermore, we have grouped these firms into 32 industries including:

Agriculture, Forestry & Fishery Apparel & Textiles Automobile Building Materials Chemicals, Petroleum & Plastics Computer & Computer Services Conglomerate Construction Electronics & Electrical Appliance Entertainment & Recreation Finance Food & Beverage Footwear & Leather Products Healthcare Services Hotels & Travel Services Household Goods Machinery & Equipment Metal Products Mineral Products Mining Miscellaneous (Consumer) Miscellaneous (Industrial) Miscellaneous (Services) Paper Products Pharmaceuticals Printing & Publishing Property Retail Telecommunications Transportation & Logistics Utilities Wood Products

Sample Selection Criteria Our sample only consists of dividend-paying publicly listed firms. A sample firm must have a typical dividend payout ratio ranging from 0 to 1. Data for calculating all the variables must be available for that firm.

METHODOLOGY Objective 1 Statistical Test This particular objective will be tested using multiple regressions on three different sets of samples namely: (1) pooled [No. of sample = 1],

For the pooled regression, we have excluded firms in the finance industry due to fundamental differences of these firms. One of the independent variables for the regression is free cash flow to assets ratio. However, for finance firms, the balance sheet and income statement of such firms are structured in a different manner and there are certain comparability issues which we found hard to overcome. (2) country specific [No. of samples = 8] For the country specific regressions, we have also similarly excluded finance firms in the regressions (3) industry specific [No. of samples = 32] Regression divpay = 0 + 1 (tobinq) + 2 (histgrwt) + 3 (mktcap) + 4 (leverage) + 5 (fcf) + 6 (colasset) + 7 (beta) + e Dependent Variable: Dividend payout ratio (divpay) In order to prevent the problem of multi-collinearity between the independent variables, we run a correlation test between the independent variables. No instance of high correlation is found between the independent variables. [Refer to Correlation Matrix in Appendix A: Objective 1] Independent Variables & Hypothesis No Expected Direction Independent Variable Leverage Abbreviation leverage

H0 : i = 0 H1 : i ? 0 Expected Direction : Positive Independent Variable Market capitalization Free cash flow to assets ratio Abbreviation mktcap fcf

Fixed assets to assets ratio

colasset

H0 : i = 0 H1 : i > 0 Expected Direction : Negative Independent Variable Tobins Q Historical revenue growth Beta Abbreviation tobinq histgrwt beta

H0 : i = 0 H1 : i < 0 Control Variables We have also introduced certain qualitative variables in our regression model to control for some factors which we would like to isolate from our research. CONTROL VARIABLES Pooled Regression Year Age Industry Country

Country-Specific Regression Year Age Industry Specific Regression Year Age Variables that are qualitative in nature are quantified into n-1 binary variables where n refers to the number of categories existing for the qualitative variable. - Year refers to the financial year of that particular observation and there are 10 categories consistent to the period covered by our study. Country Industry

- Age refers to the age of the company and we have grouped this variable into 3 categories, namely: old (age>= 24), medium (5<age<24) and young (age<=5). - Industry refers to the industry the firm is categorized under according to the 32 defined industry categories. - Country refers to the country the firm is in and there are 8 different countries in our study. Assumptions (1) For each observation, the error term e is normally distributed with mean zero and standard deviation s and is independent of the error terms associated with all other observations. That is, ? ~ N(0, s 2) for all j = 1, 2, .., n independent of other errors. (2) The variables Xj are considered fixed quantities. Xj are independent of the error term e and the only randomness comes from the error term e. Level of Significance The level of significance used in our regression tests is 5% Objectives 2 & 3 Statistical Tests Objective 2 & 3 will be carried out using similar statistical tests applied to different categories. 3 sets of statistical tests are carried out namely (1) one-way ANOVA, (2) Kruskal Wallis, (3) posthoc Turkey Pairwise-Comparisons test. ANOVA Firstly, we test for any differences in the means of the dividend payout ratios by using the parametric ANOVA test. Hypothesis Ho: The means of dividend payout ratios of the different countries are equal.

H1: Not all the means of dividend payout ratios of the different countries are equal. Assumptions (1) We assume independent random sampling from each of the r populations. (2) We assume r populations under study are normally distributed, with means i that may or may not be equal, but with equal variances 2. Populations For objective 2, there are 32 populations, which adhere to the number of industries in our study. For objective 3, there are 8 populations, which adhere to the number of countries covered in our study. The financial firms are excluded from the populations due to incomparability issues. Level of Significance The level of significance used in the ANOVA tests is 5% Kruskal Wallis We also use the non-parametric alternative of ANOVA- Kruskal Wallis test to find out if the populations have the same distribution. Hypothesis Ho: All k populations have the same distribution H1: Not all k poupulations have the same distribution Assumptions (1) The k samples are random and are independently drawn from the respective populations. (2) The random variables under the study are continuous and the measurement scale used is at least ordinal. Populations As mentioned earlier, there are 32 populations for objective 2 and 8 populations for objective 3. Level of Significance The level of significance used in the Kruskal Wallis tests is 5%

Turkey Pairwise Comparisons If Ho of the ANOVA test is being rejected, we can only conclude that not all the means are the same but cannot conclude exactly which means are different. Thus, we have to carry out further analysis to find out the exact pairwise differences. Hypothesis H0 : i - j = 0 H1 : i - j 0 Level of Significance The level of significance used in the Turkey tests is 5% Limitations There are 465 unique pairwise comparisons for objective 2 and 28 unique pairwise comparisons for objective 3. Thus, we might not be able to comment on each and every significant pairwise test. We will seek to summarize the results and provide certain general reasons to explain the differences as well as homogeneity between certain populations. RESULTS & ANALYSIS Objective 1: To investigate the determinants of dividend policy. Summary of Results TABLE 1: REGRESSION SUMMARY
VARIABLES

q
Expected Direction Full Regression By Industry Finance Property Construction Utilities Transportation & Logistics Telecommunications Agriculture, Forest& Fishery 8 Mining 1 2 3 4 5 6 7 -

histgr
-

mktcap
+

lev
? +

fcf
+

asset
+ +

beta
+

Model Sig?

YES

+ + + + + + +

+ + +

YES YES YES YES YES YES YES YES

tobinq histgrwt Expected Direction

9 10 11 12 13 14 15 16 17 18 19 20 21 22 Computer & Comp Services Electronics & E Appliance Food & Beverage Automobile Household Goods Apparel & Textiles Footwear & Leather Wood Products Miscellaneous (Consumer) Chem, Petroleum & Plastics Machinery & Equipment Paper Products Metal Products Mineral Products

mktcap

VARIABLES leverag fcf colasset e

beta

Model Sig?

?
+ + + + -

+
+

+
+ +

+ + + + YES YES YES YES YES YES YES YES YES YES YES YES YES YES

+ -

+ + + + + + + + + + + + + +

23 24 25 26 27 28 29 30 31 32

Building Materials Misc (Industrial) Entertainment & Recre Healthcare Services Pharmaceuticals Hotel & Travel Serv Printing & Publishing Retail Misc (Services) Conglomerate By Country Taiwan Korea Thailand Malaysia Hong Kong Singapore Indonesia Japan

+ + + + + + + + + + + + + +

YES YES YES YES YES NO YES YES YES YES

1 2 3 4 5 6 7 8

+ + + + +

+ + + + + + +

+ +

+ -

YES YES YES YES YES YES YES YES

Scale
Cell Entry +/- Significant at 5% (1-tail except for Leverage) + Positive Standardized Coefficient - Negative Standardized Coefficient Insignificant +/- Different from Expectation

All the regression models are significant except for Hotel & Travel Services. Table 1 provides a summary of the direction of relationship between dividend payout and the various independent variables. Only significant results are shown in the table. In the following sections, we will elaborate on the results of each of these variables.

