You are on page 1of 108

THE INFLUENCE OF OWNERSHIP STRUCTURE ON THE FIRMS DIVIDEND POLICY BASED ON LINTNER MODEL

NOHASNIZA BINTI MOHD HASAN ABDULLAH

MASTER SCIENCE FINANCE UNIVERSITI UTARA MALAYSIA NOVEMBER 2009

THE INFLUENCE OF OWNERSHIP STRUCTURE ON THE FIRMS DIVIDEND POLICY BASED ON LINTNER MODEL

by

NOHASNIZA BINTI MOHD HASAN ABDULLAH 801918

A Dissertation Submitted in Partial Fulfillment of the Requirements for the Degree of Master of Science in Finance at the Graduate School of Management, Universiti Utara Malaysia

DECLARATION

I hereby declare that the project paper is based on my original work except for quotations and citations that have been duly acknowledge. I also declare it has not been previously or concurrently submitted for any other Masters programme at Universiti Utara Malaysia or other institutions.

_____________________________________________ NORHASNIZA BINTI MOHD HASAN ABDULLAH Date: 23 NOVEMBER 2009

ACKNOWLEDGEMENT

All my praises and gratitude to Allah, the Merciful, for His kindness and for meeting me with many wonderful people who, with His Grace, have had helped me tremendously in the successful completion of this research.

This research would not have been possible without the constructive comments, suggestion and encouragement received from my supervisor who has read the various draft. In particular, I would like to acknowledge my debt to Associate Professor Dr. Yusnidah Ibrahim, without, of course, holding her responsible for any deficiencies remains in this research.

I would like to thank my parents, who have been a continuous source of inspiration and encouragement. Thanks for giving a great support throughout the duration of my studies and unceasing prayers for my success.

In addition, thanks to all my friends that helped, support and provided insight and useful ideas, constructive comments, criticism and suggestion throughout the duration of completing this research.

Thank you.

iii

PERMISSION TO USE

In presenting this dissertation as a partial fulfillment of the requirements for a postgraduate degree from Universiti Utara Malaysia, I agree that the universitys library may take it freely available for inspection. I further agree that permission for copying of this dissertation in any manner, in whole or in part, for scholarly purposes may be granted by my supervisor or in other absence by the Dean, Postgraduate Studies, and College of Business. It is understood that any copying or publication or use of this dissertation or parts thereof for financial gain shall not be allowed without my written permission. It is also understood that due to recognition shall be given to me and to Universiti Utara Malaysia for any scholarly use which may be made of any material from my dissertation.

Request for permission to copy or to make other use of materials in this dissertation, in whole or in parts should be addressed to:

Dean, Postgraduate Studies College of Business Universiti Utara Malaysia O6100 Sintok Kedah Darul Aman

ABSTRAK Kajian ini meneliti hubungan antara jenis struktur pemilikan dan bayaran dividen daripada syarikat yang berdaftar di Malaysia. Silang kajian analisis digunakan ke atas 150 sampel syarikat yang disenaraikan di papan utama Bursa Malaysia pada tahun 2007. Kajian menguji kekuatan tiga alternatif dividen model, penyesuaian penuh model, model pelarasan separa dan Waud model yang diuruskan oleh kesan mungkin lima jenis struktur pemilikan, iaitu pemusatan pemilikan, penyebaran pemilikan, institusi pemilikan, pengurusan pemilikan dan pemilikan asing. Pemusatan pemilikan diukur oleh dua proksi, yang Herfmdahl Indeks dan bentuk yang baru diukur dengan indeks penjumlahan peratusan saham dikendalikan oleh dua pemegang saham utama. Penyebaran pemilikan diukur dengan nisbah jumlah pemegang saham terhadap jumlah saham, pemilikan institusi diukur dengan peratusan ekuiti yang dimiliki oleh pelabur institusi, sementara, pengurusan pemilikan diukur dengan menambah peratusan jumlah saham secara langsung diselenggarakan oleh non-eksekutif independen pengarah di syarikat, dan pemilikan asing diukur dengan jumlah semua saham di tangan pemegang saham asing dalam senarai pemegang saham terbesar tiga puluh, baik yang diselenggarakan melalui calon syarikat atau syarikat lain pemilikan saham asing. Kedua-dua pembolehubah pemilikan pemusatan ditemui untuk secara positif dan secara statistik signifikan dalam mempengaruhi dividen dalam setiap jenis model dividen. Temuan ini konsisten dengan teori agensi kerana pembayaran dividen yang tinggi boleh digunakan untuk mengurangkan konflik agensi kerana dividen boleh digantikan pemantauan pemegang saham. Oleh kerana itu, pemegang saham besar mempunyai insentif yang kuat untuk meminta bayaran dividen yang lebih tinggi untuk mengurangkan kos pemantauan. Meskipun demikian, kajian ini menunjukkan bahawa keputusan dividen syarikat Malaysia tidak dipengaruhi oleh struktur pemilikan.

Kata kunci: dividen, struktur pemilikan

ABSTRACT This study investigates the relationship between types of ownership structure and dividend payments of Malaysian listed companies. A cross-sectional analysis of 150 sample firms listed on the main board of Bursa Malaysia for the years 2007 is utilized. The study examines the explanatory power of three alternative models of dividend policy, the full adjustment model, the partial adjustment model and the Waud model modified which are moderated by the possible effects of five types of ownership structure, namely ownership concentration, ownership dispersion, institutional ownership, managerial ownership and foreign ownership. Ownership concentration is measured by two proxies, the Herfindahl Index and a newly form index measured by the summation of the percentage of shares controlled by two major shareholders. Ownership dispersion is measured by ratio of the number of shareholders to total outstanding shares, institutional ownership is measured by a percentage of equity owned by institutional investors, while, managerial ownership is measured by adding the total percentage of shares directly held by non-independent executive directors in the company, and foreign ownership is measured by the sum of all shares in the hands of foreign shareholders in the list of thirty largest shareholders, either held through nominee companies or other corporate foreign share holdings. Both ownership concentration variables are found to be positively and statistically significant in influencing dividends in every type of dividend model. The finding is consistent with agency theory since high dividend payments can be used for mitigating agency conflict as dividends can be substituted for shareholder monitoring. Hence, large shareholders have strong incentives to require higher dividend payments in order to reduce monitoring costs. Nevertheless, this study shows that dividend decisions of Malaysian companies are not influenced by the structure of ownership.

Keywords: dividends; ownership structure

ii

TABLE OF CONTENTS

DECLARATION PERMISSION TO USE ABSTRACT (BAHASA MELAYU) ABSTRACT (ENGLISH) ACKNOWLEDGEMENT TABLE OF CONTENTS LIST OF TABLES LIST OF ABBREVIATIONS i ii iii iv vii viii

CHAPTER 1: BACKGROUND OF STUDY 1.1 1.2 1.3 1.4 1.5 1.6 Introduction Problem Statement Objective of the Study Significance of the Study Limitation of the Study Conclusion 1 4 8 9 10 11

CHAPTER 2:LITERATURE REVIEW 2.1 2.2 Introduction Theoretical Literature 2.2.1 M&M Irrelevant Dividend Theory and other Related Theories/Models 13 12 12

2.2.2 Lintner Dividend Stability Theory and other Related Theories/Models 2.3 2.4 Empirical Literature Conclusion 17 20 49

iv

CHAPTER 3: RESEARCH METHOD 3.1 3.2 Introduction Research Framework 3.2.1 Ownership Concentration 3.2.2 Ownership Dispersion 3.2.3 Institutional Ownership 3.2.4 Managerial Ownership 3.2.5 Foreign Ownership 3.3 3.4 Sample Description and Data Collection Models on Dividend Policy 3.4.1 The Full Adjustment Model 3.4.2 The Partial Adjustment Model 3.4.3 The Waud Model 3.5 Measurement of Variable 3.5.1 Dividends 3.5.2 Earnings 3.5.3 Ownership Concentration 3.5.4 Ownership Dispersion 3.5.5 Institutional Ownership 3.5.6 Managerial Ownership 3.5.7 Foreign Ownership 3.6 Conclusion 50 50 51 51 51 52 52 53 54 54 55 56 57 57 58 58 58 58 59 59 59

CHAPTER 4: ANALYSIS AND FINDINGS 4.1 4.2 4.3 4.4 Introduction Descriptive Analysis Correlation Analysis Regression Analysis 4.4.1 Multicollinearity 4.4.2 Serial Correlation and Heteroscedasticity Test 4.4.3 Regression Results 4.5 Conclusion
v

61 61 64 67 67 69 71 75

CHAPTER 5: CONCLUSION 5.1 5.2 5.3 5.4 5.5 5.6 Introduction Overview of the Research Process Summary of Findings Implications of the Study Direction for Further Studies Conclusion 76 76 78 80 81 82

REFERENCES APPENDICES

83

vi

LIST OF TABLES

Table 2.1: Table 4.1: Table 4.2: Table 4.3:

Summary of Empirical Literatures Summary Descriptive Statistic Pearson Correlation Matrix among the Variables Variance Inflation Factor of Variables (tolerance value is given in the parentheses)

Table 4.4: Table 4.5:

Durbin-Watson and Heteroscedasticity Diagnostic Test Results of Multiple Regression Analysis of Dividend Policy Models

vii

LIST OF ABBREVIATIONS

Earnings Earning Change Dividends Ownership Concentration Ownership Dispersion Institutional Ownership Managerial Ownership Foreign Ownership The Full Adjustment Model The Partial Adjustment Model The Waud Model

ECHG : D :

CONC : DISP : INST : MNG : FOR :

FAM : PAM : WM :

viii

CHAPTER ONE

BACKGROUND OF STUDY

1.1

INTRODUCTION

Incomes are earned by successful companies. These incomes can be invested in operating assets, used to retire debt or repurchase shares, or distributed to shareholders in the form of dividends. When investors buy an ordinary share in a company, they become a shareholder of the business and to that extent they will have certain entitlements, including the right to receive dividend payments. Dividends are defined as a form of rational income distribution offering to shareholders (Baker et al, 2007). Dividends are a way for companies to reward shareholders for their investment and risk-bearing. Besides, dividends also give shareholders additional returns in addition to capital gains. Normally, dividends will be distributed in the form of cash, though it can also come in the form of stock dividends.

Dividends are decided upon and declared by the board of directors. Nevertheless, this pay-out is not guaranteed and the amount that shareholders will receive varies from company to company and year to year.

Generally, there are two types of cash dividends, which are interim dividends and final dividends. Interim dividends are declared and distributed before the companys annual earnings are known. These interim dividends are paid out of undistributed profits brought from previous periods. A company may choose to pay interim dividends quarterly or half yearly as long as it has adequate undistributed profits brought forward from previous periods. These dividends usually accompany the companys interim financial statements. On the other hand, final dividends are declared at the end of the financial period at the time when the directors are aware of the companys profitability and financial health. Normally, final dividends are declared before the books are closed and will be paid the following year. Thus final dividends will appear as dividends payable or proposed dividends under current liabilities in the balance sheet of that period.

In Malaysia, companies are free to decide when and how much to pay out in dividends for a specific financial business year as long as they comply with the Companies Act, 1965. According to Section 365 of the Act, No dividend shall be payable to the shareholders of any company except out of profits or pursuant to Section 60. In other words, the Act requires that dividends of a company can only be distributed from the profits of the company except pursuant to Section 60 of the Act.

Besides, the unique characteristic of dividends in Malaysia is the tax exemption feature. With effect from the year of assessment 2008, a single-tier income tax system will replace the imputation system. Under the imputation system, a Malaysian

resident company is required to deduct taxes at the prevailing corporate tax rate on taxable dividends paid to its shareholders. This tax is already accounted for through the tax paid by the company on its taxable profits, which is accumulated as dividend franking credits (Section 108 credits). When shareholders receive taxable dividends, they are entitled to a tax credit for the tax already paid by the company in respect of the income. Those credits are then used to offset the shareholders tax liability.
However, under the single-tier system, profits are only taxed at the company level;

thus, dividends paid under this system will be tax-exempt in the hands of shareholders.

Since Modigliani and Millers seminal studies (1958, 1961), dividend policy has been an issue of great interest in the finance literature. Following their irrelevance dividend policy hypothesis many explanations have been provided in order to solve the socalled dividend puzzle. Despite a large body of literature on dividends and payout policy, researchers have yet to reach a consensus on why firms pay dividends and what determines the payout ratio. Some of the theoretical principles underlying the dividend policy of firms can be described either in terms of information asymmetries, the tax-adjusted theory, or behavioral factors. The information asymmetries encompass several aspects, including the signaling models, agency costs and the free cash flow hypothesis.

1.2

PROBLEM STATEMENT

Dividends are payments made by a company to its shareholders, usually after a company earns a profit. Thus, dividends are not considered as a business expense but are a sharing of recognized assets among shareholders. Dividends are either paid regularly or can be called out anytime. Consequently, a dividend policy is a set of company rules and guidelines used to decide how much the company will pay out to its shareholders. Dividend policy is an essential financial decision made by the board of directors and the management and this decision is one of the fundamental components of corporate policy.

Dividend policy has been viewed as an issue of interest in the financial literature and one of the most controversial topics in finance. Despite a large body of literature on dividends and payout policy, researchers have yet to reach a consensus on why firms pay dividends and what determines the payout ratio. The extent literature on dividend payout ratios provides firms with no generally accepted prescription for the level of dividend payment that will maximize share value. Some researchers believe that dividends increase shareholder wealth (Gordon, 1959) while many others believe otherwise. Miller and Modigliani (1961) in their irrelevant dividend hypothesis,

asserts that under perfect market conditions, characterized among others by the nonexistence of taxes, transaction costs and asymmetric information, dividends are irrelevant since shareholders can create homemade dividends by selling a portion of

their portfolio of equities if they want cash and that there is a tradeoff between current dividends and future capital gain.

Taking into consideration various capital market imperfections, a considerable amount of theory and model are suggested to explain the dividend policy of companies. Signaling models are based on the assumption that managers have more information about the companys future cash flow than do individuals outside the company, and they have incentives to signal that information to investors (Gugler, 2003). Unexpected changes in dividend policy are used to mitigate information asymmetries between managers and owners (Frankfurter and Wood Jr., 2002). On the other hand, agency theory posits that by distributing resources in the form of cash dividends, internally generated cash flows are no longer sufficient to satisfy the needs of the companies. As a result, companies will visit the capital market more frequently for financing needs, thereby bring them under the greater scrutiny of the capital market (Easterbrook, 1984). Therefore, the payment of dividends provides the incentive for managers to reduce the costs associated with the principal/agent relationship.

