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 Master’s programme at

Universiti Utara Malaysia or other institutions.

_____________________________________________

NORHASNIZA BINTI MOHD HASAN ABDULLAH

Date: 23 NOVEMBER 2009

iii

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.

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 university’s

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

i

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

ii

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

iv

TABLE OF CONTENTS

DECLARATION

PERMISSION TO USE

ABSTRACT (BAHASA MELAYU) i

ABSTRACT (ENGLISH) ii

ACKNOWLEDGEMENT iii

TABLE OF CONTENTS iv

LIST OF TABLES vii

LIST OF ABBREVIATIONS viii

CHAPTER 1: BACKGROUND OF STUDY

1.1 Introduction 1

1.2 Problem Statement 4

1.3 Objective of the Study 8

1.4 Significance of the Study 9

1.5 Limitation of the Study 10

1.6 Conclusion 11

CHAPTER 2: LITERATURE REVIEW

2.1 Introduction 12

2.2 Theoretical Literature 12

2.2.1 M&M Irrelevant Dividend Theory and

other Related Theories/Models 13

2.2.2 Lintner Dividend Stability Theory and

other Related Theories/Models 17

2.3 Empirical Literature 20

2.4 Conclusion 49

v

CHAPTER 3: RESEARCH METHOD

3.1 Introduction 50

3.2 Research Framework 50

3.2.1 Ownership Concentration 51

3.2.2 Ownership Dispersion 51

3.2.3 Institutional Ownership 51

3.2.4 Managerial Ownership 52

3.2.5 Foreign Ownership 52

3.3 Sample Description and Data Collection 53

3.4 Models on Dividend Policy 54

3.4.1 The Full Adjustment Model 54

3.4.2 The Partial Adjustment Model 55

3.4.3 The Waud Model 56

3.5 Measurement of Variable 57

3.5.1 Dividends 57

3.5.2 Earnings 58

3.5.3 Ownership Concentration 58

3.5.4 Ownership Dispersion 58

3.5.5 Institutional Ownership 58

3.5.6 Managerial Ownership 59

3.5.7 Foreign Ownership 59

3.6 Conclusion 59

CHAPTER 4: ANALYSIS AND FINDINGS

4.1 Introduction 61

4.2 Descriptive Analysis 61

4.3 Correlation Analysis 64

4.4 Regression Analysis 67

4.4.1 Multicollinearity 67

4.4.2 Serial Correlation and Heteroscedasticity Test 69

4.4.3 Regression Results 71

4.5 Conclusion 75

vi

CHAPTER 5: CONCLUSION

5.1 Introduction 76

5.2 Overview of the Research Process 76

5.3 Summary of Findings 78

5.4 Implications of the Study 80

5.5 Direction for Further Studies 81

5.6 Conclusion 82

REFERENCES 83

APPENDICES

vii

LIST OF TABLES

Table 2.1: Summary of Empirical Literatures

Table 4.1: Summary Descriptive Statistic

Table 4.2: Pearson Correlation Matrix among the Variables

Table 4.3: Variance Inflation Factor of Variables (tolerance value is given in the

parentheses)

Table 4.4: Durbin-Watson and Heteroscedasticity Diagnostic Test

Table 4.5: Results of Multiple Regression Analysis of Dividend Policy Models

viii

LIST OF ABBREVIATIONS

E : Earnings

ECHG : Earning Change

D : Dividends

CONC : Ownership Concentration

DISP : Ownership Dispersion

INST : Institutional Ownership

MNG : Managerial Ownership

FOR : Foreign Ownership

FAM : The Full Adjustment Model

PAM : The Partial Adjustment Model

WM : The Waud Model

1

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.

2

Generally, there are two types of cash dividends, which are interim dividends and

final dividends. Interim dividends are declared and distributed before the company’s

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

company’s 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 company’s 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

3

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 shareholder’s 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 Miller’s 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 so-

called 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.

4

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 non-

existence of taxes, transaction costs and asymmetric information, dividends are

irrelevant since shareholders can create homemade dividends by selling a portion of

5

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 company’s 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

6

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 manager’s

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 company’s

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 manager’s 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

7

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 company’s 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

8

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.

9

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,

10

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.

11

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.

12

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).