Expected Growth TABLE 2: REGRESSION COEFFICIENTS OF TOBINS Q Regression Standardized Coefficient p-value
Expected Direction (-) Property Agriculture, Forestry & Fishery Computer & Computer Services Automobile Electronics & Electrical Appliances Apparel & Textiles Retail Electronics & Electrical Appliances Machinery & Equipment Metal Products Paper Products Miscellaneous (Consumer) Miscellaneous (Industrial) Taiwan Japan Opposite Direction (+) Healthcare Services Conglomerate Indonesia -0.054 -0.189 -0.279 -0.075 -0.201 -0.120 -0.213 -0.201 -0.130 -0.081 -1.190 -0.260 -.0200 -0.173 -0.076 0.351 0.106 0.224 0.099 0.071 0.007 0.028 0.000 0.002 0.000 0.000 0.000 0.091 0.009 0.008 0.014 0.050 0.000 0.050 0.009 0.001

Expected Direction Table 2 illustrates the regression coefficients for Tobins Q which is the proxy for expected growth in our study. The results obtained are analogous to our hypothesis that dividend payout is negatively related to future growth due to investment needs for sustaining growth and the costly external funding [Rozeff, 1982].

Property firms that foresee growth will need to devote substantial amount of financial resources for future property developments in order to meet future increase in demand for property. Property development entails huge expenses in purchase of land as well as the actual development of property. Hence, anticipation of future growth necessitates a low payout ratio for property firms. Firms in the Agriculture, Forestry & Fishery industry constituent the primary industries that usually require huge capital outlays for land and equipment and other related expenses. Thus, such firms will retain more earnings to finance their expenses incurred to support future growth. For Computer & Computer Services, Automobile, Apparel & Textile and Retail firms, future growth corresponds to increased investments in inventory. It is vital that such firms maintain an adequate level of inventory to meet demands and prevent loss of business opportunities. Firms in the Computer industry, in particular, will require increased investment in computer peripherals as well as specialized personnel to provide computer services, both entailing large expenses. Hence, such firms will adopt a low payout ratio to ensure that they have adequate liquidity for their financial needs. For Electronics & Electrical Appliances, Machinery & Equipment and Metal Product firms, a high growth rate is typically matched by high investment in long-term assets. In addition, firms producing metal products are vulnerable to the price of metal, which is their core raw material. Hence, in anticipation of future growth, such firms are likely to retain more earnings to finance their long-term investments resulting in low payout ratio.

For firms producing Paper Products, future growth may entail further investments in long-term assets as well as increase in manpower requirements. In view of the increase in financial needs, firms in this industry will adopt low payout ratio meet such needs. Although there may exist fundamental differences in firms categorized under Miscellaneous (Consumer) and Miscellaneous (Industrial), such firms would typically need extra financing to invest in long-term assets or to support anticipated increase in variable costs when they foresee growth. Hence, it may be reasonable to expect such firms have low payout ratios in anticipation of future growth. Capital gains are tax-free in Taiwan while dividends distributed from earnings are tax as ordinary income and 25% withholding tax for local investors and foreign investors respectively (Lin C.H. & Shui C.Y., 2003). Due to the above tax structure, we can infer that most investors will prefer capital gains to dividend. Hence, Taiwanese companies will have more incentive to retain more earnings to finance future growth. It is commonly believed that the extensive holdings of industrial corporations by financial institutions constituent a strong preference towards long-term capital appreciation and less towards current (dividend) income [Michel and Shaked, 1986]. In view of this preference, Japanese firms will tend to retain much of their earnings for investment rather than distribute them as dividend. Opposite Direction The seemingly perplexing direction of relationship between dividend payout and future revenue growth for Healthcare Services may be due to the small number of firms included in sample. There are only 15 such firms out of the total of 3905 firms. Hence, the results obtained may not be representative of the industry.

For Conglomerate firms, there may be certain inconsistencies in the results due to fundamental differences in the firms categorized under this grouping. Hence, such inconsistencies may be the cause of the deviation of categorys results from the expected results. Indonesian firms rank banks to be the most important source of funds, followed closed by retained earnings [Ang, Fatemi and Tourani-Rad, 1997]. This result may shed some light on the puzzling relationship between future revenue growth and dividend payout. Indonesia firms perceive that interest rates on loans are reasonable if not relatively low and they believe that banks are committed to provide resources to the firm in the event of insolvency. In addition, these firms may not want to place too much reliance on retained earnings due to the uncertainty in future earnings. The extent of institutional ownership of shares may be another possible explanation for the positive relationship between revenue growth and dividend payout. However, this reason is not conclusion due to the lack of ownership data of sample firms. Historical Revenue Growth TABLE 3: REGRESSION COEFFICIENTS OF 3-YR REVENUE GROWTH Regression Standardized Coefficient p-value
Expected Direction (-) Finance Property Construction Transportation & Logistics Electronics & Electrical Appliance Automobile Machinery & Equipment Paper Products Printing & Publishing Telecommunications Apparel & Textiles Chemicals, Petroleum & Plastics Metal Products Mineral Products -0.126 -0.091 -0.202 -0.102 -0.196 -0.270 -0.286 -0.138 -0.198 -0.581 -0.083 -0.118 -0.201 -0.138 0.000 0.006 0.000 0.003 0.000 0.000 0.000 0.048 0.061 0.002 0.008 0.000 0.000 0.053

Healthcare Services Hotels & Travel Services Retail Miscellaneous (Industrial) Conglomerate Korea Thailand Malaysia Hong Kong Singapore Japan Taiwan Opposite Direction (+) Household Goods

-0.609 -0.174 -0.097 -0.190 -0.174 -0.156 -0.081 -0.059 -0.083 -0.149 -0.276 -0.184 -0.030

0.001 0.009 0.004 0.031 0.000 0.000 0.020 0.046 0.002 0.000 0.000 0.000 0.012

Expected Direction As mentioned, firms establish lower payout when they experienced higher revenue growth as this entails higher investment expenditures [Rozeff, 1982]. Firms do so to avoid the more costly external funding. With reference to Table 3, there exists significant evidence to support the hypothesis that dividend payout is negatively related to historical growth. Ability of Finance firms to pay dividend is restricted by laws pertaining to capital and liquidity requirements to ensure the safety of investors funds. In addition, finance firms would retain more earnings in historical periods of high growth to meet the increase in demand for loans. Hence, payout ratio during such periods would be low. To support past growth, property firms would have devoted substantial amount of funds in property development activities, such as purchase of land and construction of property. Due to the high cost of developing property, such firms would adopt low payout in historical periods of growth. Historical periods of growth for Construction firms would be translated into increased financial commitments in raw materials and labour. This would explain their low payout ratio during these past periods of growth.