Agency theory seeks to explain corporate capital structure as a result of attempts to maximize shareholder wealth since dividends can act as a bonding mechanism to reduce the agency costs arising from the conflict between managers and shareholders. Starting with Jensen and Meckling (1976), researchers have been addressing the agency problem in finance from many angles. Nowadays, extensive research has been

carried out regarding the issue of agency costs of dividends and the standard findings shows that dividends mitigate the free cash flow and therefore limit the managers ability to enlarge his or her own perks. However, this finding is still inconclusive since other studies have questioned the validity of this finding. For example, Noronha et al. (1996) had regressed five factors as a proxy for agency costs on the dividend payout ratio, but they found that the dividend policy is not the product of an attempt to mitigate the free cash flow problem.

Agency costs happen because of conflicts of interest between agents and shareholders. Therefore, agency costs are zero in a 100% owner-managed firm. As a companys ownership structure changes and ownership is separated from control, incentive alignment problems become more important. It is assumed that if managers and shareholders are left alone, they will attempt to act in his or her own self-interest. Self-motivated management behavior includes direct expropriation of funds by the manager, consumption of excessive perquisites, shirking and suboptimal investment. The nature of monitoring and bonding contracts, the managers taste for no pecuniary benefits and the cost of replacing the manager make the actual magnitude and impact of this self-seeking behavior vary across company and country (Jensen and Meckling, 1976).

Agency theory has also brought various external and internal monitoring and bonding mechanisms to the forefront of theoretical discussion and empirical research. Recent studies emphasize the potential conflicts of interest between controlling shareholders

and other shareholders. For example, Shleifer and Vishny (1997), Faccio et al. (2001) and Holderness (2003) argued that when large owners gain nearly full control of the corporation, they prefer to generate private benefits of control that are not shared by minority shareholders. Hence, firms with large controlling shareholders may exhibit a different type of agency conflict, namely the expropriation of minority shareholders by majority shareholders. On the other hand, in the presence of large shareholders, managerial discretion can be restrained to some extent and agency costs between managers and shareholders are reduced because large shareholders have the ability and the incentives to monitor and discipline management (Shleifer and Vishny, 1986). However, this would imply a lesser role for corporate payout policy to address agency problems between corporate insiders and outside shareholders.

Despite a great deal of prior research on the subject, few studies investigated the agency and ownership-based explanations of dividend policy. It is also important to note that the extent to which the companys dividend payout policy is effective in reducing the expected agency costs may also depend on its ownership and control structure. Nevertheless, one study by Mat Nor and Sulong (2007) had examined the relationship between ownership structure and dividends in Malaysia. They had used four types of ownership, namely ownership concentration, government ownership, foreign ownership and managerial ownership. However, their findings show a low explanatory power (between 0.118 and 0.124). On the other hand, a study in UK by Short, Zhang and Keasey (2002) that examined the link between corporate dividend policy and the ownership of shares by institutional investors and managers, using four

models of dividend policy, the full adjustment model, the partial adjustment model, the Waud model and the earnings trend model found a very high explanatory power (between 0.843 and 0.993). Their study is the first example of using well-established dividend payout models to examine the potential association between ownership structures and dividend policy. These four models, which describe the adjustment of dividends to changes in several measures of corporate earnings, have been modified by the addition of dummy variables representing institutional and managerial ownership, in order to determine whether the presence of the specific classes of investors in the ownership structure affect the process of determination of the level of the earnings that are being distributed. Thus, this situation brings up a question whether it is true that ownership structure has a low impact on corporate dividend policy in Malaysia. Therefore, this study attempts to examine the hypothesized relationship between corporate dividend policy and the various types of ownership structure by using dividend payout models.

1.3

OBJECTIVE OF THE STUDY

Main Objective:

To investigate the adoption of agency costs theory in explaining dividend policy in Malaysian listed companies.

Specific Objective:

To examine the relationship between various ownership structures based agency cost proxies on dividend policy.

To identify which agency cost proxy is dominant in influencing dividend policy over the company.

To identify which dividend model is superior in explaining the corporate dividend policy with variables associated with ownership structures.

1.4

SIGNIFICANCE OF THE STUDY

This study contributes to the growing body of survey research on dividend policy. For example, the current study not only updates previous research by Mat Nor and Sulong (2007) but is also applied in a different model, namely, the Full Adjustment Model, the Partial Adjustment Model and the Waud Model. These three types of dividend models had been modified to account for the possible effect of ownership structure and dividend policy. This study utilizes these three types of dividend models since it was found from previous research that dividend models can have the significant effect on ownership structure.

In addition, this study is expected to support the agency theory, especially in explaining the ownership structure policy to reduce agency conflict. Consequently,

this study would assist each ownership class to understand the explanation of the agency relationship. Shareholders with respect to stock investment in companies should be concerned with the agency conflict between ownership classes. Therefore, shareholders should justify that dividend policies are better control mechanisms for the agency conflict. Lastly, this study is also important in helping policy makers and companies to appropriately address the issues of agency costs.

1.5

LIMITATIONS OF THE STUDY

The main limitation of this study is that the data period covers only on the year 2007. The shorter period of study may not be representative of the way companies operate their business cycle. Thus, a longer period of study might be good to provide better results for this research.

The data for ownership structure was gathered from the list of the thirty largest shareholders disclosed in the company annual report. Consequently, the data may not be representative of the entire company.

The study only covers 150 public-listed companies in the selected sectors. Hence, the results cannot be treated as conclusive for all sectors. Besides that, since the study was limited to publicly-held companies, the results may not necessarily be applicable to privately-held companies.

10

1.6

CONCLUSION

Dividends distribution is one of the simplest ways for companies to communicate their financial well-being and shareholder value. Dividends send a clear, powerful message about future prospects and performance. Dividends are important for more than income generation since it also provides a way for investors to assess a company as an investment prospect.

This study tests the relationship of ownership structure and corporate dividend policy via three types of dividend models, namely, the Full Adjustment Model, the Partial Adjustment Model (Litner, 1956) and the Waud Model (1966). It examined the adoption of agency costs theory through ownership structure and dividend policy. Significant results could act as guidance for companies and policy makers to appropriately address the issues of agency costs.

The next section of the study briefly reviews the theoretical and empirical literature. Then, the third chapter describes the data, develops the theoretical model and also discusses the research framework. Chapter Four will reveal the empirical results while the summary and conclusion of the study are presented in Chapter Five.

11

CHAPTER 2

LITERATURE REVIEW

2.1

INTRODUCTION

This chapter discusses the theoretical and empirical literature on dividend policy. Section 2.2 reviews the theoretical literature of dividend policy beginning with the irrelevance proposition by Miller and Modigliani, followed by agency cost argument. Then, Section 2.3 discusses the empirical literature which is arranged in chronological order while Section 2.4 concludes the chapter.

2.2

THEORETICAL LITERATURE

Theoretical literature on corporate dividend policy centered around two classic works; the first is the Lintner Dividend Stability Model by Linter (1956) and second is that by M&M Dividend Irrelevant Theory by Miller and Modigliani (1961).

12

2.2.1 M&M Irrelevant Dividend Theory and Other Related Theories/Models

The seminal work on dividend policy was initiated in 1961 by Miller and Modigliani (M&M), proposed that dividend policy was irrelevant. Therefore, any changes made in dividend policy make no different to firm value since a stockholder can replicate any desired stream of payments by purchasing and selling equity. However, several assumptions were made, including: no personal or corporate taxes; no stock flotation or transaction costs; financial leverage has no effect on the cost of capital; investors and managers have asymmetry information about the firms future prospect; and distribution of income between dividends and retained earnings has no effect on the firms cost of equity (Foong, Zakaria and Tan, 2007). The main conclusion of this paper is that firms capital budgeting policy is independent of its dividend policy. M&Ms proposition was strongly supported by Friend and Phuket (1964) and Black and Scholes (1974).

Nevertheless, subsequent literature advances several theoretical justifications for firms payout choices. One branch of this literature has focused on an agency-related rationale for paying dividend policy. The agency models of payout relax the original M&Ms assumption about the independence of dividend and investment policies of the firm. According to Jensen and Meckling (1976), the origin of agency theory lies on the separation of ownership and control. The discrepancy between the value of the 100 percent owner-managed and less than 100 percent owner-managed firm is a measure of the agency cost. Jensen and Meckling defined agency relationship as a

13

contract under which one or more persons (principal) engage another person (agent) to perform some service on their behalf which involves delegating some decision making authority to the agent. If both parties to the relationships are utility maximizers, there is good reason to believe that the agent will not always act in the best
interests of the principal.

Meanwhile, Mahadwartha (2002) draws on Fama (1980) and Eisenhardt (1989) had augmented that agency theory concerned with resolving two problems that occur in agency relationship. The first is the agency problem that arises when the interest or goals of the principal and agent conflicted and it is difficult for the principle to verify what the agent is actually doing. The second is the problem of risk sharing that arises when the principle and agent have different attitudes towards risk. Different in risk preferences leads to different policy decisions and disregard the value maximizing activity as the economics pursued.

According to Mohd, Perry and Rimbey (1995), agency theory relates to dividend policy stems from the works of Rozeff (1982) and Easterbrook (1984). Rozeff adapt the agency theory argument of Jensen and Meckling by constructing a model in which dividends serve as a mechanism for reducing agency costs, thus offering a rationale for the distribution of cash resources to shareholders. According to Rozeff, if a firm is forced to raise external capital to replenish funds paid out in dividends, then managers must reduce agency costs and reveal new information in order to secure the new funding. Moreover, a dividend payment may act as one form of bonding mechanism

14

to lessen agency costs because it reduces the opportunity for managers to use firm cash flow for perquisites activities.

Besides that, Easterbrook (1984) develop an argument that outside shareholders are active in seeking to draw funds from the firm to force managers to subjects themselves to the scrutiny of capital markets. Easterbrook lists some of the mechanism by which dividends and capital raising exercises can control agency costs. Agency costs are less serious if the firm is constantly in the market for new capital since it continuously put the management under scrutiny by investment banks, security exchange and capital suppliers. Therefore, the payment of dividends causes the firm to undergo a third-party audit, which serves to motivate managers to both reveal new information and reduce agency costs in order to secure needed funds. Shareholders are willing to bear the costs of new funding to realize the greater benefits associated with the reduction in both agency costs and information asymmetries.

Additionally, Jensen (1986) free cash flow hypothesis asserts that funds remaining after financing all positive net present value projects have a tendency to have high agency costs. Thus, a commitment to pay out funds to shareholders as dividends might decrease the agency costs since it reduces the amount of free cash flows that managers could otherwise be wasted through over-investment and/or projects that provide personal benefits to managers.

15

While disbursing excess cash may reduce the agency problem between managers and shareholders, there may be alternative means of controlling this problem. Schooled and Barney (1994) draw on Jensen and Meckling (1976) argument that the agency problem is less severe when managers hold a large fraction of the outstanding shares in the company. If managers hold a small fraction, they work less vigorously or consume excessive perquisites because they bear a relatively small portion of the resulting costs. Therefore, agency theory argues that managerial ownership is a self monitoring mechanism and also bonding mechanism. Managerial stock ownership can reduce agency costs by aligning the interests of a firms management with its shareholders. Managerial ownership bonded management personal wealth to firm value (shareholders wealth) (Jensen and Meckling, 1976; Rozeff, 1982; and Easterbrook, 1984).

Furthermore, institutional stock ownership can also decrease agency costs by monitoring firm. According to Shleifer and Vishny (1986), ownership concentration creates the incentives for large shareholders to monitor the firms management, which overcomes the free-rider problem associated with dispersed ownership whereby small shareholders have not enough incentives to incur monitoring costs for the benefits of other shareholders. Due to active monitoring from shareholders, managers are better aligned towards the objective of delivering shareholder value. In addition, institutional investors also finding it increasingly difficult to sell large portions of stock without depressing stock prices. Therefore, many institutional investors choose to monitor the actions of firm managers more effectively to increase stock performance rather than

16

selling their holding at a loss. Consequently, institutional investors are actively working to effect corporate policy decisions.

2.2.2 Lintner Dividend Stability Theory and other Related Theories/Models

Lintner (1956) is among the pioneers to theorise on corporate dividend behavior through Lintner stability dividend theory. Lintner had conducted a classic series of interviews with 28 corporate managers about their dividend policy. He then proceeded to formulate a seemingly logical model of how companies decide on dividend payments. Dorsman, Montfort and Vink (1999) summarized Lintners survey in four stylised facts. First, firms have long-term target dividend payout ratios. Second, managers focus more on dividend changes than on absolute levels. Third, dividend changes follow shifts in long-term, sustainable earnings. This trend implies that managers tend to smooth dividends so that changes in transitory earnings are unlikely to affect dividend payments over the short term, and lastly, managers are reluctant to make changes to dividends that might have to be reversed. They are particularly concerned about having to rescind a dividend increase.

Based on these conclusions Linter developed a model, which has become known as the Lintner model, to explain the change in dividends each year. One assumption in this model is that managers will try to pay an amount of dividends that is an optimal percentage of the profit made. This is explains for the equation: D*ti = rEti (1)

17

with, D*ti r Eti = = = the target level dividend of dividend for fund i year t. the optimal amount of dividend as a percentage of the profit, for fund i. the profit company i made in year t.

The value of r will be between 0 and 1 since companies usually wont pay more dividends then that there was profit.

