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

thanks to all my friends that helped. constructive comments. This research would not have been possible without the constructive comments.ACKNOWLEDGEMENT All my praises and gratitude to Allah. I would like to thank my parents. have had helped me tremendously in the successful completion of this research. criticism and suggestion throughout the duration of completing this research. Thanks for giving a great support throughout the duration of my studies and unceasing prayers for my success. I would like to acknowledge my debt to Associate Professor Dr. for His kindness and for meeting me with many wonderful people who. holding her responsible for any deficiencies remains in this research. who have been a continuous source of inspiration and encouragement. of course. suggestion and encouragement received from my supervisor who has read the various draft. without. the Merciful. Yusnidah Ibrahim. In particular. iii . In addition. support and provided insight and useful ideas. with His Grace. Thank you.

PERMISSION TO USE In presenting this dissertation as a partial fulfillment of the requirements for a postgraduate degree from Universiti Utara Malaysia. Postgraduate Studies. for scholarly purposes may be granted by my supervisor or in other absence by the Dean. 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. in whole or in part. Request for permission to copy or to make other use of materials in this dissertation. I further agree that permission for copying of this dissertation in any manner. and College of Business. 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. Postgraduate Studies College of Business Universiti Utara Malaysia O6100 Sintok Kedah Darul Aman . I agree that the university’s library may take it freely available for inspection. in whole or in parts should be addressed to: Dean.

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

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

2 1.2 Lintner Dividend Stability Theory and other Related Theories/Models 2.4 1.TABLE OF CONTENTS DECLARATION PERMISSION TO USE ABSTRACT (BAHASA MELAYU) ABSTRACT (ENGLISH) ACKNOWLEDGEMENT TABLE OF CONTENTS LIST OF TABLES LIST OF ABBREVIATIONS i ii iii iv vii viii CHAPTER 1: BACKGROUND OF STUDY 1.1 M&M Irrelevant Dividend Theory and other Related Theories/Models 13 12 12 2.1 1.1 2.2.3 2.3 1.4 Empirical Literature Conclusion 17 20 49 iv .2.6 Introduction Problem Statement Objective of the Study Significance of the Study Limitation of the Study Conclusion 1 4 8 9 10 11 CHAPTER 2:LITERATURE REVIEW 2.5 1.2 Introduction Theoretical Literature 2.

4 Sample Description and Data Collection Models on Dividend Policy 3.4.1 4.1 Ownership Concentration 3.4 Ownership Dispersion 3.5 Conclusion v 61 61 64 67 67 69 71 75 .4 Introduction Descriptive Analysis Correlation Analysis Regression Analysis 4.2 Earnings 3.5.2 Serial Correlation and Heteroscedasticity Test 4.3 3.2.3 Ownership Concentration 3.6 Managerial Ownership 3.2 4.7 Foreign Ownership 3.3 Institutional Ownership 3.1 Dividends 3.2.5.5.4.5.2.4 Managerial Ownership 3.4.5 Institutional Ownership 3.1 The Full Adjustment Model 3.3 The Waud Model 3.4.3 Regression Results 4.2 Introduction Research Framework 3.2.4.CHAPTER 3: RESEARCH METHOD 3.5 Foreign Ownership 3.5.3 4.5.2 Ownership Dispersion 3.5 Measurement of Variable 3.6 Conclusion 50 50 51 51 51 52 52 53 54 54 55 56 57 57 58 58 58 58 59 59 59 CHAPTER 4: ANALYSIS AND FINDINGS 4.1 3.4.5.2.2 The Partial Adjustment Model 3.1 Multicollinearity 4.

2 5.6 Introduction Overview of the Research Process Summary of Findings Implications of the Study Direction for Further Studies Conclusion 76 76 78 80 81 82 REFERENCES APPENDICES 83 vi .4 5.1 5.3 5.CHAPTER 5: CONCLUSION 5.5 5.

5: Durbin-Watson and Heteroscedasticity Diagnostic Test Results of Multiple Regression Analysis of Dividend Policy Models vii .1: Table 4.LIST OF TABLES Table 2.3: Summary of Empirical Literatures Summary Descriptive Statistic Pearson Correlation Matrix among the Variables Variance Inflation Factor of Variables (tolerance value is given in the parentheses) Table 4.1: Table 4.2: Table 4.4: Table 4.

LIST OF ABBREVIATIONS E : Earnings Earning Change Dividends Ownership Concentration Ownership Dispersion Institutional Ownership Managerial Ownership Foreign Ownership The Full Adjustment Model The Partial Adjustment Model The Waud Model ECHG : D : CONC : DISP : INST : MNG : FOR : FAM : PAM : WM : viii .

they become a shareholder of the business and to that extent they will have certain entitlements. used to retire debt or repurchase shares.CHAPTER ONE BACKGROUND OF STUDY 1. including the right to receive dividend payments. or distributed to shareholders in the form of dividends. 1 . When investors buy an ordinary share in a company. Besides. Dividends are a way for companies to reward shareholders for their investment and risk-bearing. 2007). dividends will be distributed in the form of cash. Dividends are defined as a form of rational income distribution offering to shareholders (Baker et al. Nevertheless. dividends also give shareholders additional returns in addition to capital gains. Dividends are decided upon and declared by the board of directors. this pay-out is not guaranteed and the amount that shareholders will receive varies from company to company and year to year. though it can also come in the form of stock dividends. Normally.1 INTRODUCTION Incomes are earned by successful companies. These incomes can be invested in operating assets.

Under the imputation system. According to Section 365 of the Act. final dividends are declared before the books are closed and will be paid the following year. 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. On the other hand.” 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. 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. These interim dividends are paid out of undistributed profits brought from previous periods. In Malaysia. a Malaysian 2 . the unique characteristic of dividends in Malaysia is the tax exemption feature. 1965. These dividends usually accompany the company’s interim financial statements.Generally. Thus final dividends will appear as dividends payable or proposed dividends under current liabilities in the balance sheet of that period. With effect from the year of assessment 2008. there are two types of cash dividends. a single-tier income tax system will replace the imputation system. Besides. 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. “No dividend shall be payable to the shareholders of any company except out of profits or pursuant to Section 60. Interim dividends are declared and distributed before the company’s annual earnings are known. which are interim dividends and final dividends.

Following their irrelevance dividend policy hypothesis many explanations have been provided in order to solve the socalled dividend puzzle. the tax-adjusted theory. which is accumulated as dividend franking credits (Section 108 credits). researchers have yet to reach a consensus on why firms pay dividends and what determines the payout ratio. dividend policy has been an issue of great interest in the finance literature. profits are only taxed at the company level. under the single-tier system. they are entitled to a tax credit for the tax already paid by the company in respect of the income. 3 . thus. dividends paid under this system will be tax-exempt in the hands of shareholders. The information asymmetries encompass several aspects. When shareholders receive taxable dividends. or behavioral factors.resident company is required to deduct taxes at the prevailing corporate tax rate on taxable dividends paid to its shareholders. Despite a large body of literature on dividends and payout policy. 1961). agency costs and the free cash flow hypothesis. This tax is already accounted for through the tax paid by the company on its taxable profits. including the signaling models. Those credits are then used to offset the shareholder’s tax liability. Some of the theoretical principles underlying the dividend policy of firms can be described either in terms of information asymmetries. Since Modigliani and Miller’s seminal studies (1958. However.

a dividend policy is a set of company rules and guidelines used to decide how much the company will pay out to its shareholders.2 PROBLEM STATEMENT Dividends are payments made by a company to its shareholders. Miller and Modigliani (1961) in their irrelevant dividend hypothesis. 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. usually after a company earns a profit. dividends are irrelevant since shareholders can create homemade dividends by selling a portion of 4 . Consequently. characterized among others by the nonexistence of taxes. 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.1. asserts that under perfect market conditions. 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. researchers have yet to reach a consensus on why firms pay dividends and what determines the payout ratio. Dividend policy has been viewed as an issue of interest in the financial literature and one of the most controversial topics in finance. 1959) while many others believe otherwise. Some researchers believe that dividends increase shareholder wealth (Gordon. Despite a large body of literature on dividends and payout policy. Thus. transaction costs and asymmetric information.

thereby bring them under the greater scrutiny of the capital market (Easterbrook.their portfolio of equities if they want cash and that there is a tradeoff between current dividends and future capital gain.. As a result. Starting with Jensen and Meckling (1976). 1984). the payment of dividends provides the incentive for managers to reduce the costs associated with the principal/agent relationship. 2003). 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. a considerable amount of theory and model are suggested to explain the dividend policy of companies. internally generated cash flows are no longer sufficient to satisfy the needs of the companies. Taking into consideration various capital market imperfections. 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. researchers have been addressing the agency problem in finance from many angles. Therefore. extensive research has been 5 . Unexpected changes in dividend policy are used to mitigate information asymmetries between managers and owners (Frankfurter and Wood Jr. Nowadays. companies will visit the capital market more frequently for financing needs. and they have incentives to signal that information to investors (Gugler. 2002). On the other hand. agency theory posits that by distributing resources in the form of cash dividends.

agency costs are zero in a 100% owner-managed firm. this finding is still inconclusive since other studies have questioned the validity of this finding. As a company’s ownership structure changes and ownership is separated from control. The nature of monitoring and bonding contracts. (1996) had regressed five factors as a proxy for agency costs on the dividend payout ratio. shirking and suboptimal investment. For example. they will attempt to act in his or her own self-interest. However. incentive alignment problems become more important. 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. Self-motivated management behavior includes direct expropriation of funds by the manager. Agency costs happen because of conflicts of interest between agents and shareholders. Noronha et al. Recent studies emphasize the potential conflicts of interest between controlling shareholders 6 . Therefore. consumption of excessive perquisites. It is assumed that if managers and shareholders are left alone. but they found that the dividend policy is not the product of an attempt to mitigate the free cash flow problem. 1976). Agency theory has also brought various external and internal monitoring and bonding mechanisms to the forefront of theoretical discussion and empirical research.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.

(2001) and Holderness (2003) argued that when large owners gain nearly full control of the corporation. Nevertheless. this would imply a lesser role for corporate payout policy to address agency problems between corporate insiders and outside shareholders. Faccio et al. in the presence of large shareholders.and other shareholders. firms with large controlling shareholders may exhibit a different type of agency conflict.124). their findings show a low explanatory power (between 0. they prefer to generate private benefits of control that are not shared by minority shareholders. On the other hand. foreign ownership and managerial ownership. few studies investigated the agency and ownership-based explanations of dividend policy. However.118 and 0. namely ownership concentration. They had used four types of ownership. 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. using four 7 . For example. However. On the other hand. Shleifer and Vishny (1997). one study by Mat Nor and Sulong (2007) had examined the relationship between ownership structure and dividends in Malaysia. namely the expropriation of minority shareholders by majority shareholders. 1986). Zhang and Keasey (2002) that examined the link between corporate dividend policy and the ownership of shares by institutional investors and managers. 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. a study in UK by Short. Hence. government ownership. Despite a great deal of prior research on the subject.