For firms in Transportation & Logistics, Electronics & Electrical Appliances, Automobiles and Machinery & Equipment, Paper Products and Printing & Publishing, past growth rates must be matched by substantial investments in long-term assets such as machinery and equipment. In most instances, increase their manpower was also necessary. Due to the increase in financing needs for investments and payroll expenses, coupled with the high cost of external financing, these firms typically has a low payout ratio when they experienced growth in the past.. Firms involved in Telecommunications would have devoted high non-recurring fixed costs in the construction and upgrading of infrastructure, as well as investments in new technology when faced with growth in past periods. Hence, this necessitates a low payout ratio in these periods. Firms in the Petroleum, Chemicals & Plastics, Metal Products and Mineral Products maintain low payout in past periods of growth due to their dependence on raw materials and their vulnerability to the volatility in the prices of raw materials, which makes up the bulk of their variable costs. Labour cost would constituent another expense component, particularly for labour intensive firms. Hence, these factors contribute to the low payout by these firms in past periods of growth. Firms in the Healthcare industry require specialised personnel to cater to the needs of customers and such payroll expenses contribute the bulk of financial needs. Medical supplies as well as investment in medical equipment also amount to a hefty sum for such firms. Hence, a low payout ratio was observed in historical high growth periods in order to finance such huge increase in these expenses.

Firms in Hotel & Travel Services operate in a service industry that relies heavily on human capital. For hotel operators, the cost outlay extends beyond payroll expenses to include maintenance and construction or expansion of accommodation facilities. Hence, these firms normally have a low payout ratio in periods of growth due to the huge capital involved to support historical growth,. For Retail firms as well as Apparels & Textiles, historical growth was matched by increased investments in inventory. Hence, such firms have adopted a low payout ratio to in order to finance the increase in inventory level. Although there may exist fundamental differences in firms categorised under Miscellaneous (Industrial) and Conglomerate, such firms would typically have devoted extra financial resources for investment in long-term assets or for support of increased variable costs in order to sustain historical growth. Hence, such firms have low payout ratios in periods when they experienced growth. All countries except Indonesia exhibit a negative relationship between dividend payout and historical growth. This result is not surprising as there is overwhelming evidence from Table 4 that indicate that firms that have experienced rapid historical growth typically have low payout ratio. More than half (19 out of 33 groups) of the industrial groups, representing about 81% of total sample, exhibit this trend. In addition, more than 65% of the firms in each country fall into the category of firms that exhibit this trend. TABLE 4: BREAKDOWN OF COMPANIES
Industry/ Country Finance Property Construction Transport & Logistics Telecommunications Elec & Appliance Taiwan 15 0 9 14 0 23 Korea 81 1 39 11 2 66 Thailand 67 27 0 9 5 14 Malaysia 32 33 32 11 0 10 Hong Kong 20 80 4 1 1 8 Singapore 18 21 12 10 1 8 Japan 155 28 154 81 9 211 Total 423 191 258 140 18 360

Automobile Apparel & Textiles Misc (Consumer) Machinery & Equip Paper Products Metal Products Mineral Products Misc (Industrial) Healthcare Services Hotels & Travel Printing & Publishing Retail Conglomerate Total

4 11 18 6 2 11 3 0 0 4 0 0 1 121 (84%)

34 64 70 39 23 50 33 0 1 1 3 8 2 528 (83%)

5 27 9 4 3 0 0 13 10 11 8 7 1 220 (68%)

6 7 22 2 5 24 2 0 2 2 4 21 4 219 (67%)

2 1 7 4 4 1 0 6 1 12 3 0 129 307 (88%)

2 1 0 0 1 4 0 1 1 14 1 4 25 154 (81%)

113 70 195 156 28 135 18 27 0 5 14 119 0 1518 (83%)

173 200 337 213 69 233 56 47 15 58 33 161 162 3147

(81%)

For Japan, due to preference for long-term capital appreciation as opposed to dividends [Michel and Shaked, 1986], payout ratio tends to be low. This is particularly true when Japanese firms need the financial resources to invest in various assets to support their growth. This same reasoning applies to Taiwan due to their tax structure as previously mentioned. Due to the existence of tax on dividend, retained earnings would probably be a favored choice in sustaining their historical growth, leading to the low payout in periods when Taiwanese companies experience growth. Opposite Direction The inconsistent results yielded by Household Goods may be due to the small number of firms grouped in that category. As only 20 firms out of a total of 3905 are categorized under Household Goods, the results may not be representative for the industry. Firm Size TABLE 5: REGRESSION COEFFICIENTS OF MARKET CAPITALIZATION Regression Standardized Coefficient p-value Expected Direction (+)
Utilities Metal Products Finance 0.173 0.058 0.128 0.002 0.038 0.000

Transportation & Logistics Machinery & Equipment

0.091 0.058 -0.161 -0.137 -0.182 -0.085 -0.084 -0.310 -0.117 -0.052 -0.635 -0.165 -0.443 -0.175

0.010 0.037 0.026 0.051 0.048 0.000 0.003 0.044 0.000 0.010 0.003 0.051 0.056 0.009

Opposite Direction (-)

Singapore Pharmaceuticals Printing & Publishing Chemicals, Petroleum & Plastics Food & Beverage Wood Products Malaysia Korea Household Goods Entertainment & Recreation Healthcare Services Miscellaneous (Services)

Expected Direction With reference to Table 5, firm size is positively related to dividend payout for five industries namely Finance, Machinery & Equipment, Metal Products, Utilities and Transportation & Logistics. As the size of a firm increases, information asymmetry may increase due to increased complexity of the firm and increased dispersion of ownership. [Kenneth, 1996] As such, shareholders are not able to monitor the large firms closely and this results in weak control of the management. High dividend payout leads to the increased need of external financing which in turn leads to increased monitoring of these firms by both existing and potential creditors. [Mozes and Rapaccioli, 1995] Thus, dividend payout acts as an indirect monitoring tool. [La Porta et al., 1998] In addition, further expansions for existing large firms will be increasingly difficult. For utilities firms in general, there is a limit to the consumer demand. Thus, when utilities firms increase to a certain extent, the market becomes saturated and further expansion is close to impossible. The same reasoning applies for metal products, which is basically a

primary product. 1The market demand for such metal products has been shrinking over the years due to the increased usage of synthetic substitutes so the need for further expansion is minimal. Furthermore, there is also a reduced need for such firms to intensify their investments to reinforce their market visibility. In fact, a further increase in size may be detrimental if increased size leads to higher tax brackets or increased scrutiny by investors, legal authorities and competitors. Finance firms, in particular, are governed by laws pertaining to capital and liquidity requirements to ensure the safety of investors funds e.g. deposits, insurance premiums, investment funds] in these firms. Thus, large firms will choose to distribute a larger proportion of their earnings to shareholders through dividends rather than pumping more financial resources into expansion plans. The same reasoning holds for the Machinery & Equipment and Transportation & Logistics industries that exhibit positive relationship between dividend payout and firm size.