When the profit changes the actual amount of dividend paid differs from the optimal amount that follows out of (1). To compensate for this difference the company will gradually adjust the dividends. This is what can be seen in the next equation: D t i D ( t - 1 ) i = c(D* t i - D (t - 1 ) i ) with, c = Velocity at which a company adjusts the dividend (2)

The velocity (C) will be between 0 and 1. Higher values of C correspond to higher velocity in adjusting the dividends. Lintner also introduced a constant term. Because it is assumed that corporations are reluctant to decrease dividends, this constant term would have to be positive. This constant term together with equations (1) and (2) form the Lintner partial adjustment model: D t i D ( t - 1 ) i = a + i 1 D ( t - 1 ) i + i 2 E t i + t i (3)

18

with, i 1 i 2 ti = = = -Ci Ci ri The random disturbance

Besides the Lintner partial adjustment model, Waud (1966) proposes a second order rational distributed lag function for detailed derivation of the model. According this model, dividends are the results of a `the partial adjustment' and `the adaptive expectations'. It assumes that the target dividends are proportional to the long-run expected earnings (E*). Thus, the equation: D*ti = rE*ti (4)

On one hand, the actual dividend change will follow a partial adjustment model: D t i - D ( t - 1 ) i = a +c(D* t i D ( t - 1 ) i ) + t i (5)

As a result, the formation of expectations follows an adaptive expectation model: E* t i E ( t - 1 ) i = d(E t i E* ( t - 1 ) i ) (6)

19

2.3

EMPIRICAL LITERATURE

Miller and Modiglianis (1961) hypothesis on the irrelevance of dividend policy is not compatible with empirical evidence. This fact implies there must be additional factors that compel firms to pursue a consistent policy of paying dividend. Rozeff (1982) had initiated the adoption of agency cost in dividend determinant. He develop a model of optimal dividend payout in which increased dividends lower agency costs but raise the transaction costs. The optimal dividend payout minimizes the sum of these two costs. Rozeff use two independent variables as proxies for agency cost which are percent of stock held by insiders and the natural logarithm of the number of shareholders. Based on 1000 sample of companies from 1974 until 1980, he shows that dividend payout is negatively related to the percentage of stock held by insiders. Besides that, he also found that outside shareholders demand a higher dividend payout if they own a higher fraction of the common equity and if their ownership is more disperse.

Llyod, Jahera and Page (1985) try to confirm and expand the work of Rozeff in introducing agency theory as an explanatory factor in dividend payout ratios. The researchers had replicate Rozeffs study using more recent data. An OLSQ cross sectional regression is applied to 1984 data on 957 US firms, and the conclusions reached support and strengthen the results of Rozeff. They provide a strong support for their hypothesis of dividends as a partial solution to agency problems.

20

Jensen, Solberg and Zorn (1992) examine the determinants of cross-sectional differences in insider ownership and dividend policies in the U.S. They analyzed firm data at two points in time, 1982 and 1987on 565 and 632 firms respectively. These policies are found related directly and indirectly through their relationship with operating characteristics of firms. Their empirical results support the hypothesis that levels of insider ownership differ systematically across firms. The results of the analysis support the proposition that financial decisions and insider ownership are interdependent. Specifically, insider ownership has a negative influence on firms dividend levels. Therefore, this observation supports Rozeffs proposition that the benefits of dividends in reducing agency costs are smaller for firms with higher insider ownership.

Alli et.al (1993) re-examine the dividend policy issues by conducting a simultaneous test of the alternative explanations of corporate payout policy using a two-step procedure that involves factor analysis and multiple regression. The sample of 150 firms came from 34 industries, with the largest share from the chemical and allied products industry (13.9 percent). The average firm size and capitalization of the final sample was representative of New York Stock Exchange (NYSE) listed firms. The results reveal that six significant factors can be used to explain corporate payout policies which include agency cost factor. Although the results shows that ownership dispersion does not affect dividend but the significant positive coefficient of institutional and insider ownership indicates that dividends are used to mitigate agency problem which is consistent with the findings of Rozeff (1982).

21

Agrawal and Jayaraman (1994) use the sample of all-equity and levered firms which consist of 71 matched pairs. All equity firms are defined as those that use no long term debt throughout a continuous five year period. Their results indicate that dividend yields and payout ratios of all-equity firms are significantly higher than those of levered firms. These results are robust to the choice of the time period used for measuring these variables. They also found that within the group of all-equity firms, firms with higher managerial holdings have lower dividend payout ratios because they are substitute mechanisms for controlling the agency costs of free cash flow. This relationship is more pronounced in all-equity firms since they lack one mechanism for controlling these agency costs. Therefore, their findings support the Jensens (1986) hypothesis that dividends can be viewed as a substitute mechanism in mitigating the agency costs of free cash flow.

Hansen et.al (1994) tests the relevance of monitoring theory for explaining the dividend policies of regulated electric utilities. They focus on this industry partly because relative to industrial firms, utilities are arguably somewhat more insulated from the discipline of other monitoring mechanism for controlling agency costs. Their tests are conducted in each of two recent five year periods, the first five year period ending in 1985, which is characterized by high but declining industry wide investment growth and financing and the more recent five year period ending in 1990, which is characterized by secular asset growth yet low industry-wide growth. Their findings show that utilities faced with higher regulatory and managerial conflict, lower flotation costs and lower asset growth pay proportionally greater dividends. Their

22

findings are consistent with the monitoring hypothesis that these utilities firms use dividend induced equity financing to control equity agency costs that arise out of the stockholder-regulator and stockholder-manger conflicts.

More support and further contribution to the agency theory of dividend debate, is provided by Mohd, Perry and Rimbey (1995). These authors introduce a number of modifications to the cost minimization model including industry dummies, institutional holdings and a lagged dependent variable to the RHS of the equation to address possible dynamics. The results of a Weighted Least Squares regression, employing panel data on 341 US firms over 18 years from 1972 to 1989 support the view that the dividend process is of a dynamic nature. Higher dividend payouts are observed when managers hold a low percentage of firm shares, and as the outside ownership becomes more dispersed. This adds support for both Rozeffs and Easterbrooks hypotheses that stockholders seek greater dividend payout as they perceive their level of control to diminish.

The first study that examines the determination of financial policy variables in light of agency concerns in the banking industry is by Mendez and Willey (1995). Their study examines agency theory arguments in the banking industry by analyzing the effect of four variables that proxy for agency costs namely earnings volatility, managers portfolio diversification losses, bank size and standard deviation of bank equity returns on the three financial policy variables of managerial stock ownership, leverage and dividend yield. The study examines the largest 104 US banks during the period

23

1985-1989. Evidence support the view that bank managers consider agency costs while trying to determine the most appropriate financial policies (managerial stock ownership, dividends and leverage).

Noronha, Shome and Morgan (1996) develop an agency-cost framework for the simultaneous determination of a firms capital structure and dividend decisions. In the model, simultaneity is contingent on the applicability of Easterbrooks (1984) monitoring rationale for paying dividends, which, in turn is hypothesized to depend on the existence of alternative sources of monitoring. Estimations of the Rozeff (1982) specification for dividend payout for subsamples stratified according to the prevalence of non-dividend monitoring mechanisms and growth-induced capital market monitoring, confirm the sample-specific validity of the monitoring rationale. A simultaneous system of equations is then estimated and the results reveal that only for the subsample with lower availability of alternative mechanisms the payout rate is related to agency variables. For the subsample with alternative mechanisms in place the payout rates of firms are not related to proxies for agency cost variables.

D Sauza and Saxene (1999) examine the effects of agency costs on an international firms dividend policy. He used sample of 349 firms worldwide to determine the relationship between dividend payout and agency cost. The dividend policy of a firm is defined as its dividend payout ratio (the ratio of dividends per share and earnings per share) while the percentage of institutional holdings of a firms common stock is used as a proxy for controlling agency costs. The dividend payout variable used in the

24

study is a three year average for the period 1995 to 1997, while the institutional holdings pertain to the year 1997. Multiple regression analysis was performed and the result reveals the statistically significant and negative relationship of dividend payout with the explanatory variable institutional holdings. Therefore, these findings are consistent with those of prior studies using United States data.

Han, Lee and Suk (1999) also empirically examine the effect of institutional on corporate dividend policy. They utilize a sample of 303 firms during the 1988 to 1992 period. They had control seven factors believed to influence dividend policy namely insider ownership, revenue growth, capital expenditures on plant and equipment, ratio of debt to assets, standard deviation of return on assets, operating income to assets and target dividend yield. Nevertheless, using the Tobit analysis, they found a contradict results with agency cost hypothesis but supporting tax based hypothesis. According to tax based hypotheses, dividend payout is positively related to institutional ownership because institutions prefer dividends prefer dividends over capital gains under the differential tax treatment.

Ang, Cole and Lin (2000) measure absolute agency costs by observing a zero agencycost base case as a reference point of comparison for all other cases of ownership and management structures. Based on the Jensen and Meckling agency theory, the zero agency cost base is the firm owned solely by a single owner-manager. When management owns less than 100 percent of the firms equity, shareholders incur agency costs resulting from managements shirking and perquisite consumption. They

25

employ a sample of 1708 small corporations and provide a direct confirmation of the predictions made by Jensen and Meckling (1976). Agency costs are indeed higher among firms that are not 100 percent owned by their managers, and these costs increase as the equity share of the owner-manager declines. Hence, agency costs increase with a reduction in managerial ownership, as predicted by Jensen and Meckling.

Manos (2002) had investigated the agency theory of dividend policy in the context of an emerging economy, India. He had modified the Rozeffs cost minimization model by introducing a business group affiliation namely foreign ownership, institutional ownership, insider ownership and ownership dispersion as a proxy for agency cost theory. The model is estimated and tested on a cross-section of 661 non-financial companies listed on the Bombay Stock Exchange. The results reveal a positive impact of all business group affiliation to payout decisions. The positive relationship

between foreign and payout indicates that the greater the percentage held by foreign institutions, the greater the need to induce capital market monitoring. Besides that, capital market monitoring is also important when the dispersion of ownership increases since the more widely the ownership spread, the more acute the free rider problem, hence, the greater need for outside monitoring. Further, the evidence of a positive relationship between institutional and the payout ratio is consistent with the preference for dividends related prediction.

26

Study by Short, Zhang and Keasey (2002) is the first example of using wellestablished dividend payout models to examine the potential association between ownership structures and dividend policy. They had modified the Full Adjustment Model, the Partial Adjustment Model (Lintner, 1956), the Waud Model (Waud, 1966) and the Earnings Trend Model. Moreover, the paper presents the first results for the UK, where the institutional framework and ownership structures are different from the US. This study is conducted on a sample of 211 firms listed on the London Stock Exchange Official List for the period 1988 to 1992. The result from the four dividends models consistently shows positive and statistically significant associations between institutional ownership and dividend payout ratios and thus suggests a link between institutional ownership and dividend policy.

The study by Khan (2006) investigates how the ownership structure of firms affects their dividends policies. His sample period covers the period of 1985-1997 and the sample size reaches a maximum of 281 firms in 1989 and a minimum of 126 firms in 1985. A key contribution of this article is that it exploits extremely rich ownership data on all beneficial owners (individuals, insurance companies, pension funds and other financial institutions) holding more than 0.25% of any given firms equity. A significantly negative relation between dividends and ownership concentration result appear to corroborate Rozeffs model, dividends fall when the degree of ownership of ownership concentration increase, which is generally associated with better incentives to monitor. However, the positive relationship between dividends and insurance companies would suggest that they are relatively poor at monitoring compared to

27

individual investors. These results imply particularly acute agency problems when insurance company shareholdings is high and provide some support for the views expressed in the various governance reports.

Harada and Nguyen (2006) analyze the effect of ownership concentration on the dividend policy of Japanese firms from April 1995 to March 2002. Consistent with Khan (2006), they find that firms with high ownership concentration pay lower dividends. Their analysis uncovers a number of agency conflicts. First, tightly controlled firms are less likely to increase dividends when profitability increases and when operating profits are negative. This pattern is consistent with their lower payout and the assumption that dominant shareholder extract private benefits from resources under their control. Second, they also find that tightly controlled firms are more likely to omit dividends when investment opportunities improve which protect the interest of current shareholders. Clearly, this decision reduces the likelihood of requiring further funding that would benefit outside investors.

Mancinelli and Ozkan (2006) reports on empirical investigations into the relationship between the ownership structure of firms and the firms dividend policy using a sample of 139 listed Italian companies. Ownership structure in Italy is highly concentrated; hence the relevant agency problem of concern seems to be the one that arises from the conflicting interests of large shareholders and minority shareholders. The Tobit regression results support the prediction that higher level of ownership concentration is associated with a higher probability of expropriation of outside

28

shareholders. There are private benefits to the larger shareholders of holding larger amounts of cash; lower dividend payouts will increase the ability of the large shareholders to expropriate the outside minority shareholders. Furthermore, their findings also provide some support for the prediction that managers prefer to hold resources under their control rather than distributing returns to shareholders.

Cook and Jeon (2006) investigate the determinants of foreign and domestic ownership and a firms payout policy. Their empirical study based on a sample of 507 firms out of the 683 firms listed on Korea Exchange (KRX) for the period 1999 to 2004. The results support the agency model, higher foreign ownership is associated with a greater dividend payout. Domestic intuitional investors, however, do not play a prominent role in a firms payout policy. Thus, they conclude that foreign investors are more active monitors of corporate by reducing agency problems and leading firms to increase the level of payouts.

The study by Mollah, Rafiq and Sharp (2007) investigates the influence of agency cost variables on dividend policy during the pre and post of the 1998 financial crisis. Using cross-sectional and pooled regression, the paper measures the effect of the percentage of insider ownership, dispersion of stockholders, free cash flow and degree of collateralizable assets on the dividend payout ratio. The pre-crisis sample includes 153 companies for ten years from 1988 through 1997 while the post-crisis sample includes 153 companies for five years from 1999 through 2003. The crisis year of 1998 is omitted. The study finds agency cost variables to have only a modest

29

explanatory power during the pre-crisis period and none in the post-crisis period on the Dhaka Stock Exchange. This result might be due Bangladesh firms having highly concentrated ownership structure, thus an agency cost is insignificant in influencing the dividend policy. The failure of agency cost variables to influence dividends may indicate an impediment to efficient capital information. This failure captures an aspect of an emerging market such as Dhaka that differs fundamentally from more evolved markets.

Mat Nor and Sulong (2007) investigate the relationship between types of ownership structure and dividends on the main board of Bursa Malaysia for the years 2002 and 2005. This data from a sample of 406 firms employs a multiple regression analysis since the data are cross sectional. The results reveal that concentration ownership has a significant positive effect on dividends for both years, but with minimum impact. Results of foreign and managerial ownership on dividends show insignificant relationship in the year 2002, but the results are significant effect on dividends in 2005. The significant positive relationship of managerial ownership with dividends implies that insider shareholdings provide greater incentives for the alignment of management and shareholders interest resulting in higher dividends. The results also suggest that managerial ownership does play an active monitoring role in Malaysia, one of the emerging economies to mitigate potential managerial discretionary behavior and free cash flow problems. Nevertheless, the negative significant effect of foreign ownership on dividends fails to support the agency argument.