13

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 firm’s future prospect; and

distribution of income between dividends and retained earnings has no effect on the

firm’s cost of equity (Foong, Zakaria and Tan, 2007). The main conclusion of this

paper is that firm’s capital budgeting policy is independent of its dividend policy.

M&M’s 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&M’s 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

14

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 Moh’d, 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

15

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.

16

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 firm’s 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 firm’s 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

17

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 Lintner’s 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* t i = rE t i (1 )

18

with,

D* t i = the target level dividend of dividend for fund i year t.

r = the optimal amount of dividend as a percentage of the profit, for fund i.

E t i = the profit company i made in year t.

The value of r will be between 0 and 1 since companies usually won’t 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) (2)

with,

c = Velocity at which a company adjusts the dividend

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 + β i 1 D ( t -1 ) i + β i 2 E t i + µ t i (3)

19

with,

β i 1 = -Ci

β i 2 = Ci ri

µ t i = 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* t i = rE* t i (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) + µ 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)

20

2.3 EMPIRICAL LITERATURE

Miller and Modigliani’s (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 Rozeff’s 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.

21

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 firm’s

dividend levels. Therefore, this observation supports Rozeff’s 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).

22

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

Jensen’s (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

23

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 Moh’d, 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 Rozeff’s and

Easterbrook’s 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

24

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 firm’s capital structure and dividend decisions. In the

model, simultaneity is contingent on the applicability of Easterbrook’s (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

firm’s 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 firm’s common stock is

used as a proxy for controlling agency costs. The dividend payout variable used in the

25

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 agency-

cost 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 firm’s equity, shareholders incur

agency costs resulting from management’s shirking and perquisite consumption. They

26

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 Rozeff’s 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.

27

Study by Short, Zhang and Keasey (2002) is the first example of using well-

established 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 firm’s equity. A

significantly negative relation between dividends and ownership concentration result

appear to corroborate Rozeff’s 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

28

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 firm’s 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

29

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 firm’s 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 firm’s 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

30

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.

31

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.

32

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.

Table 2.1: Summary of Empirical Literatures

AUTHOR

SAMPLE

METHOD

MEASUREMENT OF VARIABLE

FINDINGS

DEPENDENT

INDEPENDENT

CONTROL

Michael S. Rozeff (1982)

Size: 1000

Multiple Regression

DPR: 7 year average over a period of 1974 to 1980

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 Line’s forecast of growth rate of revenues. BETA: 5 year value-weighted beta of weekly data on the stock’s returns. STOCK: natural logarithm of number of common stockholders

-

INS: negative, significant GROW 1: negative, significant GROW 2: negative, significant BETA: negative, significant STOCK: positive, significant

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: percentage of common stock held by insiders. GROW 1: 5 year average of growth rate of revenues.

-

INS: negative, significant GROW 1: negative, significant

and Daniel E. Page (1985)

July – September 1984 Size: 957

GROW 2: 5 year average of Value Line’s forecast of growth rate of revenues. BETA: 5 year value-weighted beta of weekly data on the stock’s returns. STOCK: natural logarithm of number of common stockholders. SIZE: natural logarithm of sales for 1983.

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)

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.

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.

negative, significant

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: firm’s 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.

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. α: 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.

LEVERED: negative, significant α: 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: regulatory commission rank. OWNSHIP: stockholder-ownership concentration, measured by Herfindahl Index. FLOTCOST: firm’s historical average flotation cost incurred in selling common stock, expressed as a percentage of the gross proceeds. TAGROW: growth rate in total assets.

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

Mahmoud A. Moh’d, 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. β0: 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

SIZE: natural log of firm sales.

GROW 1: negative, significant GROW 2: negative, significant β0: negative, significant OLRISK: negative, significant FLRISK: negative, significant INSTINV: positive, significant INSD: negative, significant STKHLDR: positive, significant SIZE: positive, significant

Jose Mercado-Mendez and Thomas

Location: U.S Period: 1985 -

Regression

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

EARNING VOLATILITY: standard deviation of the income before depreciation and amortization on total assets. MANAGERS’

EARNING VOLATILITY: positive, insignificant MANAGERS’ DIVERSIFICATION

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.

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 firm’s dollar dividend divided by the 5 year average of income available to common stockholders.

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.