Plastics served as substitutes for wood, glass and metal during the hardship times of World Wars I & II. After World War II, newer plastics, such as polyurethane, polyester, silicones, polypropylene, polycarbonate and polymethyl methacrylate joined polystyrene and PVC in widespread applications. By the 1960s, plastics were within everyone's reach due to their inexpensive cost. Plastics had thus come to be considered 'common' - a symbol of the consumer society. http://www.plasticsresource.com/disposal/life_cycle_feature/step1.html

Opposite Direction Dividend payout and firm size are negatively related for 3 countries namely Korea, Malaysia and Singapore and 9 industries namely Food & Beverage, Household Goods, Wood Products, Chemicals Petroleum & Plastics, Entertainment & Recreation, Healthcare Services, Pharmaceuticals, Printing & Publishing and Miscellaneous (Services). In Singapore, many of the larger firms are actually government-linked companies. There are already disclosure regulations in place to prevent misuse of funds so the problem of information asymmetry in these larger firms is in fact a smaller problem. As a result, the dividend payout ratio may be smaller for these large firms. Pharmaceutical firms focus on research and development and are by nature large. A natural line of argument following this is that larger pharmaceutical firms are more reputable and thus have a larger customer base, which includes medical centers, universities and research centers. This may lead to more R&D projects, which generally require large amount of financial resources. Thus, dividend payout will be lower for larger pharmaceutical firms. The Printing & Publishing and Chemicals, Petroleum & Plastics industries require high start-up costs due to the large amount of fixed costs required in developing such printing press and chemical plants. As a result, the large firms dominate and may be involved in greater scale production, and thus distribute less cash dividends as compared to their smaller counterparts. The Food & Beverage industry provide products, which are basic necessities to consumers. In our study, Food & Beverage is ranked 6th for comprising the largest

number of firms. This industry constitutes numerous small firms. The larger firms might be the stronger firms, which seek further expansion and thus leads to lower dividend payout. The mean market capitalization for Food & Beverage is USD757 million whereas the mean market capitalization for all firms included in our sample is USD1.36 billion. Information asymmetry may not be a deciding factor in this industry as the average size of these firms is rather small. Similarly for the Wood Products industry, the mean market capitalization is only USD329 million, which is a mere fraction of the mean market capitalization of all firms. Thus, information asymmetry of larger firms might not be a big problem as the size of all the firms are actually rather small when discussed in absolute terms. The same reasoning applies for Malaysia and South Korea. The mean market capitalization is only USD474 million for Malaysia and USD178 million for Korea in our sample. The seemingly perplexing direction of relationship between dividend payout and firm size for Household Goods, Entertainment & Recreation and Healthcare Services may be due to the small number of firms included in this particular regression. In our sample only 32 firms are categorized under Entertainment & Recreation, 21 firms under Household Goods and 16 firms under Healthcare Services. Lastly, for Miscellaneous (Services), there may be certain inconsistencies in the results due to fundamental differences in the firms categorized under this industry grouping.

Leverage TABLE 6: REGRESSION COEFFICIENTS OF DEBT TO ASSETS RATIO Regression Standardized Coefficient p-value Negative Direction (-)
Footwear & Leather Malaysia -0.344 -0.151 0.273 0.274 0.320 0.132 0.182 0.211 0.181 0.215 0.148 0.113 0.143 0.092 0.217 0.163 0.164 0.173 0.223 0.008 0.000 0.000 0.000 0.017 0.000 0.000 0.000 0.000 0.017 0.000 0.000 0.016 0.003 0.019 0.000 0.016 0.000 0.000

Positive Direction (+)

Construction Transportation & Logistics Mining Electronics & Electrical Appliance Food & Beverage Automobile Apparel & Textiles Miscellaneous (Consumer) Chemicals, Petroleum & Plastics Machinery & Equipment Paper Products Metal Products Entertainment & Recreation Retail Miscellaneous (Services) Korea Japan

Negative Direction It was found that firms trade off dividend payments with fixed financial charges. A highly leveraged firm with a high level of fixed financial charges would lower its payout ratio. Jensen (1992) Dividend payments are then substitutes for these charges. A firm with higher fixed charges depends more greatly on costly external financing, hence other things being equal, this firm will choose to pay lower dividends. Rozeff (1982) However, with reference to Table 6, only the industry of Footwear & Leather and Malaysia yield this negative relationship.

Positive Direction Our results support the positive relationship between dividend payout and leverage. About half of the industries (47%) yielded this positive relationship. Increased indebtedness or leverage leads to increased contacts with external financiers, which results in closer monitoring and an increased dividend payout. Greater external

monitoring pressures managers away from the consumption of perquisites and to use corporate assets more efficiently. In this way, more cash is disgorged, and dividends increase with indebtedness. Robert (1999) The industries whose leverage is positively related to dividend payout are Construction, Transportation & Logistics, Mining, Electronics & Electrical Appliance, Food & Beverage, Automobile, Apparel & Textiles, Miscellaneous (consumer), Chemicals, Petroleum & Plastics, Machinery & Equipment, Paper Products, Metal products, Entertainment & Recreation, Retail, Miscellaneous (Services). The countries whose leverage is positive related to dividend payout are Korea and Japan. Cash Flow Adequacy TABLE 7: REGRESSION COEFFICIENTS OF FREE CASH FLOW TO ASSETS RATIO Regression Standardized Coefficient p-value Expected Direction (+)
Machinery & Equipment Entertainment & Recreation Miscellaneous (Services) Japan Singapore Taiwan Chemicals, Petroleum & Plastics Electronics & Electrical Appliance Building Materials Utilities 0.072 0.343 0.110 0.051 0.068 0.105 0.051 0.077 0.138 0.140 -0.058 0.009 0.000 0.064 0.000 0.074 0.035 0.016 0.000 0.001 0.039 0.056

Opposite Direction (-)

Apparel & Textiles

Expected Direction Cash flow adequacy is vital in deciding the dividend payout as a firm will not distribute cash dividends without the means to do so. [Brigham and Daves, 2002] With reference to Table 7, our study further affirms this logic. Firms involved in Machinery and Equipment usually require high non-recurring fixed costs in the acquirement of the assets but low daily cash requirements. If these firms have no intention of acquiring new fixed assets in the near future, they may be left with large amount of idle cash, which generally yield low returns. Firms can reduce their cash balances by distributing cash dividends. This will lower the occurrence of the management using idle cash in non value-adding investments, thus reducing agency costs. [Mollah, Keasy and Short, 2000] For the Entertainment & Recreation and Miscellaneous (Services) industries, the daily cash exchanges are high, as most of these services are cash settled at the point of purchase. Thus, there will be idle cash balances if these firms have more cash inflows than outflows. Following the argument for Machinery and Equipment, these firms will then choose to distribute cash dividends if they have no specific use for these cash to prevent misuse of funds. The same logic follows for the rest of the countries and industries that exhibit a positive relationship between dividend payout and cash flow adequacy. There may not be a need to pinpoint any special reason specific to the country or industry to explain this typical relationship between dividend payout and cash flow adequacy, which adheres to past literature. Opposite Direction

Firms in the Apparel & Textiles industry requires large variable expenses, which includes wages, costs of materials etc to meet with their daily production. When the need arises, these firms may need to temporarily increase working scale and time to meet with the demand of customers especially during peak. Thus this is an industry which requires a lot of flexibility. In light of this, the free cash flow to asset ratio may mean the firms are intentionally pursuing such a high cash flow policy to maintain flexibility and meet with customer demands. In order to maintain a high level of free cash, the firms distribute fewer cash dividends. In the case of a positive relationship, our argument is based on the assumption that there is no specific use for the free cash on hand which then increases the likelihood of misuse of funds. However for the case of the Apparel & Textiles, a possible conjecture is that there is a need for high free cash on hand and thus low dividend payout is reasonable for firms with high cash flow adequacy. Collateralizable Assets TABLE 8: REGRESSION COEFFICIENTS OF NET FIXED ASSETS TO ASSETS RATIO Regression Standardized Coefficient p-value Expected Direction (+)
Transportation & Logistics Electronics & Electrical Appliance Food & Beverage Apparel & Textiles Chemicals, Petroleum & Plastics Machinery & Equipment Entertainment & Recreation Hotel & Travel Services Retail Miscellaneous (Services) Conglomerate Taiwan Korea 0.168 0.045 0.081 0.135 0.057 0.052 0.274 0.143 0.083 0.120 0.086 0.169 0.078 0.000 0.058 0.003 0.000 0.015 0.067 0.002 0.047 0.013 0.052 0.030 0.006 0.000

Thailand Malaysia
Singapore

0.077 0.167
0.091

0.046 0.000
0.055

Indonesia Japan

0.126 0.075 -0.098

0.077 0.000 0.095

Opposite Direction (-)

Paper Products

Expected Direction The above results support the positive relationship between dividend payout and collateralizable assets. If a firm holds more collateralizable assets, they have fewer agency problems between their bondholders and stockholders because these assets may serve as collateral against borrowing and consequently pay more dividends. Mollah, Kevin and Helen (2000) Fixed assets are long-term tangible assets held for business use and not expected to be converted to cash in the current or upcoming fiscal year, such as manufacturing equipment, real estate and furniture.