30

Obema, El-Masry and Elsegini (2008), examine the effect of ownership structure on corporate dividend policies of a sample of top Egyptian listed companies. Ownership structure is measured by four variables namely managerial ownership ratio, blockholder ownership ratio, institutional ownership ratio and free float ratio. The results show that only institutional ownership has a significant relationship with dividend policy. One explanation could be that the institutional blockholders voted for higher payout ratios to enhance managerial monitoring by external capital markets.

The study by Kouki and Guizani (2009) analyze the influence of shareholder ownership identity on dividend policy for a panel of Tunisian firms from 1995 to 2001. This study uses dividend per share as a dependent variable and ownership classes as an independent variables. The results indicate that there is a significantly negative correlation between institutional ownership with the level of dividend distributed to shareholders. This is due to most of cases, institutional investors are banks, and they are either shareholders or debt holders. They prefer paying interests to themselves than distribute dividend to all shareholders. Further, the results also show that the higher ownership of the five largest shareholders leads to the higher of dividend payment. They conclude that dividend rates are higher in Europe when there are multiple large shareholders suggesting that these large shareholders dampen expropriation in Europe. This evidence in Tunisian context strengthens the argument of the positive role of multiple large shareholders in corporate control.

31

Harjito (2009) examine the influences of agency factors to dividend payout ratio. This research tries to define an appropriate mechanism to decreasing agency cost which represent by dividend payout ratios policy. This study takes data from companies listed in JSX from year 2001 to 2005. The results reveal a significant negative effect of insider ownership on dividend policy. This implies that dividend payment is rise in order to decrease agency problem when there is separation function between corporate ownership and corporate control. Nevertheless, institutional ownership influence dividend payout negatively which is contradict with the agency argument. This might be due to institutional ownership tend to do other investment or expand their business that to pay shareholders. This condition is supported by the better economic atmosphere of Indonesia, which offers good opportunities to invest.

32

Table 2.1: Summary of Empirical Literatures MEASUREMENT OF VARIABLE DEPENDENT Michael S. Size: 1000 Rozeff (1982) Multiple Regression DPR: 7 year average over a period of 1974 to 1980 INDEPENDENT INS: percentage of common stock held by insiders. GROW 1: 5 year average of growth rate of revenues. GROW 2: 5 year average of Value Lines forecast of growth rate of revenues. BETA: 5 year value-weighted beta of weekly data on the stocks returns. STOCK: natural logarithm of number of common stockholders INS: percentage of common stock held by insiders. GROW 1: 5 year average of growth rate of revenues. CONTROL INS: negative, significant GROW 1: negative, significant GROW 2: negative, significant BETA: negative, significant STOCK: positive, significant

AUTHOR

SAMPLE

METHOD

FINDINGS

William P.Lloyd, John S. Jahear, Jr

Location: U.S Period:

Ordinary Least Squares Regression

DPR: average payout ratio last 7 years from value line

INS: negative, significant GROW 1: negative, significant

and Daniel July E. Page September (1985) 1984 Size: 957

GROW 2: 5 year average of Value Lines forecast of growth rate of revenues. BETA: 5 year value-weighted beta of weekly data on the stocks returns. STOCK: natural logarithm of number of common stockholders. SIZE: natural logarithm of sales for 1983. Three-Stage Least Square (3SLS) DIVIDEND: ratio of dividends to operating income INSIDER: percentage of shares held by insiders DEBT: ratio of long term debt to the book value of total assets. BUSINESS RISK: standard deviation of the first difference in operating income divided by total assets. PROFITABILITY: ratio of operating to total assets. GROWTH: 5 year growth rate in sales.

GROW 2: negative, significant BETA: negative, significant STOCK: positive, significant SIZE: positive, significant

Gerald R. Jensen, Donald P. Solberg and Thomas S. Zorn (1992)

Period: 1982 and 1987 Size: 565 (1982), 632 (1987)

INSIDER: negative, significant DEBT: negative, but significant only in 1987 BUSINESS RISK: negative, but significant only in 1987 PROFITABILITY: positive, significant GROWTH: negative, significant INVESTMENT:

INVESTMENT: expenditure for plant, equipment, and R&D as a percentage of total assets. Kasim L. Alli, A. Qayyum Khan and Gabriel G. Ramirez (1993) Location: U.S Period: 1983 1987 Size: 105 Factor analysis and multiple regression DPR : 3 year arithmetic average over a period of 1983 to 1985 LNTA: natural log of total assets (size) of the firm. BETA: firms beta calculated against the CRSP equally weighted index using sixty months of data. STDCDE: standard deviation of changes in the debt equity ratio using nine years data. EXCAP: average realized value of capital expenditure for 1986, 1987 and 1988 was used as a proxy for expected capital expenditure in 1985. INSTHOL: proportion of institutional holdings as a proportion of total outstanding shares. INSIDER: proportion of insider holdings as a percent of total outstanding shares.

negative, significant

LNTA: insignificant BETA: negative, significant STDCDE: positive, significant EXCAP: negative, significant INSTHOL: positive, significant INSIDER: positive, significant HOLDING: insignificant INTANG: negative, significant GROWTH: negative, significant CFV: insignificant SLACK: negative, significant STAB: positive,

HOLDING: dispersion of ownership as given by the number of shareholders to total outstanding shares. INTANG: collaterizable value of assets as represented by net plant to total assets. GROWTH: 5 year annualized growth rate in earnings. CFV: coefficient of variation of cash flow using nine years data. SLACK: sum of cash and marketable securities scaled by market value of equity and unused debt capacity (different between the industry and firm leverage ratios). STAB: dummy coded variable that takes a value of 1 if dividend payout in 1985 is 90 percent or more of the past five years dividend and 0 otherwise.

significant

Anup Agrawal and Narayan Jayaraman (1994)

Period: 1979 1983 Size: 71

Ordinary Least Squares Regression

DPR : ratio of dividends per share to earnings per share DIVYLD: cash dividends divided by stock price

LEVERED: dummy variable, levered firm = 1, all-equity firm = 0. : percentage of outstanding equity owned by directors and officers. FCF: free cash flow divided by total assets. GROWTH: average annual growth rate of sales over the previous 5 years. COMMRANK: regulatory commission rank. OWNSHIP: stockholderownership concentration, measured by Herfindahl Index. FLOTCOST: firms historical average flotation cost incurred in selling common stock, expressed as a percentage of the gross proceeds. TAGROW: growth rate in total assets.

LEVERED: negative, significant : negative, significant FCF: insignificant GROWTH: insignificant

Robert S.Hansen, Raman Kumar and Dilip K.Shome (1994)

Location: U.S Period: 1981 1985 and 1986 1990 Size: 81 (1985) and 70 (1990)

Ordinary Least Squares Regression

DPR: sum of all dividends paid during 5 year prior to and including the ending year, over the sum of all stockholder earnings over the same period

COMMRANK: negative, significant OWNSHIP: negative, significant FLOTCOST: negative, significant TAGROW: negative, significant

Mahmoud A. Mohd, Larry G. Perry and James N. Rimbey (1995)

Time-series cross sectional analysis

DPR: common dividend payment divided by net income

GROW 1: average rate of revenue growth over previous five years. GROW 2: Value line forecast of future five-year revenue growth. 0: intrinsic business risk OLRISK: operating leverage risk. FLRISK: financial leverage risk. INSTINV: percent of common stock held by institutions. INSD: percent of common stock held by insiders. STKHLDR: natural log of the number of shareholders EARNING VOLATILITY: standard deviation of the income before depreciation and amortization on total assets. MANAGERS

SIZE: natural log of firm sales.

GROW 1: negative, significant GROW 2: negative, significant 0: negative, significant OLRISK: negative, significant FLRISK: negative, significant INSTINV: positive, significant INSD: negative, significant STKHLDR: positive, significant SIZE: positive, significant EARNING VOLATILITY: positive, insignificant MANAGERS DIVERSIFICATION

Jose MercadoMendez and Thomas

Location: U.S Period: 1985 -

Regression

Dividend: 4 year average of the ratio of total dividends divided by the

Willey (1995)

1989 Size: 104

total market value of the common stock

DIVERSIFICATION LOSSES: equity risk premium divided by total equity risk. SIZE: 5 year average total assets. FLOTATION COST: standard deviation of he average stock return. INS: percentage insider holding. LNSH: logarithm of the number shareholders. VRET: variance of stock returns. LNAST: logarithm of asset size. GR: ratio of forecast growth in book value equity to forecast return on equity.

LOSSES: negative, insignificant SIZE: positive, significant FLOTATION COST: negative, insignificant

Gregory M. Noronha and George E. Morgan (1996)

Location: U.S Period: 1986 1988 Size: 400

Ordinary Least Squares Regression

DPR: ratio of the 5 year arithmetic average of a firms dollar dividend divided by the 5 year average of income available to common stockholders. DPR

INS: insignificant LNSH: insignificant VRET: insignificant LNAST: insignificant GR: insignificant

Juliet DSauza and Atul

Period: 1995 1997

Multiple Regression

BETA: beta GROWTH: past 3 years sales growth.

BETA: negative, significant GROWTH:

K Saxena (1999)

Size: 349

INSH: percentage of institutional. MTBV: market to book value.

insignificant INSH: negative, significant MTBV: insignificant ISD: insider ownership in percentage for firm GRO: firm geometric average of revenue during the 5 year period CXA: firm capital expenditures on plant and equipment as a percentage of assets DTA: ratio of debt to assets INT: positive, significant ISD: insignificant GRO: negative, significant CXA: insignificant DTA: insignificant SDR: negative, significant OIA: insignificant TDY: negative, significant

Ki C. Han, Suk Hun Lee and David Y.Suk (1999)

Location: US Period: 1988 1992 Size: 303

Tobit Analysis

DY: Dividend Yield

INT: institutional ownership in percentage for firm

SDR: standard deviation of return on assets during the five period OIA: operating income to assets for firm TDY: target dividend yield Ronny Manos (2002) Location: India Period: 1997 2001 Size: 661 OLSQ regression Homosceda stic Tobit regression Heckmans two-step regression DPR: 1 year period of DPR GROW: 5 year annual growth rate in sales. RISK: standard deviation of daily stock return over the 365 days. LIQUID: percentage of days the companys stock traded. FOREIGN: percentage of equity shares held by foreigners in one year. INST: percentage of equity shares held by institutions in GROW: negative RISK: negative LIQUID: negative FOREIGN: positive INST: positive DIRS: positive PUBLIC: positive

one year. DIRS: percentage of equity shares held by directors of the company in one year. PUBLIC: percentage of equity shares held by the public in one year. Helen Short, Hao Zhang and Kevin Keasey (2002) Location: London Period: 1988 1992 Size: 211 The Full Adjustment Model, The Partial Adjustment Model, The Waud Model and Earnings Trend Model. Tobit Regression Dti D(t-1)i: changes in dividend INST: Institutional dummy variable MDUM: Managerial dummy variable INST: positive, significant MDUM: negative, significant

Kimie Harada and Pascal Nguyen (2006)

Location: Japan Period: April 1995

DPR: dividends to operating income DIVEQTY:

LHH: ownership concentration, measured by an approximation of the Herdindahl Index, calculated by summing the squared percentage of shares

SIZE: natural log of total assets ROA: annual operating

LHH: negative Q2H: negative SIZE: negative ROA: positive Q: positive

March 2002 Observatio n: 6397

total dividend payments to book value of equity

controlled by five major shareholders. Q2H: dummy variable indicating that ownership concentration is above the sample median.

profits scaled GROW: positive by total assets DEBT: negative KD: positive Q: market to book value of assets GROW: percentage change in total assets DEBT: long term debt plus short term debt over total assets KD: dummy for affiliation with a business group SIZE: natural logarithm of sales. LEVERAGE: VOTING_RIGHTS1: negative, significant 2_LARGE_SHAREHO LDER: insignificant

Luciana Location: Mancinelli Itali and Aydin Ozkan Period:

Tobit Regression

Dividend/ Earnings Ratio and Dividend/ Market

VOTING_RIGHTS1: voting rights of the largest shareholder. 2_LARGE_SHAREHOLDER: takes value 1 when the second

(2006)

2001 Observatio n: 139

Capitalization

largest shareholder owns a fraction larger than 5% of the firms voting rights and zero otherwise. OTHER_LARGE_SHAREHO LDERS: takes value 1 when all shareholders but the largest one exceeding the 2% disclosure threshold own a fraction larger than 5% and zero otherwise. VOTING_RIGHTS _ALL: voting rights of all shareholders who own more than 2% disclosure. AGREEMENTS: Foreign Ownership: ratio of total equity owned by foreign investors to the total equity. Institutional Ownership: ratio of total equity owned by domestic investors to the total equity.

ratio of the total debt to total assets. MKT-TOBOOK: Ratio of the market value of equity to the book value equity

OTHER_LARGE_SH AREHOLDERS: insignificant VOTING_RIGHTS _ALL: insignificant AGREEMENTS: insignificant SIZE: positive, significant LEVERAGE: negative, significant MKT-TO-BOOK: negative, significant

Douglas O. Cook and Jin Q. Jeon (2006)

Location: Korea Period: 1999 2004 Size: 507

Regression

DPR: total amount of cash dividends divided by the net income after taxes.

Foreign Ownership: positive, significant. Institutional Ownership: insignificant.

A Sabur Mollah, Rfiqul Bhuyan Rafiq and Peter A Sharp (2007)

Location: Banglades h Period: 1999 2003 Size: 153

Crosssectional and pooled Ordinary Least Square regression

DPR: dividends paid divided by operating income

INSIDE: proportion of held by insiders. DOWNER: natural log of number of common stockholder. FCF: (net profit after tax dividend + depreciation) / Total Assets. COLLAS: Ratio of net asset to fixed assets. HI5: Herfindahl Index 5 that is the squared sum of shares in the hands of largest 5 shareholders. GOWN: sum of all shares held by government-controlled companies in the list 30 largest shareholders. FOWN: sum of all shares held by foreign shareholders in the list 30 largest shareholders. MOWN: percentage of shares directly held by non independent executive directors LOGMCAP: logarithm function of market capitalisation. AGE: age of listing EPS: net income divided by number of share. TS: trade and

INSIDE: negative, significant DOWNER: positive, significant FCF: positive, significant COLLAS: positive, significant

Fauzias Mat Nor and Zunaidah Sulong (2007)

Location: Malaysia Period: 2002 and 2005 Size: 406

Moderated Multiple Regression

Dividend yield: dividend per share divided by the average closing market price per share.

