Upload
doh-siew-fong
View
226
Download
3
Tags:
Embed Size (px)
Citation preview
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