With reference to Table 8, the industries whose dividend payout are positively related to collateralizable assets are Transportation & logistics, Electronics & Electrical Appliance, Food & Beverage, Apparel & Textiles, Chemicals, Petroleum & Plastics, and Machinery & Equipment. The above industries largely belong to the manufacturing sector. Their fixed assets are mostly in manufacturing equipment. The other industries which yield significantly positive relationship between collateralizable assets and dividend payout are Entertainment & Recreation, Hotel & Travel Services, Retail, Miscellaneous (Services). This category of industries has a high level of fixed assets in real estate Hotel properties, amusement parks, retail stores, servicing centers.

As for the last industry conglomerate, which comprises of holding companies and companies which deal with many businesses, would generally have a much higher level of fixed assets due to their nature of multi-business. Seven out of the eight countries except Hong Kong, under our study show similar significantly positive relationship with collateralizable assets and dividend payout. Opposite Direction Only one industry shows a negative relationship between dividend payout and collateralizable assets which is the Paper Industry.

Systematic Risk
TABLE 9: REGRESSION COEFFICIENTS OF BETA Regression Standardized Coefficient p-value Expected Direction (-)
Healthcare Services Hong Kong -0.569 -0.068 0.057 0.128 0.122 0.070 0.055 0.153 0.093 0.142 0.129 0.191 0.122 0.034 0.108 0.096 0.003 0.020 0.006 0.000 0.000 0.001 0.048 0.000 0.005 0.000 0.000 0.000 0.040 0.063 0.052 0.000

Opposite Direction (+)

Finance Construction Transportation & Logistics Electronics & Electrical Appliance Food & Beverage Automobile Apparel & Textiles Chemicals, Petroleum & Plastics Machinery & Equipment Metal Products Pharmaceuticals Korea Indonesia Japan

Expected Direction Higher betas are a reflection of the presence of high operating and financial leverage. A firm with higher operating and financial leverage would have higher transaction costs. They pay lower dividends to avoid the costs of external finance. Rozeff (1982) With

reference to Table 9, only Healthcare services and Hong Kong yielded this negative relationship. Opposite Direction Our results support the positive relationship between dividend payout and beta, which is contrary to past literature results. It might be due to increased external monitoring, an increase in dividends, and a reduction in agency problems.

With reference to Table 8, the industries whose beta is positively related to dividend payout are Finance, Construction, Transportation & Logistics, Electronics & Electrical Appliance, Food & Beverage, Automobile, Apparel & Textiles, Chemicals, Petroleum & Plastics, Machinery & Equipment, Metal Products, and Pharmaceuticals. The countries whose beta is positive related to dividend payout are Korea, Indonesia and Japan.

Objective 2: To investigate if there are industry-specific differences in dividend payout. Summary of Results
TABLE 10: ANOVA OUTPUT FOR INTER-INDUSTRY DIFFERENCES
ANOVA DIVPAY Sum of Squares 29.527 1118.676 1148.203 df 31 20317 20348 Mean Square .952 .055 F 17.299 Sig. .000

Between Groups Within Groups Total

With reference to Table 10, statistically significant differences in dividend payout are observed between the thirty-two industries, analogous to previous studies done on inter-

industry differences2. Five industries have shown radical differences in dividend payout, namely Utilities, Transport & Logistics, Telecommunications, Pharmaceuticals and Hotel and Travel Services, of which three are regulated industries. In contrast, some industries show little inter-industry difference in dividend payout. These comprise of the ten industries that have five or less significant results. The Finance industry, contrary to our expectations, shows only moderate inter-industry difference in dividend payout considering the vast difference that exists between the financial and the other sectors. With reference to the means and medians of the dividend payout of various industries, utilities yield the highest payout of 54%. This observation yields little surprise as previous studies3 have recognized this phenomenon and have derived various reasons to explain it.

Michel (1979) finds statistically significant differences in dividend payout ratio among 13 different industries during the late 1960s to mid-1970s. Baker (1988) updates the Michel study using data from 1977 to 1981, and finds support for industry effects on payout ratio.
3

Under the Smith hypothesis [Smith, 1986], utilities possess larger payout than unregulated firms. In 1992, Moyer, Rao & Tripathy (1992) test the Smith hypothesis and explore the reasons for the high observed dividend payout. They suggest that such high payout is due to the fact that utility companies use dividend policy as a mechanism for controlling or responding to regulatory risk.

With reference to Table 11, Telecommunication and Pharmaceutical industry yields the lowest dividend payout of 27% and 28% respectively. TABLE 11: Industry
Finance Property Construction Utilities Transport & Logistics Telecommunications Agriculture, Forestry & Fishery Mining Computer & Computer Services Electronics & Electrical Supplies Food & Beverage Automobile Household Goods Apparels & Textile Footwear & Leather Products Wood Products Miscellaneous (Consumer) Chemicals, Petroleum & Plastics Machinery & Equipment Paper Products Metal Products Mineral Products Building Materials Miscellaneous (Industrial) Entertainment & Recreation Healthcare Services Pharmaceuticals Hotel & Travel Services Printing & Publishing Retail Miscellaneous (Services) Conglomerate

Mean of dividend 0.391806 payout

0.402232 0.36983 0.536339 0.482685 0.271076 0.486396 0.415857 0.356678 0.387022 0.427265 0.426454 0.411156 0.394848 0.44968 0.382669 0.416936 0.435113 0.397202 0.424869 0.386466 0.405554 0.465742 0.427362 0.422899 0.479747 0.280738 0.506741 0.375193 0.415177 0.418192 0.441585

Median of dividend 0.337823 payout

0.354947 0.312274 0.564625 0.446071 0.227884 0.484776 0.359001 0.298968 0.335267 0.387466 0.389807 0.369632 0.351105 0.395319 0.283947 0.374379 0.386351 0.345498 0.382442 0.335799 0.384349 0.442404 0.408209 0.361752 0.499024 0.237071 0.477972 0.345831 0.353149 0.356548 0.417368

No. of significant results

11 7 12 26 18 25 14 3 7 11 9 9 2 9 2 5 4 12 9 4 11 6 13 3 3 2 29 20 6 8 5 10

SUMMARY OF DIVIDEND PAYOUT FOR DIFFERENT INDUSTRIES

Reasons for Vast Differences There are certain pairs of industries, which show huge difference in their dividend payout, and we seek to provide probable reasons for this phenomenon. Extent of Regulation One of the reasons proposed to explain the reasons for the radical differences in payout policy among the various industries is the extent of regulation on the industries. A number of studies have been conducted on dividend policies of regulated versus unregulated firms4 and there seem to be a consensus that difference does exist between their dividend policies. Past literature suggests that a different set of factors determine dividend policy for regulated firms than that for unregulated firms. This is probably the reason why the dividend payout for the three regulated industries is so radically different from the regulated firms. The mean dividend payout for regulated firms is significantly larger than that for unregulated firms [Saxena, 1999], which is consistent from the results obtained where utilities and transport & logistics yield a high payout ratio of 54% and 46% respectively. Telecommunications, on the other hand, yield a low payout ratio of 27%. The specific reasons for such low payout will require further research but we feel that it may be due to the unique nature of the industry or is a result of regulations. Investment Opportunities

Collins, Saxena and Wansley (1996) recognizes the potential differences in dividend policy between regulated and unregulated firms and focus on agency-cost and monitoring explanations for the relevance of dividend. Atul (1999) proceed further to explain the possible differences in the dividend policy of both regulated and unregulated firms.