These four models. the partial adjustment model. this situation brings up a question whether it is true that ownership structure has a low impact on corporate dividend policy in Malaysia. 8 . the full adjustment model.3 OBJECTIVE OF THE STUDY Main Objective: • To investigate the adoption of agency costs theory in explaining dividend policy in Malaysian listed companies. Thus. 1. have been modified by the addition of dummy variables representing institutional and managerial ownership.843 and 0. Therefore. Their study is the first example of using well-established dividend payout models to examine the potential association between ownership structures and dividend policy.models of dividend policy. which describe the adjustment of dividends to changes in several measures of corporate earnings. this study attempts to examine the hypothesized relationship between corporate dividend policy and the various types of ownership structure by using dividend payout models. the Waud model and the earnings trend model found a very high explanatory power (between 0. 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.993).

Specific Objective: • To examine the relationship between various ownership structures based agency cost proxies on dividend policy.4 SIGNIFICANCE OF THE STUDY This study contributes to the growing body of survey research on dividend policy. the Full Adjustment Model. especially in explaining the ownership structure policy to reduce agency conflict. namely. • To identify which dividend model is superior in explaining the corporate dividend policy with variables associated with ownership structures. 9 . 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. the Partial Adjustment Model and the Waud Model. this study is expected to support the agency theory. These three types of dividend models had been modified to account for the possible effect of ownership structure and dividend policy. In addition. For example. Consequently. 1. the current study not only updates previous research by Mat Nor and Sulong (2007) but is also applied in a different model. • To identify which agency cost proxy is dominant in influencing dividend policy over the company.

Lastly. The shorter period of study may not be representative of the way companies operate their business cycle. 1. 10 . 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. Besides that. • The data for ownership structure was gathered from the list of the thirty largest shareholders disclosed in the company annual report. • The study only covers 150 public-listed companies in the selected sectors. Thus. the results cannot be treated as conclusive for all sectors. a longer period of study might be good to provide better results for this research. the results may not necessarily be applicable to privately-held companies. this study is also important in helping policy makers and companies to appropriately address the issues of agency costs. Hence. since the study was limited to publicly-held companies. Consequently.this study would assist each ownership class to understand the explanation of the agency relationship.5 LIMITATIONS OF THE STUDY • The main limitation of this study is that the data period covers only on the year 2007. the data may not be representative of the entire company.

1956) and the Waud Model (1966). namely. the third chapter describes the data. develops the theoretical model and also discusses the research framework. the Partial Adjustment Model (Litner.1. It examined the adoption of agency costs theory through ownership structure and dividend policy. Then. the Full Adjustment Model. powerful message about future prospects and performance. Dividends send a clear. The next section of the study briefly reviews the theoretical and empirical literature. Dividends are important for more than income generation since it also provides a way for investors to assess a company as an investment prospect. Significant results could act as guidance for companies and policy makers to appropriately address the issues of agency costs.6 CONCLUSION Dividends distribution is one of the simplest ways for companies to communicate their financial well-being and shareholder value. This study tests the relationship of ownership structure and corporate dividend policy via three types of dividend models. 11 . Chapter Four will reveal the empirical results while the summary and conclusion of the study are presented in Chapter Five.

3 discusses the empirical literature which is arranged in chronological order while Section 2. Section 2. 2. Section 2.4 concludes the chapter. 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).1 INTRODUCTION This chapter discusses the theoretical and empirical literature on dividend policy. followed by agency cost argument.2 THEORETICAL LITERATURE Theoretical literature on corporate dividend policy centered around two classic works.CHAPTER 2 LITERATURE REVIEW 2. Then. 12 .2 reviews the theoretical literature of dividend policy beginning with the irrelevance proposition by Miller and Modigliani.

Jensen and Meckling defined agency relationship as a 13 . several assumptions were made. M&M’s proposition was strongly supported by Friend and Phuket (1964) and Black and Scholes (1974). 2007). 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. and distribution of income between dividends and retained earnings has no effect on the firm’s cost of equity (Foong. proposed that dividend policy was irrelevant. One branch of this literature has focused on an agency-related rationale for paying dividend policy.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).2. investors and managers have asymmetry information about the firm’s future prospect. the origin of agency theory lies on the separation of ownership and control. including: no personal or corporate taxes. Nevertheless. The agency models of payout relax the original M&M’s assumption about the independence of dividend and investment policies of the firm.2. no stock flotation or transaction costs. financial leverage has no effect on the cost of capital. Zakaria and Tan. The main conclusion of this paper is that firm’s capital budgeting policy is independent of its dividend policy. According to Jensen and Meckling (1976). However. 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. subsequent literature advances several theoretical justifications for firms’ payout choices. Therefore.

thus offering a rationale for the distribution of cash resources to shareholders. then managers must reduce agency costs and reveal new information in order to secure the new funding. The second is the problem of risk sharing that arises when the principle and agent have different attitudes towards risk. If both parties to the relationships are utility maximizers. Different in risk preferences leads to different policy decisions and disregard the value maximizing activity as the economics pursued. Perry and Rimbey (1995). there is good reason to believe that the agent will not always act in the best interests of the principal. 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. agency theory relates to dividend policy stems from the works of Rozeff (1982) and Easterbrook (1984).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. According to Moh’d. Moreover. Mahadwartha (2002) draws on Fama (1980) and Eisenhardt (1989) had augmented that agency theory concerned with resolving two problems that occur in agency relationship. Meanwhile. 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. According to Rozeff. if a firm is forced to raise external capital to replenish funds paid out in dividends. a dividend payment may act as one form of bonding mechanism 14 .

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

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

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

15

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

16

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

2.2.2 Lintner Dividend Stability Theory and other Related Theories/Models

Lintner (1956) is among the pioneers to theorise on corporate dividend behavior through Lintner stability dividend theory. Lintner had conducted a classic series of interviews with 28 corporate managers about their dividend policy. He then proceeded to formulate a seemingly logical model of how companies decide on dividend payments. Dorsman, Montfort and Vink (1999) summarized 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*ti = rEti (1)

17

1 ) i = a + β i 1 D ( t . When the profit changes the actual amount of dividend paid differs from the optimal amount that follows out of (1). the profit company i made in year t. This is what can be seen in the next equation: D t i –D ( t .1 ) i ) with. Lintner also introduced a constant term. Higher values of C correspond to higher velocity in adjusting the dividends. To compensate for this difference the company will gradually adjust the dividends. c = Velocity at which a company adjusts the dividend (2) The velocity (C) will be between 0 and 1. the optimal amount of dividend as a percentage of the profit. D*ti r Eti = = = the target level dividend of dividend for fund i year t.D (t .with. This constant term together with equations (1) and (2) form the Lintner partial adjustment model: D t i –D ( t .1 ) i + β i 2 E t i + µ t i (3) 18 . Because it is assumed that corporations are reluctant to decrease dividends. this constant term would have to be positive. The value of r will be between 0 and 1 since companies usually won’t pay more dividends then that there was profit.1 ) i = c(D* t i . for fund i.

It assumes that the target dividends are proportional to the long-run expected earnings (E*). Waud (1966) proposes a second order rational distributed lag function for detailed derivation of the model.1 ) i ) (6) 19 . According this model.with.1 ) i ) + µ t i (5) As a result.D ( t .1 ) i = a +c(D* t i – D ( t . the formation of expectations follows an adaptive expectation model: E* t i –E ( t . dividends are the results of a `the partial adjustment' and `the adaptive expectations'.1 ) i = d(E t i – E* ( t . βi 1 βi 2 µ ti = = = -Ci Ci ri The random disturbance Besides the Lintner partial adjustment model. the equation: D*ti = rE*ti (4) On one hand. Thus. the actual dividend change will follow a partial adjustment model: D t i .

and the conclusions reached support and strengthen the results of Rozeff. The optimal dividend payout minimizes the sum of these two costs.3 EMPIRICAL LITERATURE Miller and Modigliani’s (1961) hypothesis on the irrelevance of dividend policy is not compatible with empirical evidence. 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. Llyod. Besides that. They provide a strong support for their hypothesis of dividends as a partial solution to agency problems. Rozeff (1982) had initiated the adoption of agency cost in dividend determinant. 20 . 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. he shows that dividend payout is negatively related to the percentage of stock held by insiders. This fact implies there must be additional factors that compel firms to pursue a consistent policy of paying dividend. 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.2. 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. He develop a model of optimal dividend payout in which increased dividends lower agency costs but raise the transaction costs. Based on 1000 sample of companies from 1974 until 1980.

The results reveal that six significant factors can be used to explain corporate payout policies which include agency cost factor. Alli et. Solberg and Zorn (1992) examine the determinants of cross-sectional differences in insider ownership and dividend policies in the U.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. Their empirical results support the hypothesis that levels of insider ownership differ systematically across firms. 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).9 percent).Jensen. They analyzed firm data at two points in time. The average firm size and capitalization of the final sample was representative of New York Stock Exchange (NYSE) listed firms. 1982 and 1987on 565 and 632 firms respectively. this observation supports Rozeff’s proposition that the benefits of dividends in reducing agency costs are smaller for firms with higher insider ownership. The results of the analysis support the proposition that financial decisions and insider ownership are interdependent.S. Specifically. with the largest share from the chemical and allied products industry (13. Therefore. The sample of 150 firms came from 34 industries. 21 . insider ownership has a negative influence on firm’s dividend levels. These policies are found related directly and indirectly through their relationship with operating characteristics of firms.

Their tests are conducted in each of two recent five year periods. utilities are arguably somewhat more insulated from the discipline of other monitoring mechanism for controlling agency costs. This relationship is more pronounced in all-equity firms since they lack one mechanism for controlling these agency costs. Hansen et. 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. Their results indicate that dividend yields and payout ratios of all-equity firms are significantly higher than those of levered firms.Agrawal and Jayaraman (1994) use the sample of all-equity and levered firms which consist of 71 matched pairs. They also found that within the group of all-equity firms.al (1994) tests the relevance of monitoring theory for explaining the dividend policies of regulated electric utilities. These results are robust to the choice of the time period used for measuring these variables. which is characterized by high but declining industry wide investment growth and financing and the more recent five year period ending in 1990. Their findings show that utilities faced with higher regulatory and managerial conflict. All equity firms are defined as those that use no long term debt throughout a continuous five year period. Their 22 . They focus on this industry partly because relative to industrial 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. the first five year period ending in 1985. which is characterized by secular asset growth yet low industry-wide growth. Therefore. lower flotation costs and lower asset growth pay proportionally greater dividends.