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

Juliet D’Sauza and Atul

Period: 1995 - 1997

Multiple Regression

DPR

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

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

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

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 Homoscedastic Tobit regression Heckman’s 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 company’s 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.

Dt i – 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

Tobit Regression

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 Observation: 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 by total assets 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

GROW: positive DEBT: negative KD: positive

Luciana Mancinelli and Aydin Ozkan

Location: Itali 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

SIZE: natural logarithm of sales. LEVERAGE:

VOTING_RIGHTS1: negative, significant 2_LARGE_SHAREHOLDER: insignificant

(2006) 2001 Observation: 139

Capitalization

largest shareholder owns a fraction larger than 5% of the firms voting rights and zero otherwise. OTHER_LARGE_SHAREHOLDERS: 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:

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

OTHER_LARGE_SHAREHOLDERS: 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: 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.

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

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

Location: Bangladesh Period: 1999 – 2003 Size: 153

Cross-sectional 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.

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: 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

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 Abdelsalam, Ahmed El-Masry and Sabri Elsegini (2008)

Location: Egypt Period: 2003 – 2005 Size: 50

Regression

DIVDECISION: 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

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.

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

D. Agus Harjito and

Location: Jakrta

Multiple Regression

DPR: percentage of profit paid in

INSDO: share percentage owned by director and commissioner.

INSDO: negative, significant OFO: negative,

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 year’s income growth level β: return volatility measurement of security toward market-risk

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

33

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.

50

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:

51

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.

52

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

director’s 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

EXPECTED RELATION

Ownership Concentration

Positive

Ownership Dispersion

Positive

Institutional Ownership

Positive

Managerial Ownership

Negative

Foreign Ownership

Positive

53

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.

54

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 = α + r(E t i – E ( t - 1 )) + µ 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:

55

D t i –D ( t - 1) i = α + r(E t i–E ( t - 1 )) + rCO NC(E t i–E( t - 1))*CONC

+ rD IS P(E t i–E ( t - 1 ))*DISP + r I NS T(E t i–E ( t -

1 ))*INST + rMNG(E t i–E ( t - 1 ))*MNG +

rFO 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 t i –D ( t -1) i= a + c(D* t i - D (t -1)i) + µ 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.

56

Assuming that companies with significant ownership classes have different

target payout ratios (r), the model becomes:

Dti – D(t-1)i = α + crEti + crCONCEti*CONC + crDISPEti*DISP +

crINSTEti*INST + crMNGEti*MNG + crFOREti*FOR –

cD(t-1)i + µti

(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 long-

run expected earnings, E*:

D* t i = rE* t i

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) + µ t i

The formation of expectations follows an adaptive expectation model:

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

57

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 + cdrCO NCE t i*CONC +

cdrD IS PE t i*DISP + cdrINS TE t i*INST +

cdrMNG E t i*MNG + cdrFO RE t i*FOR +

(1-d-c)D ( t - 1 ) i – (1 -d)(1-c)D ( t - 2 ) i - µ 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.

58

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 Moh’d et al. (1995), Amidu (2006) and Kouki and

Guizani (2009) defined institutional ownership as a percentage of equity

59

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

60

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

2007.

61

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

62

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.

63

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

ME

AN

MIN

IMU

M

MA

XIM

UM

ST

D.

DE

V

SK

EW

NE

SS

KU

RT

OS

IS

Dividend

Earnings

CONC1

CONC2

DISP

INST

MNG

FOR

5.8688

15.8671

0.1482

0.4185

0.0000

0.2417

0.1114

0.0822

0.0000

-107.300

0.0062

0.0848

0.0000

0.0016

0.0000

0.0000

38.1000

121.0400

0.5362

0.7811

0.0001

0.9593

0.5549

0.5981

7.4203

23.6610

0.1212

0.1716

0.0000

0.2024

0.1558

0.1188

2.281

0.245

1.216

0.213

1.820

1.369

1.496

2.455

5.510

8.402

0.878

-0.717

4.172

1.821

1.021

6.541

64

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

65

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.