Another possible reason for the difference in payout ratio between industries may be the investment opportunities available for each industry. Michel (1979, p. 24) suggests that investment opportunities within industries may account partially for the industry effect. Firms or industries that have good investment opportunities typically have lower dividend payout ratio. This phenomenon can be explained by the pecking order which firms exhibit greater preference for retained earnings rather than external source of funds. However, regulated firms have their growth prospects determined by regulators, who directly or indirectly, dictate how much dividend the firm can pay [Saxena, 1998]. Since regulated firms do not base their dividend payout in relation to their growth opportunities, it is no surprise that their payout ratio is exceptionally different from unregulated firms. The pharmaceutical industry is characterized by their heavy investments in R&D, which differentiates it from other industries. Vast investment opportunities that exist in this industry probably led to low dividend payout. Cyclical Nature of Industry The cyclical movements of investment opportunities and earnings also affect the dividend policy of firms. [Dempsey, Laber and Rozeff, 1993] Firms in the Hotel & Travel Services industry operate in a cyclical environment, which involves peaks and off-peak periods in their demand. Hence, this probably explains the reason behind the industrys vast difference in dividend payout.

Reasons for Homogeneity

On the other hand, certain industries maintain typical payout ratios which are not significantly different from other industries when pairwise comparisons are made. Number of Firms in Category

Certain industries that exhibit minimal inter-industry differences in dividend payout may be due to the fact that the firms included in these industries are relatively small and may not be representative of the industry. Five out of the ten such industries comprises of less than 30 firms in each category, with Healthcare and Mining firms not even exceeding twenty. Thus, the mean dividend payout of the industry is likely to be of a typical range, not too different from the rest of the industries. Fundamental Differences of firms within industries For the three miscellaneous categories, namely for consumer, industrial and services, the minimal inter-industry differences in their dividend payout with respect to other industries may be due to the fundamental differences in firms classified under such categories. Since firms of the firms categorized under these categories might have a mixture of characteristics of other industries, there is unlikely to be huge inter-industry differences. Individual firms in the Paper Products and Wood Products industry may be very different with regards to their range of products and/or the type of customers they serve. Thus, the mean dividend payout of the industry will not be too extreme after the averaging of the dividend payout of all the firms in that industry. The pairwise comparisons of such countries with the rest of the industries will naturally show insignificant results. Conclusions In conclusion, our results show that differences in dividend payout among industries do exist. However, it is not possible to go into details to account for the differences among each and every industry. We have identified those industries either exhibit vast

differences or minimum differences and did a closer examination to propose some reasons resulting in the two extremes. The main reasons we proposed that cause the vast differences are the extent of regulations, investment opportunities and the cyclical nature of industry. As for the industries that show little difference in dividend payout, the reason may be that the firms included these industries are relatively small in number or have fundamental differences among them.

Objective 3: To investigate if there are country-specific differences in dividend payout. Summary of Results
Significant inter-country differences in dividend payout are found between the countries at 5% level of significance. Table 12 summarizes the mean difference of the dividend payout between the various countries. TABLE 12: SUMMARY OF MEAN DIFFERNCE BETWEEN COUNTRIES Significance of Mean Difference
Taiwan Taiwan Korea Thailand Malaysia Hong Kong Singapore Indonesia Japan 0.000 0.043 0.001 0.613 0.097 0.000 0.132 Korea 0.000 0.000 0.000 0.000 0.000 1.000 0.000 Thailand 0.043 0.000 0.000 0.000 0.000 0.000 0.000 Malaysia 0.001 0.000 0.000 0.024 0.925 0.000 0.036 Hong Kong 0.613 0.000 0.000 0.024 0.777 0.000 0.945 Singapore 0.097 0.000 0.000 0.925 0.777 0.000 0.974 Indonesia 0.000 1.000 0.000 0.000 0.000 0.000 0.000 Japan 0.132 0.000 0.000 0.036 0.945 0.974 0.000

With reference to Table 13, Thailand displayed interesting results, having the highest mean dividend payout ratios among all the countries in our study. Also, its dividend payout ratio is significantly different from all the other countries at 5% level of significance.

TABLE 13: SUMMARY OF DIVIDEND PAYOUT OF DIFFERENT COUNTRIES Country Mean of dividend payout Median of dividend payout
Indonesia Korea Malaysia Singapore Japan Hong Kong Taiwan Thailand 0.312 0.313 0.406 0.420 0.429 0.436 0.460 0.503 0.269 0.266 0.357 0.378 0.374 0.407 0.423 0.497

Indonesia and South Korea exhibit homogeneity in their dividend payout ratios. The two countries have the lowest mean dividend payout ratios. Also, their dividend payout ratios are significantly different from all other countries except between themselves. Tax Structures In order to explain the differences in dividend payout of these different countries, it is essential to look at the tax structures of these 8 countries. [Table 14] TABLE 14: SUMMARY OF APPROXIMATE TAX RATES OF DIFFERENT COUNTRIES Tax Rates Country Capital Gains Tax Dividend Taxes
Taiwan Korea Thailand Malaysia Hong Kong Singapore Indonesia Japan Ordinary Income (</=25%) NA NA NA NA NA Deemed Gains (0.1% of transaction value) Ordinary Income (0% to 55%) Corporate Level (25%) 21.5% NA Ordinary Income (</=28%) NA Imputation (</=26%) NA 35%

In South Korea, capital gains are tax-free. Dividends are taxed at a flat rate of 21.5%. Thus, investors in South Korea will definitely favour capital gains over dividends. [Table 15] In Indonesia, deemed gains on sale of shares traded on Indonesian Stock Exchange are subject to 0.1% of the transaction value. On the other hand, dividends are not taxed. Although this might suggest a higher dividend payout, the difference in tax rates between dividends and capital gains is rather insignificant. Moreover, capital gains are taxed only when the shares are sold. Thus, with a long-term horizon, capital gains might still be preferred over dividends, which constitute current income. Investors will need to reinvest their dividend income and might even incur more transaction costs than it would be if the dividends are not distributed and used to generate long-term capital gains which will only be taxed at the end of the investment horizon. TABLE 15: RELATIVE ADVANTAGE OF CAPITAL GAINS OVER DIVIDENDS Tax Korea Indonesia Dividend (Max) 21.5% 0% Capital Gains (Min) 0% 0.1% Advantage 21.5% -0.1% Dividend (Min) 21.5% 0% Capital Gains (Max) 0% 0.1% Advantage 21.5% -0.1% In Thailand, there is no tax imposed on dividends. With the exception of Hong Kong and Indonesia, the tax-free dividends are very much favourable as compared to all the other countries. Thus, higher dividend payout observed in Indonesia is justifiable. Hong Kongs mean dividend payout is significantly different from that of Malaysia. Hong Kong is a tax haven with no taxes on both dividends and capital gains. Whereas in Malaysia, capital gains are not taxed but dividends are taxed as normal income. Thus,

higher dividend payout observed in Hong Kong is a logical outcome with comparison of the various tax rates on relative grounds. TABLE 16: APPROXIMATE MAXIMUM TAX RATES Country TH HK ID KS TW SG MY Dividend Taxes [max] 0% 0% 0% 21.5% 25% 26% 28%