HI5: insignificant GOWN: negative, insignificant in 2002 but significant in 2005 FOWN: negative, insignificant in 2002 but significant in 2005 MOWN: positive, insignificant in 2002 but significant in 2005 LOGMCAP: positive, significant AGE: negative,

in the company

services (dummy variable) IP: industrial products (dummy variable) CP: consumer products (dummy variable).

significant EPS: insignificant TS: insignificant IP: insignificant CP: positive, significant

Omneya Abdelsala m, Ahmed El-Masry and Sabri Elsegini (2008)

Location: Egypt Period: 2003 2005 Size: 50

Regression

DIVDECISIO N: dummy variable that set to one if company paid dividends. DIVRATIO: dividend yield

INDEPENDENCE: ratio of independent directors DUALROLE: dual role BOARDSIZE: board size MANOWN: ratio of directors ownership BLOCKOWN: ratio of block ownership INSTOWN: ratio of institutional ownership FREEFLOAT: percentage of shares held by outsiders

INDEPENDENCE: insignificant DUALROLE: insignificant BOARDSIZE: insignificant MANOWN: insignificant BLOCKOW N: insignificant INSTOWN: positive, significant

ROE: return on equity PE: price earnings ratio

FREEFLOAT: insignificant ROE: positive, significant PE:insignificant SIZE: log of total assets FCF: positive, significant GROWT: positive, significant LEV: negative, significant MAJ: positive, significant INST: negative, significant ETA: positive, significant SIZE: negative, significant INSDO: negative, significant OFO: negative,

Mondher Kouki and Moncef Guizani (2009)

Location: Tunisian Period: 1995 2001 Size: 203

Regression

DPS: total dividend divided by the number of outstanding equity.

FCF: cash flow per unit of asset. GROWT: ratio of market to book value LEV: long term debt deflated by the book value of equity MAJ: dummy variable for ownership concentration INST: percentage of equity owned by institutional ownership. ETA: percentage of equity owned by state.

D. Agus Harjito and

Location: Jakrta

Multiple Regression

DPR: percentage of profit paid in

INSDO: share percentage owned by director and commissioner.

Ambang Aries Yudanto (2009)

Period: 2001 2005

the form of dividend.

OFO: percentage of shares owned by other outside firms. SD: variance of percentage of share ownership data. FG: upcoming years income growth level : return volatility measurement of security toward market-risk

significant SD: positive, significant FG: negative, significant : insignificant

2.4

CONCLUSION

Miller and Modgliani (M&M) claim that under assumption of perfect capital market, dividends are irrelevant and they have no influence on the share price. Nevertheless, when capital markets are imperfect and when the assumptions made by M&M are relaxed, some researchers have argued that dividends do matter; hence firms should pursue an appropriate dividend policy. A difference set of explanatory variables has been hypothesized to distinguish the companies specific characteristic that influence on the dividend distribution.

33

CHAPTER THREE

RESEARCH METHOD

3.1

INTRODUCTION

This study looks at the impact of ownership structure on corporate dividend policy. This chapter describes the methodology of this study. These include the framework of research as presented in Section 3.2. The description of the data sample development is presented in Section 3.3. Additionally, models of dividend policy are presented in Section 3.4 while Section 3.5 provides definitions and measurements for each of the variables that are used in the study. Lastly, Section 3.6 concludes the chapter.

3.2

RESEARCH FRAMEWORK

Based on the review of literature, theoretical and empirical, the impact of ownership structure on corporate dividend policy can be examined through the relationship between selected ownership variables and dividend policy. The ownership variables identified from the literature are ownership concentration, ownership dispersion, institutional ownership, managerial ownership and foreign ownership and the expected relations as described below:

50

3.2.1 Ownership Concentration (CONC) In concentrated ownership companies, large shareholders could find less need for using dividends as a disciplining mechanism if they have strong board representation (Renneboog and Szilagyi, 2006). On the other hand, according to La Porta et al. (2000a) larger controlling shareholders could expropriate corporate wealth from other minority shareholders and enjoy private benefits instead of distributing dividends to shareholders. Therefore, to circumvent the problem a positive relationship was expected between ownership

concentration and dividends.

3.2.2 Ownership Dispersion (DISP) The greater the number of shareholders will lead to the greater dispersion of ownership. Hence, agency costs will increase and the need for monitoring managerial action also increases. If dividends can alleviate this problem, a positive relationship between ownership dispersion and dividend is expected.

3.2.3 Institutional Ownership (INST) Large institutional investors are more willing and able to monitor corporate management than are smaller and more diffuse owners since the benefits of monitoring are more likely to exceed the costs for these shareholders. Thus, a positive relationship was anticipated between institutional ownership and dividends.

51

3.2.4 Managerial Ownership (MNG) According to agency theory, managerial ownership has a potential to align the interest between managers and shareholders (Jensen and Meckling, 1976). However, if a larger percentage of common shares are in the hands of managers, there will be less influence from outsiders. In such case, management will tend to increase their own benefits such as increase directors fees, employees salaries and bonuses, rather than pay dividends. Besides, since the purpose of managerial ownership is the same as dividend policy, which is to reduce agency costs, it will be ineffective to use two tools at the same time for the same problem. Hence, dividends will be hypothesized to be negatively related with managerial ownership.

3.2.5 Foreign Ownership (FOR) According to agency theory, foreign investors who are well-informed and hold a substantial share can play their monitoring role on management and reducing the agency costs, and therefore, companies are more likely to increase dividends (Easterbrook, 1984; Jensen, 1986). Thus, a positive relationship was therefore expected between foreign ownership and dividends.

VARIABLE Ownership Concentration Ownership Dispersion Institutional Ownership Managerial Ownership Foreign Ownership

EXPECTED RELATION Positive Positive Positive Negative Positive

52

3.3

SAMPLE DESCRIPTION AND DATA COLLECTION

The sample for the study includes 150 companies from four of the largest sectors (consumers, industrial, trading and services and properties) on the Main Board of Bursa Malaysia whose annual reports are available for the year 2007. These

companies are selected based on proportionate stratified random sampling. Therefore, these companies are expected to be a representative of the four largest sectors in Bursa Malaysia.

This study utilised dividends, earnings and different types of ownership structure data. The dividend and earnings variables were retrieved from DataStream financial database. In addition, data on ownership was hand-collected from sample companies annual reports. These annual reports are gathered from the website of Bursa Malaysia and individual companies.

This pooled cross-sectional study employs annual data from 2005 to 2007. According to Mat Nor and Sulong (2009), annual data allows the study to capture more discretionary rather than autonomous behaviour. Annual data is also more systematic since it represents the highest periodicity.

53

3.4

MODELS ON DIVIDEND POLICY

Following the methodology of Short, Zhang and Keasey (2002), three dividend models were used to test the hypothesis of positive link between ownership structure and dividend policy: the Full Adjustment Model, the Partial Adjustment Model (Litner, 1956) and the Waud Model (1966). These models describe the adjustment of dividends to changes in several measures of corporate earnings. Nevertheless, these models have been modified to account for the possible effects of ownership structure in determining the level of the corporate dividend.

3.4.1 The Full Adjustment Model (FAM) According to the full adjustment model, changes in earnings are considered as permanent. Therefore, companies will adjust their dividends (D) to the new level of earnings (E) to achieve the companies desired payout ratio (r). Consequently, the relationship between the changes in earnings and changes in dividends, for company i at time t, is given by:

D t i D ( t - 1 ) i = + r(E t i E ( t - 1 ) ) + t i
The hypothesis that ownership structures affect dividend policy means that companies target payout ratio (r) for different levels of ownership classes. Therefore, in this case, the model becomes:

54

D t i D (t - 1 ) i = + r(E t i E (t - 1 ) ) + r C O N C (E t i E (t - 1 ) )*CON C + r D I S P (E t i E (t - 1 ) )*DIS P + + r I NS T (E t i E (t +

1 ) )*INST

r M N G (E t i E (t - 1 ) )*MNG

r FO R (E t i E (t - 1 ) )* FOR
(Model 1 , FAM)

3.4.2 The Partial Adjustment Model (PAM) The partial adjustment model assumes that the desired level of dividends (D*) for company i at time t is related to its earnings (E), according to the target payout ratio (r):

D * t i = rE t i
Nevertheless, the company adjusts only partially to the target dividend level. In contrast, firms move towards the desired level of distribution gradually and dividends adjust only partially to the changes in earnings. As a result, the model takes the form:

D ti D ( t - 1) i = a + c(D* t i - D (t - 1)i ) + t i
Where a is a coefficient representing the refusal of managers to reduce dividends, whereas c is the speed of an adjustment coefficient that represents the extent to which the management wishes to play-safe by not amending to the new target immediately.

55

Assuming that companies with significant ownership classes have different target payout ratios (r), the model becomes:

Dti D(t-1)i = + crEti + crCONCEti*CONC + crDISPEti*DISP + crINSTEti*INST + crMNGEti*MNG + crFOREti*FOR cD(t-1)i + ti


(Model 2, PAM)

3.4.3 The Waud Model (WM) The Waud model integrates elements of the both partial and full adjustment model. It believes that the target dividends D* are the proportional to the longrun expected earnings, E*:

D * t i = rE * t i

On one hand, the actual dividend change will follow a partial adjustment model:

D ti - D ( t- 1 ) i = a +c(D * t i D ( t - 1) i ) + t i

The formation of expectations follows an adaptive expectation model:

E* ti E (t - 1 )i = d(E t i E* (t - 1) i )

56

According to this model, dividends are the results of a `partial adjustment' and the `adaptive expectations'. Therefore, assuming a possible difference in payout ratio for firms with different ownership classes, the model becomes:

D t i D (t - 1 ) i =

ad + cdrE t i + cdr C O N C E t i *CONC + cdr D I S P E t i *DIS P + cdr I N S T E t i *INS T + cdr M NG E t i *MNG + cd r FO R E t i *FO R + (1-d -c)D (t - 1 ) i (1 -d)(1-c)D (t - 2 ) i - t i
(Model 3, WM)

3.5

MEASUREMENT OF VARIABLE

This section outlines each variable employed in this study in order to examine empirically the dividend models discussed in Section 3.3.

3.5.1 Dividends (D) The dividends variable was calculated as the total amount of distributed dividend divided by the number of ordinary outstanding equity shares relating to the accounting year.

57

3.5.2 Earnings (E) Earnings variable was calculated as net profit derived from normal trading activities after depreciation and other operating provisions divided by the number of ordinary outstanding shares.

3.5.3 Ownership Concentration (CONC) Following Hansen et al. (1994), Harada and Nguyen (2006) and Khan (2006), ownership concentration was measured by the Herfindahl Index 5 (HI5), that is, the squared sum of shares in the hands of the five largest shareholders (referred as CONC1). In addition, considering that Malaysian companies have highly concentrated ownership (Nor and Sulong, 2009), a new proxy was used, which is the summation of the percentage of shares controlled by two major shareholders (referred as CONC2). However, since both CONC1 and CONC2 are used to measure a similar variable, thus, CONC1 and CONC2 will be used one at a time.

3.5.4 Ownership Dispersion (DISP) Following Alli et al. (1993), ownership dispersion is defined as the ratio of the number of shareholders to total outstanding shares.

3.5.5 Institutional Ownership (INST) Alli et al.(1993) and Mohd et al. (1995), Amidu (2006) and Kouki and Guizani (2009) defined institutional ownership as a percentage of equity

58

owned by institutional investors such as insurance companies, unit trusts, mutual funds, pension funds and financial companies. Nevertheless, this study used the total percentage of institutional ownership in a list of the thirty largest shareholders as the measure of INST.

3.5.6 Managerial Ownership (MNG) Following Mat Nor and Sulong (2007), managerial ownership was measured by adding the total percentage of shares directly held by non-independent executive directors in the company.

3.5.7 Foreign Ownership (FOR) Following Mat Nor and Sulong (2007), the sum of all shares in the hands of foreign shareholders in the list of thirty largest shareholders, either held through nominee companies or other corporate foreign share holdings, was identified to calculate the total percentage of foreign shareholdings (FOR).

3.6

CONCLUSION

This study utilized income statements and ownership structure data for 150 companies listed in the four largest sectors (consumers, industrial, trading and services and properties) on the Main Board of Bursa Malaysia. Three types of dividend models were used to see whether specific ownership classes would influence the dividends

59

distribution of the company. The period of observation for this study was only in 2007.

60

CHAPTER FOUR

ANALYSIS AND FINDINGS

4.1

INTRODUCTION

This chapter discusses the findings of this study on ownership structure effect on the corporate dividend policy based on three dividend payout models, namely, the full adjustment model, the partial adjustment model and the Waud model. The analysis commenced with a data and variable description as presented in Section 4.2. Subsequently, the correlation analysis is illustrated in Section 4.3 to reveal the strength of the relationship between the variables that are utilized in the study. This is followed by Section 4.4 which discusses the result of the regression analysis, which constitutes the main findings of the study. Lastly, this chapter ends with the conclusion in Section 4.5.

4.2

DESCRIPTIVE ANALYSIS

Table 4.1 presents a summary of the descriptive statistics for each of the hypothesised variables for the 150 companies covered in this study. Focusing on the dependent variable, it can been seen that the standard deviation for dividends is 7.42 percent

61

which can be considered as high, thus, it indicates a substantial variation in the amount of dividend distribution in Malaysia. This is due to some companies not disbursing any dividend while some companies distribute their dividend as high as RM 0.381 per share. The average dividend distributed among the companies in the sample is RM 0.059 per share.

The earnings per share show an average of RM 0.159, with a minimum value of -RM 1.073 and a maximum value of RM 1.210.