More support and further contribution to the agency theory of dividend debate. managers’ portfolio diversification losses. 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). These authors introduce a number of modifications to the cost minimization model including industry dummies. bank size and standard deviation of bank equity returns on the three financial policy variables of managerial stock ownership. The results of a Weighted Least Squares regression.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. leverage and dividend yield. is provided by Moh’d. Higher dividend payouts are observed when managers hold a low percentage of firm shares. and as the outside ownership becomes more dispersed. Perry and Rimbey (1995). 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. institutional holdings and a lagged dependent variable to the RHS of the equation to address possible dynamics. 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. The study examines the largest 104 US banks during the period 23 .

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.1985-1989. which. simultaneity is contingent on the applicability of Easterbrook’s (1984) monitoring rationale for paying dividends. in turn is hypothesized to depend on the existence of alternative sources of monitoring. Noronha. For the subsample with alternative mechanisms in place the payout rates of firms are not related to proxies for agency cost variables. 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. In the model. confirm the sample-specific validity of the monitoring rationale. He used sample of 349 firms worldwide to determine the relationship between dividend payout and agency cost. dividends and leverage). Shome and Morgan (1996) develop an agency-cost framework for the simultaneous determination of a firm’s capital structure and dividend decisions. The dividend payout variable used in the 24 . Evidence support the view that bank managers consider agency costs while trying to determine the most appropriate financial policies (managerial stock ownership. 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. D’ Sauza and Saxene (1999) examine the effects of agency costs on an international firm’s dividend policy.

Multiple regression analysis was performed and the result reveals the statistically significant and negative relationship of dividend payout with the explanatory variable institutional holdings. capital expenditures on plant and equipment.study is a three year average for the period 1995 to 1997. They had control seven factors believed to influence dividend policy namely insider ownership. Nevertheless. dividend payout is positively related to institutional ownership because institutions prefer dividends prefer dividends over capital gains under the differential tax treatment. Ang. they found a contradict results with agency cost hypothesis but supporting tax based hypothesis. operating income to assets and target dividend yield. these findings are consistent with those of prior studies using United States’ data. standard deviation of return on assets. Based on the Jensen and Meckling agency theory. They 25 . revenue growth. shareholders incur agency costs resulting from management’s shirking and perquisite consumption. Cole and Lin (2000) measure absolute agency costs by observing a zero agencycost base case as a reference point of comparison for all other cases of ownership and management structures. using the Tobit analysis. Lee and Suk (1999) also empirically examine the effect of institutional on corporate dividend policy. Therefore. According to tax based hypotheses. the zero agency cost base is the firm owned solely by a single owner-manager. Han. ratio of debt to assets. They utilize a sample of 303 firms during the 1988 to 1992 period. When management owns less than 100 percent of the firm’s equity. while the institutional holdings pertain to the year 1997.

He had modified the Rozeff’s cost minimization model by introducing a business group affiliation namely foreign ownership. and these costs increase as the equity share of the owner-manager declines. Hence. capital market monitoring is also important when the dispersion of ownership increases since the more widely the ownership spread. the greater the need to induce capital market monitoring. Manos (2002) had investigated the agency theory of dividend policy in the context of an emerging economy. insider ownership and ownership dispersion as a proxy for agency cost theory. agency costs increase with a reduction in managerial ownership. 26 .employ a sample of 1708 small corporations and provide a direct confirmation of the predictions made by Jensen and Meckling (1976). the greater need for outside monitoring. The positive relationship between foreign and payout indicates that the greater the percentage held by foreign institutions. institutional ownership. India. hence. the more acute the free rider problem. as predicted by Jensen and Meckling. The model is estimated and tested on a cross-section of 661 non-financial companies listed on the Bombay Stock Exchange. the evidence of a positive relationship between institutional and the payout ratio is consistent with the preference for dividends related prediction. Agency costs are indeed higher among firms that are not 100 percent owned by their managers. The results reveal a positive impact of all business group affiliation to payout decisions. Further. Besides that.

Study by Short. the paper presents the first results for the UK. 1956). the positive relationship between dividends and insurance companies would suggest that they are relatively poor at monitoring compared to 27 . This study is conducted on a sample of 211 firms listed on the London Stock Exchange Official List for the period 1988 to 1992. 1966) and the Earnings Trend Model. where the institutional framework and ownership structures are different from the US. 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. The study by Khan (2006) investigates how the ownership structure of firms affects their dividends policies. pension funds and other financial institutions) holding more than 0. the Waud Model (Waud. However. They had modified the Full Adjustment Model. A significantly negative relation between dividends and ownership concentration result appear to corroborate Rozeff’s model. insurance companies. A key contribution of this article is that it exploits extremely rich ownership data on all beneficial owners (individuals. Moreover. Zhang and Keasey (2002) is the first example of using wellestablished dividend payout models to examine the potential association between ownership structures and dividend policy. 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. which is generally associated with better incentives to monitor.25% of any given firm’s equity. the Partial Adjustment Model (Lintner. dividends fall when the degree of ownership of ownership concentration increase.

hence the relevant agency problem of concern seems to be the one that arises from the conflicting interests of large shareholders and minority shareholders. 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. they find that firms with high ownership concentration pay lower dividends. Second. Their analysis uncovers a number of agency conflicts. This pattern is consistent with their lower payout and the assumption that dominant shareholder extract private benefits from resources under their control. The Tobit regression results support the prediction that higher level of ownership concentration is associated with a higher probability of expropriation of outside 28 . Ownership structure in Italy is highly concentrated. tightly controlled firms are less likely to increase dividends when profitability increases and when operating profits are negative. 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). First. they also find that tightly controlled firms are more likely to omit dividends when investment opportunities improve which protect the interest of current shareholders.individual 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. this decision reduces the likelihood of requiring further funding that would benefit outside investors. Clearly.

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.shareholders. Furthermore. however. There are private benefits to the larger shareholders of holding larger amounts of cash. Using cross-sectional and pooled regression. Thus. free cash flow and degree of collateralizable assets on the dividend payout ratio. The results support the agency model. dispersion of stockholders. 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. The crisis year of 1998 is omitted. 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. 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. do not play a prominent role in a firm’s payout policy. the paper measures the effect of the percentage of insider ownership. Domestic intuitional investors. lower dividend payouts will increase the ability of the large shareholders to expropriate the outside minority shareholders. Rafiq and Sharp (2007) investigates the influence of agency cost variables on dividend policy during the pre and post of the 1998 financial crisis. The study finds agency cost variables to have only a modest 29 . higher foreign ownership is associated with a greater dividend payout.

This failure captures an aspect of an emerging market such as Dhaka that differs fundamentally from more evolved markets. 30 . one of the emerging economies to mitigate potential managerial discretionary behavior and free cash flow problems. This data from a sample of 406 firms employs a multiple regression analysis since the data are cross sectional.explanatory power during the pre-crisis period and none in the post-crisis period on the Dhaka Stock Exchange. but the results are significant effect on dividends in 2005. but with minimum impact. thus an agency cost is insignificant in influencing the dividend policy. 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. The results also suggest that managerial ownership does play an active monitoring role in Malaysia. The failure of agency cost variables to influence dividends may indicate an impediment to efficient capital information. Nevertheless. the negative significant effect of foreign ownership on dividends fails to support the agency argument. This result might be due Bangladesh firms having highly concentrated ownership structure. The results reveal that concentration ownership has a significant positive effect on dividends for both years. 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. Results of foreign and managerial ownership on dividends show insignificant relationship in the year 2002.

institutional ownership ratio and free float ratio. Further. El-Masry and Elsegini (2008). They conclude that dividend rates are higher in Europe when there are multiple large shareholders suggesting that these large shareholders dampen expropriation in Europe. Ownership structure is measured by four variables namely managerial ownership ratio. and they are either shareholders or debt holders. This is due to most of cases.Obema. blockholder ownership ratio. This evidence in Tunisian context strengthens the argument of the positive role of multiple large shareholders in corporate control. The results indicate that there is a significantly negative correlation between institutional ownership with the level of dividend distributed to shareholders. 31 . The results show that only institutional ownership has a significant relationship with dividend policy. institutional investors are banks. 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. One explanation could be that the institutional blockholders voted for higher payout ratios to enhance managerial monitoring by external capital markets. the results also show that the higher ownership of the five largest shareholders leads to the higher of dividend payment. This study uses dividend per share as a dependent variable and ownership classes as an independent variables. They prefer paying interests to themselves than distribute dividend to all shareholders. examine the effect of ownership structure on corporate dividend policies of a sample of top Egyptian listed companies.

This condition is supported by the better economic atmosphere of Indonesia. 32 . institutional ownership influence dividend payout negatively which is contradict with the agency argument. This implies that dividend payment is rise in order to decrease agency problem when there is separation function between corporate ownership and corporate control. This might be due to institutional ownership tend to do other investment or expand their business that to pay shareholders. The results reveal a significant negative effect of insider ownership on dividend policy. Nevertheless. 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.Harjito (2009) examine the influences of agency factors to dividend payout ratio. which offers good opportunities to invest.

Table 2.1: Summary of Empirical Literatures MEASUREMENT OF VARIABLE DEPENDENT Michael S. significant STOCK: positive. GROW 1: 5 year average of growth rate of revenues. significant . 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. significant BETA: negative. significant GROW 1: negative. Size: 1000 Rozeff (1982) Multiple Regression DPR: 7 year average over a period of 1974 to 1980 INDEPENDENT INS: percentage of common stock held by insiders. significant GROW 1: negative.Lloyd. significant GROW 2: negative. John S. Jr Location: U. BETA: 5 year value-weighted beta of weekly data on the stock’s returns. Jahear.S Period: Ordinary Least Squares Regression DPR: average payout ratio last 7 years from value line - INS: negative. CONTROL INS: negative. significant AUTHOR SAMPLE METHOD FINDINGS William P. STOCK: natural logarithm of number of common stockholders INS: percentage of common stock held by insiders.

significant DEBT: negative. but significant only in 1987 PROFITABILITY: positive. Zorn (1992) Period: 1982 and 1987 Size: 565 (1982). significant SIZE: positive. but significant only in 1987 BUSINESS RISK: negative. GROWTH: 5 year growth rate in sales. BETA: 5 year value-weighted beta of weekly data on the stock’s returns. significant BETA: negative. PROFITABILITY: ratio of operating to total assets. significant INVESTMENT: . significant Gerald R. SIZE: natural logarithm of sales for 1983. BUSINESS RISK: standard deviation of the first difference in operating income divided by total assets. Page September (1985) 1984 Size: 957 GROW 2: 5 year average of Value Line’s forecast of growth rate of revenues.and Daniel July – E. Solberg and Thomas S. Jensen. GROW 2: negative. significant STOCK: positive. STOCK: natural logarithm of number of common stockholders. Donald P. significant GROWTH: negative. 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. 632 (1987) INSIDER: negative.