JP 35%

Subscribing to the same reasoning, the significant difference in dividend payout between companies in Taiwan and Malaysia is due to the different tax rates on dividends. In Malaysia, the maximum tax rate on dividends is 28% whereas in Taiwan, a flat rate of 25% at the corporate level is imposed. [Table 16]

Block Holdings
Another probable reason behind the difference in dividend payout ratio might be due to the difference in ownership structure in these various countries. As such information is considered proprietary, they may not be readily available especially when there is no regulation for the companies to disclose such information. We have thus chosen to take a closer look at the available information summarized by the various sources. TABLE 17: CONTROL OF PUBLICLY LISTED COMPANIES IN 1996 Percentage of Control Countries No. of Family Widely The State Widely Held Widely Held Nonsample Held Financial Financial companies Institutions Corporations Indonesia 178 67.3 6.6 15.2 2.5 8.4 Korea 345 24.6 51.1 19.9 0.2 4.3 Malaysia 238 42.6 16.2 34.8 1.1 5.3 Thailand 167 51.9 8.2 24.1 6.3 9.5
Weighted by market capitalization

As shown in Table 17, many of +the public listed companies in East Asia are familyheld. It is especially so in Indonesia with 67.3% held by family and only 6.6% widely held. Since shareholders are made up of family members, they are most likely to be the

management of the companies as well. Assuming there is no expropriation and moral hazard, it can be argued that there will be little agency problem. These companies will have a lower dividend payout, which constitutes current income and focus more on longterm growth, which comes in the form of capital gains. Thus, the low dividend payout in Indonesia might be due to the high concentration of family owned corporations. TABLE 18: OWNERSHIP CONCENTRATION Concentration Ratio (Percent) Largest Top Five As of End of the Company Coverage Shareholder Shareholders Year 48.2 67.5 1997 All PLCs 20.4 38.5 1998 81 non-financial PLCs 30.3 58.8 1998 All PLCs 28.5 56.6 1997 All PLCs

Country Indonesia Korea Malaysia Thailand

Concentration ratio is defined as the percentage of total outstanding shares of an average PLC owned by the largest or top five shareholders. The percentages are not weighted by market capitalization

Also, the ownership is highly concentrated with top shareholders owning a large proportion of the companies. Thus, this further illuminates the high decision power of the family who holds ownership of these companies. [Table 18] In the case of Singapore, the mean dividend payout is not significantly different from other countries with the exception of Indonesia, South Korea and Thailand mentioned earlier. Many of the large companies are actually government-linked companies. Thus, there are reasonable monitoring tools and legislation governing such companies. With this in mind, the dividend payout ratio may assume a more typical range and will not be as low as that of Indonesia or as high as that of Thailand as the reasons supporting the extreme dividend payout ratios may not be that relevant in the Singapore context.

Socio-Cultural & Political Influence

We must bear in mind that socio-cultural has a great deal of influence on the way the corporations are structured. In the East Asian context, political and cultural influences are important factors that may explain the differences in dividend payout. In Japan, there is a unique system whereby firms and banks- kereitsus have close banking relationships. In contrast there is no such practice in Malaysia. Thus, additional financing might be easier for companies in Japan as compared to Malaysia due to the established banking relationships in Japan. As a result, firms in Japan are able to distribute higher cash dividends, as there is greater ease of obtaining funds when the need arises. Another interesting observation is that the 3 countries with extreme dividend payout namely Thailand, Indonesia and South Korea are the worst hit countries during the Asian financial crisis. These countries have considerable less conservative lending practices e.g. directed loans. In the case of Thailand, due to the relaxed lending practices, the companies might have less consideration about their ability to maintain the high dividend payout since they will be able to get loans easily. As for Indonesia and Thailand, it will be illogical to use the same line of reasoning since they have the lowest dividend payout. A possible explanation might be due to the overall lack of monitoring of these companies. Thus, the shareholders do not actually demand a higher dividend payout to exert indirect monitoring by creditors as the indirect monitoring mechanism simply does not work. Thus, we conclude that there are indeed differences in the general dividend payout adopted by firms in different countries. It is impossible to explore the differences that lie between each and every country by the sheer number of possibilities but we seek to summarize our findings and provide some probable explanations that generally explain

these differences. The 3 main reasons we propose are tax structure, ownership structure and lastly socio-cultural & political

CONCLUSION Summary of Findings

This study examined the relationship between dividend payout and several variables including expected growth, historical revenue growth, firm size, indebtedness, cash flow adequacy, collateralizable assets and market risk. Our uniqueness lies in the East Asian context of our study. Past studies have rarely investigated dividend policy with regards to the East Asian context and do not usually extend their sample to eight countries as we have done in our study. The results show that there are significant relationships between dividend payout and the tested variables. However, the degree of significance for different variables varies among countries and industries. Interestingly, the proxy for market risk- beta has a positive relationship with dividend payout contrary to that found in previous studies. One probable reason cited is that investors will demand more monitoring on firms with high firm risk. However, it might be inequitable for each individual investors to monitor the firm risk on their own especially if they do not own large stakes in the risky firms. Thus, indirect monitoring tools such as a high dividend payout may prove to be useful in serving this purpose. Other tested variables such as firm size and cash flow adequacy showed inconsistency in their directions of relationship with dividend payout for different countries and industries. This seemingly puzzling observation can be explained by the unique characteristics of these countries and industries. With respect to past literature, indebtedness proves to have different directions of relationship with dividend payout in

different studies. Our study has shown that indebtedness is positively related to the dividend payout. Our inter-industry study of dividend payout showed that there are significant differences in the dividend payout of different industries. Inter-industry differences are due to the different operating nature as well as degree of government surveillance of these firms. Dividend policy has been viewed as an indirect monitoring tool and its usefulness in serving this purpose has to be weighted against the fundamental nature of these industries. Certain monitoring systems may be in place for certain industries. Also, certain industries may have other more important objectives in mind, thus choose a certain suitable level of dividend payout. Similar tests carried out for inter-country differences also showed significant results. Such differences are attributed to the different tax rates, institutional structures and socio-cultural and political influences. Among the countries in our study, Indonesia and South Korea displayed the lowest dividend payout whereas Thailand showed the highest dividend payout. Interestingly, we have made a side remark that these three countries with extreme dividend payout are the worst hit countries during the Asian Financial Crisis.

References
Aczel, A.D., Complete Business Statistics, Mc-Graw Hill, (1999), pp. 498-598. Allen, F. and R. Michaely, Payout Policy, The Wharton Financial Institutions Center, (June 2001). Alli, K.L., A.Q. Khan and G.G. Ramirez, Determinants of Corporate Dividend Policy: A Factorial Analysis, The Financial Review, Vol. 28 Issue 4, (November 1993), pp. 523547.