In terms of ownership variables, the range of firm ownership concentration represented by the percentage of ownership owned by five largest shareholders (CONC1) is from 7.87 percent to 73.23 percent, resulted the standard deviation of 34.81 percent. The mean percentage of the CONC1 is 38.50 percent which implies that almost 40 percent of shares ownership is concentrated in hands of five largest shareholders among Malaysian firms. However, in the study by Mat Nor and Sulong (2009) found the mean percentage of ownership concentration is about 57 percent.

Nevertheless, for the second ownership concentration variable (CONC2), which measured by the summation of the percentage of shares controlled by two major shareholders, the mean is 41.85 percent, and the range is from 8.48 percent to 78.11 percent. Thus, the substantial mean value of CONC2 implies that Malaysian firms are highly concentrated.

62

In addition, the mean for ownership dispersion of zero percent and ranging from 0 percent to 0.01 percent, is another indication of highly concentrated feature of Malaysian firms.

Table 4.1: Summary Descriptive Statistic SKEWNESS MAXIMUM

Dividend

5.8688

0.0000

38.1000

7.4203

2.281

5.510

Earnings

15.8671

-107.300

121.0400

23.6610

0.245

8.402

CONC1

0.1482

0.0062

0.5362

0.1212

1.216

0.878

CONC2

0.4185

0.0848

0.7811

0.1716

0.213

-0.717

DISP

0.0000

0.0000

0.0001

0.0000

1.820

4.172

INST

0.2417

0.0016

0.9593

0.2024

1.369

1.821

MNG

0.1114

0.0000

0.5549

0.1558

1.496

1.021

FOR

0.0822

0.0000

0.5981

0.1188

2.455

6.541

63

KURTOSIS

MINIMUM

STD. DEV

MEAN

For institutional ownership (INST), the mean percentage is about 24 percent which implies that more than 20 percent of share ownership is in the hands of institutional shareholders such as insurance companies, unit trusts, mutual funds, pension funds and financial companies. The range is from 0.16 percent to 95.93 percent and showed a 20.24 percent standard deviation.

Further, managerial ownership (MNG) has a mean percentage of 11.14, which ranges from a low of zero percent to a 55.49 percent. Thus, a standard deviation of 15.58 percent had been recorded.

The foreign ownership (FOR) has an average value of 8.22 percent, an increase from the 6.12 percent found by Mat Nor and Sulong (2009) in 2005. The increment in foreign shareholdings could be partly due to new government initiatives to build up investors confidence after being badly affected by the Asian Financial Crisis in 1997.

4.3

CORRELATION ANALYSIS

Pearson correlation coefficients for the primary variables are provided in Table 4.2. Correlation is a single number that describes the degree of relationship between two variables. The coefficient has a range of possible value from -1.00 to +1.00. The value indicates the strength of the relationship, while the sign (+ or ) indicates the

64

direction. In this study, eight interval-level variables are studied and the relationships among all of variables are estimated.

There is a positive significant correlation (corr = 0.500, p-value = 0.000) between dividend and earnings. The positive correlations are consistent with the signaling theory, which argues that an increment in dividends will lead to earnings increasing. Besides that, dividends are also significantly positively correlated with CONC1 (corr = 0.424, p-value = 0.000), CONC2 (corr = 0.374, p-value = 0.000) and INST (corr = 0.191, p-value = 0.019) indicating the possibility of these three variables having predictive power on dividends and the positive relationship as theorized by the literature. Nevertheless, the negative, significant correlation between MNG (corr = 0.256, p-value = 0.001) and dividends contradicts the theoretical literature.

Among the independent variables, there is a positive correlation (CONC1 = 0.137, CONC2 = 0.095) between earnings and ownership concentration. This is probably because highly concentrated companies will lead to a good awareness of the company progress. Thus, the result reveals a negative and significant correlation (corr = -0.166, p-value = 0.021) between earnings and ownership dispersion, as expected. Besides that, earnings also have a positive significant correlation with institutional (corr = 0.253, p-value = 0.001) and foreign (corr = 0.173, p-value = 0.017) ownership since profitable companies are an attractive place for investors to invest. However, a negative correlation (-0.128) between managerial ownership and earnings was surprising.

65

Table 4.2: Pearson Correlation Matrix among the Variables D E 0.500** (0.000) E CONC1 CONC2 DISP INST MNG

CONC1

0.424** (0.000)

0.137* (0.047)

CONC2

0.374** (0.000)

0.095 (0.125)

0.933** (0.000)

DISP

-0.145 (0.077)

-0.166* (0.021)

-0.174* (0.016)

-0.192** 0.009

INST

0.191* (0.019)

0.253** (0.001)

-0.019 (0.411)

-0.008 (0.460)

-0.048 (0.281)

MNG

-0.256** (0.001)

-0.128 (0.060)

-0.339** (0.000)

-0.338** (0.000)

-0.088 0.143

-0.052 (0.264)

FOR

-0.020 (0.808)

0.173* (0.017)

-0.132 0.054

-0.110 (0.091)

-0.146** 0.037

0.056 (0.247)

-0.126 (0.062)

* Correlation is significant at the 0.05 level


** Correlation is significant at the 0.01 level

66

Both ownership concentration variables, CONC1 and CONC2, are significantly positively correlated (corr = 0.933, p-value = 0). The significant positive correlation between these two variables is expected since both are measurements of similar variables. Further, the results also reveal that both ownership concentration variables have a significant negative relationship with ownership dispersion. The degree of correlation between CONC1 and DISP is -0.174 (p-value = 0.016), and between CONC2 and DISP is -0.192 (p-value = 0.009). Besides that, both ownership concentration variables also recorded a significant negative correlation with managerial ownership, CONC1 (corr = -0.339, p-value = 0.000) and CONC2 (corr = 0.338, p-value = 0.000). The significant correlation among the independent variables indicates a need for particular attention to circumvent potential multicollinearity problems during the regression analysis.

4.4

REGRESSION ANALYSIS

4.4.1 Multicollinearity

The regression process commences with the identification of multicollinearity problems. Multicollinearity problems arise when one or more of the explanatory variables are exact or near exact linear combinations of other explanatory variables. Multicollinearity problems could be detected from the correlation matrix for the independent variables. If the variance inflation factor (VIF) value is larger than ten and the tolerance value is below 0.1, multicollinearity problem is said to exist among

67

the independent variables. Table 4.3 shows that there are some multicollinearity problems in the regression models. As a consequence, a multiple regression analysis was executed again, but some of the variables were dropped.

Table 4.3: Variance Inflation Factor of Variables (tolerance value is given in the parentheses) FA M 1 ERNCHG 28.907 (0.035) 5.049 (0.198) 18.418 (0.054) 3.651 (0.274) 5.863 (0.171) 3.976 (0.251) 3.499 (0.286) 3.440 (0.291) 6.017 (0.166) 6.162 (0.162) 3.221 (0.310) 13.192 (0.076) 4.475 (0.223) 23.104 (0.043) 13.333 (0.075) 4.520 (0.221) 23.715 (0.042) 2 46.340 (0.022) 1 PA M 2 1 WAUD 2

ECHGCONC1

ECHGCONC2

ECHGDISP

ECHGINST

ECHGMNG

ECHGFOR

ERN

ERNCONC1

68

ERNCONC2

12.738 (0.079) 1.965 (0.509) 4.049 (0.247) 2.287 (0.437) 2.477 (0.404) 1.561 (0.641) 1.946 (0.514) 4.078 (0.245) 3.319 (0.301) 2.342 (0.427) 1.487 (0.672) 1.967 (0.508) 4.050 (0.247) 2.338 (0.428) 2.509 (0.339) 1.918 (0.521) 1.403 (0.713)

13.075 (0.076) 1.948 (0.513) 4.078 (0.245) 3.446 (0.290) 2.383 (0.420) 1.829 (0.547) 1.426 (0.701)

ERNDISP

ERNGINST

ERNGMNG

ERNFOR

D(t-1)

D(t-2)

4.4.2 Serial Correlation and Heteroscedasticity Test

Subsequently, the models were tested for serial correlation and heteroscedasticity. Serial correlation occurs when a long series of observations are correlated with each other. This problem emerged when the residuals are not free from one observation to other observation.

69

Table 4.4: Serial Correlation and Heteroscedasticity Diagnostic Test FA M 1 Serial Correlationa 3.0197 (0.082) 0.331 (0.565) 2 3.160 (0.075) 0.103 (0.749) 1 2.960 (0.085) 0.477 (0.490) PAM 2 3.010 (0.083) 0.905 (0.342) 1 2.962 (0.085) 0.475 (0.491) WAUD 2 3.010 (0.083) 0.904 (0.342)

Heteroscedasticityb

Lagrange multiplier test of residual serial correlation Based on the regression of squared residuals on squared fitted values

On the other hand, the purpose of the heteroscedasticity test is to test whether the regression model meets the assumption of homoscedasticy, or in other words, whether there is any unequal variance of the residual between one to the other observation in the regression model. Homoscedasticity refers to the model where the variance of residual from one to the other observation is constant, while heteroscedasticity refers to the situation where the variances of residuals vary.

Both of tests were done by using Microfit software, and it can be detected from the serial correlation and heteroscedasticity diagnostic test. As a result, from Table 4.4, the diagnostic test for serial correlation and heteroscedasticity shows that treatment for the problem is not required since the p-values indicate that the null hypothesis of no serial correlation and equal variance cannot be rejected.

70

4.4.3 Regression Results

Table 4.5: Results of Multiple Regression Analysis of Dividend Policy Models FA M 1 Constant 0.604* (0.14) 0.305* (0.016) 0.152* (0.026) -213.696 (0.684) -0.015 (0.736) 0.065 (0.404) 0.145 (0.173) -566.575 (0.328) -0.029 (0.567) 0.064 (0.421) 0.063 (0.595) 0.215* (0.007) 0.150* (0.001) 43.520 (0.909) 0.044 (0.133) -237.758 (0.551) 0.014 (0.667) 44.023 (0.909) 0.045 (0.134) 0.215* (0.007) 0.149* (0.001) -237.638 (0.553) 0.014 (0.669) 2 0.616* (0.12) 1 0.478 (0.141) PA M 2 0.389 (0.230) 1 0.477 (0.144) WAUD 2 0.388 (0.233)

ECHGCONC1

ECHGCONC2

ECHGDISP

ECHGINST

ECHGMNG

ECHGFOR

ERNCONC1

ERNCONC2

ERNDISP

ERNGINST

71

ERNGMNG

-0.029 (0.664) 0.020 (0.790) -0.108* (0.018)

-0.023 (0.717) -0.057 (0.449) -0.110* (0.012)

-0.029 (0.663) 0.020 (0.793) -0.109* (0.031) 0.001 (0.960)

-0.023 (0.719) -0.057 (0.451) -0.110* (0.025) 0.218 (0.993)

ERNFOR

D(t-1)

D(t-2)

R2 Adjusted R2 F-statistic

0.150 0.120 5.081 (0.000)

0.145 0.115 4.874 (0.000)

0.196 0.163 5.827 (0.000)

0.213 0.180 6.449 (0.000)

0.196 0.157 4.960 (0.000)

0.213 0.174 5.489 (0.000)

*Significant at the 0.05 level

The F- tests, a measure for the strength of the regression, reveals that each dividend model is significant at 5 percent (p-value = 0.000). Therefore, it can be concluded that ownership classes are vital in determining a dividend policy. In terms of the adjusted coefficient of variation (R2), the partial adjustment model is better in explaining the variation of corporate dividend policy. The explanatory power for partial adjustment model is 16.3 percent if CONC1 is used and 18.0 percent if CONC2 is used. Whereas, for the Waud model is 15.7 percent if CONC1 is used and 17.4 percent if CONC2 is used, while for the full adjustment model is only 12.0 percent if CONC1 is used and 11.5 percent if CONC2 is used.
72

T- tests show that only the concentrated ownership variable is significant for every type of dividend model. Both CONC1 and CONC2 were positively and significant in influencing dividends at the 5 percent critical value. This finding is consistent with the results presented by Easterbrook (1984) and Mat Nor and Sulong (2009). High dividend payments can be used for mitigating agency conflicts since dividends can be substituted for shareholder monitoring. Therefore, large shareholders have strong incentives to require higher dividend payments in order to reduce monitoring costs.

Further, managerial ownership has a negative coefficient in the partial adjustment model and the Waud model, but the critical values are insignificant. While the full adjustment model did not only produce the unexpected sign, it is also insignificant. The insignificant value for managerial ownership implies that Malaysian companies do not use dividends as a mechanism to reduce the agency costs between managers and shareholders. Nevertheless, this finding is consistent with the study by Mat Nor and Sulong (2009).

Institutional ownership had been found to be positively and significantly related to dividends in Alli et al. (1993), Mohd et al. (1995) and Manos (2002). In this study, although the results reveal the expected sign in the partial adjustment model and the Waud model, it was insignificant. Therefore, it shows that dividends in Malaysia do not have any significant relationship with institutional ownership. However, this finding is similar to the results found by Noronha and George (1996). They show that

73

if there are alternative devices to control for agency costs, the payout rates are not related to proxies for agency cost variables.

The full adjustment model records a positive relationship between dividend payouts and foreign holdings but, the relationship is insignificant. Besides that, the partial adjustment model and the Waud model also showed similar results if CONC1 is used. Similar results were also found by Mat Nor and Sulong (2009). Hence, this study rejects the agency argument that foreign investors are more active monitors of corporations to reduce agency problems and leading firms to increase the level of payouts.

For ownership dispersion, most of the regression models do not only produce the unexpected sign but is also insignificant relationship. This result is contrasts to that of Rozeff (1982) and Mohd et al. (1995) which concluded that the more widely the ownership spread, the more acute the free rider problem; hence to minimize the agency problem, the greater the need for dividend distribution as outsider monitoring.

Interestingly, this study reveals that D(t-1) is significant in influencing dividends but in a negative form. Although it shows that the last year dividend is vital in determining current dividends, but the direction of relationship contrasts with that suggested by the Lintners (1956) theory of dividend smoothing by which claims that managers adopt a
policy of progressiveness in order to stabilize dividend distributions and to avoid erratic rates. Thus, dividends are smoothed and rarely decreased.

74

4.5

CONCLUSION

The empirical results reveal that the partial adjustment model is better in compared to the full adjustment model and the Waud model in explaining the variation in dividends with variables associated with ownership classes. Furthermore, the findings also reveal that only ownership concentration had significant influence on Malaysian corporate dividend policy. Besides that, this study also reveals that Malaysian dividend behavior contrasts with the theory of dividend smoothing proposed by Lintner (1956).