. 1987 and 1988 was used as a proxy for expected capital expenditure in 1985. significant INSIDER: positive. INSTHOL: proportion of institutional holdings as a proportion of total outstanding shares.INVESTMENT: expenditure for plant.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. significant EXCAP: negative. EXCAP: average realized value of capital expenditure for 1986. INSIDER: proportion of insider holdings as a percent of total outstanding shares. significant STAB: positive. negative. significant GROWTH: negative. A. Ramirez (1993) Location: U. significant LNTA: insignificant BETA: negative. Alli. and R&D as a percentage of total assets. significant HOLDING: insignificant INTANG: negative. Kasim L. BETA: firm’s beta calculated against the CRSP equally weighted index using sixty months of data. significant STDCDE: positive. Qayyum Khan and Gabriel G. STDCDE: standard deviation of changes in the debt equity ratio using nine years data. significant CFV: insignificant SLACK: negative. equipment. significant INSTHOL: positive.

significant . 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.HOLDING: dispersion of ownership as given by the number of shareholders to total outstanding shares. 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). CFV: coefficient of variation of cash flow using nine years data. GROWTH: 5 year annualized growth rate in earnings. INTANG: collaterizable value of assets as represented by net plant to total assets.

COMMRANK: regulatory commission rank. significant OWNSHIP: negative. over the sum of all stockholder earnings over the same period COMMRANK: negative. OWNSHIP: stockholderownership concentration.Hansen.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. levered firm = 1. LEVERED: negative. significant . significant TAGROW: negative. GROWTH: average annual growth rate of sales over the previous 5 years. expressed as a percentage of the gross proceeds.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. FLOTCOST: firm’s historical average flotation cost incurred in selling common stock. significant FLOTCOST: negative.Shome (1994) Location: U. significant FCF: insignificant GROWTH: insignificant Robert S. significant α: negative. all-equity firm = 0. FCF: free cash flow divided by total assets. α: percentage of outstanding equity owned by directors and officers. measured by Herfindahl Index. Raman Kumar and Dilip K. TAGROW: growth rate in total assets.

STKHLDR: natural log of the number of shareholders EARNING VOLATILITY: standard deviation of the income before depreciation and amortization on total assets. significant GROW 2: negative. INSD: percent of common stock held by insiders.S Period: 1985 - Regression Dividend: 4 year average of the ratio of total dividends divided by the . β0: intrinsic business risk OLRISK: operating leverage risk. FLRISK: financial leverage risk. Larry G. significant INSD: negative. Perry and James N. significant INSTINV: positive. significant OLRISK: negative. significant β0: negative. Moh’d. significant EARNING VOLATILITY: positive. GROW 2: Value line forecast of future five-year revenue growth. 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. MANAGERS’ SIZE: natural log of firm sales.Mahmoud A. GROW 1: negative. INSTINV: percent of common stock held by institutions. significant STKHLDR: positive. significant FLRISK: negative. insignificant MANAGERS’ DIVERSIFICATION Jose MercadoMendez and Thomas Location: U. significant SIZE: positive.

LNAST: logarithm of asset size. Noronha and George E. significant FLOTATION COST: negative. Morgan (1996) Location: U. INS: percentage insider holding. insignificant SIZE: positive. SIZE: 5 year average total assets. VRET: variance of stock returns. significant GROWTH: . LOSSES: negative. BETA: negative.Willey (1995) 1989 Size: 104 total market value of the common stock DIVERSIFICATION LOSSES: equity risk premium divided by total equity risk. GR: ratio of forecast growth in book value equity to forecast return on equity. insignificant Gregory M. LNSH: logarithm of the number shareholders. DPR INS: insignificant LNSH: insignificant VRET: insignificant LNAST: insignificant GR: insignificant Juliet D’Sauza and Atul Period: 1995 1997 Multiple Regression BETA: beta GROWTH: past 3 years’ sales growth.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. FLOTATION COST: standard deviation of he average stock return.

MTBV: market to book value.K Saxena (1999) Size: 349 INSH: percentage of institutional. significant CXA: insignificant DTA: insignificant SDR: negative. significant Ki C. significant ISD: insignificant GRO: negative. significant OIA: insignificant TDY: negative. Suk Hun Lee and David Y. Han. insignificant INSH: negative.Suk (1999) Location: US Period: 1988 – 1992 Size: 303 Tobit Analysis DY: Dividend Yield INT: institutional ownership in percentage for firm . significant MTBV: insignificant ISD: insider ownership in percentage for firm GRO: firm geometric average of revenue during the 5 year period CXA: firm capital expenditures on plant and equipment as a percentage of assets DTA: ratio of debt to assets INT: positive.

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

The Partial Adjustment Model.one year. Hao Zhang and Kevin Keasey (2002) Location: London Period: 1988 – 1992 Size: 211 The Full Adjustment Model. The Waud Model and Earnings Trend Model. 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 . significant MDUM: negative. DIRS: percentage of equity shares held by directors of the company in one year. Tobit Regression Dti – D(t-1)i: changes in dividend INST: Institutional dummy variable MDUM: Managerial dummy variable INST: positive. measured by an approximation of the Herdindahl Index. Helen Short. PUBLIC: percentage of equity shares held by the public in one year. significant Kimie Harada and Pascal Nguyen (2006) Location: Japan Period: April 1995 DPR: dividends to operating income DIVEQTY: LHH: ownership concentration.

– March 2002 Observatio n: 6397 total dividend payments to book value of equity controlled by five major shareholders. Q2H: dummy variable indicating that ownership concentration is above the sample median. profits scaled GROW: positive by total assets DEBT: negative KD: positive Q: market to book value of assets GROW: percentage change in total assets DEBT: long term debt plus short term debt over total assets KD: dummy for affiliation with a business group SIZE: natural logarithm of sales. LEVERAGE: VOTING_RIGHTS1: negative. 2_LARGE_SHAREHOLDER: takes value 1 when the second . significant 2_LARGE_SHAREHO LDER: insignificant Luciana Location: Mancinelli Itali and Aydin Ozkan Period: Tobit Regression Dividend/ Earnings Ratio and Dividend/ Market VOTING_RIGHTS1: voting rights of the largest shareholder.

significant Douglas O. . ratio of the total debt to total assets. Cook and Jin Q. Foreign Ownership: positive. OTHER_LARGE_SHAREHO LDERS: takes value 1 when all shareholders but the largest one exceeding the 2% disclosure threshold own a fraction larger than 5% and zero otherwise. significant LEVERAGE: negative. significant MKT-TO-BOOK: negative. VOTING_RIGHTS _ALL: voting rights of all shareholders who own more than 2% disclosure. Institutional Ownership: ratio of total equity owned by domestic investors to the total equity. Jeon (2006) Location: Korea Period: 1999 – 2004 Size: 507 Regression DPR: total amount of cash dividends divided by the net income after taxes. AGREEMENTS: Foreign Ownership: ratio of total equity owned by foreign investors to the total equity. MKT-TOBOOK: Ratio of the market value of equity to the book value equity OTHER_LARGE_SH AREHOLDERS: insignificant VOTING_RIGHTS _ALL: insignificant AGREEMENTS: insignificant SIZE: positive.(2006) 2001 Observatio n: 139 Capitalization largest shareholder owns a fraction larger than 5% of the firms voting rights and zero otherwise. significant. Institutional Ownership: insignificant.

AGE: age of listing EPS: net income divided by number of share. MOWN: percentage of shares directly held by non independent executive directors LOGMCAP: logarithm function of market capitalisation. . insignificant in 2002 but significant in 2005 MOWN: positive. HI5: insignificant GOWN: negative. significant DOWNER: positive. significant COLLAS: positive. COLLAS: Ratio of net asset to fixed assets. insignificant in 2002 but significant in 2005 FOWN: negative. GOWN: sum of all shares held by government-controlled companies in the list 30 largest shareholders. TS: trade and INSIDE: negative. DOWNER: natural log of number of common stockholder. FOWN: sum of all shares held by foreign shareholders in the list 30 largest shareholders. insignificant in 2002 but significant in 2005 LOGMCAP: positive.A Sabur Mollah. Rfiqul Bhuyan Rafiq and Peter A Sharp (2007) Location: Banglades h Period: 1999 – 2003 Size: 153 Crosssectional and pooled Ordinary Least Square regression DPR: dividends paid divided by operating income INSIDE: proportion of held by insiders. 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. FCF: (net profit after tax – dividend + depreciation) / Total Assets. significant FCF: positive. HI5: Herfindahl Index 5 that is the squared sum of shares in the hands of largest 5 shareholders. significant AGE: negative.

Ahmed El-Masry and Sabri Elsegini (2008) Location: Egypt Period: 2003 – 2005 Size: 50 Regression DIVDECISIO N: dummy variable that set to one if company paid dividends. significant EPS: insignificant TS: insignificant IP: insignificant CP: positive. significant . significant Omneya Abdelsala m. 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.in the company services (dummy variable) IP: industrial products (dummy variable) CP: consumer products (dummy variable).

D. FCF: cash flow per unit of asset. ETA: percentage of equity owned by state. significant LEV: negative. . significant INST: negative. significant INSDO: negative. Mondher Kouki and Moncef Guizani (2009) Location: Tunisian Period: 1995 2001 Size: 203 Regression DPS: total dividend divided by the number of outstanding equity. significant PE:insignificant SIZE: log of total assets FCF: positive. Agus Harjito and Location: Jakrta Multiple Regression DPR: percentage of profit paid in INSDO: share percentage owned by director and commissioner. significant MAJ: positive. significant SIZE: negative. significant ETA: positive.ROE: return on equity PE: price earnings ratio FREEFLOAT: insignificant ROE: positive. significant OFO: negative. significant GROWT: positive. 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.

SD: variance of percentage of share ownership data. OFO: percentage of shares owned by other outside firms.Ambang Aries Yudanto (2009) Period: 2001 2005 the form of dividend. significant FG: negative. FG: upcoming year’s income growth level β: return volatility measurement of security toward market-risk significant SD: positive. significant β: insignificant .