Ang, J.S., A. Fatemi and A. Tourani-Rad, Capital Structure and Dividend Policies of Indonesian Firms, Pacific-Basin Finance Journal, Vol. 5 Issue 1, (Feb 1997), pp. 87-103. Baker H.K., The Relationship Between Industry Classification and Dividend Policy, Southern Business Review, (Spring 1988), pp. 1-8. Brigham, E.F. and P.R. Daves, Intermediate Financial Management, seventh edition, United States of America: South-Western Thomson Learning, (2002), pp. 560-586. Buchanan, B.G., Institution Smoothing, Institutions & Agency Issues: A Group of Seven Country Comparisons, University of Georgia, (2000) Capulong, M.V., D. Edwards, D. Webb and J.Z. Zhuang, Corporate Governance and Finance in East Asia: A Study of Indonesia, Republic of Korea, Malaysia, Philippines, and Thailand: Volume One (A Consolidated Report), Asian Development Bank, (2000), pp. 23-61. Chirinko, R.S. and A.D Phillips, Dividend Policy at The Baby Bells: A Study of Septuplets, Emory University, (November 1998).

Claessens, S., S. Djankov and H.P. Lang, Who Controls East Asian Corporations?, Financial Economics Unit, Financial Sector Practice Department, World Bank, (1999). Collins, M.C., A.K. Saxena and J.W. Wansley, The role of insiders and dividend policy: A comparison of regulated and unregulated firms, Journal of Financial and Strategic Decisions, Vol. 9 Issue 2, (Summer 1996), pp. 1-9. Crutchley, C.E. and R.S. Hansen, A Test of the Agency Theory of Managerial Ownership, Corporate Leverage, and Corporate Dividends, Journal of Financial Management, Vol. 18 Issue 4, (Winter 1989), pp. 36-46. Easterbrook and H. Frank, Two agency-cost explanations of dividends, American Economic Review 74, (1984), pp. 650-659. Edward, D., and H.J. Ronald, A note on dividend policy and beta, Journal of Business Finance and accounting, Vol. 13 Issue 1, (1986), pp. 107-115. Fama, E.F. and K.R. French, Disappearing Dividends: Changing Firm Characteristics or Lower Propensity to Pay?, The Center for Research in Security Prices, (June 2000). Fung, K.H. and M. F. Theobald, Dividends and Debt under Alternative Tax Systems, The Journal of Financial and Quantitative Analysis, Vol. 9 Issue 1, (March 1984), pp. 5972.

Harkins, E. and F.Walsh, Dividend Policies and Practices, New York, The Conference Board, (1971). Harris, T.S., R.G. Hubbard and D. Kemsley, The Share Price Effects of Dividend Taxes and Tax Imputation Credits, Journal of Public Economics, (Dec 1999). Higgins, R. S., The Corporate Dividend-Saving Decision., Journal of Financial and Quantitative Analysis, Vol. 7, (March 1972), pp. 1527-1541. Houston, J.F. and C.M. James, Banking Relationships, Financial Constraints and Investment: Are Bank Dependent Borrowers More Financially Constrained?, (Dec 1995). Jensen, G.R., D.P. Solberg, and T.S. Zorn, Simultaneous Determination of Insider Ownership, Debt and Dividend Policies, The Journal of Financial and Quantitative Analysis, Vol. 27 Issue 2, (June 1992), pp. 347-263. Kang, B.S., Essays on Corporate Dividend Policy: International Investigation at the Micro and Macro Levels, Bell & Howell Information and Learning, (2001).

Kenneth J.M., The Method of Payment in Corporate Acquisitions, Investment Opportunities, and Management Ownership, The Journal of Finance, Vol. 51 Issue 4, (September 1996), pp. 1227-1246. La Porta, R., F. Lopez-de-Silanes, A. Shleifer and R.W. Vishny, Agency Problems and Dividend Policies Around the World, (October 1998). Lin, C.H. and C.Y. Shui, Foreign Ownership in Taiwan Stock Market An Empirical Analysis., Journal of Multinational Financial Management, Vol. 13 Issue 1, (Feb 2003), pp. 19-41. Lintner, J., Distribution of Incomes of Corporations Among Dividend, Retained Earnings and Taxes, The American Economic Review, (May 1956), pp. 97-113. Michel A., Industry Influence on Dividend Policy, The Journal Financial Management, Vol. 8 Issue 3, (August 1979), pp. 22-26. Michel, A. and I. Shaked, Country and Industry Influence on Dividend Policy: Evidence From Japan and the U.S.A., Journal of Business Finance & Accounting, Vol. 13 Issue 3, (Autumn 1986), pp. 365-381. Mohd, M.A., L.G. Perry and J.N. Rimbey, An investigation of the dynamic relationship between agency theory and dividend policy, The Financial Review, Vol. 30 Issue 2, (1995), pp. 367-385.

Mollah, A.S., K. Keasy and H. Short, The Influence of Agency Costs on Dividend Policy in an Emerging Market: Evidence from the Dhaka Stock Exchange, University of Leeds, (2000). Mozes, H. and D. Rapaccioli, The Relation Among Dividend Policy, Firm Size, and the Information Content of Earnings Announcements, The Journal of Financial Research, Vol. 8 Issue 1, (Spring 1995), pp. 75-88. Rozeff, M. S., Growth, Beta and Agency Costs as Determinants of Dividend Payout Ratios, The Journal of Financial Research, Vol. 5 Issue 3, (Fall 1982), pp. 249-259. Saxena, A.K., Determinants of Dividend Payout Policy: Regulated vs. Unregulated Firms, Mercer University, (1999). Titman, S. and T. Wessels, The Determinants of Capital Structure Choice, Journal of Finance, (March 1998), pp. 1-19.

URL References
American Council for Capital Formation Center, An International Comparison of Capital Gains Tax Rates, ACCF Center for Policy Research Special Reports, (August 1998), http://www.accf.org/capgains898.htm. BKR Walker Wayland Australian Offices, Asia-Pacific Tax Facts, BKR International, (1998), http://www.bkrlewbarr.com/taxfacts/Default.htm. Yahoo! Finance, Asia Tax Facts, Yahoo! International Finance Center, http://biz.yahoo.com/ifc/ National Center for Policy Analysis, "The Supply-Side Decade", Investor's Business Daily, (January 11 1996), http://www.ncpa.org/pd/economy/ecoa3.html. Plastic Resources, Introduction to Plastics, http://www.plasticsresource.com/disposal/life_cycle_feature/step1.html

APPENDIX I COMPUTATIONSs Dividend Payout Dividend Payout Ratio = Expected Growth Tobins Q = Total liabilities + Ending stock price No. of common shares Total assets Common cash dividends per share No. of common shares Net Income

Historical Revenue Growth Sale(x) 1/3 Revenue Growth = - 1 100% Sale(x-3) Firm Size Ending share price (domestic) e* where e* = Domestic Currency per USD Ending Share Price (USD) = Market Capitalization = Ending Share Price (USD) No. of Common Shares (End) Leverage Debt to assets ratio = Total liabilities 100 % Total assets

Cash Flow Adequacy As the PACAP database does not provide depreciation figures, we have resorted to an adjusted measure of free cash flow so that the measures in the various countries are comparable. Free Cash Flow (FCF) = Income from operations Income tax - ? Total current assets + ? Total current liabilities - ? Net fixed assets

Free Cash Flow to Assets ratio = Collateralizable Assets Fixed Assets To Assets Ratio = Systematic Risk

Free cash flow Total assets

Net fixed assets Total assets

Ri,t = i + i RM,t + Ri,t = the rate of return for asset i during period t RM,t = the rate of return for the market Portfolio M during period t i = the constant term of the regression i = the systematic risk of asset i = the random error term

Beta is derived by regressing 3-year monthly stock returns and the monthly market returns of the respective markets, which the firms are listed in. The monthly market returns are market value-weighted. Also, both the monthly stock returns and monthly stock returns are with the reinvestment of dividends.