75

CHAPTER FIVE

CONCLUSION

5.1

INTRODUCTION

This chapter winds up the overall study. Section 5.2 discusses the overview of the link between dividends and ownership composition research process. Then, Section 5.3 presents the summary of the findings in the analysis and Section 5.4 discusses the implications of the study. Subsequently, the directions for further research are presented in Section 5.5, and Section 5.6 concludes the chapter.

5.2

OVERVIEW OF THE RESEARCH PROCESS

This study is done to examine whether ownership structure influences dividend policy among the public-listed companies in Malaysia. Therefore, theoretical literature for dividend policy, specifically the Modigliani-Miller theorem and Agency theory has been reviewed. Besides, in-depth empirical literature about the relationship between dividend policy and ownership structure have also been reviewed. Five independent variables used as the proxies of ownership structure were identified, namely ownership concentration, ownership dispersion, institutional ownership, managerial

76

ownership and foreign ownership. Besides that, the measurement of variables was guided by the prior research.

A total of 150 companies were identified as the sample for the study. The samples are taken from the four largest sectors on the Main Board of Bursa Malaysia, namely consumer, industrial, trading and services and properties sector. These companies were selected based on proportionate stratified random sampling. Data on dividends and earnings was collected from the companies financial statements provided by DataStream financial database while data on ownership was hand-collected from sample companies annual reports. This study employed annual data from 2005 to 2007.

This study utilized three dividend models to test the hypothesis of positive links between ownership structure and dividend policy: the Full Adjustment Model, the Partial Adjustment Model and the Waud Model. These models had been modified to account for the possible effects of ownership structure in determining the level of the corporate dividend.

The analysis commenced with a discussion of the sample descriptive statistics. Subsequently, the correlation analysis was carried out to examine the degree of relationship between possible pairs of variables. The cross-sectional nature of the data called for use of the OLS multiple regression technique. The predetermined variables are hypothesised; however, the result revealed that a multicollinearity problem had

77

occurred. Therefore, as a corrective action, some of the variables were dropped from the model and regression analysis was executed again. Nevertheless, the regression model had been not encountered with serial correlation and heteroscedasticity problems.

5.3

SUMMARY OF FINDINGS

This study was designed to examine the effect of ownership structure on corporate dividend policy. 150 companies were identified as the sample. This sample is representative for Malaysian companies, since it was selected from the four largest sectors on the Main Board of Bursa Malaysia whose annual reports were available for 2007. This study had employed the Full adjustment model, the Partial adjustment model and the Waud model to examine the potential associations between ownership structures and dividend policy. Five predetermined explanatory variables, namely ownership concentration, ownership dispersion, institutional ownership, managerial ownership and foreign ownership were regressed against dividends.

After a corrective analysis was conducted, and handling for multicollinearity problems, the regression model of dividend change against all the independent variables revealed that each dividend model was significant at a 5 percent confidence level. However, the Partial Adjustment Model was superior, since it could explain up

78

to 18.0 percent of the variation in dividend compared to 17.4 percent by the Waud model and 12.0 percent by the Full adjustment model.

Nevertheless, only one explanatory variable, which is ownership concentration, was found to be statistically significant in influencing corporate dividend policy. Ownership concentration has a positive significant relationship with dividend payment. The positive relationship between ownership concentration and dividends supports the findings in Shleifer and Vishny (1986). Large share ownership provides the incentives for controlling shareholders to use their influence to maximize the value of firms by reducing resources consumed in low return projects, thus implying that more cash flows can be distributed as dividends.

Besides that, this study introduced a new measurement for ownership concentration, which is the summation of the percentage of shares controlled by two major shareholders (referred as CONC2) since according to Nor and Sulong (2009) Malaysian companies have highly concentrated ownership. The results showed that CONC2 is more significant in influencing corporate dividend policy compared that CONC1 measured by Herfindahl Index 5 as traditionally being used.

Furthermore, the results prove the insignificant relationship of managerial, institutional, foreign and ownership dispersion on dividends. Therefore, it implies that these four variables are not vital in explaining dividends, hence dividend decisions in Malaysian companies are not influenced by managerial, institutional, foreign and

79

ownership dispersion. Nevertheless, the insignificant value of these four variables in determining dividend distribution has also been found by previous researchers.

Additionally, this study reveals that D(t-1) is negative and significant in influencing dividends, which contrasts with the theory of dividends smoothing by Lintner (1956). According to Lintner, managers are reluctant to cut dividend payments because they believe that any cut in dividends may give negative signals about the firm in the market. Thus, dividends are smoothed and rarely declined. In this study, it is observed that the dividend decreasing trend, instead of dividend increasing trend, over time is taking place.

5.4

IMPLICATIONS OF THE STUDY

The research has examined the relationship between dividend policy and ownership composition among the public-listed companies in Malaysia. The positive significance of ownership concentration variables implies that the formation of ownership has an effect on the amount of dividends distribution. Besides that, the regression model of dividends against all the independent variables was also found to be significant. Nevertheless, the findings reveal that the model of research explains less than 20 percent variation of dividend phenomenon in Malaysia. Thus, it indicates the possibility that dividend policy of Malaysian companies can also be explained by other dividend theory such as signaling theory and life-cycle theory.

80

This study suggests that shareholders with respect to stock investment in companies should concern themselves with the agency conflict between ownership classes. Shareholders must realize that financial policies such as dividend policy can serve as a mechanism for reducing agency costs. Besides that, regulatory bodies should also be concerned with the formation of ownership in formulating the related regulations to better control the agency conflict.

Moreover, the findings also reveal that the Partial adjustment model is better in explaining the variation of corporate dividend policy compared to the Waud model and the Full adjustment model.

5.5

DIRECTION FOR FURTHER STUDIES

There are a rich possible number of variables that can be used to examine the determinants of dividend policy. Nevertheless, this research concentrates on the ownership structure among the companies listed on the main board of the Bursa Malaysia and focuses on the five major variables that were repeatedly used by prior researchers. However, there might be other ownership variables that can be incorporated to explain the link between dividends and ownership composition. Thus, it would be beneficial if further research would be able to include other variables such as government ownership, board of directors ownership, family ownership and many

81

other types of ownership classes. This can help to better understand Malaysian companies dividend decisions.

Moreover, the lower explanatory power of the model examined in this study suggest the need of future research to focus on other dividend theories such as signaling theory, residual theory, life-cycle theory, smoothing theory and catering theory in the pursuit to understand the influence of factors on dividend policy in Malaysia. Furthermore, future research also can use Tobit regression to get better results since some of dependent variable is zeros. Future researchers on this topic may also use survey and interview methods to gauge top management and investor perspectives on this issue. In addition, future research may also increase the observation by incorporating companies listed in other sectors that are not included in this study as well as Second Board listed companies. Besides that, the longer period of study may also enhance the predictability model of the research. The findings will provide an interesting comparison to the findings from this study.

5.6

CONCLUSION

This chapter provides a brief summary of the overall study. The research procedure and the results from the analysis were discussed. Besides, the consequences of study and the recommendation for future research were also presented.

82

REFERENCES Abdelsalam, O., El-Masry, A., & Elsegini, S. 2008. Board composition, ownership structure and dividend policies in an emerging market. Managerial Finance 12: 953-964.

Agrawal, A., Jayaraman, N. 1994. The dividend policies of all-equity firms: A direct test of the free cash flow theory. Managerial and Decision Economics 15: 139-148.

Alli, K.L., Khan, Q., & Ramirez, G. G. 1993. Determinants of corporate dividend policy: A factorial analysis. The Financial Review 28: 523-547.

Amidu and Abor. 2006. Determinants of dividend payout ratios in Ghana. The Journal of Risk Finance 7: 136-145.

Ang, J.S., Cole, R.A., Lin, J.W. 2000. Agency costs and ownership structure. The Journal of Finance 1: 81-105.

Baker, H.K., Saadi, S., Dutta, S., and Gandhi, D. 2007. The perception of dividends by Canadian managers: new survey evidence 3: 70 91.

Black, F. and Scholes, M.S. 1974. The effect of dividend yield and dividend policy on common stock prices and returns. Journal of Financial Economics 1: 1-22.

Coffee, J.C.Jr. 1991. Liquidity versus control: The institutional investor as corporate monitor. Columbia Law Review 91: 1277-1368.

Cook., D.O and Jeon., J.Q. 2006. Foreign ownership, domestic ownership and payout policy. http://fma.org/SLC/Papers/ForeignOwnership_Dividends.pdf

83

Crutchley, C.E,. Jensen, M.R.H,. Jahera, J.S,. & Raymond, J.E. 1999. Agency problem and the simultaneity of financial decision making. The role of institutional ownership. International Review of Financial Analysis 8:177-197.

Dorsman, A.B., Montfort, K.V., Vink, I. 1999. An adjusted Lintner-model for The

Netherlands. Research Paper. Available at http://hdl.handle.net/1871/1517

DSauza, J., & Saxena, A. K. 1999. Agency cost, market risk, investment opportunities and dividend policy - An international perspective. Managerial Finance 25.

Easterbrook, F. 1984. Two agency-cost explanations of dividends. American Economic Review 74: 605-659.

Eisenhardt, K.M. 1989. Agency Theory: An assessment and review. Academy of Management Review: 57-74.

Fama, E.F., Babiak, H. 1968. Dividend policy: An empirical analysis. Journal of the American Statistical Association 63: 1132-1161.

Fama, E.F. 1980. Agency problem and the theory of the firm. Journal of political economy 88: 288-307.

Faccio, M., Lang, L.H.P. and Young, L. 2001. Dividend and expropriation. American Economic Review 91: 5478.

Fluck, Z. 1999. The dynamics of the managers-shareholders conflict. Review of Financial Studies 2: 379-404.

84

Foong, S.S., Zakaria, N., Tan, H.B. 2007. Firm performance and dividend-related factors: The case of Malaysia. Labuan Bulletin of International Business & Finance 5: 97-111.

Frankfurter, G.M. and Wood, B.G., Jr. 2002. Dividend policy theories and their empirical tests. International Review of Financial Analysis 11: 111-138.

Friend, I. and Puckett, M. 1964. Dividend and stock prices. American Economic Review 54: 656-682.

Gordon, M.J. 1959. Dividends, Earnings and Stock Prices, Review of Economics and Statistics: 99-105.

Gugler, K. 2003. Corporate governance, dividend payout policy and the interrelation between dividends, R&D, and capital investment. Journal of Banking and Finance 27: 1279-1321.

Hansen, S.R., Kumar, R., & Shome, D.K. 1994. Dividend policy and corporate monitoring: Evidence from the regulated electric utility industry. Financial management: 16-22.

Han, K.C., S.H. Lee and D.Y. Suk. 1999. Institutional shareholders and dividends. Journal of financial and Strategic Decisions 12: 53-62.

Harada, Kimie & Nguyen, Pascal. 2006. Ownership concentration, agency conflicts, and dividend policy in Japan. Available at SSRN:

http://ssrn.com/abstract=953433

Harjito, D. A. 2009. The influences of agency factors and transaction costs factors to dividend payout ratio. 11th Malaysian Financial Association Conference.

85

Holderness, C.G. 2003. A survey of blockholders and corporate control, Economic Policy Review 9: 5164.

Jensen, M.C., and W.H. Meckling. 1976. Theory of the Firm: Managerial behavior Agency Costs and Ownership Structure, Journal of Financial Economics 3: 305-360.

Jensen, M. 1986. Agency costs of free cash flow, corporate finance and takeovers. American Economic Review 76: 323-329.

Jensen, G. R., Solberg, D.P. and Zorn, T.S. 1992. Simultaneous determination of insider ownership , debt and dividend policies. The Journal of Financial and Quantitative Analysis 27: 247-263.

Khan, T. 2006. Company dividends and ownership structure: evidence from UK panel data. The Economic Journal 116: 172-189.

Kouki., M. & Guizani., M. 2009. Ownership structure and dividend policy evidence from the Tunisian Stock Market. European Journal of Scientific Research 25: 42-53.

Kent Baker, H., Saadi, S., Dutta, S. 2007. The perception of dividends by Canadian managers: new survey evidence. International Journal of Managerial Finance 3: 70-91.

La Porta, R., F.Lopez-de-Silances, A. Shleiffer and R.W. Vishny. 2000a. Agency problems and dividend policies around the world. Journal of Finance 55: 133.

La Porta, R., F.Lopez-de-Silances, A. Shleiffer and R.W. Vishny. 1999. Corporate ownership around the world. Journal of Finance 54: 451-517.

86

Litner, J. 1994. Distribution of incomes of corporations among dividends, retained earnings and taxes. American Economic Review 46: 97-113.

Lloyd, W.P., J.S Jahera Jr. and D.E. Page. 1985. Agency costs and dividend payout ratios. Quarterly Journal of Business and Economics 24: 19-29.

Mahadwartha, P.A. 2002. The association of managerial ownership with dividend policy and leverage policy: Indonesian firms. Presented on National Seminar Surviving Strategies to Cope with the Future, University Atmajaya Yogyakarta-Indonesia.

Mancinelli, L., & Ozkan, A. 2006. Ownership structure and dividend policy: Evidence from Italian firms. The European Journal of Finance, 12: 262-282.

Manos, R. 2002. Dividend policy and agency theory: Evidence on Indian firms. Finance and development research programme, working paper series 41: 125.

Mendez, J.M. & Willey, T. 1995. Agency costs in the banking industry: An examination of ownership behavior, leverage and dividend policies. Journal of Economics and Finance 3: 105-117.

Mohd, M.A,. Perry, L.G,. and Rimbey, J.N,. 1995. An investigation of the dynamic relationship between agency theory and dividend policy. The Financial Review 30: 367-385.

Mollah, A.S., Rafiq, R.B., and Sharp, P.A. 2007. Relevance of agency theory fro dividend policy in an emerging economy: The case of Bangladesh. The IFCAI Journal of Applied Finance 13: 5-15.

87

Miller, M., and Modigliani, F. 1958. The cost of capital, corporation finance and the theory of investment. American Economic Review 48: 261297.

Miller, M., and Modigliani, F. 1961. Dividend policy, growth and the valuation of shares. Journal of Business 34: 411-433.