A difference set of explanatory variables has been hypothesized to distinguish the companies’ specific characteristic that influence on the dividend distribution. some researchers have argued that dividends do matter. 33 . when capital markets are imperfect and when the assumptions made by M&M are relaxed. Nevertheless.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.2. hence firms should pursue an appropriate dividend policy.

5 provides definitions and measurements for each of the variables that are used in the study. institutional ownership.1 INTRODUCTION This study looks at the impact of ownership structure on corporate dividend policy. The description of the data sample development is presented in Section 3. theoretical and empirical.CHAPTER THREE RESEARCH METHOD 3. the impact of ownership structure on corporate dividend policy can be examined through the relationship between selected ownership variables and dividend policy.2. These include the framework of research as presented in Section 3. managerial ownership and foreign ownership and the expected relations as described below: 50 .6 concludes the chapter. Lastly. 3. Section 3. ownership dispersion.3. The ownership variables identified from the literature are ownership concentration.2 RESEARCH FRAMEWORK Based on the review of literature. Additionally.4 while Section 3. models of dividend policy are presented in Section 3. This chapter describes the methodology of this study.

agency costs will increase and the need for monitoring managerial action also increases. 3.3. Hence. On the other hand.2. 51 . to circumvent the problem a positive relationship was expected between ownership concentration and dividends.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. according to La Porta et al.2. a positive relationship was anticipated between institutional ownership and dividends. 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). (2000a) larger controlling shareholders could expropriate corporate wealth from other minority shareholders and enjoy private benefits instead of distributing dividends to shareholders. a positive relationship between ownership dispersion and dividend is expected. Therefore.2 Ownership Dispersion (DISP) The greater the number of shareholders will lead to the greater dispersion of ownership. If dividends can alleviate this problem.

4 Managerial Ownership (MNG) According to agency theory. managerial ownership has a potential to align the interest between managers and shareholders (Jensen and Meckling. management will tend to increase their own benefits such as increase director’s fees. 1984.3. rather than pay dividends. Besides. foreign investors who are well-informed and hold a substantial share can play their monitoring role on management and reducing the agency costs.2. 3. VARIABLE Ownership Concentration Ownership Dispersion Institutional Ownership Managerial Ownership Foreign Ownership EXPECTED RELATION Positive Positive Positive Negative Positive 52 . In such case. 1986). if a larger percentage of common shares are in the hands of managers.5 Foreign Ownership (FOR) According to agency theory. which is to reduce agency costs. However. companies are more likely to increase dividends (Easterbrook. there will be less influence from outsiders. Hence.2. 1976). dividends will be hypothesized to be negatively related with managerial ownership. it will be ineffective to use two tools at the same time for the same problem. employees’ salaries and bonuses. Jensen. and therefore. a positive relationship was therefore expected between foreign ownership and dividends. Thus. since the purpose of managerial ownership is the same as dividend policy.

earnings and different types of ownership structure data. In addition. Therefore. This pooled cross-sectional study employs annual data from 2005 to 2007. these companies are expected to be a representative of the four largest sectors in Bursa Malaysia.3 SAMPLE DESCRIPTION AND DATA COLLECTION The sample for the study includes 150 companies from four of the largest sectors (consumers.3. 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. These companies are selected based on proportionate stratified random sampling. 53 . industrial. annual data allows the study to capture more discretionary rather than autonomous behaviour. trading and services and properties) on the Main Board of Bursa Malaysia whose annual reports are available for the year 2007. Annual data is also more systematic since it represents the highest periodicity. The dividend and earnings variables were retrieved from DataStream financial database. This study utilised dividends. According to Mat Nor and Sulong (2009).

4 MODELS ON DIVIDEND POLICY Following the methodology of Short.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. Zhang and Keasey (2002).3. the Partial Adjustment Model (Litner. these models have been modified to account for the possible effects of ownership structure in determining the level of the corporate dividend. is given by: D t i –D ( t . Therefore. 3.1 ) i = α + r(E t i – E ( t . 1956) and the Waud Model (1966). Therefore. These models describe the adjustment of dividends to changes in several measures of corporate earnings. in this case.4. changes in earnings are considered as permanent. the relationship between the changes in earnings and changes in dividends. the model becomes: 54 . three dividend models were used to test the hypothesis of positive link between ownership structure and dividend policy: the Full Adjustment Model.1 The Full Adjustment Model (FAM) According to the full adjustment model. Nevertheless. Consequently. companies will adjust their dividends (D) to the new level of earnings (E) to achieve the companies’ desired payout ratio (r). for company i at time t.

the company adjusts only partially to the target dividend level. 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.1 ) i = α + r(E t i –E (t .1 ) )*CON C + r D I S P (E t i –E (t .D t i –D (t .1 ) )*MNG r FO R (E t i – E (t . the model takes the form: D ti –D ( t .4. according to the target payout ratio (r): D * t i = rE t i Nevertheless.1)i ) + µ t i Where a is a coefficient representing the refusal of managers to reduce dividends.1 ) )* FOR (Model 1 .1 ) ) + r C O N C (E t i – E (t .D (t . FAM) 3. 55 .1 ) )*DIS P + + r I NS T (E t i –E (t + 1 ) )*INST r M N G (E t i –E (t .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).1) i = a + c(D* t i . As a result. firms move towards the desired level of distribution gradually and dividends adjust only partially to the changes in earnings. In contrast.

3 The Waud Model (WM) The Waud model integrates elements of the both partial and full adjustment model.1 ) i = a +c(D * t i – D ( t . the actual dividend change will follow a partial adjustment model: D ti .4. 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. It believes that the target dividends D* are the proportional to the longrun expected earnings. E*: D * t i = rE * t i On one hand.1 )i = d(E t i – E* (t .D ( t.Assuming that companies with significant ownership classes have different target payout ratios (r).1) i ) + µ t i The formation of expectations follows an adaptive expectation model: E* ti –E (t .1) i ) 56 . PAM) 3.

the model becomes: D t i – D (t . assuming a possible difference in payout ratio for firms with different ownership classes.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.According to this model.µ t i (Model 3.2 ) i . dividends are the results of a `partial adjustment' and the `adaptive expectations'. 3.1 ) i = ad + cdrE t i + cdr C O N C E t i *CONC + cdr D I S P E t i *DIS P + cdr I N S T E t i *INS T + cdr M NG E t i *MNG + cd r FO R E t i *FO R + (1-d -c)D (t . WM) 3. 57 .1 ) i – (1 -d)(1-c)D (t .5. Therefore.3.1 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.

the squared sum of shares in the hands of the five largest shareholders (referred as CONC1). 2009). CONC1 and CONC2 will be used one at a time. 3. ownership dispersion is defined as the ratio of the number of shareholders to total outstanding shares. a new proxy was used.4 Ownership Dispersion (DISP) Following Alli et al. (1994).3. Amidu (2006) and Kouki and Guizani (2009) defined institutional ownership as a percentage of equity 58 . considering that Malaysian companies have highly concentrated ownership (Nor and Sulong. 3. 3. In addition. However.5.3 Ownership Concentration (CONC) Following Hansen et al. thus.(1993) and Moh’d et al.5 Institutional Ownership (INST) Alli et al.5. which is the summation of the percentage of shares controlled by two major shareholders (referred as CONC2). (1993). (1995).5. that is. ownership concentration was measured by the Herfindahl Index 5 (HI5). since both CONC1 and CONC2 are used to measure a similar variable.5. Harada and Nguyen (2006) and Khan (2006).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.

5.owned by institutional investors such as insurance companies. Three types of dividend models were used to see whether specific ownership classes would influence the dividends 59 . Nevertheless. this study used the total percentage of institutional ownership in a list of the thirty largest shareholders as the measure of INST. pension funds and financial companies. industrial. managerial ownership was measured by adding the total percentage of shares directly held by non-independent executive directors in the company. either held through nominee companies or other corporate foreign share holdings. 3. was identified to calculate the total percentage of foreign shareholdings (FOR).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.5. unit trusts. 3. trading and services and properties) on the Main Board of Bursa Malaysia.6 CONCLUSION This study utilized income statements and ownership structure data for 150 companies listed in the four largest sectors (consumers. 3. mutual funds.6 Managerial Ownership (MNG) Following Mat Nor and Sulong (2007).

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

the correlation analysis is illustrated in Section 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. which constitutes the main findings of the study.4 which discusses the result of the regression analysis. the full adjustment model.42 percent 61 . the partial adjustment model and the Waud model. The analysis commenced with a data and variable description as presented in Section 4.2. this chapter ends with the conclusion in Section 4.CHAPTER FOUR ANALYSIS AND FINDINGS 4. This is followed by Section 4.2 DESCRIPTIVE ANALYSIS Table 4. Focusing on the dependent variable. Subsequently. 4. it can been seen that the standard deviation for dividends is 7. Lastly. namely.5.3 to reveal the strength of the relationship between the variables that are utilized in the study.1 presents a summary of the descriptive statistics for each of the hypothesised variables for the 150 companies covered in this study.

The mean percentage of the CONC1 is 38.11 percent. the range of firm ownership concentration represented by the percentage of ownership owned by five largest shareholders (CONC1) is from 7. resulted the standard deviation of 34. The earnings per share show an average of RM 0.159. Thus.210.81 percent.87 percent to 73.23 percent.85 percent.48 percent to 78. However.381 per share. in the study by Mat Nor and Sulong (2009) found the mean percentage of ownership concentration is about 57 percent. and the range is from 8. it indicates a substantial variation in the amount of dividend distribution in Malaysia.073 and a maximum value of RM 1. Nevertheless.059 per share.50 percent which implies that almost 40 percent of shares ownership is concentrated in hands of five largest shareholders among Malaysian firms. the mean is 41. the substantial mean value of CONC2 implies that Malaysian firms are highly concentrated. The average dividend distributed among the companies in the sample is RM 0.which can be considered as high. In terms of ownership variables. This is due to some companies not disbursing any dividend while some companies distribute their dividend as high as RM 0. with a minimum value of -RM 1. 62 . which measured by the summation of the percentage of shares controlled by two major shareholders. for the second ownership concentration variable (CONC2). thus.

878 CONC2 0.300 121.717 DISP 0.1000 7.0822 0.245 8.455 6.281 5.216 0.820 4.0001 0. Table 4.0016 0.213 -0.0000 0.In addition.1716 0.0000 1.0000 0.0000 0.7811 0.541 63 KURTOSIS MINIMUM STD.5362 0.4185 0.0000 0.1188 2.0062 0.2417 0.496 1. the mean for ownership dispersion of zero percent and ranging from 0 percent to 0.1558 1.510 Earnings 15.1212 1.8688 0.1114 0.8671 -107. DEV MEAN .0400 23.021 FOR 0.402 CONC1 0.01 percent.4203 2. is another indication of highly concentrated feature of Malaysian firms.172 INST 0.1: Summary Descriptive Statistic SKEWNESS MAXIMUM Dividend 5.9593 0.0000 38.0848 0.2024 1.6610 0.5981 0.369 1.5549 0.1482 0.821 MNG 0.