66

Table 4.2: Pearson Correlation Matrix among the Variables

D

E

CONC1

CONC2

DISP

INST

MNG

E

CONC1

CONC2

DISP

INST

MNG

FOR

0.500**

(0.000)

0.424**

(0.000)

0.374**

(0.000)

-0.145

(0.077)

0.191*

(0.019)

-0.256**

(0.001)

-0.020

(0.808)

0.137*

(0.047)

0.095

(0.125)

-0.166*

(0.021)

0.253**

(0.001)

-0.128

(0.060)

0.173*

(0.017)

0.933**

(0.000)

-0.174*

(0.016)

-0.019

(0.411)

-0.339**

(0.000)

-0.132

0.054

-0.192**

0.009

-0.008

(0.460)

-0.338**

(0.000)

-0.110

(0.091)

-0.048

(0.281)

-0.088

0.143

-0.146**

0.037

-0.052

(0.264)

0.056

(0.247)

-0.126

(0.062)

* Correlation is significant at the 0.05 level

** Correlation is significant at the 0.01 level

67

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

68

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

PA M

W A UD

1

2

1

2

1

2

ERNCHG

ECHGCONC1

ECHGCONC2

ECHGDISP

ECHGINST

ECHGMNG

ECHGFOR

ERN

ERNCONC1

28.907

(0.035)

5.049

(0.198)

3.651

(0.274)

5.863

(0.171)

3.976

(0.251)

3.499

(0.286)

46.340

(0.022)

18.418

(0.054)

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)

69

ERNCONC2

ERNDISP

ERNGINST

ERNGMNG

ERNFOR

D(t-1)

D(t-2)

1.965

(0.509)

4.049

(0.247)

2.287

(0.437)

2.477

(0.404)

1.561

(0.641)

12.738

(0.079)

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)

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.

70

Table 4.4: Serial Correlation and Heteroscedasticity Diagnostic Test

FA M

PA M

W A UD

1

2

1

2

1

2

Serial Correlationa

Heteroscedasticityb

3.0197

(0.082)

0.331

(0.565)

3.160

(0.075)

0.103

(0.749)

2.960

(0.085)

0.477

(0.490)

3.010

(0.083)

0.905

(0.342)

2.962

(0.085)

0.475

(0.491)

3.010

(0.083)

0.904

(0.342)

aLagrange multiplier test of residual serial correlation

bBased 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.

71

4.4.3 Regression Results

Table 4.5: Results of Multiple Regression Analysis of Dividend Policy Models

FA M

PA M

W A UD

1

2

1

2

1

2

Constant

ECHGCONC1

ECHGCONC2

ECHGDISP

ECHGINST

ECHGMNG

ECHGFOR

ERNCONC1

ERNCONC2

ERNDISP

ERNGINST

0.604*

(0.14)

0.305*

(0.016)

-213.696

(0.684)

-0.015

(0.736)

0.065

(0.404)

0.145

(0.173)

0.616*

(0.12)

0.152*

(0.026)

-566.575

(0.328)

-0.029

(0.567)

0.064

(0.421)

0.063

(0.595)

0.478

(0.141)

0.215*

(0.007)

43.520

(0.909)

0.044

(0.133)

0.389

(0.230)

0.150*

(0.001)

-237.758

(0.551)

0.014

(0.667)

0.477

(0.144)

0.215*

(0.007)

44.023

(0.909)

0.045

(0.134)

0.388

(0.233)

0.149*

(0.001)

-237.638

(0.553)

0.014

(0.669)

72

ERNGMNG

ERNFOR

D(t-1)

D(t-2)

-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)

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.

73

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), Moh’d 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

74

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 Moh’d 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

Lintner’s (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.

75

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).

76

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

77

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

78

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

79

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

80

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.

81

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

82

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.

83

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

84

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

D’Sauza, 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: 54–78.

Fluck, Z. 1999. The dynamics of the managers-shareholders conflict. Review of

Financial Studies 2: 379-404.

85

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.

86

Holderness, C.G. 2003. A survey of blockholders and corporate control, Economic

Policy Review 9: 51–64.

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: 1-

33.

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.

87

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: 1-

25.

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.

Moh’d, 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.

88

Miller, M., and Modigliani, F. 1958. The cost of capital, corporation finance and the

theory of investment. American Economic Review 48: 261–297.

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.

89

Shleifer, A. and Vishny, R.W. 1997. A survey on corporate governance. Journal of

Finance 52: 737–783.

Shleifer, A. and Vishny, R.W. 1986. Large stockholders and corporate control.

Journal of Political Economy 95: 461–488.

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.

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