Naceur, B., Goaied, M. and Belanes, A. 2006. On the Determinants and Dynamics of Dividend Policy. Available at SSRN: http://ssrn.com/abstract=889330

Najjar, A.B. 2009. Dividend behaviour and smoothing new evidence from Jordanian panel data. Studies in Economics and Finance 26: 182-197.

Nor, F.M., and Sulong, Z. 2007. The interaction effect of ownership structure and board governance on dividends: Evidence from Malaysian Listed Firms. Capital market review 15: 73-101.

Noronha, G.M., Shome, D.K and G.E Morgon. 1996. The monitoring rational for dividend and the interaction of capital structure and dividend decision. Journal of Banking and Finance 20: 439-454.

Renneboog, L. and P.G.. Szilagyi. 2006. How relevant is dividend policy under low shareholder protection. ECGI-Finance Working Paper 128.

http://www.ssrn.com

Rozeff, M.S. 1982. Growth, beta and agency costs as determinants of dividend payout ratios. Journal of Financial Research 5: 249-259.

Schooley, D.K. and D.L. Barney, Jr. 1994. Using dividend policy and managerial ownership to reduce agency costs. Journal of Financial Research 17: 363-373.

88

Shleifer, A. and Vishny, R.W. 1997. A survey on corporate governance. Journal of Finance 52: 737783.

Shleifer, A. and Vishny, R.W. 1986. Large stockholders and corporate control. Journal of Political Economy 95: 461488.

Short, H., H. Zhang and Keasey. 2002. The link between dividend policy and institutional ownership. Journal of Corporate Finance 8: 105-122.

Waud, R. 1966. Small sample bias due to misspecification in the partial adjustment and adapted expectations models. Journal of the American Statistical Association: 134-145.

89

APPENDICES

THE FULL ADJUSTMENT MODEL

Regression of DIVCHG on ECHGCONC1, ECHGDISP, ECHGINST, ECHGMNG and ECHGFOR after treatment for multicollinearity problem

Ordinary Least Squares Estimation ******************************************************************************* Dependent variable is DIVCHG 150 observations used for estimation from 1 to 150 ******************************************************************************* Regressor Coefficient Standard Error T-Ratio[Prob] INPT .60418 .24248 2.4917[.014] ECHGCONC1 .30463 .12477 2.4415[.016] ECHGDISP -213.6962 523.9020 -.40789[.684] ECHGINST -.014797 .043824 -.33765[.736] ECHGMNG .065265 .078047 .83623[.404] ECHGFOR .14543 .10623 1.3691[.173] ******************************************************************************* R-Squared .14997 R-Bar-Squared .12046 S.E. of Regression 2.9507 F-stat. F( 5, 144) 5.0813[.000] Mean of Dependent Variable .72460 S.D. of Dependent Variable 3.1463 Residual Sum of Squares 1253.8 Equation Log-likelihood -372.0869 Akaike Info. Criterion -378.0869 Schwarz Bayesian Criterion -387.1188 DW-statistic 2.2719 *******************************************************************************

Diagnostic Tests ******************************************************************************* * Test Statistics * LM Version * F Version * ******************************************************************************* * * * * * A:Serial Correlation*CHSQ( 1)= 3.0197[.082]*F( 1, 143)= 2.9380[.089]* * * * * * B:Functional Form *CHSQ( 1)= 1.6874[.194]*F( 1, 143)= 1.6269[.204]* * * * * * C:Normality *CHSQ( 2)= 998.6167[.000]* Not applicable * * * * * * D:Heteroscedasticity*CHSQ( 1)= .33125[.565]*F( 1, 148)= .32756[.568]* ******************************************************************************* A:Lagrange multiplier test of residual serial correlation B:Ramsey's RESET test using the square of the fitted values C:Based on a test of skewness and kurtosis of residuals D:Based on the regression of squared residuals on squared fitted values

THE FULL ADJUSTMENT MODEL

Regression of DIVCHG on ECHGCONC2, ECHGDISP, ECHGINST, ECHGMNG and ECHGFOR after treatment for multicollinearity problem

Ordinary Least Squares Estimation ******************************************************************************* Dependent variable is DIVCHG 150 observations used for estimation from 1 to 150 ******************************************************************************* Regressor Coefficient Standard Error T-Ratio[Prob] INPT .61609 .24299 2.5355[.012] ECHGCONC2 .15220 .067778 2.2455[.026] ECHGDISP -566.5753 577.2733 -.98147[.328] ECHGINST -.028723 .050049 -.57390[.567] ECHGMNG .063974 .079257 .80717[.421] ECHGFOR .063485 .11911 .53298[.595] ******************************************************************************* R-Squared .14474 R-Bar-Squared .11504 S.E. of Regression 2.9598 F-stat. F( 5, 144) 4.8738[.000] Mean of Dependent Variable .72460 S.D. of Dependent Variable 3.1463 Residual Sum of Squares 1261.5 Equation Log-likelihood -372.5476 Akaike Info. Criterion -378.5476 Schwarz Bayesian Criterion -387.5795 DW-statistic 2.2787 *******************************************************************************

Diagnostic Tests ******************************************************************************* * Test Statistics * LM Version * F Version * ******************************************************************************* * * * * * A:Serial Correlation*CHSQ( 1)= 3.1603[.075]*F( 1, 143)= 3.0777[.082]* * * * * * B:Functional Form *CHSQ( 1)= .72704[.394]*F( 1, 143)= .69649[.405]* * * * * * C:Normality *CHSQ( 2)= 1002.3[.000]* Not applicable * * * * * * D:Heteroscedasticity*CHSQ( 1)= .10264[.749]*F( 1, 148)= .10134[.751]* ******************************************************************************* A:Lagrange multiplier test of residual serial correlation B:Ramsey's RESET test using the square of the fitted values C:Based on a test of skewness and kurtosis of residuals D:Based on the regression of squared residuals on squared fitted values

THE PARTIAL ADJUSTMENT MODEL

Regression of DIVCHG on ERNCONC1, ERNDISP, ERNINST, ERNMNG, ERNFOR and D(t-1) after treatment for multicollinearity problem

Ordinary Least Squares Estimation ******************************************************************************* Dependent variable is DIVCHG 150 observations used for estimation from 1 to 150 ******************************************************************************* Regressor Coefficient Standard Error T-Ratio[Prob] INPT .47793 .32273 1.4809[.141] ERNCONC1 .21474 .077975 2.7539[.007] ERNDISP 43.5199 381.7247 .11401[.909] ERNINST .044420 .029368 1.5125[.133] ERNMNG -.028521 .065519 -.43531[.664] ERNFOR .019751 .074039 .26676[.790] DIV06 -.10789 .044946 -2.4005[.018] ******************************************************************************* R-Squared .19647 R-Bar-Squared .16276 S.E. of Regression 2.8789 F-stat. F( 6, 143) 5.8275[.000] Mean of Dependent Variable .72460 S.D. of Dependent Variable 3.1463 Residual Sum of Squares 1185.2 Equation Log-likelihood -367.8679 Akaike Info. Criterion -374.8679 Schwarz Bayesian Criterion -385.4051 DW-statistic 2.2668 *******************************************************************************

Diagnostic Tests ******************************************************************************* * Test Statistics * LM Version * F Version * ******************************************************************************* * * * * * A:Serial Correlation*CHSQ( 1)= 2.9597[.085]*F( 1, 142)= 2.8582[.093]* * * * * * B:Functional Form *CHSQ( 1)= .011544[.914]*F( 1, 142)= .010929[.917]* * * * * * C:Normality *CHSQ( 2)= 904.5238[.000]* Not applicable * * * * * * D:Heteroscedasticity*CHSQ( 1)= .47735[.490]*F( 1, 148)= .47249[.493]* ******************************************************************************* A:Lagrange multiplier test of residual serial correlation B:Ramsey's RESET test using the square of the fitted values C:Based on a test of skewness and kurtosis of residuals D:Based on the regression of squared residuals on squared fitted values

THE PARTIAL ADJUSTMENT MODEL

Regression of DIVCHG on ERNCONC2, ERNDISP, ERNINST, ERNMNG, ERNFOR and D(t-1) after treatment for multicollinearity problem

Ordinary Least Squares Estimation ******************************************************************************* Dependent variable is DIVCHG 150 observations used for estimation from 1 to 150 ******************************************************************************* Regressor Coefficient Standard Error T-Ratio[Prob] INPT .38852 .32229 1.2055[.230] ERNCONC2 .14913 .045502 3.2774[.001] ERNDISP -237.7579 397.9363 -.59748[.551] ERNINST .014456 .033542 .43097[.667] ERNMNG -.023324 .064280 -.36285[.717] ERNFOR -.057133 .075253 -.75920[.449] DIV06 -.10950 .043210 -2.5342[.012] ******************************************************************************* R-Squared .21297 R-Bar-Squared .17995 S.E. of Regression 2.8492 F-stat. F( 6, 143) 6.4494[.000] Mean of Dependent Variable .72460 S.D. of Dependent Variable 3.1463 Residual Sum of Squares 1160.9 Equation Log-likelihood -366.3115 Akaike Info. Criterion -373.3115 Schwarz Bayesian Criterion -383.8488 DW-statistic 2.2666 *******************************************************************************

Diagnostic Tests ******************************************************************************* * Test Statistics * LM Version * F Version * ******************************************************************************* * * * * * A:Serial Correlation*CHSQ( 1)= 3.0098[.083]*F( 1, 142)= 2.9076[.090]* * * * * * B:Functional Form *CHSQ( 1)= .19052[.662]*F( 1, 142)= .18058[.672]* * * * * * C:Normality *CHSQ( 2)= 1062.1[.000]* Not applicable * * * * * * D:Heteroscedasticity*CHSQ( 1)= .90475[.342]*F( 1, 148)= .89810[.345]* ******************************************************************************* A:Lagrange multiplier test of residual serial correlation B:Ramsey's RESET test using the square of the fitted values C:Based on a test of skewness and kurtosis of residuals D:Based on the regression of squared residuals on squared fitted values

THE WAUD MODEL Regression of DIVCHG on ERNCONC1, ERNDISP, ERNINST, ERNMNG, ERNFOR, D(t-1) and D (t-2) after treatment for multicollinearity problem

Ordinary Least Squares Estimation ******************************************************************************* Dependent variable is DIVCHG 150 observations used for estimation from 1 to 150 ******************************************************************************* Regressor Coefficient Standard Error T-Ratio[Prob] INPT .47685 .32459 1.4691[.144] ERNCONC1 .21454 .078352 2.7381[.007] ERNDISP 44.0229 383.1966 .11488[.909] ERNINST .044512 .029529 1.5074[.134] ERNMNG -.028871 .066125 -.43662[.663] ERNFOR .019534 .074426 .26247[.793] DIV06 -.10896 .049954 -2.1813[.031] DIV05 .0013400 .026941 .049739[.960] ******************************************************************************* R-Squared .19648 R-Bar-Squared .15687 S.E. of Regression 2.8890 F-stat. F( 7, 142) 4.9605[.000] Mean of Dependent Variable .72460 S.D. of Dependent Variable 3.1463 Residual Sum of Squares 1185.2 Equation Log-likelihood -367.8666 Akaike Info. Criterion -375.8666 Schwarz Bayesian Criterion -387.9091 DW-statistic 2.2669 *******************************************************************************

Diagnostic Tests ******************************************************************************* * Test Statistics * LM Version * F Version * ******************************************************************************* * * * * * A:Serial Correlation*CHSQ( 1)= 2.9620[.085]*F( 1, 141)= 2.8404[.094]* * * * * * B:Functional Form *CHSQ( 1)= .012220[.912]*F( 1, 141)= .011488[.915]* * * * * * C:Normality *CHSQ( 2)= 905.7561[.000]* Not applicable * * * * * * D:Heteroscedasticity*CHSQ( 1)= .47457[.491]*F( 1, 148)= .46973[.494]* ******************************************************************************* A:Lagrange multiplier test of residual serial correlation B:Ramsey's RESET test using the square of the fitted values C:Based on a test of skewness and kurtosis of residuals D:Based on the regression of squared residuals on squared fitted values

THE WAUD MODEL

Regression of DIVCHG on ERNCONC2, ERNDISP, ERNINST, ERNMNG, ERNFOR, D(t-1) and D (t-2) after treatment for multicollinearity problem

Ordinary Least Squares Estimation ******************************************************************************* Dependent variable is DIVCHG 150 observations used for estimation from 1 to 150 ******************************************************************************* Regressor Coefficient Standard Error T-Ratio[Prob] INPT .38835 .32405 1.1984[.233] ERNCONC2 .14911 .045735 3.2603[.001] ERNDISP -237.6381 399.6039 -.59468[.553] ERNINST .014474 .033738 .42903[.669] ERNMNG -.023381 .064883 -.36036[.719] ERNFOR -.057157 .075575 -.75629[.451] DIV06 -.10968 .048312 -2.2702[.025] DIV05 .2180E-3 .026671 .0081748[.993] ******************************************************************************* R-Squared .21297 R-Bar-Squared .17418 S.E. of Regression 2.8592 F-stat. F( 7, 142) 5.4894[.000] Mean of Dependent Variable .72460 S.D. of Dependent Variable 3.1463 Residual Sum of Squares 1160.9 Equation Log-likelihood -366.3115 Akaike Info. Criterion -374.3115 Schwarz Bayesian Criterion -386.3541 DW-statistic 2.2666 *******************************************************************************

Diagnostic Tests ******************************************************************************* * Test Statistics * LM Version * F Version * ******************************************************************************* * * * * * A:Serial Correlation*CHSQ( 1)= 3.0102[.083]*F( 1, 141)= 2.8875[.091]* * * * * * B:Functional Form *CHSQ( 1)= .19108[.662]*F( 1, 141)= .17984[.672]* * * * * * C:Normality *CHSQ( 2)= 1062.3[.000]* Not applicable * * * * * * D:Heteroscedasticity*CHSQ( 1)= .90395[.342]*F( 1, 148)= .89730[.345]* ******************************************************************************* A:Lagrange multiplier test of residual serial correlation B:Ramsey's RESET test using the square of the fitted values C:Based on a test of skewness and kurtosis of residuals D:Based on the regression of squared residuals on squared fitted values

You might also like