The coefficient has a range of possible value from -1. which ranges from a low of zero percent to a 55. while the sign (+ or –) indicates the 64 .12 percent found by Mat Nor and Sulong (2009) in 2005.00. 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 1997.49 percent. Thus. mutual funds. The value indicates the strength of the relationship.3 CORRELATION ANALYSIS Pearson correlation coefficients for the primary variables are provided in Table 4. managerial ownership (MNG) has a mean percentage of 11. a standard deviation of 15.For institutional ownership (INST). unit trusts.93 percent and showed a 20.16 percent to 95. 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.22 percent. Further. Correlation is a single number that describes the degree of relationship between two variables. pension funds and financial companies. The foreign ownership (FOR) has an average value of 8.00 to +1.24 percent standard deviation. an increase from the 6.14. 4. The range is from 0.2.58 percent had been recorded.

Among the independent variables. The positive correlations are consistent with the signaling theory. In this study. dividends are also significantly positively correlated with CONC1 (corr = 0.001) and dividends contradicts the theoretical literature. p-value = 0. However. the result reveals a negative and significant correlation (corr = -0.173. 65 . Nevertheless.019) indicating the possibility of these three variables having predictive power on dividends and the positive relationship as theorized by the literature. p-value = 0.166.500.001) and foreign (corr = 0. a negative correlation (-0. CONC2 (corr = 0. p-value = 0. Thus.021) between earnings and ownership dispersion.256. CONC2 = 0. significant correlation between MNG (corr = 0. which argues that an increment in dividends will lead to earnings increasing.253.017) ownership since profitable companies are an attractive place for investors to invest. Besides that.000) and INST (corr = 0.424. earnings also have a positive significant correlation with institutional (corr = 0.095) between earnings and ownership concentration. p-value = 0. Besides that. This is probably because highly concentrated companies will lead to a good awareness of the company progress.191. eight interval-level variables are studied and the relationships among all of variables are estimated.128) between managerial ownership and earnings was surprising. there is a positive correlation (CONC1 = 0.direction.374.000). p-value = 0. p-value = 0. There is a positive significant correlation (corr = 0.000) between dividend and earnings. p-value = 0. as expected. the negative. p-value = 0.137.

174* (0.000) -0.253** (0.808) 0.062) * Correlation is significant at the 0.000) 0.126 (0.424** (0.374** (0.077) -0.145 (0.091) -0.338** (0.001) -0.020 (0.056 (0.128 (0.000) 0.019 (0.016) -0.009 INST 0.047) CONC2 0.052 (0.460) -0.192** 0.054 -0.132 0.019) 0.339** (0.037 0.021) -0.125) 0.247) -0.146** 0.000) E CONC1 CONC2 DISP INST MNG CONC1 0.095 (0.2: Pearson Correlation Matrix among the Variables D E 0.411) -0.933** (0.000) -0.137* (0.173* (0.281) MNG -0.500** (0.110 (0.001) -0.166* (0.143 -0.017) -0.Table 4.060) -0.008 (0.256** (0.01 level 66 .048 (0.000) DISP -0.264) FOR -0.05 level ** Correlation is significant at the 0.088 0.191* (0.

4. the results also reveal that both ownership concentration variables have a significant negative relationship with ownership dispersion. are significantly positively correlated (corr = 0. p-value = 0. Multicollinearity problems could be detected from the correlation matrix for the independent variables. p-value = 0).1.192 (p-value = 0.016).000). Further.174 (p-value = 0. If the variance inflation factor (VIF) value is larger than ten and the tolerance value is below 0.1 Multicollinearity The regression process commences with the identification of multicollinearity problems.338. Besides that. The degree of correlation between CONC1 and DISP is -0. CONC1 and CONC2.000) and CONC2 (corr = 0. both ownership concentration variables also recorded a significant negative correlation with managerial ownership.4 REGRESSION ANALYSIS 4.009). multicollinearity problem is said to exist among 67 . CONC1 (corr = -0. and between CONC2 and DISP is -0. 4. The significant correlation among the independent variables indicates a need for particular attention to circumvent potential multicollinearity problems during the regression analysis. The significant positive correlation between these two variables is expected since both are measurements of similar variables.Both ownership concentration variables. p-value = 0. Multicollinearity problems arise when one or more of the explanatory variables are exact or near exact linear combinations of other explanatory variables.933.339.

3: Variance Inflation Factor of Variables (tolerance value is given in the parentheses) FA M 1 ERNCHG 28.the independent variables.054) 3.171) 3.340 (0.976 (0.166) 6.221) 23. a multiple regression analysis was executed again. Table 4.651 (0. Table 4.3 shows that there are some multicollinearity problems in the regression models.286) 3.076) 4.221 (0.520 (0. As a consequence.198) 18.049 (0. but some of the variables were dropped.907 (0.043) 13.291) 6.274) 5.475 (0.104 (0.035) 5.440 (0.333 (0.017 (0.310) 13.715 (0.042) 2 46.499 (0.075) 4.162 (0.251) 3.418 (0.863 (0.022) 1 PA M 2 1 WAUD 2 ECHGCONC1 ECHGCONC2 ECHGDISP ECHGINST ECHGMNG ECHGFOR ERN ERNCONC1 68 .223) 23.162) 3.192 (0.

404) 1.287 (0.521) 1.4.508) 4.383 (0.713) 13.049 (0.487 (0.547) 1.301) 2.247) 2.319 (0.ERNCONC2 12.290) 2.509 (0.437) 2.339) 1.561 (0.079) 1.672) 1.967 (0.738 (0.427) 1.428) 2.513) 4.078 (0.514) 4.245) 3.342 (0.078 (0. This problem emerged when the residuals are not free from one observation to other observation.338 (0.446 (0.965 (0.509) 4.076) 1.918 (0. Serial correlation occurs when a long series of observations are correlated with each other.948 (0.477 (0.420) 1.829 (0.641) 1.245) 3.701) ERNDISP ERNGINST ERNGMNG ERNFOR D(t-1) D(t-2) 4.403 (0.2 Serial Correlation and Heteroscedasticity Test Subsequently.426 (0.946 (0.050 (0. 69 . the models were tested for serial correlation and heteroscedasticity.247) 2.075 (0.

477 (0. 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. Both of tests were done by using Microfit software.475 (0.075) 0.083) 0.103 (0.083) 0.4: Serial Correlation and Heteroscedasticity Diagnostic Test FA M 1 Serial Correlationa 3.010 (0.085) 0. the purpose of the heteroscedasticity test is to test whether the regression model meets the assumption of homoscedasticy. while heteroscedasticity refers to the situation where the variances of residuals vary.160 (0.342) Heteroscedasticityb a Lagrange multiplier test of residual serial correlation Based on the regression of squared residuals on squared fitted values b On the other hand.565) 2 3.331 (0.962 (0.905 (0. As a result.904 (0.010 (0.749) 1 2.491) WAUD 2 3. 70 . from Table 4.Table 4. whether there is any unequal variance of the residual between one to the other observation in the regression model.342) 1 2.082) 0.490) PAM 2 3. Homoscedasticity refers to the model where the variance of residual from one to the other observation is constant.085) 0.960 (0. or in other words.4. and it can be detected from the serial correlation and heteroscedasticity diagnostic test.0197 (0.

909) 0.015 (0.520 (0.014 (0.063 (0.065 (0.736) 0.4.133) -237.215* (0.684) -0.669) 2 0.023 (0.134) 0.173) -566.044 (0.575 (0.12) 1 0.141) PA M 2 0.604* (0.5: Results of Multiple Regression Analysis of Dividend Policy Models FA M 1 Constant 0.595) 0.230) 1 0.567) 0.029 (0.014 (0.145 (0.758 (0.696 (0.478 (0.4.477 (0.016) 0.007) 0.616* (0.421) 0.388 (0.328) -0.001) 43.638 (0.3 Regression Results Table 4.215* (0.305* (0.152* (0.14) 0.001) -237.551) 0.045 (0.909) 0.553) 0.150* (0.389 (0.667) 44.233) ECHGCONC1 ECHGCONC2 ECHGDISP ECHGINST ECHGMNG ECHGFOR ERNCONC1 ERNCONC2 ERNDISP ERNGINST 71 .026) -213.144) WAUD 2 0.149* (0.404) 0.007) 0.064 (0.

081 (0.451) -0.0 percent if CONC2 is used.960) -0.000) *Significant at the 0.449 (0.110* (0.5 percent if CONC2 is used.018) -0.001 (0.000).489 (0.213 0.213 0.023 (0. a measure for the strength of the regression.180 6.057 (0.174 5.ERNGMNG -0.029 (0. for the Waud model is 15.020 (0. In terms of the adjusted coefficient of variation (R2).157 4. 72 .717) -0.163 5.108* (0.874 (0. reveals that each dividend model is significant at 5 percent (p-value = 0.218 (0.719) -0.827 (0.031) 0. Therefore. The explanatory power for partial adjustment model is 16.115 4.110* (0.000) 0.663) 0.960 (0.023 (0.150 0.000) 0.057 (0.664) 0.000) 0.196 0.109* (0.3 percent if CONC1 is used and 18.020 (0.196 0.05 level The F.000) 0.000) 0.993) ERNFOR D(t-1) D(t-2) R2 Adjusted R2 F-statistic 0.449) -0.7 percent if CONC1 is used and 17. while for the full adjustment model is only 12.tests.025) 0.793) -0. the partial adjustment model is better in explaining the variation of corporate dividend policy.790) -0.0 percent if CONC1 is used and 11.120 5.145 0. it can be concluded that ownership classes are vital in determining a dividend policy.4 percent if CONC2 is used.012) -0. Whereas.029 (0.

Moh’d et al. Therefore. Nevertheless. managerial ownership has a negative coefficient in the partial adjustment model and the Waud model. They show that 73 . Therefore.tests show that only the concentrated ownership variable is significant for every type of dividend model. although the results reveal the expected sign in the partial adjustment model and the Waud model. Both CONC1 and CONC2 were positively and significant in influencing dividends at the 5 percent critical value. High dividend payments can be used for mitigating agency conflicts since dividends can be substituted for shareholder monitoring. this finding is similar to the results found by Noronha and George (1996). Institutional ownership had been found to be positively and significantly related to dividends in Alli et al. but the critical values are insignificant. Further. large shareholders have strong incentives to require higher dividend payments in order to reduce monitoring costs. This finding is consistent with the results presented by Easterbrook (1984) and Mat Nor and Sulong (2009). However. In this study. While the full adjustment model did not only produce the unexpected sign. (1995) and Manos (2002).T. it shows that dividends in Malaysia do not have any significant relationship with institutional ownership. it was insignificant. it is also insignificant. this finding is consistent with the study by Mat Nor and Sulong (2009). (1993). 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.

dividends are smoothed and rarely decreased. the payout rates are not related to proxies for agency cost variables. Interestingly. 74 . the more acute the free rider problem. Besides that. the relationship is insignificant. the greater the need for dividend distribution as outsider monitoring. 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. 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. most of the regression models do not only produce the unexpected sign but is also insignificant relationship.if there are alternative devices to control for agency costs. this study reveals that D(t-1) is significant in influencing dividends but in a negative form. For ownership dispersion. hence to minimize the agency problem. Although it shows that the last year dividend is vital in determining current dividends. This result is contrasts to that of Rozeff (1982) and Moh’d et al. Hence. (1995) which concluded that the more widely the ownership spread. Similar results were also found by Mat Nor and Sulong (2009). The full adjustment model records a positive relationship between dividend payouts and foreign holdings but. the partial adjustment model and the Waud model also showed similar results if CONC1 is used. Thus.

Furthermore. this study also reveals that Malaysian dividend behavior contrasts with the theory of dividend smoothing proposed by Lintner (1956). Besides that.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. the findings also reveal that only ownership concentration had significant influence on Malaysian corporate dividend policy. 75 .4.

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. namely ownership concentration.5. Section 5.1 INTRODUCTION This chapter winds up the overall study. Then. specifically the Modigliani-Miller theorem and Agency theory has been reviewed.6 concludes the chapter.2 discusses the overview of the link between dividends and ownership composition research process. Subsequently. the directions for further research are presented in Section 5. ownership dispersion. institutional ownership.CHAPTER FIVE CONCLUSION 5. Five independent variables used as the proxies of ownership structure were identified. Therefore. in-depth empirical literature about the relationship between dividend policy and ownership structure have also been reviewed. Besides. 5.3 presents the summary of the findings in the analysis and Section 5. Section 5. managerial 76 . and Section 5.4 discusses the implications of the study. theoretical literature for dividend policy.

however. The samples are taken from the four largest sectors on the Main Board of Bursa Malaysia. industrial.ownership and foreign ownership. the correlation analysis was carried out to examine the degree of relationship between possible pairs of variables. The predetermined variables are hypothesised. These models had been modified to account for the possible effects of ownership structure in determining the level of the corporate dividend. the Partial Adjustment Model and the Waud Model. These companies were selected based on proportionate stratified random sampling. The analysis commenced with a discussion of the sample descriptive statistics. the measurement of variables was guided by the prior research. trading and services and properties sector. Subsequently. A total of 150 companies were identified as the sample for the study. This study utilized three dividend models to test the hypothesis of positive links between ownership structure and dividend policy: the Full Adjustment Model. This study employed annual data from 2005 to 2007. The cross-sectional nature of the data called for use of the OLS multiple regression technique. 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. namely consumer. Besides that. the result revealed that a multicollinearity problem had 77 .

Five predetermined explanatory variables.occurred. the Partial Adjustment Model was superior. managerial ownership and foreign ownership were regressed against dividends. the regression model of dividend change against all the independent variables revealed that each dividend model was significant at a 5 percent confidence level. institutional ownership. and handling for multicollinearity problems. the regression model had been not encountered with serial correlation and heteroscedasticity problems. since it was selected from the four largest sectors on the Main Board of Bursa Malaysia whose annual reports were available for 2007. the Partial adjustment model and the Waud model to examine the potential associations between ownership structures and dividend policy.3 SUMMARY OF FINDINGS This study was designed to examine the effect of ownership structure on corporate dividend policy. ownership dispersion. as a corrective action. After a corrective analysis was conducted. Nevertheless. 150 companies were identified as the sample. Therefore. This study had employed the Full adjustment model. since it could explain up 78 . 5. However. some of the variables were dropped from the model and regression analysis was executed again. namely ownership concentration. This sample is representative for Malaysian companies.

Nevertheless. 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. institutional. thus implying that more cash flows can be distributed as dividends. Besides that. it implies that these four variables are not vital in explaining dividends. foreign and ownership dispersion on dividends. foreign and 79 . only one explanatory variable. The positive relationship between ownership concentration and dividends supports the findings in Shleifer and Vishny (1986). this study introduced a new measurement for ownership concentration. which is ownership concentration.4 percent by the Waud model and 12. 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. Ownership concentration has a positive significant relationship with dividend payment. institutional. Furthermore. was found to be statistically significant in influencing corporate dividend policy. hence dividend decisions in Malaysian companies are not influenced by managerial. the results prove the insignificant relationship of managerial.0 percent by the Full adjustment model. 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. Therefore.0 percent of the variation in dividend compared to 17.to 18.

Thus. which contrasts with the theory of dividends smoothing by Lintner (1956). the regression model of dividends against all the independent variables was also found to be significant. Nevertheless. this study reveals that D(t-1) is negative and significant in influencing dividends. the findings reveal that the model of research explains less than 20 percent variation of dividend phenomenon in Malaysia. According to Lintner. Besides that. the insignificant value of these four variables in determining dividend distribution has also been found by previous researchers. In this study. dividends are smoothed and rarely declined. The positive significance of ownership concentration variables implies that the formation of ownership has an effect on the amount of dividends distribution. 80 . 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. Additionally. Nevertheless.ownership dispersion. 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.4 IMPLICATIONS OF THE STUDY The research has examined the relationship between dividend policy and ownership composition among the public-listed companies in Malaysia. over time is taking place. instead of dividend increasing trend. 5. it is observed that the dividend decreasing trend. Thus.

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. Nevertheless. Moreover. Thus. 5. Besides that. However. Shareholders must realize that financial policies such as dividend policy can serve as a mechanism for reducing agency costs. regulatory bodies should also be concerned with the formation of ownership in formulating the related regulations to better control the agency conflict. board of directors’ ownership.5 DIRECTION FOR FURTHER STUDIES There are a rich possible number of variables that can be used to examine the determinants of dividend policy. there might be other ownership variables that can be incorporated to explain the link between dividends and ownership composition. it would be beneficial if further research would be able to include other variables such as government ownership.This study suggests that shareholders with respect to stock investment in companies should concern themselves with the agency conflict between ownership classes. 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. family ownership and many 81 .

6 CONCLUSION This chapter provides a brief summary of the overall study. Besides that. Future researchers on this topic may also use survey and interview methods to gauge top management and investor perspectives on this issue. 5. smoothing theory and catering theory in the pursuit to understand the influence of factors on dividend policy in Malaysia. In addition. residual theory. the longer period of study may also enhance the predictability model of the research. Moreover. The findings will provide an interesting comparison to the findings from this study. future research also can use Tobit regression to get better results since some of dependent variable is zeros. life-cycle theory. 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. 82 . Furthermore. Besides. The research procedure and the results from the analysis were discussed. the consequences of study and the recommendation for future research were also presented. This can help to better understand Malaysian companies’ dividend decisions. 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.other types of ownership classes.

Ang. The Financial Review 28: 523-547.O and Jeon. Lin. F. El-Masry. D.L. Baker. Amidu and Abor. The Journal of Risk Finance 7: 136-145. 1993... G. Determinants of corporate dividend policy: A factorial analysis.K. and Gandhi. Foreign ownership. R. S..Q. S. domestic ownership and payout policy. Managerial Finance 12: 953-964. A. 1991. ownership structure and dividend policies in an emerging market...A. Agency costs and ownership structure. 2000..REFERENCES Abdelsalam. J.. 2008. http://fma.pdf 83 . The effect of dividend yield and dividend policy on common stock prices and returns. The Journal of Finance 1: 81-105. A. Determinants of dividend payout ratios in Ghana. Cole. Black. G.S. Khan. M. & Ramirez. 1974. The dividend policies of all-equity firms: A direct test of the free cash flow theory. Board composition. Liquidity versus control: The institutional investor as corporate monitor.. K. J. Jayaraman. The perception of dividends by Canadian managers: new survey evidence 3: 70 – 91. Agrawal.W.C.S. Columbia Law Review 91: 1277-1368. Managerial and Decision Economics 15: 139-148. Dutta. Cook. Journal of Financial Economics 1: 1-22. Saadi. 2006. 2006. H.Jr. Q. 2007. and Scholes. & Elsegini.. S.. Coffee. J. J. O. N. 1994. D.org/SLC/Papers/ForeignOwnership_Dividends.. Alli..

1999. International Review of Financial Analysis 8:177-197. E.. 84 . K. American Economic Review 74: 605-659. K. F. 1984. 1980.H. Fluck. Faccio. H. 2001.. Montfort. Available at http://hdl. Easterbrook. J. Fama. The role of institutional ownership. K. M. L. American Economic Review 91: 54–78. and Young.. E. I. Dividend and expropriation.S. Academy of Management Review: 57-74.E. C.Crutchley. Research Paper. M. L. J. Vink..H. A. 1999. Agency cost. Jahera. Two agency-cost explanations of dividends. Journal of the American Statistical Association 63: 1132-1161. Lang.. Dividend policy: An empirical analysis. Fama.F.net/1871/1517 D’Sauza. Dorsman.. Agency Theory: An assessment and review. Jensen.R. J.F. Z. 1989. 1968.handle. Eisenhardt.V. Babiak.M.E.P. 1999.An international perspective. A. Agency problem and the theory of the firm. Review of Financial Studies 2: 379-404. & Saxena.B. An adjusted Lintner-model for The Netherlands. Journal of political economy 88: 288-307. The dynamics of the managers-shareholders conflict.. & Raymond. Managerial Finance 25.. market risk. Agency problem and the simultaneity of financial decision making. investment opportunities and dividend policy . 1999.

Kumar. R.R. M. Available at SSRN: http://ssrn. 85 . Jr.B.Foong. S. Gugler. R&D. Zakaria.. I. International Review of Financial Analysis 11: 111-138. Lee and D. Financial management: 16-22. Friend. Dividend and stock prices. M. D. G. Review of Economics and Statistics: 99-105. B. 1999. Dividends. S. and dividend policy in Japan. Dividend policy theories and their empirical tests. Dividend policy and corporate monitoring: Evidence from the regulated electric utility industry. Earnings and Stock Prices. American Economic Review 54: 656-682. and Puckett. K.J. Pascal.. agency conflicts. Kimie & Nguyen. Journal of Banking and Finance 27: 1279-1321. 2007. 11th Malaysian Financial Association Conference. 1964.H. A. Corporate governance.G.C.com/abstract=953433 Harjito. & Shome. Journal of financial and Strategic Decisions 12: 53-62. Ownership concentration. D. 2002..K. H. Harada. N. S. Frankfurter. 1994. Han. The influences of agency factors and transaction costs factors to dividend payout ratio..S.M.. Gordon. Institutional shareholders and dividends.Y. and capital investment. 2003. dividend payout policy and the interrelation between dividends. 2006.. 2009. Tan. K. Hansen. 1959. and Wood. Labuan Bulletin of International Business & Finance 5: 97-111. Suk. Firm performance and dividend-related factors: The case of Malaysia.

1999. & Guizani.. C. European Journal of Scientific Research 25: 42-53. Jensen.. Agency costs of free cash flow.. Journal of Financial Economics 3: 305-360.Holderness.. 86 . Jensen. Company dividends and ownership structure: evidence from UK panel data. Economic Policy Review 9: 51–64. Theory of the Firm: Managerial behavior Agency Costs and Ownership Structure. Vishny.H. La Porta. M. A survey of blockholders and corporate control. The perception of dividends by Canadian managers: new survey evidence. A. International Journal of Managerial Finance 3: 70-91. M. La Porta. and W. Simultaneous determination of insider ownership .G. S. D. F.W.. 2007. Dutta.W. 2009. American Economic Review 76: 323-329. Journal of Finance 55: 133. A. Kent Baker. M. debt and dividend policies. S. Khan. Saadi. Shleiffer and R. Jensen.. 1976. Kouki. 2000a. T. G. 2006.P. Vishny.Lopez-de-Silances. H. 2003. and Zorn. M.S. R. Solberg. Meckling.Lopez-de-Silances. F.. 1986. T. Agency problems and dividend policies around the world. R.. R.C. 1992. Ownership structure and dividend policy evidence from the Tunisian Stock Market. corporate finance and takeovers. Journal of Finance 54: 451-517. The Journal of Financial and Quantitative Analysis 27: 247-263. The Economic Journal 116: 172-189. Shleiffer and R. Corporate ownership around the world.

Journal of Economics and Finance 3: 105-117. & Willey. L. Mancinelli.N. and Rimbey. M. and Sharp.S.. J. Mendez. Distribution of incomes of corporations among dividends. A. University Atmajaya Yogyakarta-Indonesia. & Ozkan. Lloyd. 2002.P. W. The Financial Review 30: 367-385. Rafiq. 2006.E. Perry. 12: 262-282.. Moh’d. J.. Presented on National Seminar Surviving Strategies to Cope with the Future. leverage and dividend policies. Relevance of agency theory fro dividend policy in an emerging economy: The case of Bangladesh. The association of managerial ownership with dividend policy and leverage policy: Indonesian firms.. 1985. Agency costs and dividend payout ratios.S Jahera Jr. and D.A.. American Economic Review 46: 97-113.M. retained earnings and taxes. T. A.B..G. Page. An investigation of the dynamic relationship between agency theory and dividend policy. 1995.A. 87 . R. L. Mollah. The European Journal of Finance. The IFCAI Journal of Applied Finance 13: 5-15. 2007.A. P. 2002. working paper series 41: 125. Quarterly Journal of Business and Economics 24: 19-29. Mahadwartha. J. R. J. Dividend policy and agency theory: Evidence on Indian firms. Agency costs in the banking industry: An examination of ownership behavior. 1994.. Ownership structure and dividend policy: Evidence from Italian firms.Litner. 1995. Finance and development research programme. P. Manos.

The monitoring rational for dividend and the interaction of capital structure and dividend decision. D. ECGI-Finance Working Paper 128. Shome. Journal of Business 34: 411-433. The cost of capital. A. D. Szilagyi.S. Capital market review 15: 73-101. beta and agency costs as determinants of dividend payout ratios. 2009. The interaction effect of ownership structure and board governance on dividends: Evidence from Malaysian Listed Firms. F.. 1982. How relevant is dividend policy under low shareholder protection. Noronha. Naceur. Studies in Economics and Finance 26: 182-197. M. Journal of Financial Research 5: 249-259.. F.L. Journal of Banking and Finance 20: 439-454. A. M. Growth. L. On the Determinants and Dynamics of Dividend Policy.M.. Schooley. Barney. F. Goaied. Nor.M. Jr.. Dividend policy. B. and P.com/abstract=889330 Najjar. and Belanes. M. Using dividend policy and managerial ownership to reduce agency costs.K and G. growth and the valuation of shares. 2006. Available at SSRN: http://ssrn. 1994. and D.. Z.E Morgon. and Modigliani. http://www. and Modigliani..B.K. Renneboog. American Economic Review 48: 261–297. Miller. 1961. 88 . Dividend behaviour and smoothing new evidence from Jordanian panel data.com Rozeff.Miller. corporation finance and the theory of investment. and Sulong. Journal of Financial Research 17: 363-373. M. 1958.ssrn. 2007. 2006. G.G. 1996.

1986. 1997. Journal of Finance 52: 737–783.W. Journal of the American Statistical Association: 134-145. A. Waud. Small sample bias due to misspecification in the ‘partial adjustment’ and ‘adapted expectations’ models. R. H. A survey on corporate governance. The link between dividend policy and institutional ownership.. A. and Vishny. 89 . Shleifer. Short. Large stockholders and corporate control. H.W. Zhang and Keasey.Shleifer. 1966. R. R. Journal of Political Economy 95: 461–488. Journal of Corporate Finance 8: 105-122. 2002. and Vishny.

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

32229 1.057133 .090]* * * * * * B:Functional Form *CHSQ( 1)= . 143) 6.0098[.19052[.672]* * * * * * C:Normality *CHSQ( 2)= 1062.075253 -. Criterion -373.10950 .38852 . 142)= 2.90475[.8488 DW-statistic 2.000] Mean of Dependent Variable .THE PARTIAL ADJUSTMENT MODEL Regression of DIVCHG on ERNCONC2.342]*F( 1.033542 .17995 S.014456 .551] ERNINST .045502 3.1[. 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 .4494[.59748[.717] ERNFOR -. ERNMNG.43097[.662]*F( 1.345]* ******************************************************************************* 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 .89810[.064280 -.1463 Residual Sum of Squares 1160.7579 397.2774[.9363 -. F( 6. 148)= .2666 ******************************************************************************* Diagnostic Tests ******************************************************************************* * Test Statistics * LM Version * F Version * ******************************************************************************* * * * * * A:Serial Correlation*CHSQ( 1)= 3.449] DIV06 -. of Dependent Variable 3.230] ERNCONC2 . ERNINST.000]* Not applicable * * * * * * D:Heteroscedasticity*CHSQ( 1)= .18058[.E.083]*F( 1.001] ERNDISP -237.D.75920[. of Regression 2.72460 S.3115 Schwarz Bayesian Criterion -383.8492 F-stat.36285[.667] ERNMNG -.9 Equation Log-likelihood -366.3115 Akaike Info.2055[.14913 .043210 -2. 142)= .9076[.012] ******************************************************************************* R-Squared .023324 .21297 R-Bar-Squared .5342[. ERNDISP.

ERNDISP.078352 2. 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 .134] ERNMNG -.491]*F( 1.46973[.E. ERNINST.793] DIV06 -.915]* * * * * * C:Normality *CHSQ( 2)= 905.2 Equation Log-likelihood -367.012220[.0229 383.7381[.044512 .960] ******************************************************************************* R-Squared .000]* Not applicable * * * * * * D:Heteroscedasticity*CHSQ( 1)= .144] ERNCONC1 .074426 . 148)= .19648 R-Bar-Squared . 141)= 2.026941 .21454 . of Regression 2.9620[.47685 .019534 .1463 Residual Sum of Squares 1185. Criterion -375.085]*F( 1.8404[.909] ERNINST .D.663] ERNFOR .THE WAUD MODEL Regression of DIVCHG on ERNCONC1.494]* ******************************************************************************* 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 .1966 .049954 -2.7561[.011488[.11488[. ERNMNG.4691[.031] DIV05 .028871 .26247[.72460 S.2669 ******************************************************************************* Diagnostic Tests ******************************************************************************* * Test Statistics * LM Version * F Version * ******************************************************************************* * * * * * A:Serial Correlation*CHSQ( 1)= 2. ERNFOR.8890 F-stat.066125 -.049739[. 142) 4.8666 Akaike Info.5074[.10896 .47457[.8666 Schwarz Bayesian Criterion -387.007] ERNDISP 44.0013400 .1813[.029529 1.9091 DW-statistic 2.9605[. of Dependent Variable 3.094]* * * * * * B:Functional Form *CHSQ( 1)= . F( 7.15687 S.912]*F( 1.32459 1.000] Mean of Dependent Variable . 141)= .43662[.

ERNDISP. of Regression 2.38835 .000] Mean of Dependent Variable .17984[.001] ERNDISP -237. 141)= .59468[.026671 .2603[.9 Equation Log-likelihood -366.89730[.THE WAUD MODEL Regression of DIVCHG on ERNCONC2.8875[.75629[.014474 .D.025] DIV05 .033738 .3[.36036[.057157 .19108[.2180E-3 .14911 . ERNMNG.21297 R-Bar-Squared .553] ERNINST . 141)= 2.6381 399. F( 7. of Dependent Variable 3.8592 F-stat.048312 -2.90395[.719] ERNFOR -.4894[. 148)= .669] ERNMNG -. Criterion -374.993] ******************************************************************************* R-Squared . 142) 5. ERNFOR.2702[.1984[.075575 -.32405 1.3541 DW-statistic 2.672]* * * * * * C:Normality *CHSQ( 2)= 1062.083]*F( 1.3115 Schwarz Bayesian Criterion -386.0081748[.10968 .064883 -.342]*F( 1.1463 Residual Sum of Squares 1160.6039 -.451] DIV06 -.E.000]* Not applicable * * * * * * D:Heteroscedasticity*CHSQ( 1)= . 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 .233] ERNCONC2 .72460 S.045735 3.17418 S.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 .023381 .42903[.091]* * * * * * B:Functional Form *CHSQ( 1)= . ERNINST.3115 Akaike Info.2666 ******************************************************************************* Diagnostic Tests ******************************************************************************* * Test Statistics * LM Version * F Version * ******************************************************************************* * * * * * A:Serial Correlation*CHSQ( 1)= 3.0102[.662]*F( 1.

Sign up to vote on this title
UsefulNot useful